Consolidated 9 September 2026. Start with Where influence becomes a decision, then inspect the claim and evidence index. The current visual/portable edition is Red Threads r33.
\nWe have identified specific financial, governance, campaign, policy and operational relationships. The strongest organising question is how an assessment becomes a condition of participation: who supplies it, who acts on it, which institution retains discretion, and who can reverse the result. Funding, selection, classification, action and remedy recur across the AI, online-safety, identity, advertising, payments and research-access branches, with different instruments in each.
\nThe evidence supports substantial organised resources for AI advocacy, specific creator production/selection mechanisms, identifiable policy transmission, and consequential classification and access systems. It also shows divergent positions among recipients of common funding, declined or narrowed proposals, terminated distribution relationships and actual constraints on recipient discretion. These observations strengthen a distributed institutional account while leaving a common cross-sector controller unestablished.
\nThe consolidation integrates six accepted packets previously outside the graph: CDDP evaluation, OSDI/PUBLIC, EU researcher-access rules, AliExpress commitments, audit/trustee evidence and voluntary application outcomes. It also brings the seven r32 GDI/rights packets into the current visuals. The dated public ThreatExchange correction review and stablecoin check remain clearly identified supplements.
\nThe most consequential remaining gaps are specific instruments and outcomes: signed grant conditions and exercised approvals; actual evaluation/acceptance and funding decisions; a classification correction received and acted on by a customer; a real application paired with its response and usable access; an appeal linked to a corrected practical result. The inquiry has not established the broader Fabian/WEF/owner-bank/medical-data instruction chain suggested earlier.
\nThis is a consolidation of retained research through 9 September 2026, not a new external investigation or a fresh audit of every original. Source records retain their reading scopes and attribution; counts are not independent corroboration. Original reports and earlier revisions remain historical snapshots. Construction and verification status are recorded in current work; the earlier summary is preserved in the consolidation folder.
", "path": "campaign-summary.md", "sourceIds": [], "markdown": "Consolidated 9 September 2026. Start with [Where influence becomes a decision](current-synthesis.md), then inspect the [claim and evidence index](consolidation-wave-2026-09-09-synthesis-claim-index.md). The current visual/portable edition is Red Threads r33.\n\nWe have identified specific financial, governance, campaign, policy and operational relationships. The strongest organising question is how an assessment becomes a condition of participation: who supplies it, who acts on it, which institution retains discretion, and who can reverse the result. Funding, selection, classification, action and remedy recur across the AI, online-safety, identity, advertising, payments and research-access branches, with different instruments in each.\n\nThe evidence supports substantial organised resources for AI advocacy, specific creator production/selection mechanisms, identifiable policy transmission, and consequential classification and access systems. It also shows divergent positions among recipients of common funding, declined or narrowed proposals, terminated distribution relationships and actual constraints on recipient discretion. These observations strengthen a distributed institutional account while leaving a common cross-sector controller unestablished.\n\nThe consolidation integrates six accepted packets previously outside the graph: CDDP evaluation, OSDI/PUBLIC, EU researcher-access rules, AliExpress commitments, audit/trustee evidence and voluntary application outcomes. It also brings the seven r32 GDI/rights packets into the current visuals. The dated public ThreatExchange correction review and stablecoin check remain clearly identified supplements.\n\nThe most consequential remaining gaps are specific instruments and outcomes: signed grant conditions and exercised approvals; actual evaluation/acceptance and funding decisions; a classification correction received and acted on by a customer; a real application paired with its response and usable access; an appeal linked to a corrected practical result. The inquiry has not established the broader Fabian/WEF/owner-bank/medical-data instruction chain suggested earlier.\n\nThis is a consolidation of retained research through 9 September 2026, not a new external investigation or a fresh audit of every original. Source records retain their reading scopes and attribution; counts are not independent corroboration. Original reports and earlier revisions remain historical snapshots. Construction and verification status are recorded in current work; the earlier summary is preserved in the consolidation folder.\n", "file": "research/campaign-summary.md" }, { "id": "current-synthesis", "title": "Where influence becomes a decision", "description": "Consolidated account of research through 9 September 2026. This account accompanies r33 and two explicitly identified supplements. It is an interpretation of the research record, not a fresh audit of every original or a claim that the institutions below share one controller.", "html": "Consolidated account of retained research through 9 September 2026. This account accompanies r33 and two explicitly identified supplements. It is an interpretation of the acquired record, not a fresh audit of every original or a claim that the institutions below share one controller.
\nThe investigation has moved beyond a list of people who know one another. We can identify recurring ways in which resources become influence: financing particular work, selecting the people who perform it, defining qualifying categories, supplying assessments, and attaching practical consequences to those assessments. An advertisement, a payment relationship, a platform account or access to research data can depend on decisions made across several institutions.
\nThe most useful unit is the handoff. Who supplied an assessment? Under which agreement, rule or duty did someone else act on it? Who retained discretion? What actually happened? Who could reverse it? The record contains real financial and policy connections, along with shared methods and analogous structures. Those are different kinds of connection, and the coherent account depends on keeping them distinct.
\nOur strongest overall interpretation is that practical power is often distributed across these decision points. A central command is unnecessary for a restrictive result: independently motivated institutions can combine to produce it. That is an explanatory possibility supported by several bounded examples, not proof that every restriction arose this way. Common funders also support divergent positions, regulators decline proposals, distributors exit relationships, and courts or regulators impose remedies. Those observations constrain the stronger theory of a unified programme.
\n\nThe original question remains important. The retained record shows substantial organised resources behind AI advocacy and several concrete routes into creator production. FLI's 2025 standalone audit reports US$45,197,838 in grants from Lightcone Foundation. Earlier transfers involving Lightcone Foundation and Future of Life Europe concern assets and organisational allocation; they cannot be added together as annual persuasion expenditure. Lightcone Foundation is also a different entity from Lightcone Infrastructure. Scale is real, but the accounting categories matter. A1
\nCreator influence is not one arrangement. People First's published terms provide for approval, revisions, compliance and conditional payment; an individual campaign statement of work was not acquired. Frame describes an independent creator track and an institutional-brief track, names funding sources, and denies funder control over content or selection. A Future Tense creator publicly acknowledged support from FLI, Frame and BlueDot while asserting editorial independence. Rational Animations and ControlAI acknowledged a partnership, but the fee and agreement remain unacquired. The sponsor behind the initial Sabine Hossenfelder account remains unidentified. A2
\nSelection can affect public communication without script approval. Supporting people who already share a concern can increase that concern's reach. That inference requires less than editorial command, but it still cannot identify an undisclosed payment or instruction. ControlAI's acquired UK articles locate formal management and membership authority in its directors. A small related-organisation payable in its US return does not name the counterparty and cannot complete an assumed UK–US funding chain. A3
\nThe Global Call for AI Red Lines supplies a documented campaign-production chain: drafting, management, partner feedback and signatory work have named roles. CeSIA, The Future Society and CHAI are organisers; ControlAI and FLI provided partner feedback. The Call is distinct from ControlAI's A Narrow Path. Evan Hubinger's signature does not assign him every proposal advanced by another participant. Advocacy and summit access also remain separate from government adoption. This branch supplies real coordination, with a bounded object and purpose. A4
\n\nOak's grants make an instructive comparison. Its published 2025 awards include three-year support to Check My Ads, Digital Freedom Fund and Statewatch. Statewatch endorsed objections to broad scanning and possible identity requirements in the EU child-abuse proposal; the WeProtect Secretariat supported mandatory detection with safeguards in a different, earlier version of that debate. The comparison is evidence of divergent interventions, not a same-draft controlled experiment.
\nDFF reports support for Liberty's challenge concerning the alleged Apple technical-capability notice. Liberty filed in March 2025, before the reported June 2026 grant. Funding could sustain existing litigation; it did not originate that filing. Acquired procedural results also do not determine the merits. A common donor is therefore an important financial connection, but not a sufficient explanation of every recipient's choices. A5
\nDFF's selection reforms further separate donor resources, project selection and award execution. Public descriptions assign project selection to external peers and information/facilitation roles to staff, with confidentiality and conflict commitments. They do not reveal every operative recusal, signature or payment decision. The question “who chooses?” needs an answer at each stage. A6
\n\nThe UK–Australia duty-of-care lane supplies stronger transmission evidence than a shared-event roster. Carnegie-associated work, collaboration with Reset and citations in Australia's statutory review connect a developed model to official consideration. Commissioned research and programme funding sustained parts of that policy infrastructure; later funding cannot be assigned authorship of an earlier idea.
\nThe Australian exposure draft dated 8 September 2026 would allocate consequential discretion among Parliament's framework, ministerial rules and eSafety. It remains a proposal. Reset requested access for academic and third-sector researchers; the acquired draft instead requires employment by an Australian university and approval under prospective rules. That is a meaningful narrowing, not an awarded NGO entitlement. The private adult-consent provision limits what the duty requires; it does not create a universal adult veto. Nothing inspected establishes universal identification. A7
\nThe UK record also distinguishes required outcomes from particular identity suppliers. Ofcom recognises different age-assurance methods and accepted email-based estimation after supplied evidence; that was a methodological choice, not vendor certification or procurement. Check My Ads later sought customer/identity checks for monetising creators and advertisers and pre-monetisation clearance where risk indicators appear. Those proposals could affect pseudonymity and income, but do not concern every ordinary viewer. Ofcom expressly considered its crisis-revenue request and did not recommend specific monetisation practices in that measure. Other recommendations retain their own unresolved dispositions. A8
\nFull Fact offers unusually precise evidence of input and institutional response. Ofcom attributes a notice-indicator amendment to its submission and questions from an earlier DSIT meeting. A notice becomes an indicator to consider; the provider retains its criteria-based crisis determination. The record supports a specific contribution while leaving the meeting's underlying records and other decisions unacquired. Influence is visible without assuming wholesale control. A8
\nThe newly integrated procurement evidence makes the same distinction financially concrete. CDDP's Oliver Wyman evaluation commission has a disclosed £656,000 price excluding VAT, proposed criteria and a Treasury-linked feasibility gate. These establish commissioning and conditions, not payment, the final evaluation or an exercised funding decision. The announced £8.4 million programme allocation cannot be added to supplier prices as separate expenditure. Redactions and inconsistent contract dates remain unresolved. A23
\nOSDI had a CDEI-convened advisory group, PA Consulting programme evaluation and supplier participation by Faculty and PUBLIC. These are distinct roles. PUBLIC's account of a TRL4 sandbox describes prototype work; its later STCF2 assessment explicitly credits OSDI practices while retaining limited testing and supplier-reported performance. PUBLIC's claim to have contributed to CDDP does not identify it in a redacted subcontract list. Its current website-controller legal identity is not proof of the historical contracting party. The record now connects programmes, methods and personnel more clearly while preserving the missing agreements and acceptance decisions. A24
\n\nThe Lantern case study reports Meta sharing signals followed by separate investigations and action by Cash App and Xbox. The underlying case files were not inspected; the actions remain operator-reported. Programme rules include recourse, removal of no-longer-relevant information and administrator notification. Nothing in that case establishes that Lantern caused Steam removals. The live question is how information becomes a recipient decision, and what follows when the information changes. A9
\nPayment records make the recipient's role especially clear. Mastercard's retained August 2026 MATCH rules require inquiry but expressly permit onboarding a listed merchant. Stripe publishes a generally disqualifying practice with possible exceptions. The list and a restrictive underwriting decision are connected, but are not identical or a universal prohibition on banking. The 2025 eGumball appellate opinion applies FilmOn's reasoning about audience, purpose and context to confidential commercial reporting: an actual judicial bridge between advertising classification and payment reporting. Its procedural decision does not adjudicate the underlying report's truth or restore a merchant. A11
\nOther infrastructure gates allocate authority differently. Nominet's documented IWF cooperation predates the 2021 UK Safer Internet Centre grant, weakening the claim that the grant created that relationship. PIPCU combines public funding and police assessment with private intelligence; its announced industry co-funding pilot starts in April 2027 and does not establish current contributor command. Operation Creative's dated procedure allocates police tactical judgment, while advertising recipients choose how to implement information. Broader KYBC requests sought identification and consequences for commercial customers; requested reach across hosting and payments is not enacted universal user-ID law. A13
\nAdvertising classification adds a question about scale: is the object a page, channel or whole domain? February 2025 Senate letters questioned financing a website even where advertisements avoided prohibited material. DoubleVerify subsequently announced default Highly Illicit and P2P Sharing and Streaming categories and later reported implementation. The chronology does not establish that the letters uniquely caused the change. NCMEC notifications, DV's assessment and its later IWF relationship are different inputs. Announced defaults do not constitute our observation of customer settings, and general customer-choice language does not prove an override in every configuration. A14
\nCounterpressure belongs in this account. TAG reported no violations of the investigated brand-safety guidelines: its determination, not independent exoneration. DV's complaint challenged critics' assertions: allegations, not judicial findings. Its disclosed fee model does not establish a bounty for blacklisting. The consequences can be significant without a proven private motive or a single upstream instruction.
\nGDI illustrates why time matters. Its advertising-risk data had a documented Oracle distribution route, but Oracle's relevant service ended in September 2024. IAS later said it discontinued limited past GDI inputs. Historical relationships cannot establish present delivery. A publisher-review procedure may revise GDI's assessment, but the acquired record does not show automatic propagation to every buyer or an obligation to resume advertising. A16
\n\nGDI's public funding and research work reveal further roles without completing an alleged government blacklist. FCDO described supervised activities outside the United States; the signed grant and commercial-reuse conditions remain unacquired. CORDIS identifies a German GDI entity and a €385,275 EU contribution to AI4TRUST, not the whole consortium budget or an assumed UK guarantee. Project documents show keywords and annotations, model-assisted triage and human review. A seed term is not a verdict; a research warning is not an advertising ban. GDI's request for mandatory tool/API compatibility is a bid for an institutional role, not proof of its award. A17
\nThe integrated EU access rules separate three mechanisms. Article 40(4) provides project-specific vetting and a reasoned request through the relevant Digital Services Coordinator. Article 40(12) concerns public-data access. Article 22 trusted-flagger priority concerns notices and is different again. Funding and independence disclosures can form part of eligibility; public overviews need not name every applicant or funder. Neither Horizon funding nor a failed exact-name search establishes whether GDI received platform access. A18
\nAliExpress's June 2025 binding commitments supply concrete procedures and clocks. The company nominates and pays a monitoring trustee; the Commission approves and supervises the role. Paying the trustee does not confer the Commission's approval power, and appointment does not establish delivery of a report. The substantive commitment term and external reporting schedule differ. A later illegal-product fine concerns a separate issue; it is not evidence of breach of the research-access commitments. A19
\n\nThe retained FTI audit offers a useful stress test. It gives a positive Article 40(12) opinion while reporting no approved applications in the year ending June 2025. It says all examined applications were incomplete or invalid. The record supports a narrower conclusion about assessed readiness. It does not establish either that meaningful access was delivered or that requests were improperly refused. Its period also precedes the monitoring trustee's appointment. A20
\nTwo explanations remain possible: legitimate screening of deficient requests, or an interpretation/cure process that makes an access right difficult to use. Without application files, responses and actual delivery, the public audit cannot decide between them. Likewise, the first monitoring report and the Commission's response were not acquired; absence from the acquired record is not proof of non-delivery.
\nThe researcher-side material is informative but not an adjudication. The DSA40 Collaboratory collects voluntary reports, assists applicants and advocates support for intermediaries. Its September 2026 tracker headline reports 70 applications, while its retained table sums to 68; other subtotals also differ. Its one open AliExpress entry lacks the identity and dates needed to connect it to GDI or the FTI cohort. Organisers' judgments that rejection reasons are invalid are leads, not acquired rulings. Disclosed project and host funding establish resources, not donor control of findings. A21
\nThis is the same evidentiary problem seen in the procurement branch: an evaluator commissioned, a test described, a positive opinion, an accepted application and a useful result are distinct events. Counting one as another can make both success and obstruction appear more certain than the record permits.
\n\nThe retained public ThreatExchange client provides a substantive counterexample to technical permanence. At the inspected commit, deletion markers can remove local records; a stale-checkpoint path calls for clearing and a fresh fetch; opinion structures can represent disagreement. These are static capabilities in particular code paths, not a demonstrated end-to-end deployed execution trace. Lantern's use of hosted ThreatExchange does not prove its participants use that client or version. No account restoration was observed. TCAP's separately inspected hash interface also documents deletion and feedback; a historical Altitude importer is another interface and version. A10
\nThere are also actual obligations constraining recipient discretion. The January 2025 CFPB order requires Block to provide specified restriction/appeal notices and to decide an appeal and change status as appropriate within ten business days. That does not guarantee every appellant restored access or constitute an admission of every finding. It does materially narrow the earlier remedy gap. Acquired July settlement terms likewise require careful separation of assent, court entry, credited redress, reimbursement and actual payment. A12
\nThe unresolved question is therefore sharper than “is there an appeal?” Does a correction reach the recipient? Does it reach exported copies and enforcement systems? Is a completed case reopened? Is access, money or reputation restored where warranted? The most revealing missing artifact would be an ordinary operational trace connecting those events—not an implausible confession of hidden intent.
\nThe record now supports several connected institutional stories, with specific cross-domain bridges and consequential local powers. It also supports a comparative model: resources sustain capacity; selection and definitions shape what is assessed; assessments can condition access; recipients may add their own restrictions; and correction must cross the same institutional boundaries as the original action. These roles are a way of investigating the systems, not proof that all of them form one coordinated pipeline. A22
\nThe original broad hypothesis of common direction remains open. Evidence of local coordination has strengthened considerably, while evidence of divergence and retained discretion makes a single-controller account less adequate as a default explanation. No acquired chain establishes the broader Fabian/WEF/owner-bank/medical-data coordination suggested earlier. Those remain hypotheses or unresolved leads, not conclusions smuggled in through visual proximity.
\nThe dated stablecoin supplement reaches a similar boundary: BlackRock's reserve management, BNY's custody and Circle's issuance are different functions. The retained check examines actual control surfaces but establishes no instruction chain from them to Steam's restrictions. It remains a bounded 9 September check, with no live-key audit or new current-law verification. A15
\nThe next most discriminating evidence lies between the roles already identified: a donor's actual terms and exercised approval; a redacted supplier's real agreement; a classification revision received and acted on by a customer; an application paired with its response and usable data; an assurance report paired with corrective action. That is where the investigation can turn a plausible pattern into a demonstrated mechanism—or rule one out.
\nInspect the claim and evidence index. The full atlas retains detailed relationships that this account cannot reasonably narrate. Source counts describe records, not independent witnesses; several edges can rest on one document. This consolidation reconciles accepted material and selected stale questions. It does not claim every surviving lead has been exhausted.
", "path": "CURRENT-SYNTHESIS.md", "sourceIds": [], "markdown": "Consolidated account of retained research through **9 September 2026**. This account accompanies r33 and two explicitly identified supplements. It is an interpretation of the acquired record, not a fresh audit of every original or a claim that the institutions below share one controller.\n\nThe investigation has moved beyond a list of people who know one another. We can identify recurring ways in which resources become influence: financing particular work, selecting the people who perform it, defining qualifying categories, supplying assessments, and attaching practical consequences to those assessments. An advertisement, a payment relationship, a platform account or access to research data can depend on decisions made across several institutions.\n\nThe most useful unit is **the handoff**. Who supplied an assessment? Under which agreement, rule or duty did someone else act on it? Who retained discretion? What actually happened? Who could reverse it? The record contains real financial and policy connections, along with shared methods and analogous structures. Those are different kinds of connection, and the coherent account depends on keeping them distinct.\n\nOur strongest overall interpretation is that practical power is often distributed across these decision points. A central command is unnecessary for a restrictive result: independently motivated institutions can combine to produce it. That is an explanatory possibility supported by several bounded examples, not proof that every restriction arose this way. Common funders also support divergent positions, regulators decline proposals, distributors exit relationships, and courts or regulators impose remedies. Those observations constrain the stronger theory of a unified programme.\n\n## AI advocacy: capacity, selection and editorial authority\n\nThe original question remains important. The retained record shows substantial organised resources behind AI advocacy and several concrete routes into creator production. FLI's 2025 standalone audit reports **US$45,197,838 in grants from Lightcone Foundation**. Earlier transfers involving Lightcone Foundation and Future of Life Europe concern assets and organisational allocation; they cannot be added together as annual persuasion expenditure. Lightcone Foundation is also a different entity from Lightcone Infrastructure. Scale is real, but the accounting categories matter. [A1](consolidation-wave-2026-09-09-synthesis-claim-index.md#a1)\n\nCreator influence is not one arrangement. People First's published terms provide for approval, revisions, compliance and conditional payment; an individual campaign statement of work was not acquired. Frame describes an independent creator track and an institutional-brief track, names funding sources, and denies funder control over content or selection. A Future Tense creator publicly acknowledged support from FLI, Frame and BlueDot while asserting editorial independence. Rational Animations and ControlAI acknowledged a partnership, but the fee and agreement remain unacquired. The sponsor behind the initial Sabine Hossenfelder account remains unidentified. [A2](consolidation-wave-2026-09-09-synthesis-claim-index.md#a2)\n\nSelection can affect public communication without script approval. Supporting people who already share a concern can increase that concern's reach. That inference requires less than editorial command, but it still cannot identify an undisclosed payment or instruction. ControlAI's acquired UK articles locate formal management and membership authority in its directors. A small related-organisation payable in its US return does not name the counterparty and cannot complete an assumed UK–US funding chain. [A3](consolidation-wave-2026-09-09-synthesis-claim-index.md#a3)\n\nThe Global Call for AI Red Lines supplies a documented campaign-production chain: drafting, management, partner feedback and signatory work have named roles. CeSIA, The Future Society and CHAI are organisers; ControlAI and FLI provided partner feedback. The Call is distinct from ControlAI's **A Narrow Path**. Evan Hubinger's signature does not assign him every proposal advanced by another participant. Advocacy and summit access also remain separate from government adoption. This branch supplies real coordination, with a bounded object and purpose. [A4](consolidation-wave-2026-09-09-synthesis-claim-index.md#a4)\n\n## Shared resources do not settle the policy position\n\nOak's grants make an instructive comparison. Its published 2025 awards include three-year support to Check My Ads, Digital Freedom Fund and Statewatch. Statewatch endorsed objections to broad scanning and possible identity requirements in the EU child-abuse proposal; the WeProtect Secretariat supported mandatory detection with safeguards in a different, earlier version of that debate. The comparison is evidence of divergent interventions, not a same-draft controlled experiment.\n\nDFF reports support for Liberty's challenge concerning the alleged Apple technical-capability notice. Liberty filed in March 2025, before the reported June 2026 grant. Funding could sustain existing litigation; it did not originate that filing. Acquired procedural results also do not determine the merits. A common donor is therefore an important financial connection, but not a sufficient explanation of every recipient's choices. [A5](consolidation-wave-2026-09-09-synthesis-claim-index.md#a5)\n\nDFF's selection reforms further separate donor resources, project selection and award execution. Public descriptions assign project selection to external peers and information/facilitation roles to staff, with confidentiality and conflict commitments. They do not reveal every operative recusal, signature or payment decision. The question “who chooses?” needs an answer at each stage. [A6](consolidation-wave-2026-09-09-synthesis-claim-index.md#a6)\n\n## Policy travels through documents and decisions\n\nThe UK–Australia duty-of-care lane supplies stronger transmission evidence than a shared-event roster. Carnegie-associated work, collaboration with Reset and citations in Australia's statutory review connect a developed model to official consideration. Commissioned research and programme funding sustained parts of that policy infrastructure; later funding cannot be assigned authorship of an earlier idea.\n\nThe Australian exposure draft dated 8 September 2026 would allocate consequential discretion among Parliament's framework, ministerial rules and eSafety. It remains a proposal. Reset requested access for academic and third-sector researchers; the acquired draft instead requires employment by an Australian university and approval under prospective rules. That is a meaningful narrowing, not an awarded NGO entitlement. The private adult-consent provision limits what the duty requires; it does not create a universal adult veto. Nothing inspected establishes universal identification. [A7](consolidation-wave-2026-09-09-synthesis-claim-index.md#a7)\n\nThe UK record also distinguishes required outcomes from particular identity suppliers. Ofcom recognises different age-assurance methods and accepted email-based estimation after supplied evidence; that was a methodological choice, not vendor certification or procurement. Check My Ads later sought customer/identity checks for monetising creators and advertisers and pre-monetisation clearance where risk indicators appear. Those proposals could affect pseudonymity and income, but do not concern every ordinary viewer. Ofcom expressly considered its crisis-revenue request and did not recommend specific monetisation practices in that measure. Other recommendations retain their own unresolved dispositions. [A8](consolidation-wave-2026-09-09-synthesis-claim-index.md#a8)\n\nFull Fact offers unusually precise evidence of input and institutional response. Ofcom attributes a notice-indicator amendment to its submission and questions from an earlier DSIT meeting. A notice becomes an indicator to consider; the provider retains its criteria-based crisis determination. The record supports a specific contribution while leaving the meeting's underlying records and other decisions unacquired. Influence is visible without assuming wholesale control. [A8](consolidation-wave-2026-09-09-synthesis-claim-index.md#a8)\n\nThe newly integrated procurement evidence makes the same distinction financially concrete. CDDP's Oliver Wyman evaluation commission has a disclosed **£656,000 price excluding VAT**, proposed criteria and a Treasury-linked feasibility gate. These establish commissioning and conditions, not payment, the final evaluation or an exercised funding decision. The announced **£8.4 million programme allocation** cannot be added to supplier prices as separate expenditure. Redactions and inconsistent contract dates remain unresolved. [A23](consolidation-wave-2026-09-09-synthesis-claim-index.md#a23)\n\nOSDI had a CDEI-convened advisory group, PA Consulting programme evaluation and supplier participation by Faculty and PUBLIC. These are distinct roles. PUBLIC's account of a TRL4 sandbox describes prototype work; its later STCF2 assessment explicitly credits OSDI practices while retaining limited testing and supplier-reported performance. PUBLIC's claim to have contributed to CDDP does not identify it in a redacted subcontract list. Its current website-controller legal identity is not proof of the historical contracting party. The record now connects programmes, methods and personnel more clearly while preserving the missing agreements and acceptance decisions. [A24](consolidation-wave-2026-09-09-synthesis-claim-index.md#a24)\n\n## A classification matters because someone acts on it\n\nThe Lantern case study reports Meta sharing signals followed by separate investigations and action by Cash App and Xbox. The underlying case files were not inspected; the actions remain operator-reported. Programme rules include recourse, removal of no-longer-relevant information and administrator notification. Nothing in that case establishes that Lantern caused Steam removals. The live question is how information becomes a recipient decision, and what follows when the information changes. [A9](consolidation-wave-2026-09-09-synthesis-claim-index.md#a9)\n\nPayment records make the recipient's role especially clear. Mastercard's retained August 2026 MATCH rules require inquiry but expressly permit onboarding a listed merchant. Stripe publishes a generally disqualifying practice with possible exceptions. The list and a restrictive underwriting decision are connected, but are not identical or a universal prohibition on banking. The 2025 eGumball appellate opinion applies FilmOn's reasoning about audience, purpose and context to confidential commercial reporting: an actual judicial bridge between advertising classification and payment reporting. Its procedural decision does not adjudicate the underlying report's truth or restore a merchant. [A11](consolidation-wave-2026-09-09-synthesis-claim-index.md#a11)\n\nOther infrastructure gates allocate authority differently. Nominet's documented IWF cooperation predates the 2021 UK Safer Internet Centre grant, weakening the claim that the grant created that relationship. PIPCU combines public funding and police assessment with private intelligence; its announced industry co-funding pilot starts in April 2027 and does not establish current contributor command. Operation Creative's dated procedure allocates police tactical judgment, while advertising recipients choose how to implement information. Broader KYBC requests sought identification and consequences for commercial customers; requested reach across hosting and payments is not enacted universal user-ID law. [A13](consolidation-wave-2026-09-09-synthesis-claim-index.md#a13)\n\nAdvertising classification adds a question about scale: is the object a page, channel or whole domain? February 2025 Senate letters questioned financing a website even where advertisements avoided prohibited material. DoubleVerify subsequently announced default Highly Illicit and P2P Sharing and Streaming categories and later reported implementation. The chronology does not establish that the letters uniquely caused the change. NCMEC notifications, DV's assessment and its later IWF relationship are different inputs. Announced defaults do not constitute our observation of customer settings, and general customer-choice language does not prove an override in every configuration. [A14](consolidation-wave-2026-09-09-synthesis-claim-index.md#a14)\n\nCounterpressure belongs in this account. TAG reported no violations of the investigated brand-safety guidelines: its determination, not independent exoneration. DV's complaint challenged critics' assertions: allegations, not judicial findings. Its disclosed fee model does not establish a bounty for blacklisting. The consequences can be significant without a proven private motive or a single upstream instruction.\n\nGDI illustrates why time matters. Its advertising-risk data had a documented Oracle distribution route, but Oracle's relevant service ended in September 2024. IAS later said it discontinued limited past GDI inputs. Historical relationships cannot establish present delivery. A publisher-review procedure may revise GDI's assessment, but the acquired record does not show automatic propagation to every buyer or an obligation to resume advertising. [A16](consolidation-wave-2026-09-09-synthesis-claim-index.md#a16)\n\n## Research access has its own gatekeepers\n\nGDI's public funding and research work reveal further roles without completing an alleged government blacklist. FCDO described supervised activities outside the United States; the signed grant and commercial-reuse conditions remain unacquired. CORDIS identifies a German GDI entity and a **€385,275 EU contribution** to AI4TRUST, not the whole consortium budget or an assumed UK guarantee. Project documents show keywords and annotations, model-assisted triage and human review. A seed term is not a verdict; a research warning is not an advertising ban. GDI's request for mandatory tool/API compatibility is a bid for an institutional role, not proof of its award. [A17](consolidation-wave-2026-09-09-synthesis-claim-index.md#a17)\n\nThe integrated EU access rules separate three mechanisms. Article 40(4) provides project-specific vetting and a reasoned request through the relevant Digital Services Coordinator. Article 40(12) concerns public-data access. Article 22 trusted-flagger priority concerns notices and is different again. Funding and independence disclosures can form part of eligibility; public overviews need not name every applicant or funder. Neither Horizon funding nor a failed exact-name search establishes whether GDI received platform access. [A18](consolidation-wave-2026-09-09-synthesis-claim-index.md#a18)\n\nAliExpress's June 2025 binding commitments supply concrete procedures and clocks. The company nominates and pays a monitoring trustee; the Commission approves and supervises the role. Paying the trustee does not confer the Commission's approval power, and appointment does not establish delivery of a report. The substantive commitment term and external reporting schedule differ. A later illegal-product fine concerns a separate issue; it is not evidence of breach of the research-access commitments. [A19](consolidation-wave-2026-09-09-synthesis-claim-index.md#a19)\n\n## Assurance is not the same evidence as a delivered result\n\nThe retained FTI audit offers a useful stress test. It gives a positive Article 40(12) opinion while reporting **no approved applications** in the year ending June 2025. It says all examined applications were incomplete or invalid. The record supports a narrower conclusion about assessed readiness. It does not establish either that meaningful access was delivered or that requests were improperly refused. Its period also precedes the monitoring trustee's appointment. [A20](consolidation-wave-2026-09-09-synthesis-claim-index.md#a20)\n\nTwo explanations remain possible: legitimate screening of deficient requests, or an interpretation/cure process that makes an access right difficult to use. Without application files, responses and actual delivery, the public audit cannot decide between them. Likewise, the first monitoring report and the Commission's response were not acquired; absence from the acquired record is not proof of non-delivery.\n\nThe researcher-side material is informative but not an adjudication. The DSA40 Collaboratory collects voluntary reports, assists applicants and advocates support for intermediaries. Its September 2026 tracker headline reports **70 applications**, while its retained table sums to **68**; other subtotals also differ. Its one open AliExpress entry lacks the identity and dates needed to connect it to GDI or the FTI cohort. Organisers' judgments that rejection reasons are invalid are leads, not acquired rulings. Disclosed project and host funding establish resources, not donor control of findings. [A21](consolidation-wave-2026-09-09-synthesis-claim-index.md#a21)\n\nThis is the same evidentiary problem seen in the procurement branch: an evaluator commissioned, a test described, a positive opinion, an accepted application and a useful result are distinct events. Counting one as another can make both success and obstruction appear more certain than the record permits.\n\n## Can correction travel as far as the consequence?\n\nThe retained public ThreatExchange client provides a substantive counterexample to technical permanence. At the inspected commit, deletion markers can remove local records; a stale-checkpoint path calls for clearing and a fresh fetch; opinion structures can represent disagreement. These are static capabilities in particular code paths, not a demonstrated end-to-end deployed execution trace. Lantern's use of hosted ThreatExchange does not prove its participants use that client or version. No account restoration was observed. TCAP's separately inspected hash interface also documents deletion and feedback; a historical Altitude importer is another interface and version. [A10](consolidation-wave-2026-09-09-synthesis-claim-index.md#a10)\n\nThere are also actual obligations constraining recipient discretion. The January 2025 CFPB order requires Block to provide specified restriction/appeal notices and to decide an appeal and change status as appropriate within ten business days. That does not guarantee every appellant restored access or constitute an admission of every finding. It does materially narrow the earlier remedy gap. Acquired July settlement terms likewise require careful separation of assent, court entry, credited redress, reimbursement and actual payment. [A12](consolidation-wave-2026-09-09-synthesis-claim-index.md#a12)\n\nThe unresolved question is therefore sharper than “is there an appeal?” Does a correction reach the recipient? Does it reach exported copies and enforcement systems? Is a completed case reopened? Is access, money or reputation restored where warranted? The most revealing missing artifact would be an ordinary operational trace connecting those events—not an implausible confession of hidden intent.\n\n## The coherent picture—and its remaining breaks\n\nThe record now supports several connected institutional stories, with specific cross-domain bridges and consequential local powers. It also supports a comparative model: resources sustain capacity; selection and definitions shape what is assessed; assessments can condition access; recipients may add their own restrictions; and correction must cross the same institutional boundaries as the original action. These roles are a way of investigating the systems, not proof that all of them form one coordinated pipeline. [A22](consolidation-wave-2026-09-09-synthesis-claim-index.md#a22)\n\nThe original broad hypothesis of common direction remains open. Evidence of local coordination has strengthened considerably, while evidence of divergence and retained discretion makes a single-controller account less adequate as a default explanation. No acquired chain establishes the broader Fabian/WEF/owner-bank/medical-data coordination suggested earlier. Those remain hypotheses or unresolved leads, not conclusions smuggled in through visual proximity.\n\nThe dated stablecoin supplement reaches a similar boundary: BlackRock's reserve management, BNY's custody and Circle's issuance are different functions. The retained check examines actual control surfaces but establishes no instruction chain from them to Steam's restrictions. It remains a bounded 9 September check, with no live-key audit or new current-law verification. [A15](consolidation-wave-2026-09-09-synthesis-claim-index.md#a15)\n\nThe next most discriminating evidence lies between the roles already identified: a donor's actual terms and exercised approval; a redacted supplier's real agreement; a classification revision received and acted on by a customer; an application paired with its response and usable data; an assurance report paired with corrective action. That is where the investigation can turn a plausible pattern into a demonstrated mechanism—or rule one out.\n\n[Inspect the claim and evidence index](consolidation-wave-2026-09-09-synthesis-claim-index.md). The full atlas retains detailed relationships that this account cannot reasonably narrate. Source counts describe records, not independent witnesses; several edges can rest on one document. This consolidation reconciles accepted material and selected stale questions. It does not claim every surviving lead has been exhausted.\n", "file": "research/current-synthesis.md" }, { "id": "discord-age-assurance-verification-2026-09-08", "title": "Discord age assurance: transcript verification", "description": "The video identifies a real expansion of age-gated access and real financial/digital-identity infrastructure. It overstates the evidence when it describes universal identity uploads or treats a newly completed global rollout as established. The implementation's trust boundaries matter: Discord's age decision, third-party evidence processing, upstream credent", "html": "The video identifies a real expansion of age-gated access and real financial/digital-identity infrastructure. It overstates the evidence when it describes universal identity uploads or treats a newly completed global rollout as established. The implementation's trust boundaries matter: Discord's age decision, third-party evidence processing, upstream credential creation, and old support-ticket retention are separate data surfaces.
\nDiscord announced global teen defaults on 9 February 2026, initially for March. The announcement now carries a 24 February correction postponing global age assurance to the second half of 2026. It explicitly excludes universal ID uploads/face scans and predicts more than 90% need no additional action. The original announced restrictions include adult-only Stage speaking. Narrow searches of current official sources found no September 8 relaunch notice: the latest central help page inspected (updated August 10) still describes delayed broader global expansion, and limited June–August experiments in Australia/Brazil, with Google Wallet also tested in the UK. This is evidence of a continuing plan and tests, not proof nothing changed in individual accounts after the page update.
\nSources: http[local research file] (February 9; February 24 correction, Teen Safety Settings and clarification); http[local research file] (February 24; March 9 Brazil update); http[local research file] (August 10 update, opening note).
\nDiscord's February 24 FAQ describes internal age inference from account tenure, payment method on file and activity patterns, and says message contents are excluded. The 90% figure is a company projection, not an independently measured result. http[local research file]
\nThe current UK page (July 28 update) covers every new/existing UK account. Unconfirmed users retain accounts, ordinary servers, friends, DMs and voice. They cannot access age-restricted channels/servers, enable age-restricted app commands or NSFW-server access on iOS, unblur flagged media, select Show in sensitive-content filters, or disable Message Requests. Flagged media is blurred with friends/in servers and blocked in stranger DMs. Before accepting a message request, attachments are unavailable and links inert; accepting it permits normal communication subject to existing permissions. Large-server activity visibility can be changed without age assurance. Friend-request warnings cannot currently be disabled. This is not a ban on ordinary voice or all DMs. Stage-speaking restrictions are explicitly part of the announced global package and central help trigger list; the UK page does not separately list them.
\nSource: http[local research file] (Default Safety Settings and FAQ); Stage: global announcement above.
\nThe August 10 central help page says:
\n| Route | \nProcessing / what Discord receives | \nStatus | \n
|---|---|---|
| k-ID facial age estimation | \nVideo remains on device; Discord gets estimated age | \nEstablished method | \n
| ID scan plus matching selfie | \nDocument path names k-ID and Veratad; Discord says it never sees images and receives age; images deleted after confirmation | \nEstablished method | \n
| Credit card | \nk-ID routes to Stripe; card details unseen/unretained by Discord, unretained by k-ID; Discord gets adult yes/no | \nLimited test | \n
| Google Wallet | \nDiscord gets age group and country, other ID information stays in Wallet | \nLimited test | \n
Credit route excludes debit/prepaid; a possible sub-US$0.50 charge is refunded within 14 business days. Central privacy wording qualifies deletion as quick/in most cases immediate, so avoid an absolute universal instantaneous-deletion guarantee. Source: http[local research file] (Choose Your Method, privacy).
\nk-ID independently documents Stripe credit-card verification and Veratad ID authentication. Card verification proves adult status by the card-ownership proxy, particularly where credit ownership is adult-restricted; it is not a document-based proof of the person typing. Its wider catalogue includes other methods, which must not be presumed deployed on Discord. http[local research file] (Credit card verification; ID scan verification).
\nk-ID's privacy policy (updated July 10, 2026) specifies that proof data is processed by providers, k-ID receives/stores the verification result, and its on-device FAE provider is Privately. It also describes jurisdiction information, device identifiers/analytics in its wider services, and session retention determined by customers. Thus 'no retained ID image' does not mean 'no retained metadata.' Crucially, its introduction says customer-provided data processed for integrated products is governed by customer contracts; the public policy covers its controller activities. Do not apply every Family Connect or website-analytics category to Discord users without the integration-specific terms. Its generic policy says k-ID itself does not receive proof information, while Discord describes data going directly to k-ID and Veratad: best read k-ID as an orchestrated service path unless a technical integration proves which endpoint receives bytes. http[local research file] (information required for age assurance; sections 2, 5, 9).
\nStripe's general policy allows transaction and verification data processing and retention after completed transactions for fraud, legal and financial obligations. Neither the Discord nor k-ID page establishes Stripe's exact retention/configuration for this integration. A refund is not evidence of data deletion. http[local research file] and http[local research file] (Security and retention).
\nGoogle's ID-pass setup requires a supported passport, its information-page and NFC-chip scans, and a face video. Automated biometric comparison or human review may be used; biometric templates are deleted following verification. The resulting pass is encrypted locally; the user reviews app-requested information before agreeing to share it. Therefore a minimal age signal to Discord can still depend on prior identity proofing at Google. http[local research file] (Create an ID pass; automated face matching; Use your ID pass; Manage your ID pass).
\nDocument-version trap: search returned a cached June 23 variant of Discord's central page saying Google shares birthdate transiently with Discord and listing an Incode test. The live canonical August 10 page instead says age group + country and omits Incode. Use the live version for current claims; do not silently mix them or infer the engineering change/date from wording alone.
\nDiscord's October 3 statement, updated October 9, says approximately 70,000 users MAY have had government-ID photos exposed through the customer-service ticketing incident; these related to age appeals. It also lists support communications, contact details, IP addresses and limited billing, excluding full card numbers and general chats. This does not establish 70,000 selfies, that every image was exfiltrated, or that all were publicly distributed. http[local research file]
\nDiscord calls it compromise of support vendor 5CA. 5CA's October 14 statement (updated October 27) disputes that its own systems were breached and says one employee's actions enabled access to the client's third-party ticketing system; it also denies handling government IDs. Preserve that disagreement rather than resolve it by choosing one company's wording. Both accounts support a support-access failure; they do not identify k-ID/Stripe as breached. http[local research file]
\nThe substantive pattern is attribute-gated participation relying on several credential providers, rather than evidence that every platform directly builds a passport database. Card infrastructure can authorize access without being the content-moderation authority. Minimal disclosure reduces one exposure while preserving dependency on upstream proofing, eligibility rules and authentication. The old appeals channel demonstrates that exception handling and support permissions may retain identity exposure even when the normal verification flow advertises minimization.
\nUnknown: September account-by-account rollout coverage; Discord-specific Stripe retention and metadata linkage; implemented Wallet disclosure at the time of an individual test; audit proof of deletion/isolation; precise resolution of the 5CA disagreement. Neither a data-sale motive nor a common NGO/funder directing these Discord changes is established by these records. Discord/k-ID expressly deny selling the relevant user data; those are policy representations, not a forensic audit.
", "path": "discord-age-assurance-verification-2026-09-08.md", "sourceIds": [ "s2-discord-methods", "s2-kid-methods", "s2-kid-privacy", "s2-wallet", "s2-discord-incident", "s2-5ca-response" ], "markdown": "## Result\n\nThe video identifies a real expansion of age-gated access and real financial/digital-identity infrastructure. It overstates the evidence when it describes universal identity uploads or treats a newly completed global rollout as established. The implementation's trust boundaries matter: Discord's age decision, third-party evidence processing, upstream credential creation, and old support-ticket retention are separate data surfaces.\n\n## Rollout versus tests\n\nDiscord announced global teen defaults on 9 February 2026, initially for March. The announcement now carries a 24 February correction postponing global age assurance to the second half of 2026. It explicitly excludes universal ID uploads/face scans and predicts more than 90% need no additional action. The original announced restrictions include adult-only Stage speaking. Narrow searches of current official sources found no September 8 relaunch notice: the latest central help page inspected (updated August 10) still describes delayed broader global expansion, and limited June–August experiments in Australia/Brazil, with Google Wallet also tested in the UK. This is evidence of a continuing plan and tests, not proof nothing changed in individual accounts after the page update.\n\nSources: http[local research file] (February 9; February 24 correction, Teen Safety Settings and clarification); http[local research file] (February 24; March 9 Brazil update); http[local research file] (August 10 update, opening note).\n\nDiscord's February 24 FAQ describes internal age inference from account tenure, payment method on file and activity patterns, and says message contents are excluded. The 90% figure is a company projection, not an independently measured result. http[local research file]\n\n## What is restricted\n\nThe current UK page (July 28 update) covers every new/existing UK account. Unconfirmed users retain accounts, ordinary servers, friends, DMs and voice. They cannot access age-restricted channels/servers, enable age-restricted app commands or NSFW-server access on iOS, unblur flagged media, select Show in sensitive-content filters, or disable Message Requests. Flagged media is blurred with friends/in servers and blocked in stranger DMs. Before accepting a message request, attachments are unavailable and links inert; accepting it permits normal communication subject to existing permissions. Large-server activity visibility can be changed without age assurance. Friend-request warnings cannot currently be disabled. This is not a ban on ordinary voice or all DMs. Stage-speaking restrictions are explicitly part of the announced global package and central help trigger list; the UK page does not separately list them.\n\nSource: http[local research file] (Default Safety Settings and FAQ); Stage: global announcement above.\n\n## Methods and data boundaries\n\nThe August 10 central help page says:\n\n| Route | Processing / what Discord receives | Status |\n| --- | --- | --- |\n| k-ID facial age estimation | Video remains on device; Discord gets estimated age | Established method |\n| ID scan plus matching selfie | Document path names k-ID and Veratad; Discord says it never sees images and receives age; images deleted after confirmation | Established method |\n| Credit card | k-ID routes to Stripe; card details unseen/unretained by Discord, unretained by k-ID; Discord gets adult yes/no | Limited test |\n| Google Wallet | Discord gets age group and country, other ID information stays in Wallet | Limited test |\n\nCredit route excludes debit/prepaid; a possible sub-US$0.50 charge is refunded within 14 business days. Central privacy wording qualifies deletion as quick/in most cases immediate, so avoid an absolute universal instantaneous-deletion guarantee. Source: http[local research file] (Choose Your Method, privacy).\n\nk-ID independently documents Stripe credit-card verification and Veratad ID authentication. Card verification proves adult status by the card-ownership proxy, particularly where credit ownership is adult-restricted; it is not a document-based proof of the person typing. Its wider catalogue includes other methods, which must not be presumed deployed on Discord. http[local research file] (Credit card verification; ID scan verification).\n\nk-ID's privacy policy (updated July 10, 2026) specifies that proof data is processed by providers, k-ID receives/stores the verification result, and its on-device FAE provider is Privately. It also describes jurisdiction information, device identifiers/analytics in its wider services, and session retention determined by customers. Thus 'no retained ID image' does not mean 'no retained metadata.' Crucially, its introduction says customer-provided data processed for integrated products is governed by customer contracts; the public policy covers its controller activities. Do not apply every Family Connect or website-analytics category to Discord users without the integration-specific terms. Its generic policy says k-ID itself does not receive proof information, while Discord describes data going directly to k-ID and Veratad: best read k-ID as an orchestrated service path unless a technical integration proves which endpoint receives bytes. http[local research file] (information required for age assurance; sections 2, 5, 9).\n\nStripe's general policy allows transaction and verification data processing and retention after completed transactions for fraud, legal and financial obligations. Neither the Discord nor k-ID page establishes Stripe's exact retention/configuration for this integration. A refund is not evidence of data deletion. http[local research file] and http[local research file] (Security and retention).\n\nGoogle's ID-pass setup requires a supported passport, its information-page and NFC-chip scans, and a face video. Automated biometric comparison or human review may be used; biometric templates are deleted following verification. The resulting pass is encrypted locally; the user reviews app-requested information before agreeing to share it. Therefore a minimal age signal to Discord can still depend on prior identity proofing at Google. http[local research file] (Create an ID pass; automated face matching; Use your ID pass; Manage your ID pass).\n\nDocument-version trap: search returned a cached June 23 variant of Discord's central page saying Google shares birthdate transiently with Discord and listing an Incode test. The live canonical August 10 page instead says age group + country and omits Incode. Use the live version for current claims; do not silently mix them or infer the engineering change/date from wording alone.\n\n## The 2025 breach\n\nDiscord's October 3 statement, updated October 9, says approximately 70,000 users MAY have had government-ID photos exposed through the customer-service ticketing incident; these related to age appeals. It also lists support communications, contact details, IP addresses and limited billing, excluding full card numbers and general chats. This does not establish 70,000 selfies, that every image was exfiltrated, or that all were publicly distributed. http[local research file]\n\nDiscord calls it compromise of support vendor 5CA. 5CA's October 14 statement (updated October 27) disputes that its own systems were breached and says one employee's actions enabled access to the client's third-party ticketing system; it also denies handling government IDs. Preserve that disagreement rather than resolve it by choosing one company's wording. Both accounts support a support-access failure; they do not identify k-ID/Stripe as breached. http[local research file]\n\n## Structural implications and unresolved checks\n\nThe substantive pattern is attribute-gated participation relying on several credential providers, rather than evidence that every platform directly builds a passport database. Card infrastructure can authorize access without being the content-moderation authority. Minimal disclosure reduces one exposure while preserving dependency on upstream proofing, eligibility rules and authentication. The old appeals channel demonstrates that exception handling and support permissions may retain identity exposure even when the normal verification flow advertises minimization.\n\nUnknown: September account-by-account rollout coverage; Discord-specific Stripe retention and metadata linkage; implemented Wallet disclosure at the time of an individual test; audit proof of deletion/isolation; precise resolution of the 5CA disagreement. Neither a data-sale motive nor a common NGO/funder directing these Discord changes is established by these records. Discord/k-ID expressly deny selling the relevant user data; those are policy representations, not a forensic audit.\n", "file": "research/discord-age-assurance-verification-2026-09-08.md" }, { "id": "intake-steam-age-infrastructure-2026-09-08", "title": "Steam transcript intake: age credentials, policy sponsorship, and data custody", "description": "Research date: 2026-09-08. Nova. This is an addendum to the existing Influence Atlas campaign, not a new completed investigation or an explorer data release.", "html": "Research date: 2026-09-08. Nova. This is an addendum to the existing Influence Atlas campaign, not a new completed investigation or an explorer data release.
\nThe useful extension is the relation among payment permission, age-credential supply, and identity-data custody. These are different dependencies with different operators. An app may receive an age band without possessing an identity document; a verifier or support provider may nevertheless retain source documents elsewhere. A remedy at the operating-system layer need not resolve a storefront's separate obligations. The existence of several dependencies does not establish common direction.
\nA newly documented institutional bridge deserves follow-up: ICMEC publicly lists payment companies as donors to its financial coalition, while California's official April 2025 committee analysis identifies ICMEC and Children Now as sponsors of AB1043. ICMEC programme funding, legislative sponsorship, research citation, and software implementation must retain their separate relation types. The present record does not show that those donations financed the legislation, whitepaper, or Steam restrictions.
\nCalifornia's April 22, 2025 committee analysis, concerning the March 28 version of AB1043, names the International Centre for Missing and Exploited Children (ICMEC) and Children Now as sponsors. This is sponsorship of that legislative proposal, not a grant amount or authorship of every eventual provision.\nhttp[local research file]
\nWicks's March 25, 2025 announcement cites a 2021 5Rights report and quotes ICMEC policy-engagement director Bob Cunningham describing continuing work with Wicks and her staff on drafting. This is a specific citation and publicly stated collaboration; it is not proof of 5Rights drafting or funding AB1043.\nhttp[local research file]
\nWicks's September 9, 2025 announcement publishes endorsements by Google, Meta and Internet Works, alongside Children Now and Common Sense Media. Meta's Dan Sachs explicitly supports centralizing age verification within app stores and operating systems. Kareem Ghanem is Google's quoted policy executive. These are dated advocacy statements, not proof of a shared contract or identical commercial interests.\nhttp[local research file]
\nICMEC's supporters page identifies Mastercard, PayPal and Visa, among others, as donors to its Financial Coalition Against Child Sexual Exploitation. That subsection supplies neither amounts nor donation dates. A separate general donor list is expressly a $25,000-plus roster as of May 1, 2023; do not describe it as a fresh 2026 donor list. No earmark to the age-assurance initiative is established.\nhttp[local research file]
\nMicrosoft's August 27, 2026 documentation describes Windows age-range and verification-status APIs. Applications receive a bracket rather than a birthdate. The five brackets span under 10 through 18+. Microsoft-account session and registered application identity are requirements. The page says the APIs are not yet enabled and are planned for later 2026; unknown/unavailable responses require fallbacks. This is documented architecture, not evidence that every Windows user currently needs to upload government ID. Who may obtain meaningful signals, acceptable proof methods, refusal, correction and appeal remain consequential implementation questions.\nhttp[local research file]
\nAs checked September 8, AB1856 is enrolled and was presented to the Governor September 4; the official record does not yet show signature. Its text excludes qualifying open-source distributors from the OS-provider definition, preserves an age-declaration interface and app/store signal duties, and restricts unnecessary signal requests and reuse. It does not impose a universal government-ID upload requirement. The separate OS/store definitions mean an OS exemption cannot simply be treated as an exemption for the entire Steam ecosystem; precise application to Valve's mixed distribution remains unresolved. Earlier drafts adding general browser/website duties should not be passed off as the enrolled text.\nhttp[local research file]\nhttp[local research file]\nhttp[local research file]
\nKrebs's September 1 report identifies IDScan.net as the apparent source of a service advertising over 153 million driver-license records. That is not an independently established count of unique people. The alleged year-long ongoing exfiltration is the seller's claim; capture timestamps are not intrusion timestamps. His participant interviews and timestamp comparisons are attributed investigative evidence. His September 2 update records Caesars saying it stopped using VeriScan in February 2025, had no active accounts and did not authorize retention; the vendor told Caesars no impact was expected. A customer-logo roster is not an affected-client list.\nhttp[local research file]
\nIDScan.net's own September 4 notice says an unauthorized party may have accessed/copied information within customer accounts in its cloud. It names possible exposure of names and government-ID numbers, says the investigation continues and confirms cooperation with federal law enforcement. It does not confirm 153 million individuals or a full image inventory. A claim that the company has issued no statement is now outdated.\nhttp[local research file]
\nThe company's May 23, 2022 Planet13 case study describes merchant scanners, six-image capture, centralized VeriScan visitor management and parsed ID data/images passed into Dutchie's point-of-sale system. This documents a commercial data path. It does not establish that Dutchie was breached or that the particular incident arose from online age-assurance mandates.\nhttp[local research file]
\nICMEC's age-assurance project budget and restricted-grant terms could connect or separate its financial-coalition funding from policy work. Legislative sponsor correspondence, versioned drafts and technical-submission records could identify who wrote or altered the operative architecture. Microsoft app-registration terms and verification-provider criteria could establish practical admission and refusal powers. Vendor data-processing agreements, retention schedules, subprocessor lists and deletion audit evidence could establish where an identity check leaves durable copies. IDScan's forensic account inventory and breach notices would clarify actual incident reach.
\nThe video also raises LG telemetry, litigation against Valve, ownership predictions and a broad data-monetization motive. Those were not established by this bounded intake. They remain separate leads, not new factual links in the atlas.
\nPrimary pages and public original reporting were inspected through the web tool. No stolen records or criminal service were accessed. Direct archival capture of the Wicks page failed (web open 502; Python certificate-chain failure); indexed full-page text was available and read. No local raw-page capture is claimed. This intake does not modify the tested explorer or its portable bundle, and does not restart the paused heartbeat. Additional agent findings are appended below when received.
\nThe supplied policy-site PDF is also hosted on ICMEC's official CDN. Agent inspection found byte-identical copies and no named writer, project funder or credited corporate partner. The February 7, 2025 filename is not independently verified publication timing. The proposal calls for device-based verification, local or other secure OS-managed storage, and OS review/approval of website API integrations before age data is shared. Apple, Google and Microsoft are named as needed collaborators rather than proven participants. This supplies a policy-design record concerning API admission authority; it does not prove Microsoft implemented that paper or that its privacy claims have been validated.\nhttp[local research file]\nhttp[local research file]
\nThe legal operator of AgeVerificationPolicy.org remains unresolved. Its displayed AV Policy Source copyright and links to ICMEC/Free Speech Coalition do not establish ownership or a formal partnership.
\nRead discord-age-assurance-verification-2026-09-08.md for the separately sourced review. Discord's August 10 primary methods page names k-ID/Veratad for document checks, k-ID/Stripe for the credit-card experiment and Google Wallet for an age-group/country signal. Credit and Wallet routes remain labelled limited tests; this pass found no official September 8 global-launch confirmation. The October 2025 age-appeals support incident concerns a different custody path and must not be assigned to those verification providers. These are vendor relationships and distinct rollout states, not evidence that Stripe directs Discord policy.
\nRetain these findings as a sourced intake pending the next atlas content revision. Highest-priority new node: ICMEC, with distinct financial-programme donor, legislative-sponsor and technical-proposal relationships. New infrastructure nodes include Windows age-signaling, k-ID, Veratad, Google Wallet and IDScan.net/VeriScan. Never connect IDScan to Discord without a separate source. Priority questions concern programme earmarks, drafting provenance, provider admission, credential refusal/correction, retention and deletion duties. No public release or automated restart occurred.
", "path": "intake-steam-age-infrastructure-2026-09-08.md", "sourceIds": [ "s2-wicks-drafting", "s2-wicks-support", "s2-age-paper-copy", "s2-windows", "s2-idscan-notice", "s2-idscan-casestudy" ], "markdown": "Research date: 2026-09-08. Nova. This is an addendum to the existing Influence Atlas campaign, not a new completed investigation or an explorer data release.\n\n## What changes the model\n\nThe useful extension is the relation among payment permission, age-credential supply, and identity-data custody. These are different dependencies with different operators. An app may receive an age band without possessing an identity document; a verifier or support provider may nevertheless retain source documents elsewhere. A remedy at the operating-system layer need not resolve a storefront's separate obligations. The existence of several dependencies does not establish common direction.\n\nA newly documented institutional bridge deserves follow-up: ICMEC publicly lists payment companies as donors to its financial coalition, while California's official April 2025 committee analysis identifies ICMEC and Children Now as sponsors of AB1043. ICMEC programme funding, legislative sponsorship, research citation, and software implementation must retain their separate relation types. The present record does not show that those donations financed the legislation, whitepaper, or Steam restrictions.\n\n## Primary policy and technical evidence\n\nCalifornia's April 22, 2025 committee analysis, concerning the March 28 version of AB1043, names the International Centre for Missing and Exploited Children (ICMEC) and Children Now as sponsors. This is sponsorship of that legislative proposal, not a grant amount or authorship of every eventual provision.\nhttp[local research file]\n\nWicks's March 25, 2025 announcement cites a 2021 5Rights report and quotes ICMEC policy-engagement director Bob Cunningham describing continuing work with Wicks and her staff on drafting. This is a specific citation and publicly stated collaboration; it is not proof of 5Rights drafting or funding AB1043.\nhttp[local research file]\n\nWicks's September 9, 2025 announcement publishes endorsements by Google, Meta and Internet Works, alongside Children Now and Common Sense Media. Meta's Dan Sachs explicitly supports centralizing age verification within app stores and operating systems. Kareem Ghanem is Google's quoted policy executive. These are dated advocacy statements, not proof of a shared contract or identical commercial interests.\nhttp[local research file]\n\nICMEC's supporters page identifies Mastercard, PayPal and Visa, among others, as donors to its Financial Coalition Against Child Sexual Exploitation. That subsection supplies neither amounts nor donation dates. A separate general donor list is expressly a $25,000-plus roster as of May 1, 2023; do not describe it as a fresh 2026 donor list. No earmark to the age-assurance initiative is established.\nhttp[local research file]\n\nMicrosoft's August 27, 2026 documentation describes Windows age-range and verification-status APIs. Applications receive a bracket rather than a birthdate. The five brackets span under 10 through 18+. Microsoft-account session and registered application identity are requirements. The page says the APIs are not yet enabled and are planned for later 2026; unknown/unavailable responses require fallbacks. This is documented architecture, not evidence that every Windows user currently needs to upload government ID. Who may obtain meaningful signals, acceptable proof methods, refusal, correction and appeal remain consequential implementation questions.\nhttp[local research file]\n\nAs checked September 8, AB1856 is enrolled and was presented to the Governor September 4; the official record does not yet show signature. Its text excludes qualifying open-source distributors from the OS-provider definition, preserves an age-declaration interface and app/store signal duties, and restricts unnecessary signal requests and reuse. It does not impose a universal government-ID upload requirement. The separate OS/store definitions mean an OS exemption cannot simply be treated as an exemption for the entire Steam ecosystem; precise application to Valve's mixed distribution remains unresolved. Earlier drafts adding general browser/website duties should not be passed off as the enrolled text.\nhttp[local research file]\nhttp[local research file]\nhttp[local research file]\n\n## IDScan.net: a distinct custody branch\n\nKrebs's September 1 report identifies IDScan.net as the apparent source of a service advertising over 153 million driver-license records. That is not an independently established count of unique people. The alleged year-long ongoing exfiltration is the seller's claim; capture timestamps are not intrusion timestamps. His participant interviews and timestamp comparisons are attributed investigative evidence. His September 2 update records Caesars saying it stopped using VeriScan in February 2025, had no active accounts and did not authorize retention; the vendor told Caesars no impact was expected. A customer-logo roster is not an affected-client list.\nhttp[local research file]\n\nIDScan.net's own September 4 notice says an unauthorized party may have accessed/copied information within customer accounts in its cloud. It names possible exposure of names and government-ID numbers, says the investigation continues and confirms cooperation with federal law enforcement. It does not confirm 153 million individuals or a full image inventory. A claim that the company has issued no statement is now outdated.\nhttp[local research file]\n\nThe company's May 23, 2022 Planet13 case study describes merchant scanners, six-image capture, centralized VeriScan visitor management and parsed ID data/images passed into Dutchie's point-of-sale system. This documents a commercial data path. It does not establish that Dutchie was breached or that the particular incident arose from online age-assurance mandates.\nhttp[local research file]\n\n## Next records with high explanatory value\n\nICMEC's age-assurance project budget and restricted-grant terms could connect or separate its financial-coalition funding from policy work. Legislative sponsor correspondence, versioned drafts and technical-submission records could identify who wrote or altered the operative architecture. Microsoft app-registration terms and verification-provider criteria could establish practical admission and refusal powers. Vendor data-processing agreements, retention schedules, subprocessor lists and deletion audit evidence could establish where an identity check leaves durable copies. IDScan's forensic account inventory and breach notices would clarify actual incident reach.\n\nThe video also raises LG telemetry, litigation against Valve, ownership predictions and a broad data-monetization motive. Those were not established by this bounded intake. They remain separate leads, not new factual links in the atlas.\n\n## Custody and limits\n\nPrimary pages and public original reporting were inspected through the web tool. No stolen records or criminal service were accessed. Direct archival capture of the Wicks page failed (web open 502; Python certificate-chain failure); indexed full-page text was available and read. No local raw-page capture is claimed. This intake does not modify the tested explorer or its portable bundle, and does not restart the paused heartbeat. Additional agent findings are appended below when received.\n\n## ICMEC technical proposal provenance\n\nThe supplied policy-site PDF is also hosted on ICMEC's official CDN. Agent inspection found byte-identical copies and no named writer, project funder or credited corporate partner. The February 7, 2025 filename is not independently verified publication timing. The proposal calls for device-based verification, local or other secure OS-managed storage, and OS review/approval of website API integrations before age data is shared. Apple, Google and Microsoft are named as needed collaborators rather than proven participants. This supplies a policy-design record concerning API admission authority; it does not prove Microsoft implemented that paper or that its privacy claims have been validated.\nhttp[local research file]\nhttp[local research file]\n\nThe legal operator of AgeVerificationPolicy.org remains unresolved. Its displayed AV Policy Source copyright and links to ICMEC/Free Speech Coalition do not establish ownership or a formal partnership.\n\n## Discord verification addendum\n\nRead discord-age-assurance-verification-2026-09-08.md for the separately sourced review. Discord's August 10 primary methods page names k-ID/Veratad for document checks, k-ID/Stripe for the credit-card experiment and Google Wallet for an age-group/country signal. Credit and Wallet routes remain labelled limited tests; this pass found no official September 8 global-launch confirmation. The October 2025 age-appeals support incident concerns a different custody path and must not be assigned to those verification providers. These are vendor relationships and distinct rollout states, not evidence that Stripe directs Discord policy.\n\n## Integration decision\n\nRetain these findings as a sourced intake pending the next atlas content revision. Highest-priority new node: ICMEC, with distinct financial-programme donor, legislative-sponsor and technical-proposal relationships. New infrastructure nodes include Windows age-signaling, k-ID, Veratad, Google Wallet and IDScan.net/VeriScan. Never connect IDScan to Discord without a separate source. Priority questions concern programme earmarks, drafting provenance, provider admission, credential refusal/correction, retention and deletion duties. No public release or automated restart occurred.\n", "file": "research/intake-steam-age-infrastructure-2026-09-08.md" }, { "id": "stablecoin-control-check-2026-09-09", "title": "Stablecoin, payment and California discussion check", "description": "Checked 9 September 2026. Bounded factual follow-up; no canonical case or reader revision. Earlier campaign records were used to recover the Steam/payment source chain, then relevant documents were checked currently. No outreach or paid access.", "html": "Checked 9 September 2026. Bounded factual follow-up; no canonical case or reader revision. Earlier campaign records were used to recover the Steam/payment source chain, then relevant documents were checked currently. No outreach or paid access.
\nCalifornia: AB1043 is chaptered, operative1January2027; it requires declared age/DOB and bracket signals, not government-ID upload. AB1856 remains enrolled at Governor as of9September; its proposed licence-based OS-provider exclusion is separate from app-store duties. No SteamOS-specific legal determination acquired.\nSources: http[local research file] ; http[local research file] ; http[local research file]
\nCircle:2025 SEC filing reports a four-year March2025 agreement allocating at least90% of U.S.-managed fiat USDC reserves excluding bank deposits to BlackRock products, subject exceptions. BlackRock manages the reserve fund; BNY custodies assets; Circle issues USDC. This does not assign BlackRock the token-blocklisting authority.\nSource: http[local research file] (Partnership with BlackRock; reserve infrastructure).
\nCircle terms21(15) include further transaction restrictions notified by bank/payment partners. Terms13/14 and21 address blocking/prevention and account/redemption consequences. On-chain blocking is expressly subject to separate access-denial policy: ordinary commercial objection to goods is not a separately enumerated trigger. A bank restriction does not by itself establish bank control of the on-chain key. Official EVM code documents privileged blacklist,pause and upgrade roles; no live key audit performed.\nSources: http[local research file] ; http[local research file] ; http[local research file]
\nGENIUS is enacted; Treasury17August2026 identifies18January2027 as expected general effective date. Sections2(16),4(a)(5)-(6) specify lawful-order and blocking capability. Reserve eligibility includes multiple liquid-asset classes, not exclusively Treasury bills. CLARITY is not enacted: Senate8September record schedules cloture for15September. Treasury has publicly identified increased Treasury demand as intended; Treasury repayment obligations remain distinct from token redemption obligations.\nSources: http[local research file] ; http[local research file] ; http[local research file] ; http[local research file] ; http[local research file]
\nSteam: reread recipient-published Valve statement describing Mastercard communications through processors/acquiring banks and brand-rule5.12.7, alongside Mastercard's denial of requiring game restrictions. No named acquiring-bank owner or BlackRock instruction established. Official Steam purchasing FAQ search index contains an acquiring-bank/PayPal currency passage, but root web open and ordinaryGET did not yield readable relevant body; do not upgrade that indexed passage to a full inspected record. No additional route attempted.\nSources: http[local research file] ; http[local research file]
\nThe examined instruments do not establish the proposed wider bank-owner/BlackRock/Cayman/WEF/medical-data instruction chain or a shared origin for Steam lawsuits. These remain unestablished connections in this bounded check, not a finding that no such connection could exist. No political policy preference or recommendation is made.
", "path": "stablecoin-control-check-2026-09-09.md", "sourceIds": [], "markdown": "Checked 9 September 2026. Bounded factual follow-up; no canonical case or reader revision. Earlier campaign records were used to recover the Steam/payment source chain, then relevant documents were checked currently. No outreach or paid access.\n\nCalifornia: AB1043 is chaptered, operative1January2027; it requires declared age/DOB and bracket signals, not government-ID upload. AB1856 remains enrolled at Governor as of9September; its proposed licence-based OS-provider exclusion is separate from app-store duties. No SteamOS-specific legal determination acquired.\nSources: http[local research file] ; http[local research file] ; http[local research file]\n\nCircle:2025 SEC filing reports a four-year March2025 agreement allocating at least90% of U.S.-managed fiat USDC reserves excluding bank deposits to BlackRock products, subject exceptions. BlackRock manages the reserve fund; BNY custodies assets; Circle issues USDC. This does not assign BlackRock the token-blocklisting authority.\nSource: http[local research file] (Partnership with BlackRock; reserve infrastructure).\n\nCircle terms21(15) include further transaction restrictions notified by bank/payment partners. Terms13/14 and21 address blocking/prevention and account/redemption consequences. On-chain blocking is expressly subject to separate access-denial policy: ordinary commercial objection to goods is not a separately enumerated trigger. A bank restriction does not by itself establish bank control of the on-chain key. Official EVM code documents privileged blacklist,pause and upgrade roles; no live key audit performed.\nSources: http[local research file] ; http[local research file] ; http[local research file]\n\nGENIUS is enacted; Treasury17August2026 identifies18January2027 as expected general effective date. Sections2(16),4(a)(5)-(6) specify lawful-order and blocking capability. Reserve eligibility includes multiple liquid-asset classes, not exclusively Treasury bills. CLARITY is not enacted: Senate8September record schedules cloture for15September. Treasury has publicly identified increased Treasury demand as intended; Treasury repayment obligations remain distinct from token redemption obligations.\nSources: http[local research file] ; http[local research file] ; http[local research file] ; http[local research file] ; http[local research file]\n\nSteam: reread recipient-published Valve statement describing Mastercard communications through processors/acquiring banks and brand-rule5.12.7, alongside Mastercard's denial of requiring game restrictions. No named acquiring-bank owner or BlackRock instruction established. Official Steam purchasing FAQ search index contains an acquiring-bank/PayPal currency passage, but root web open and ordinaryGET did not yield readable relevant body; do not upgrade that indexed passage to a full inspected record. No additional route attempted.\nSources: http[local research file] ; http[local research file]\n\nThe examined instruments do not establish the proposed wider bank-owner/BlackRock/Cayman/WEF/medical-data instruction chain or a shared origin for Steam lawsuits. These remain unestablished connections in this bounded check, not a finding that no such connection could exist. No political policy preference or recommendation is made.\n", "file": "research/stablecoin-control-check-2026-09-09.md" }, { "id": "tegmark-lightcone-followup", "title": "Tegmark and Lightcone follow-up — 5 September 2026", "description": "Tegmark is currently identified as FLI founder and chair, and its biography describes ongoing AI research. FLI's current published AI position calls for a frontier-experiment pause and an ASI moratorium of at least15years. The page bears a2024 publication stamp but does not establish when that specific wording was adopted. Treat this as the organization's cu", "html": "Tegmark is currently identified as FLI founder and chair, and its biography describes ongoing AI research. FLI's current published AI position calls for a frontier-experiment pause and an ASI moratorium of at least15years. The page bears a2024 publication stamp but does not establish when that specific wording was adopted. Treat this as the organization's current position, not an independently established technical conclusion or a ban on all AI.\nhttp[local research file]\nhttp[local research file]
\nLightcone Foundation is the Nevada entity EIN92-0636259, formed2022 in the FY2022 IRS-derived profile; tax-exempt ruling June2023. The FY2024 ProPublica extraction reports$174,320,788 net assets, with Tegmark board president, Aguirre secretary and Mihaela Chita-Tegmark treasurer. It is a grantmaking/reserve affiliate, not established as Lightcone Infrastructure/LessWrong. Neither the original Lightcone charter nor complete current filing was recovered.\nhttp[local research file]\nhttp[local research file]
\nImpala's separate Mission Statement field reproduces:\n“TO SUPPORTING THE CHARITABLE MISSION OF THE FUTURE OF LIFE FOUNDATION, PRIMARILY VIA GRANTMAKING.”\nThis is an aggregator's tax-data reproduction, not a directly recovered charter. It names Future of Life FOUNDATION (FLF, EIN88-0926067). Impala has an AI-generated summary nearby and inconsistent fiscal-year labels; do not rely on its summary for new transaction amounts.\nhttp[local research file]
\nFLI's primary2022 return ScheduleA PartVI says its articles were amended December14 2022 to become a supporting organization to FLF. ScheduleO, PDFp51, reproduces amended Article3, limiting its operation to supporting FLF's purposes within charitable rules. This is more than a shared-name association, but the group's ScheduleR cross-control fields do not justify a simple unverified parent/subsidiary tree.\nhttp[local research file]
\nFLI's primary2024 ScheduleR identifies Lightcone as a related grantmaker and records$21.05m received from it. These records support actual affiliation and financial movements while leaving specific grant decision rights and all inter-entity contracts unresolved.\nhttp[local research file]
\nFLF's own site describes actively creating organizations and initiatives, recruiting founders, running fellowships and commissioning prototypes. It focuses on AI tools for collective human agency; current priorities address AI-supported reasoning, collaboration and reliable knowledge. Its approach is selective/proactive rather than an open grants process. This is concrete self-described activity, not demonstrated effectiveness.\nhttp[local research file]
\nA lower-readiness lead: philanthropy.org's FY2022 transcription lists a$10m Lightcone grant to SVCF; Impala labels a matching amount2023. The underlying alternate PDF failed retrieval. Do not connect this to SVCF's2023 CAIS grant without donor-account or onward-allocation evidence.\nhttp[local research file]
\nTegmark's February2 2026 Canadian parliamentary appearance supplies direct evidence of advocacy for mandatory safety testing while favoring beneficial AI tools. Other empirical claims in that testimony were not validated by this follow-up; testimony proves what a witness argued, not the accuracy of every assertion.\nhttp[local research file]
\nInterpretation: distinguish the breadth of a nonprofit's short mission from its current programs, legal relationships, and exercised grantmaking discretion. The reporting focus is who decides which approaches get resources. A broad mandate or ominous-sounding name does not by itself establish improper conduct.
", "path": "tegmark-lightcone-followup.md", "sourceIds": [], "markdown": "Tegmark is currently identified as FLI founder and chair, and its biography describes ongoing AI research. FLI's current published AI position calls for a frontier-experiment pause and an ASI moratorium of at least15years. The page bears a2024 publication stamp but does not establish when that specific wording was adopted. Treat this as the organization's current position, not an independently established technical conclusion or a ban on all AI.\nhttp[local research file]\nhttp[local research file]\n\nLightcone Foundation is the Nevada entity EIN92-0636259, formed2022 in the FY2022 IRS-derived profile; tax-exempt ruling June2023. The FY2024 ProPublica extraction reports$174,320,788 net assets, with Tegmark board president, Aguirre secretary and Mihaela Chita-Tegmark treasurer. It is a grantmaking/reserve affiliate, not established as Lightcone Infrastructure/LessWrong. Neither the original Lightcone charter nor complete current filing was recovered.\nhttp[local research file]\nhttp[local research file]\n\nImpala's separate Mission Statement field reproduces:\n“TO SUPPORTING THE CHARITABLE MISSION OF THE FUTURE OF LIFE FOUNDATION, PRIMARILY VIA GRANTMAKING.”\nThis is an aggregator's tax-data reproduction, not a directly recovered charter. It names Future of Life FOUNDATION (FLF, EIN88-0926067). Impala has an AI-generated summary nearby and inconsistent fiscal-year labels; do not rely on its summary for new transaction amounts.\nhttp[local research file]\n\nFLI's primary2022 return ScheduleA PartVI says its articles were amended December14 2022 to become a supporting organization to FLF. ScheduleO, PDFp51, reproduces amended Article3, limiting its operation to supporting FLF's purposes within charitable rules. This is more than a shared-name association, but the group's ScheduleR cross-control fields do not justify a simple unverified parent/subsidiary tree.\nhttp[local research file]\n\nFLI's primary2024 ScheduleR identifies Lightcone as a related grantmaker and records$21.05m received from it. These records support actual affiliation and financial movements while leaving specific grant decision rights and all inter-entity contracts unresolved.\nhttp[local research file]\n\nFLF's own site describes actively creating organizations and initiatives, recruiting founders, running fellowships and commissioning prototypes. It focuses on AI tools for collective human agency; current priorities address AI-supported reasoning, collaboration and reliable knowledge. Its approach is selective/proactive rather than an open grants process. This is concrete self-described activity, not demonstrated effectiveness.\nhttp[local research file]\n\nA lower-readiness lead: philanthropy.org's FY2022 transcription lists a$10m Lightcone grant to SVCF; Impala labels a matching amount2023. The underlying alternate PDF failed retrieval. Do not connect this to SVCF's2023 CAIS grant without donor-account or onward-allocation evidence.\nhttp[local research file]\n\nTegmark's February2 2026 Canadian parliamentary appearance supplies direct evidence of advocacy for mandatory safety testing while favoring beneficial AI tools. Other empirical claims in that testimony were not validated by this follow-up; testimony proves what a witness argued, not the accuracy of every assertion.\nhttp[local research file]\n\nInterpretation: distinguish the breadth of a nonprofit's short mission from its current programs, legal relationships, and exercised grantmaking discretion. The reporting focus is who decides which approaches get resources. A broad mandate or ominous-sounding name does not by itself establish improper conduct.\n", "file": "research/tegmark-lightcone-followup.md" }, { "id": "ab621-origins-wave-2026-09-09-march-analysis", "title": "AB 621: enforcement experience informs the proposed law", "description": "9 September 2026. Root read all18 pages of the independently published 18 March2025 Assembly Privacy and Consumer Protection analysis(httplocal research file This describes the introduced bill and proposed amendments; it is not the final statute.", "html": "9 September 2026. Root read all18 pages of the independently published [18 March2025 Assembly Privacy and Consumer Protection analysis](http[local research file] This describes the introduced bill and proposed amendments; it is not the final statute.
\nThe analysis identifies San Francisco City Attorney David Chiu as sponsor and Rebecca Bauer-Kahan as author. It connects the proposal to Chiu's2024 litigation against16 websites. Staff explicitly identifies payment processors as potential enabling-service defendants and explains the difficulty of reaching underlying operators. Chiu's reproduced support statement argues for broader facilitator liability, explicit prosecutor standing and larger penalties to strengthen deterrence. These are documented design aims, not inferred motives. See PDF1,7-8,17.
\nThe dated support list names the City Attorney's Office, California District Attorneys Association, Consumer Attorneys of California and Joyful Heart Foundation. California Civil Liberties Advocacy is listed as opposing unless amended, with free-expression, federal-law and service-provider-burden objections. The analysis says the author offered the primary-purpose narrowing and prints its proposed redline. It does not establish which outside party obtained that concession or its later adoption. See PDF12-13,17-18. SAG-AFTRA's sponsorship is attributed to predecessor AB602, not this bill.
\nThe payment-provider rationale is especially consequential for the investigation: the proposed extension reaches a commercial service needed by the underlying operator. Current law supplies a related conditional mechanism, but connecting particular wording to a legislator's decision requires dated amendment evidence. Financial penalties and fee recovery are proposed litigation remedies; no source here establishes funds received, budget dependence or revenue-driven target selection.
\nThe analysis's constitutional discussion, including its forecasts about section230, is staff reasoning as of March2025. Root read it without adopting its forecasts as judgments or freshly verified case law. The present inquiry does not audit every cited harm statistic or underlying lawsuit. Its dated claims about prevalence, studies and institutional practice remain source-attributed.
\nReading combined full official web-reader text and retained PDF extraction; no page omitted. Redline extraction merges deleted and inserted words, so proposed-change conclusions here rely on the explicit explanatory prose. Carver's independent version lane must not use an ignored historical URL parameter as proof of old text. Later committee analysis and current statutory text are distinct sources with distinct dates.
\nOriginal retained: assembly-privacy-march18.pdf,627,224bytes,18pages, SHA256602bea00e4c4bd481ca1bbdf3ae372c11c7289af7e8012fc55eb728f5a4a4aaf. root-captures.json records direct ordinaryGET200. No original support letters, lobbying correspondence, drafting instructions or fiscal allocation records were acquired in this root lane. Acquisition of later independently published analysis continues; this run has not yet reached its forest.
", "path": "ab621-origins-wave-2026-09-09/march-analysis.md", "sourceIds": [ "w25-sf-s-march-analysis" ], "markdown": "9 September 2026. Root read all18 pages of the independently published 18 March2025 Assembly Privacy and Consumer Protection analysis (see source references) This describes the introduced bill and proposed amendments; it is not the final statute.\n\nThe analysis identifies San Francisco City Attorney David Chiu as sponsor and Rebecca Bauer-Kahan as author. It connects the proposal to Chiu's2024 litigation against16 websites. Staff explicitly identifies payment processors as potential enabling-service defendants and explains the difficulty of reaching underlying operators. Chiu's reproduced support statement argues for broader facilitator liability, explicit prosecutor standing and larger penalties to strengthen deterrence. These are documented design aims, not inferred motives. See PDF1,7-8,17.\n\nThe dated support list names the City Attorney's Office, California District Attorneys Association, Consumer Attorneys of California and Joyful Heart Foundation. California Civil Liberties Advocacy is listed as opposing unless amended, with free-expression, federal-law and service-provider-burden objections. The analysis says the author offered the primary-purpose narrowing and prints its proposed redline. It does not establish which outside party obtained that concession or its later adoption. See PDF12-13,17-18. SAG-AFTRA's sponsorship is attributed to predecessor AB602, not this bill.\n\nThe payment-provider rationale is especially consequential for the investigation: the proposed extension reaches a commercial service needed by the underlying operator. Current law supplies a related conditional mechanism, but connecting particular wording to a legislator's decision requires dated amendment evidence. Financial penalties and fee recovery are proposed litigation remedies; no source here establishes funds received, budget dependence or revenue-driven target selection.\n\nThe analysis's constitutional discussion, including its forecasts about section230, is staff reasoning as of March2025. Root read it without adopting its forecasts as judgments or freshly verified case law. The present inquiry does not audit every cited harm statistic or underlying lawsuit. Its dated claims about prevalence, studies and institutional practice remain source-attributed.\n\nReading combined full official web-reader text and retained PDF extraction; no page omitted. Redline extraction merges deleted and inserted words, so proposed-change conclusions here rely on the explicit explanatory prose. Carver's independent version lane must not use an ignored historical URL parameter as proof of old text. Later committee analysis and current statutory text are distinct sources with distinct dates.\n\nOriginal retained: assembly-privacy-march18.pdf,627,224bytes,18pages, SHA256602bea00e4c4bd481ca1bbdf3ae372c11c7289af7e8012fc55eb728f5a4a4aaf. root-captures.json records direct ordinaryGET200. No original support letters, lobbying correspondence, drafting instructions or fiscal allocation records were acquired in this root lane. Acquisition of later independently published analysis continues; this run has not yet reached its forest.\n", "file": "research/ab621-origins-wave-2026-09-09-march-analysis.md" }, { "id": "ab621-origins-wave-2026-09-09-version-changes", "title": "AB 621 version comparison — bounded access result", "description": "Research cutoff: 9 September 2026. No historical textual delta was established by this lane. The official version selector and bill history were verified, and the enacted baseline was already read in full. The attempted introduced-version download returned the chaptered version. Historical browser selection then encountered an explicit administrator-policy r", "html": "Research cutoff: 9 September 2026. No historical textual delta was established by this lane. The official version selector and bill history were verified, and the enacted baseline was already read in full. The attempted introduced-version download returned the chaptered version. Historical browser selection then encountered an explicit administrator-policy refusal. Live LegInfo access is closed; this note does not convert index labels or a proposed committee redline into an adopted bill version.
\nThe actual selector in the retained [official chaptered-bill page](http[local research file] lists:
\n| Selector date and stage | \nExact option value | \nSubstantive text state in this lane | \n
|---|---|---|
| 13 February 2025 — Introduced | \n20250AB62199INT | \nNot acquired; attempted request returned v93 chaptered text. | \n
| 24 March 2025 — Amended Assembly | \n20250AB62198AMD | \nNot read. | \n
| 9 April 2025 — Amended Assembly | \n20250AB62197AMD | \nNot read. | \n
| 3 July 2025 — Amended Senate | \n20250AB62196AMD | \nNot read. | \n
| 5 September 2025 — Amended Senate | \n20250AB62195AMD | \nNot read. | \n
| 12 September 2025 — Enrolled | \n20250AB62194ENR | \nNot read. | \n
| 13 October 2025 — Chaptered | \n20250AB62193CHP | \nComplete operative text read in the prior statutory lane; reused baseline, not a new historical finding. | \n
The retained [official history](http[local research file] corroborates the amendment dates and records approval/chaptering on 13 October 2025. It separately records enrollment and presentation to the Governor on 22 September; that event must not be substituted for the selector’s 12 September enrolled-text date. The index/history evidence establishes dated publication or procedural stages, not their substantive changes.
\nThe enacted baseline contains the primary-purpose definition at 1708.86(a)(3); substantive third-party knowing-facilitation/reckless-aiding liability at (b)(3); the conditional presumption, sufficient evidence, required particulars and delivery route at (c)(2)(A); receipt-based 30 days and court extension at (c)(2)(B); federal protection at (i); and specific ISP boundaries at (j). This lane cannot say when those words first entered or changed. The full enacted scope remains in statutory-mechanism.md. Root’s independently published March 18 committee proposal/redline is a separate primary artifact: it can establish what was offered or explained at that point, but without an actual later version it does not independently establish adoption on March 24.
\nThe selector values were taken from the actual retained HTML control, not invented. An ordinary request to http[local research file] returned HTTP 200, but the document’s hidden vid was 20250AB62193CHP, and its body identified AB621:v93#DOCUMENT, Chapter 673, approved/filed 13 October 2025. It therefore did not return the requested introduced text.
Those bytes are retained as version-query-returned-chaptered.html, SHA-256 66c94ef928832a7373a41a203dd41584072523e273efb516b29f030dc180633f, and explicitly excluded as historical-version evidence. The derivative is version-query-returned-chaptered.txt. Neither is represented as an introduced bill, and the HTTP success code is not treated as successful version acquisition. Earlier requested-version filenames were renamed to avoid that implication. Before this direct request, the web reader returned a non-retryable URL-opening error for the historical query; it did not provide historical text either.
Tool: mcp__cua_repl.js. Action: cua.createBrowserTab("chrome", "http[local research file] { sessionName: "📜 AB 621 versions" }). Target: the official bill’s ordinary version-selection page. The tool returned this exact refusal:
\n\nBrowser Use could not complete this action because a browser security check was unavailable. Reason: The admin-enforced policy could not be verified, so access was not granted. Browser use could not verify the admin-enforced policy before accessing http[local research file] This failure may be temporary. The agent may retry after the issue is resolved, but must not bypass browser security controls or use an indirect workaround.
\n
This is an administrator-policy/browser access refusal, not an automatic sandbox approval-review rejection. No live LegInfo request, mirror, proxy, other-agent acquisition or indirect workaround followed it. The previously failed 10 September Assembly Floor Analysis form was not retried. Already retained local originals remain available for reading; their custody does not establish access to unacquired historical versions.
\nThe reopening condition is restored verified access to the official historical controls, or legitimately provided original historical versions already in lawful custody. A later independent official publication can be assessed on its own provenance; it must not be sought as a proxy around this refusal. Root owns any next acquisition decision after the forest review.
\nMetadata, exact selector values and the refusal are in version-captures.json. Selector source is the prior official sf-notice-wave-2026-09-09/statutory-chaptered-ab621.html, SHA-256 9c3f9f8bd1019fd5c941781001d3bc2083745a3c09c4546768172c6e3d91d3e8; history source SHA-256 4467ecb0b50d8942d1bcc9366f3015f524f38d4ef8767f844839bf0ca42ea40a. No organization is credited with winning an amendment on the basis of timing, and no canonical case or other-law branch was opened.
Bounded acquisition closed, 9 September 2026. One new substantive primary original was acquired and read. Two final deliverables were located through the official results index but their download service returned technical errors. No dataset, model or live platform was accessed. Existing D2.1 was neither reacquired nor reread in this lane; root owns its selected annex reading.
\nThe [CORDIS final-period report](http[local research file] covering 1 March 2024–28 February 2026 and last updated 3 August 2026, reports that the final Disinformation Warning System combines outputs from six AI tools, using an ensemble of LightGBM classifiers with isotonic regression calibration, producing triage categories and feature attribution. This is a considerably more specific mechanism than an unspecified risk score. Locator: “Work performed,” second reporting period, “Disinformation Warning System (WP3)”; official-reader line 141.
\nThe same report describes an enhanced Human Validation Dashboard and iterative user evaluation. It reports three evaluation rounds involving 206 users, including media professionals, fact-checkers, policy experts and students. These are project-reported results hosted by the Commission, not this investigation's reproduced performance or an independent Commission audit. Locators: WP5 and WP6, lines 143–144.
\nIts approximately 56% weighted F1 concerns the separate textual disinformation-signals tool, which detects 39 signals across six tactics and eight languages. It is not the final warning system's accuracy or a probability that a particular item is false. Locator: WP3 textual component, line 137. No model or benchmark was run here.
\nCollection, annotation, inference and action cannot be collapsed into one decision. The report separately describes data collection and ground-truth resources, component analyses, an aggregate warning system, human validation, and intended reporting to media professionals and policymakers. Its collection counts are not counts of adjudicated misinformation. Its component metrics are not measures of subsequent journalistic decisions or advertising restrictions.
\nThe published ensemble/calibration description does not disclose, in this read record, the six selected inputs, exact model weights, thresholds or triage-category definitions; who approved them; the training/validation split for the aggregate system; or how competing human assessments alter a score. “Human Validation Dashboard” is a named component, not an acquired override policy. Nor does a configurable interface establish authority to change model thresholds. The note therefore does not assert a reviewer veto, automatic score replacement, publisher appeal, correction propagation or final action right.
\nThe report identifies intended users and a research evaluation population. It does not provide an actual recipient distribution list, an executed downstream-use agreement, or an advertising-system interface. The strongest supported interpretation remains a research triage system intended to assist people. A route from its score to an ad ban, account restriction or government enforcement decision is not established by these records. That limit does not prove no subsequent use exists.
\nThe official [project results index](http[local research file] last updated 17 June 2026, exposes these two public records:
\n| Indexed record | \nActual official module | \nAccess result | \n
|---|---|---|
| Final release of disinformation detection tools and warning system | \n[Document 080166e529e5a48d](http[local research file] | \nModule HTML, followed through its own meta-refresh to technical error HDS-010; no deliverable bytes acquired. | \n
| Evaluation and impact assessment report | \n[Document 080166e529e59930](http[local research file] | \nSame technical error, with a distinct document identifier; no deliverable bytes acquired. | \n
The download HTML explicitly supplied the next URL. Following that ordinary redirect was not a bypass. Both returned-error originals and their module originals are retained. Their HTTP 200 status does not turn the error pages into the requested reports. No retry after HDS-010, alternate mirror, login or helpdesk contact followed.
\nThe public platform About page was discovered by search at http[local research file], but the reader returned 502 and the one ordinary GET failed verification of an expired certificate. That route was closed without disabling certificate checks. Search-indexed explanatory text remains discovery, not an independently read rule or current interface observation. The main project homepage reader separately returned a non-retryable unsafe-open error and was not fetched through another method.
The two indexed final records are the immediate expected instruments for the actual aggregate scoring and evaluation method. The known holders are the project consortium/coordinator and the Commission's public document service. A useful next record would expose the approved score/category thresholds and evaluation protocol, or a reviewer workflow showing how disputed labels and overrides are authorized and recorded. GDI's annotation codebook remains unacquired from the previous run. The public final-report narrative does not substitute for any of these instruments.
\nThis run produces a concrete method-level finding and a bounded document-access gap. No further source search, graph update, operational experiment or follow-on branch was undertaken.
\ndss-reporting.html and dss-reporting.derived.txt: full substantive final-period report read, from reporting-period heading through uptake needs, plus date field. Official-reader equivalents: lines 119–159. Original-source header was inspected only for record identity; no new grant-accounting conclusions are drawn. This is a current snapshot of a historical reporting period, not a claim that a live service is now operating.
dss-index.reader.txt: retained official-reader extract of deliverable listings, lines 119–199. Discovery index only; no full publication-list audit. dss-final-module.html, dss-evaluation-module.html, and the two *.response.html files retain the actual download wrappers/errors. dss-transport.json records URLs, timestamps, status and checksums. Exact hashes, source states and closed routes are in dss-captures.json.
9 September 2026. This lane reuses the accepted AI4TRUST D2.1 original in ../gdi-horizon-role-wave-2026-09-09/horizon-d21.pdf, SHA256 1222963ce874a0455e10109cafff95ad201e1a43c1d6be6cfdd1848f9a2f2dfe. No new independent source or download. Current classification-run root text scope: PDF4-5 contents,53-59 and81-82. Earlier Horizon root text scope was2,3,11,12,34,35,55,65,68,69 and visual2,11,55. These are selected pages of90, not a complete annex audit.
\nPDF53-54 makes climate, health and migration initial development/test topics, with reusable selection criteria including cross-country relevance, historicity, factual assessability and consortium feasibility. It acknowledges that misinformation need not be a coordinated campaign. Its feasibility discussion is a dated2023 design account, not a current exclusion decision. It points to D2.2 for active topic-mining elaboration.
\nPDF55 combines research/media/factchecker inputs with GDI's existing internal lists and some ChatGPT suggestions. Broad initial keyword lists are to be refined by removing irrelevant terms and adding co-occurring terms. There is no per-term provenance: individual entries cannot all be attributed to GDI. These are collection seeds, not a website-risk codebook or an automatic sanction list.
\nPDF56-58 includes neutral subjects such as weather, renewable energy, clinical trials, adverse events, immigration law and migrant rights alongside charged narratives, public figures and conspiracy language. Inclusion does not label a named person or proposition false. Duplicates and mixed capitalization occur; this document does not prove runtime configuration. The migration notes permit national or origin terms to vary by country and period. No bulk targeting list is reproduced here.
\nPDF58-59 explicitly declares an intersectional methodological framing attentive to harm, discrimination and social identity. That is an observable normative choice. Its effect on annotation labels or thresholds still needs a rule-level record; the framing alone does not prove suppression. PDF81-82 shows German annex terms are locally adapted, rather than a literal English copy. Selected pages do not identify translator approval or update authority. They do not show German GDI authored every German term. French tail/Greek opening on these pages were contextual reading, not complete language-list review.
\nThe revealing distinction is collection perimeter versus treatment rule. Broad search terms can shape a sample, but a bias finding requires inspecting resulting coverage, annotation decisions and validation. The final report separately describes aggregate model scoring and human validation. Sampling, annotation, inference, human judgment and downstream action stay separate in this case. Actual codebook, aggregate thresholds, override rules and agreements remain unacquired. No person/site score audit or model execution occurred.
", "path": "ai4trust-classification-decisions-wave-2026-09-09/root-seed-reading.md", "sourceIds": [], "markdown": "9 September 2026. This lane reuses the accepted AI4TRUST D2.1 original in ../gdi-horizon-role-wave-2026-09-09/horizon-d21.pdf, SHA256 1222963ce874a0455e10109cafff95ad201e1a43c1d6be6cfdd1848f9a2f2dfe. No new independent source or download. Current classification-run root text scope: PDF4-5 contents,53-59 and81-82. Earlier Horizon root text scope was2,3,11,12,34,35,55,65,68,69 and visual2,11,55. These are selected pages of90, not a complete annex audit.\n\nPDF53-54 makes climate, health and migration initial development/test topics, with reusable selection criteria including cross-country relevance, historicity, factual assessability and consortium feasibility. It acknowledges that misinformation need not be a coordinated campaign. Its feasibility discussion is a dated2023 design account, not a current exclusion decision. It points to D2.2 for active topic-mining elaboration.\n\nPDF55 combines research/media/factchecker inputs with GDI's existing internal lists and some ChatGPT suggestions. Broad initial keyword lists are to be refined by removing irrelevant terms and adding co-occurring terms. There is no per-term provenance: individual entries cannot all be attributed to GDI. These are collection seeds, not a website-risk codebook or an automatic sanction list.\n\nPDF56-58 includes neutral subjects such as weather, renewable energy, clinical trials, adverse events, immigration law and migrant rights alongside charged narratives, public figures and conspiracy language. Inclusion does not label a named person or proposition false. Duplicates and mixed capitalization occur; this document does not prove runtime configuration. The migration notes permit national or origin terms to vary by country and period. No bulk targeting list is reproduced here.\n\nPDF58-59 explicitly declares an intersectional methodological framing attentive to harm, discrimination and social identity. That is an observable normative choice. Its effect on annotation labels or thresholds still needs a rule-level record; the framing alone does not prove suppression. PDF81-82 shows German annex terms are locally adapted, rather than a literal English copy. Selected pages do not identify translator approval or update authority. They do not show German GDI authored every German term. French tail/Greek opening on these pages were contextual reading, not complete language-list review.\n\nThe revealing distinction is collection perimeter versus treatment rule. Broad search terms can shape a sample, but a bias finding requires inspecting resulting coverage, annotation decisions and validation. The final report separately describes aggregate model scoring and human validation. Sampling, annotation, inference, human judgment and downstream action stay separate in this case. Actual codebook, aggregate thresholds, override rules and agreements remain unacquired. No person/site score audit or model execution occurred.\n", "file": "research/ai4trust-classification-decisions-wave-2026-09-09-root-seed-reading.md" }, { "id": "ali-oversight-evidence-wave-2026-09-09-oversight-findings", "title": "AliExpress trustee oversight: the public reporting frontier", "description": "Closed 9 September 2026. This lane sought the actual approved mandate, the first monitoring trustee report due in November 2025, or a Commission response, including researcher-access results and evidence sharing with the Article 37 auditor. No such case-specific substantive record was acquired. The two new HTML captures document the public routes inspected; ", "html": "Closed 9 September 2026. This lane sought the actual approved mandate, the first monitoring trustee report due in November 2025, or a Commission response, including researcher-access results and evidence sharing with the Article 37 auditor. No such case-specific substantive record was acquired. The two new HTML captures document the public routes inspected; neither is a completed monitoring report. The adopted decision and appointment notice are reused originals from the preceding accepted run.
\nThe defensible result is a defined disclosure gap, not a finding that Ankura failed to report, that the Commission did nothing, or that AliExpress complied. The 2026 cycle ends on 30 September 2026 and its report is due within two months thereafter; it is not yet due at this research snapshot.
\nThe Commission's [supervision index](http[local research file] which displays information updated on 31 August 2026, was read specifically at the AliExpress entry. Its 18 June 2025 row links the commitments narrative, monitoring trustee notice, adopted decision and press release. There is no separately labelled mandate, work plan, completed trustee report or Commission response in that inspected entry. The later 20 July 2026 fine is also listed, but the accepted preceding packet establishes its separate illegal-product risk assessment and mitigation scope. It cannot fill this researcher-access monitoring gap. This index is an overview of main enforcement activity, not a representation that every case document is published.
\nOriginal oversight-index.html; scoped text oversight-index.derived.txt; exact relevant hrefs retained in oversight-index.links.json; official-reader representation oversight-index.reader.txt, lines 33–64. The links still resolve by identity to the already retained decision and appointment URLs: [decision 124718](http[local research file] and [appointment notice 124727](http[local research file] They were not downloaded again.
Ankura's [Monitorship & Independent Oversight page](http[local research file] describes services and a different, unnamed automotive engagement. Its substantive body does not identify an AliExpress mandate, examination, finding or delivered report. It is retained as an inspected search route and excluded as proof of case-specific work. Original oversight-ankura.html; body/links derivatives and official-reader representation retained. The page provides no exact publication or modification date in the inspected body; the reader's copyright year is not an engagement date. No contact or form was used, and no unrelated case or biography was followed.
The retained [18 June 2025 decision and annex](http[local research file] ../dsa-access-enforcement-wave-2026-09-09/ali-decision.pdf, supplies a precise record chain. This run text-read PDF 75–77, Annex C.5–17; earlier accepted scope for the wider decision is preserved in the prior manifest. No new visual reading is claimed.
| Expected record or act | \nAdopted requirement | \nWhat is established here | \n
|---|---|---|
| Approved trustee mandate | \nC.5, PDF 75: Commission approves the proposed mandate, with modifications it considers necessary; appointment follows that approved mandate. | \nThe separate appointment notice confirms approval of Ankura Consulting (Europe), Limited on 23 September 2025. It does not publish the mandate or disclose its negotiated terms. | \n
| Detailed work plan | \nC.11, PDF 76: trustee proposes its monitoring work plan to the Commission within one month of appointment. | \nRequired submission; actual plan and receipt not acquired. | \n
| First report | \nC.10, PDF 75, and C.12(iv), PDF 76: first report by 30 November 2025 covers the initial period ending 30 September 2025. Initial implementation was on a best-efforts basis; full obligations apply thereafter. | \nDue date and initial assessment standard are established; delivery, tested scope, findings and Commission reaction remain unacquired. | \n
| Researcher-access evidence | \nPrior accepted Annex B.V.7, PDF 66: semiannual material includes request totals, processing/approval/rejection, researcher feedback and technical tests. C.12(v), PDF 76 separately requires the trustee's opinion on the first internal QA semiannual report within one month. | \nThis identifies useful expected evidence. No actual request counts, feedback, test results or trustee opinion was found. The first interim report should not automatically be treated as a full-year or mature-system evaluation. | \n
| Noncompliance response | \nC.12(vi)–(viii), PDF 76–77: trustee may propose necessary measures; must promptly report reasonably concluded failure to the Commission and answer its questions. | \nDuties and permitted response, not an observed finding, recommendation or correction. | \n
The first report goes to the Commission, with a contemporaneous copy to AliExpress that removes confidential third-party material where applicable (C.12(iv)). C.13(iv), PDF 77, restricts disclosure of confidential information and prohibits trustee public statements concerning performance of its functions. Those provisions make nonpublic reporting a plausible explanation for the search result. They neither establish that the whole report is confidential nor determine whether the Commission could release a nonconfidential version. No case-specific confidentiality or access-to-documents refusal was acquired.
\nC.12(iii), PDF 76, requires the trustee to minimise repetitive overlap with the Article 37 auditor. Where information is missing, the trustee may first ask whether the auditor has already collected it, request a copy, and rely on or incorporate the resulting material. AliExpress undertakes reasonably required steps enabling that disclosure. In the other direction, C.14, PDF 77, provides that the auditor will receive access, at its request, to relevant monitoring and reporting documents prepared by the trustee for the Commission.
\nThis is an express two-way evidence route. It does not prove that either party requested documents, shared a particular test or finding, or accepted the other's conclusions. No automatic wholesale publication or transfer to researchers follows. The actual sharing request, document log or auditor reference to the first trustee report would discriminate between a permitted route and its exercise. Provider Article 37 audit discovery belongs to the parallel root lane and was not duplicated here.
\nThirteen targeted indexed queries were run in four batches, combining the exact service, trustee, case number, mandate, first-report date, researcher and Commission/Ankura domain terms. The complete query list and selected returned-path assessment are in oversight-search-log.json. Search returned the existing Commission announcement and case index, generic Ankura services, secondary summaries and unrelated results. It did not yield a new case-specific Commission or Ankura mandate, first report or response that could be read. Search result ranking and indexing are incomplete and can relax terms; this is not a comprehensive archive search.
A public professional profile snippet attributed an appointment announcement to the named trustee, and an adviser firm's snippet described support to Ankura. Neither was opened or acquired, and neither supplied a report or mandate. They remain discovery only; no personnel or subcontract decision is added. Generic other consumer-commitment reports and provider financial/audit results were excluded rather than mistaken for this monitoring instrument. Both selected original GETs returned HTTP 200; no failed transport route or fallback loop occurred in this run. Previously closed routes were not retried.
\nThe strongest rival explanation to nonperformance is that the first report was delivered privately under this recipient/confidentiality structure and has not been located publicly. Its converse remains possible too; the present record does not decide. The most useful next record is a nonconfidential version or receipt/response for Ankura's first interim report due 30 November 2025, ideally with the researcher-access section and the Article 37 auditor cross-reference. Expected holders are Ankura, the Commission and AliExpress for its copy; the auditor may have relevant material if it requested access. The approved mandate and work plan would clarify the examination design. No request, contact or further acquisition was initiated.
", "path": "ali-oversight-evidence-wave-2026-09-09/oversight-findings.md", "sourceIds": [ "w33-con-s-ankuraroute" ], "markdown": "Closed 9 September 2026. This lane sought the actual approved mandate, the first monitoring trustee report due in November 2025, or a Commission response, including researcher-access results and evidence sharing with the Article 37 auditor. No such case-specific substantive record was acquired. The two new HTML captures document the public routes inspected; neither is a completed monitoring report. The adopted decision and appointment notice are reused originals from the preceding accepted run.\n\nThe defensible result is a defined disclosure gap, not a finding that Ankura failed to report, that the Commission did nothing, or that AliExpress complied. The 2026 cycle ends on 30 September 2026 and its report is due within two months thereafter; it is not yet due at this research snapshot.\n\n## What the current public routes disclose\n\nThe Commission's supervision index (see source references) which displays information updated on 31 August 2026, was read specifically at the AliExpress entry. Its 18 June 2025 row links the commitments narrative, monitoring trustee notice, adopted decision and press release. There is no separately labelled mandate, work plan, completed trustee report or Commission response in that inspected entry. The later 20 July 2026 fine is also listed, but the accepted preceding packet establishes its separate illegal-product risk assessment and mitigation scope. It cannot fill this researcher-access monitoring gap. This index is an overview of main enforcement activity, not a representation that every case document is published.\n\nOriginal `oversight-index.html`; scoped text `oversight-index.derived.txt`; exact relevant hrefs retained in `oversight-index.links.json`; official-reader representation `oversight-index.reader.txt`, lines 33–64. The links still resolve by identity to the already retained decision and appointment URLs: decision 124718 (see source references) and appointment notice 124727 (see source references) They were not downloaded again.\n\nAnkura's Monitorship & Independent Oversight page (see source references) describes services and a different, unnamed automotive engagement. Its substantive body does not identify an AliExpress mandate, examination, finding or delivered report. It is retained as an inspected search route and excluded as proof of case-specific work. Original `oversight-ankura.html`; body/links derivatives and official-reader representation retained. The page provides no exact publication or modification date in the inspected body; the reader's copyright year is not an engagement date. No contact or form was used, and no unrelated case or biography was followed.\n\n## The existing instrument explains the expected records and their recipients\n\nThe retained 18 June 2025 decision and annex (see source references) `../dsa-access-enforcement-wave-2026-09-09/ali-decision.pdf`, supplies a precise record chain. This run text-read PDF 75–77, Annex C.5–17; earlier accepted scope for the wider decision is preserved in the prior manifest. No new visual reading is claimed.\n\n| Expected record or act | Adopted requirement | What is established here |\n| --- | --- | --- |\n| Approved trustee mandate | C.5, PDF 75: Commission approves the proposed mandate, with modifications it considers necessary; appointment follows that approved mandate. | The separate appointment notice confirms approval of Ankura Consulting (Europe), Limited on 23 September 2025. It does not publish the mandate or disclose its negotiated terms. |\n| Detailed work plan | C.11, PDF 76: trustee proposes its monitoring work plan to the Commission within one month of appointment. | Required submission; actual plan and receipt not acquired. |\n| First report | C.10, PDF 75, and C.12(iv), PDF 76: first report by 30 November 2025 covers the initial period ending 30 September 2025. Initial implementation was on a best-efforts basis; full obligations apply thereafter. | Due date and initial assessment standard are established; delivery, tested scope, findings and Commission reaction remain unacquired. |\n| Researcher-access evidence | Prior accepted Annex B.V.7, PDF 66: semiannual material includes request totals, processing/approval/rejection, researcher feedback and technical tests. C.12(v), PDF 76 separately requires the trustee's opinion on the first internal QA semiannual report within one month. | This identifies useful expected evidence. No actual request counts, feedback, test results or trustee opinion was found. The first interim report should not automatically be treated as a full-year or mature-system evaluation. |\n| Noncompliance response | C.12(vi)–(viii), PDF 76–77: trustee may propose necessary measures; must promptly report reasonably concluded failure to the Commission and answer its questions. | Duties and permitted response, not an observed finding, recommendation or correction. |\n\nThe first report goes to the Commission, with a contemporaneous copy to AliExpress that removes confidential third-party material where applicable (C.12(iv)). C.13(iv), PDF 77, restricts disclosure of confidential information and prohibits trustee public statements concerning performance of its functions. Those provisions make nonpublic reporting a plausible explanation for the search result. They neither establish that the whole report is confidential nor determine whether the Commission could release a nonconfidential version. No case-specific confidentiality or access-to-documents refusal was acquired.\n\n## Auditor sharing is a real route, not an observed exchange\n\nC.12(iii), PDF 76, requires the trustee to minimise repetitive overlap with the Article 37 auditor. Where information is missing, the trustee may first ask whether the auditor has already collected it, request a copy, and rely on or incorporate the resulting material. AliExpress undertakes reasonably required steps enabling that disclosure. In the other direction, C.14, PDF 77, provides that the auditor will receive access, at its request, to relevant monitoring and reporting documents prepared by the trustee for the Commission.\n\nThis is an express two-way evidence route. It does not prove that either party requested documents, shared a particular test or finding, or accepted the other's conclusions. No automatic wholesale publication or transfer to researchers follows. The actual sharing request, document log or auditor reference to the first trustee report would discriminate between a permitted route and its exercise. Provider Article 37 audit discovery belongs to the parallel root lane and was not duplicated here.\n\n## Search scope and limits\n\nThirteen targeted indexed queries were run in four batches, combining the exact service, trustee, case number, mandate, first-report date, researcher and Commission/Ankura domain terms. The complete query list and selected returned-path assessment are in `oversight-search-log.json`. Search returned the existing Commission announcement and case index, generic Ankura services, secondary summaries and unrelated results. It did not yield a new case-specific Commission or Ankura mandate, first report or response that could be read. Search result ranking and indexing are incomplete and can relax terms; this is not a comprehensive archive search.\n\nA public professional profile snippet attributed an appointment announcement to the named trustee, and an adviser firm's snippet described support to Ankura. Neither was opened or acquired, and neither supplied a report or mandate. They remain discovery only; no personnel or subcontract decision is added. Generic other consumer-commitment reports and provider financial/audit results were excluded rather than mistaken for this monitoring instrument. Both selected original GETs returned HTTP 200; no failed transport route or fallback loop occurred in this run. Previously closed routes were not retried.\n\nThe strongest rival explanation to nonperformance is that the first report was delivered privately under this recipient/confidentiality structure and has not been located publicly. Its converse remains possible too; the present record does not decide. The most useful next record is a nonconfidential version or receipt/response for Ankura's first interim report due 30 November 2025, ideally with the researcher-access section and the Article 37 auditor cross-reference. Expected holders are Ankura, the Commission and AliExpress for its copy; the auditor may have relevant material if it requested access. The approved mandate and work plan would clarify the examination design. No request, contact or further acquisition was initiated.\n", "file": "research/ali-oversight-evidence-wave-2026-09-09-oversight-findings.md" }, { "id": "ali-oversight-evidence-wave-2026-09-09-root-audit-findings", "title": "A positive audit with no approved public-data requests", "description": "Closed and root-reviewed 9 September 2026. This bounded lane the Commission audit index and the linked AliExpress 2024/25 audit and implementation report. It inspected the public-data access obligation, selected audit identity/method/evidence passages, and report dates. It did not read all 365 audit pages or test the live system.", "html": "Closed and root-reviewed 9 September 2026. This bounded lane acquired the Commission audit index and the linked AliExpress 2024/25 audit and implementation report. It inspected the public-data access obligation, selected audit identity/method/evidence passages, and report dates. It did not read all 365 audit pages or test the live system.
\nThe [FTI audit](http[local research file] PDF295-297, gives Article40(12) a positive conclusion with no recommendation. The explanation expressly concerns intention and capability. It describes a public application form, an SOP for evaluation/amendment/communication, Legal-team management, designated staff applying40(8)(b)-(e), and standardized approval/rejection/clarification emails. Approved access is described as DSA Ads, Items and Punishments datasets via dashboard and request-dependent API queries. This is the auditor's account of the process, not our exercised access.
\nThe consequential observation is that no applications were approved during 1 July2024-30 June2025. The request count is redacted. The auditor says it inspected all applications and verified them as incomplete or invalid on their submitted information. That statement must accompany the zero-approval finding: the record is not merely a provider assertion accepted without claimed examination. The conclusion describes preparedness for a future qualifying request; its statement that no instance required application of the obligation cannot mean that no requests existed. Neither improper rejection nor successful data delivery is demonstrated by the public report. Application files, clarification/rejection reasons and the operative eligibility SOP would test those alternatives.
\nThe stated evidence includes public/internal documents, written attestations, interviews and substantive electronic-process testing. PDF296 names EV207 (the prior audit/implementation repository) and EV208 (a prior DSA audit report), with other entries confidential. Those are explicit prior-report references; they are not the complete basis and do not prove circular assurance. No individual applications, interview records, actual tests or audit workpapers were acquired. No applicant is identified as GDI or AI4TRUST.
\nThe [implementation report](http[local research file] PDF4, names FTI Consulting LLP as engaged by Alibaba.com Singapore E-commerce Private Limited. It dates adoption to29August2025 and audit work to5May-29August2025. FTI's exact LLP identity is also visible in the audit footers. The evaluation period largely precedes the18June2025 commitments and entirely precedes Ankura's23September appointment and full implementation after30September. This audit cannot establish performance under the first November2025 trustee cycle or the current2026 access system.
\nSelected audit front matter describes conflict checks, noncontingent fees, independence and quality methods. These are the auditor's declarations, not independently verified conflict or engagement records. Its stated absence of a single global quality-management policy is accompanied by segment/professional controls; it is not by itself proof of a deficient opinion. Provider implementation contents and closing improvement statements are not an independent validation of Article40 access. The report contains no Article40 section in the inspected contents; no full recommendation audit is claimed.
\nThe parallel trustee lane closed after13 indexed queries and two original HTML route captures. Its note correctly distinguishes required reports and permitted two-way auditor sharing from reports actually delivered and exchanges actually made. The current Commission AliExpress entry is a main-activity index, not a full docket. Ankura's generic service page contains no case-specific findings and adds no proof of performed AliExpress work. Report confidentiality/recipient rules make private delivery plausible, but do not establish delivery, non-delivery or a Commission refusal to disclose. The2026 reporting cycle has not yet ended.
\nCommission index: AliExpress table rows and their actual audit/implementation links only. Audit text: PDF1,3-6,10-14,295-297; visual:295-296. Keyword locations on other pages are discovery only. Implementation text:PDF1-5,37; no visual or whole-report review. On continuation, audit295-297 and implementation4 were reread to verify the key result/date. Root also read the complete oversight note/manifest, scoped index derivative and Ankura substantive derivative; reused decision/appointment reading remains in the preceding accepted packet. Original and derivative custody is in root-captures.json and oversight-captures.json.
\nThe administrative gate has a named operating owner: the provider's Legal team, under the reported SOP. A positive access opinion can coexist with no delivered access when all requests are judged ineligible. That is a concrete mechanism for testing gatekeeping, not evidence by itself of obstruction. The best next public record is an actual request and reasoned response, preferably with the missing-information exchange; this connects the formal entitlement to the judgment that made it usable or unusable.
", "path": "ali-oversight-evidence-wave-2026-09-09/root-audit-findings.md", "sourceIds": [ "w33-con-s-auditindex", "w33-con-s-audit", "w33-con-s-implementation" ], "markdown": "Closed and root-reviewed 9 September 2026. This bounded lane acquired the Commission audit index and the linked AliExpress 2024/25 audit and implementation report. It inspected the public-data access obligation, selected audit identity/method/evidence passages, and report dates. It did not read all 365 audit pages or test the live system.\n\nThe FTI audit (see source references) PDF295-297, gives Article40(12) a positive conclusion with no recommendation. The explanation expressly concerns intention and capability. It describes a public application form, an SOP for evaluation/amendment/communication, Legal-team management, designated staff applying40(8)(b)-(e), and standardized approval/rejection/clarification emails. Approved access is described as DSA Ads, Items and Punishments datasets via dashboard and request-dependent API queries. This is the auditor's account of the process, not our exercised access.\n\nThe consequential observation is that no applications were approved during 1 July2024-30 June2025. The request count is redacted. The auditor says it inspected all applications and verified them as incomplete or invalid on their submitted information. That statement must accompany the zero-approval finding: the record is not merely a provider assertion accepted without claimed examination. The conclusion describes preparedness for a future qualifying request; its statement that no instance required application of the obligation cannot mean that no requests existed. Neither improper rejection nor successful data delivery is demonstrated by the public report. Application files, clarification/rejection reasons and the operative eligibility SOP would test those alternatives.\n\nThe stated evidence includes public/internal documents, written attestations, interviews and substantive electronic-process testing. PDF296 names EV207 (the prior audit/implementation repository) and EV208 (a prior DSA audit report), with other entries confidential. Those are explicit prior-report references; they are not the complete basis and do not prove circular assurance. No individual applications, interview records, actual tests or audit workpapers were acquired. No applicant is identified as GDI or AI4TRUST.\n\nThe implementation report (see source references) PDF4, names FTI Consulting LLP as engaged by Alibaba.com Singapore E-commerce Private Limited. It dates adoption to29August2025 and audit work to5May-29August2025. FTI's exact LLP identity is also visible in the audit footers. The evaluation period largely precedes the18June2025 commitments and entirely precedes Ankura's23September appointment and full implementation after30September. This audit cannot establish performance under the first November2025 trustee cycle or the current2026 access system.\n\nSelected audit front matter describes conflict checks, noncontingent fees, independence and quality methods. These are the auditor's declarations, not independently verified conflict or engagement records. Its stated absence of a single global quality-management policy is accompanied by segment/professional controls; it is not by itself proof of a deficient opinion. Provider implementation contents and closing improvement statements are not an independent validation of Article40 access. The report contains no Article40 section in the inspected contents; no full recommendation audit is claimed.\n\nThe parallel trustee lane closed after13 indexed queries and two original HTML route captures. Its note correctly distinguishes required reports and permitted two-way auditor sharing from reports actually delivered and exchanges actually made. The current Commission AliExpress entry is a main-activity index, not a full docket. Ankura's generic service page contains no case-specific findings and adds no proof of performed AliExpress work. Report confidentiality/recipient rules make private delivery plausible, but do not establish delivery, non-delivery or a Commission refusal to disclose. The2026 reporting cycle has not yet ended.\n\n## Actual root reading\n\nCommission index: AliExpress table rows and their actual audit/implementation links only. Audit text: PDF1,3-6,10-14,295-297; visual:295-296. Keyword locations on other pages are discovery only. Implementation text:PDF1-5,37; no visual or whole-report review. On continuation, audit295-297 and implementation4 were reread to verify the key result/date. Root also read the complete oversight note/manifest, scoped index derivative and Ankura substantive derivative; reused decision/appointment reading remains in the preceding accepted packet. Original and derivative custody is in root-captures.json and oversight-captures.json.\n\n## What this changes\n\nThe administrative gate has a named operating owner: the provider's Legal team, under the reported SOP. A positive access opinion can coexist with no delivered access when all requests are judged ineligible. That is a concrete mechanism for testing gatekeeping, not evidence by itself of obstruction. The best next public record is an actual request and reasoned response, preferably with the missing-information exchange; this connects the formal entitlement to the judgment that made it usable or unusable.\n", "file": "research/ali-oversight-evidence-wave-2026-09-09-root-audit-findings.md" }, { "id": "assurance-wave-2026-09-09-fathom-bill", "title": "Fathom and California SB 813: verification authority and its limits", "description": "Research cutoff: 9 September 2026. Bounded instrument review following the whole-case review in forest-reviews.md. No canonical case changes. All obligations below describe the enrolled bill if enacted, not obligations already established by this bill.", "html": "Research cutoff: 9 September 2026. Bounded instrument review following the whole-case review in forest-reviews.md. No canonical case changes. All obligations below describe the enrolled bill if enacted, not obligations already established by this bill.
\nThe advertised instrument is California SB 813 (McNerney), 2025–26 session. The official record shows passage by both houses on 30 August 2026 and enrollment on 1 September. As inspected on 9 September, its status is “Active Bill – Passed”; no governor approval or chapter entry appears. The senator’s 30 August release says it proceeds to the governor. The evidence supports passed/enrolled, pending enactment in the inspected record, rather than signed law. [Official status](http[local research file] [history](http[local research file] [author’s passage announcement](http[local research file]
\nThe enrolled text creates a Government Operations Agency (GovOps) framework to qualify independent verification organizations (IVOs). It expressly does not require a developer, deployer or operator to engage an IVO or obtain a covered audit as a condition of developing, deploying or operating AI in California (§8898.4(a)(3)). A qualified evaluator is not a state endorsement of an AI model or system. The bill neither appoints Fathom nor makes an IVO a universal permission gate. [Enrolled text, §§8898–8898.4](http[local research file]
\n| Owner or participant | \nEnrolled power or obligation | \nLimit and exact locator | \n
|---|---|---|
| GovOps | \nBy 1 January 2028, establish application requirements, qualification criteria and procedures for suspending or terminating IVO designation. Publish requirements and criteria; review and revise them. | \n§§8898.1(a)–(d), (f). The bill supplies considerations rather than a named vendor or fixed numerical certification benchmark. Actual implementing criteria remain future records. | \n
| GovOps selecting evaluators | \nConsider risk-assessment expertise, metrics and methods, technical staffing, financial conflicts and operational independence. Applicants supply qualifications, proposed methods and supporting documentation. | \n§8898.1(a), (c). An AI auditor may be a person, partnership, academic institution, nonprofit or corporation (§8898 definitions); the designation route is not limited to nonprofits. | \n
| GovOps supervising evaluators | \nDevelop suspension/termination procedures considering standards failures, material misrepresentation, conflicts, missing documentation, integrity/objectivity/competence and cybersecurity lapses. | \n§8898.1(b). This concerns evaluator designation. The final text does not reproduce the old model-certification/decertification system. | \n
| Assessed party and evaluator | \nReasonable market-rate payment is allowed. Payment or its amount cannot depend on the assessment outcome. The evaluator must maintain operational and management independence, including freedom from assessed-party control over conclusions and recommendations. | \n§8898.1(c)(2)(C)–(D). This permits an assessed-party-paid market subject to independence requirements; it does not establish an executed contract or actual fee schedule. | \n
| GovOps and advisory participants | \nConsult relevant stakeholders as appropriate; convene working groups including engineers from competing AI companies and AI-safety experts; report findings to the Legislature. | \n§8898.2. Advisory participation is not a private veto or an appointment of any named organization. | \n
| Designated IVO | \nBeginning no sooner than 12 months after designation, report annually on methods/standards, relevant governance or funding changes and application changes. Retain unredacted reports for five years. | \n§8898.3. Specified redactions protect trade secrets, security and other legally protected material, with justification where possible. Public reporting will not necessarily disclose all underlying material. | \n
| Developer, deployer or operator | \nMay choose whether to use this verification route. | \n§8898.4(a)(3). The bill also does not require an IVO to perform covered compliance audits to register (§8898.4(a)(4)). This conclusion concerns SB 813, not duties imposed by other laws. | \n
| Court considering an audit | \nAn audit performed according to an identified standard may be relevant but is not conclusive in a harm proceeding. | \n§8898.4(b). No affirmative defense or rebuttable presumption survives. Section 8898.4(a)(1) disclaims liability arising solely from failure to comply with a standard under this chapter. | \n
Alternative compliance material matters: §8898.1(g)–(h) calls for practicable alignment with existing auditing standards and permits substantially similar reports, assessments, audits or assurance work when the agency’s requirements are met. That is not automatic interstate reciprocity or acceptance of every existing certificate. The complete enrolled text contains no bespoke appeal, reconsideration or hearing procedure. This does not establish that ordinary administrative or judicial remedies are unavailable; the future agency procedures and any applicable general law would need separate examination.
\nThe senator’s 26 March 2025 announcement identifies Fathom as sponsoring through its advocacy arm Meridian. Fathom’s 31 August 2026 release claims that it sponsored, developed and championed the IVO approach. The latter is an organization’s account of its contribution, not independent proof that it wrote each clause or controlled revisions. The exact legal relationship between Fathom and Meridian was not resolved in this lane. [Author’s original announcement](http[local research file] [Fathom’s passage account](http[local research file]
\nThe most revealing interest is documented in an official source: the Senate Judiciary analysis for the 29 April 2025 hearing identifies Fathom as sponsor and reports that it hoped to become a model review organization (MRO), p.1. On p.9 it attributes to the sponsor the assessment that only a handful of entities were qualified and that Fathom considered itself among them. This establishes a historical prospective-beneficiary interest. It does not establish a later designation, award, monopoly or successful attempt to control an assessment. [Judiciary analysis, pp.1 and 9](http[local research file]
\n| Version or decision | \nMaterial rule or change | \nEvidentiary boundary | \n
|---|---|---|
| 26 March 2025, official version 20250SB81398AMD | \nAttorney General designation of MROs, renewable every three years; model/application certification could furnish an affirmative defense to specified personal-injury or property-damage liability, except intentional misconduct. | \nHistorical text, particularly old §8898.4. This stronger legal incentive is absent from the enrolled bill. | \n
| 29 April hearing; 1 May 2025 version 20250SB81397AMD | \nJudiciary p.16 records author agreement to replace the affirmative defense with a rebuttable presumption. The 1 May text actually makes that replacement: certification could support presumed reasonable care, overcome by contrary admissible evidence. | \nProposal/acceptance and subsequent text are separately supported. Deleted and added markup was checked, rather than treating both as operative. | \n
| 5 January 2026, official version 20250SB81396AMD | \nRecasts the institutional structure and deletes the rebuttable presumption. | \nHeader/digest and the relevant deletion markup were read; this was not a complete audit of every January provision. | \n
| 1 July 2026 committee hearing on 16 June text; 2 July version reflected in August Appropriations | \nIntermediate seven-member commission, two levels of safety standards and dependencies involving AB 1405 and potentially AB 1709. Committee pp.14–15 report Encode’s request to distribute appointment authority and the author’s agreement to give appointments to the Speaker and Senate Rules. | \nA concrete instance of accepted input, but the commission and these appointment rights do not survive in the enrolled framework. Neither referenced Assembly bill was separately investigated. | \n
| 26 August amendment; enrolled 1 September 2026 | \nGovOps qualification framework, express voluntary scope, audit relevant but not conclusive; no commission, two-tier scheme, affirmative defense or rebuttable presumption. | \nComplete enrolled operative text read. Intermediate amendment dates for each individual deletion were not separately established. | \n
Official historical texts are available through the [bill text page’s Version selector](http[local research file] An ordinary URL parameter did not change the selected version and was not used as historical evidence. Normal read-only form selection returned the verified historical headers and amendment markup.
\nThe final Senate floor analysis, dated 30 August and based on the 26 August text, records Anthropic, Encode and Fathom in support and the Business Software Alliance (BSA) opposed. Anthropic supports evaluator designation while explicitly advocating additional work to require independently evaluated developer safeguards. That additional requirement is advocacy beyond this bill. BSA objects to premature implementation, evaluator competence safeguards, international alignment and the possibility that undefined standards in litigation could make nominally voluntary audits a practical expectation. The last point is a documented prediction of indirect pressure, not evidence that insurers, procurers or courts have already made certification mandatory. [Official analysis index, Senate Floor 30 August 2026, analysisId 405892, pp.5–7](http[local research file]
\nThe Assembly floor analysis dated 27 August calls Fathom the sponsor and reproduces its strategic case for competitive advantage and a potential national de facto standard; that passage expressly concerns a previous version. It must not be silently restated as an operative legal requirement. Its fiscal comments say Unknown. [Same official index, Assembly Floor 27 August 2026, analysisId 405661, pp.2–3](http[local research file]
\nA material fiscal trap remains in the official packet. The 5 August Appropriations hearing analysis, based on the 2 July commission version, estimates costs in the tens of millions for that framework. The final Senate floor document still reproduces fiscal discussion of superseded AG/MRO, rebuttable-presumption and commission provisions. Those are not reliable estimates of the final GovOps bill. Report the latest Assembly estimate as unknown; no enacted appropriation, operational budget or actual fee revenue is established here. [Same index, Assembly Appropriations entry 3 August, hearing 5 August, analysisId 402751, pp.1–3; Senate Floor 30 August, pp.4–5](http[local research file]
\nThese are candidate assertions, not edits to the canonical graph.
\n| Candidate ID | \nDirection and typed relation | \nDate/state and limit | \n
|---|---|---|
| assurance-fathom-sponsors-sb813 | \nFathom → SB 813: sponsors legislative proposal | \nOfficial author announcement, 26 March 2025; confirmed as sponsor in committee discussion. Sponsorship does not establish exclusive drafting or approval authority. | \n
| assurance-fathom-prospective-mro | \nFathom → proposed MRO designation: expresses interest in becoming eligible operator | \nOfficial Judiciary report of sponsor position, 29 April 2025, pp.1, 9. Historical aspiration, not appointment. | \n
| assurance-sb813-govops-designation | \nEnrolled SB 813 → GovOps: would authorize evaluator qualification and suspension procedures | \nEnrolled 1 September 2026; enactment not established. Final §§8898.1–8898.2. | \n
| assurance-govops-ivo-conditions | \nGovOps → applicant/designated IVO: would impose qualification, independence and reporting conditions | \nProposed statutory delegation; detailed implementing standards not yet established in this source packet. | \n
| assurance-assessed-party-ivo-payment | \nAssessed party → IVO: may pay reasonable market-rate assessment fees, without outcome contingency | \nEnrolled §8898.1(c)(2)(C). Permitted transaction class, not observed payment. | \n
| assurance-encode-intermediate-appointments | \nEncode → author’s intermediate commission amendment: requests shared appointment authority, reported accepted | \nJuly committee analysis pp.14–15; superseded institution, no surviving current appointment right. | \n
| assurance-bsa-voluntary-pressure-objection | \nBSA → SB 813: objects that litigation treatment could create practical pressure for voluntary audits | \nFinal Senate floor pp.6–7, reported concern rather than demonstrated effect. | \n
| assurance-sb813-voluntary-operation | \nEnrolled SB 813 → AI developer/deployer/operator: preserves operation without this verification route | \n§8898.4(a)(3), proposed law. Does not negate other statutory or contractual requirements. | \n
Records were retrieved from their original publishers on 9 September 2026. The following hashes describe returned bodies inspected in memory; original bodies were not retained as separate local files in this lane. The official analysis PDFs are delivered by the index’s ordinary public form selectors, so the stable citation is the index plus dated entry, analysisId and PDF page. Hashes are corroborating retrieval identifiers, not a claim of locally preserved originals. No access restriction was bypassed.
\n| Record | \nScope actually read | \nSHA-256 where recorded | \n
|---|---|---|
| Enrolled official text, billTextClient.xhtml, published 1 September 2026 21:00 | \nHeader/digest and complete §§8898–8898.4 | \n82a81fb20e59151762055d3c7461c3fbf5048ef7e1d92a58d4403b11a98bb646 | \n
| Official status and history pages | \nFull bill-specific status/history, through enrollment | \nNot recorded | \n
| Author 26 March and 30 August announcements; Fathom 31 August account | \nComplete substantive bill announcement bodies | \nNot recorded | \n
| Senate Judiciary PDF linked above | \npp.1–3, 8–11, 14–17; decisive pp.1, 9, 16 | \nNot recorded | \n
| Official analysis index | \nDated entry list and public form selectors | \nf04803891c081237681924f75aefdaf2192d5e98062f7f94fb47b482c95300f9 | \n
| Senate Floor, analysisId 405892, 30 August 2026 | \nAll seven PDF pages | \nc3ac5b1617aef0a3770742619d2d1f1b31f01856aa3c7e778ed4a44fc5cd8414 | \n
| Assembly Floor, analysisId 405661, 27 August 2026 | \nAll three PDF pages | \n5e1efe9e6b56496c8df63dda2337a382f87ef22a98b540c09d518039025fdd9f | \n
| Assembly Appropriations, analysisId 402751, index 3 August/hearing 5 August | \nAll three PDF pages; header identifies 2 July version | \n0571f9c79fdda04b6566173d2b5764ba2c4fc78a94dc587cfd62c9a30d2453c9 | \n
| Assembly Privacy and Consumer Protection, analysisId 401958, index 30 June/hearing 1 July | \nPDF pp.1, 10–23; header identifies 16 June version and proposed amendments. Earlier output for pp.2–9 was truncated and is not claimed read. | \n6bb665073294fd314946d0d407cc30bfe51b618135e9c57fe9579755c0e942f3 | \n
| 26 March 2025 text, selector 20250SB81398AMD | \nHeader/digest and relevant designation/liability provisions | \n46962fbb523a016c058f4a3f77271c877698b422af840a284735b8f987ced7bc | \n
| 1 May 2025 text, selector 20250SB81397AMD | \nHeader/digest and liability revision, including deletion/addition markup | \n5cf00166a54d2d8267478b9844cb6c1ce908aa057d8dbd45917d2e3ca43cee3d | \n
| 5 January 2026 text, selector 20250SB81396AMD | \nHeader/digest and relevant provisions/deleted presumption markup | \n1566fc7d12befdbc0789355b0025c5c39f5291592458eea2fff933d94f116a2f | \n
This instrument supports a recognized verification market with government selection and independence rules, alongside a documented historical sponsor interest in becoming a recognized operator. It does not establish compulsory private certification, a current Fathom designation or enacted state authority. An ordinary policy-development explanation remains strong: legislators revised a sponsored proposal, removed stronger liability protections and preserved a voluntary route while opponents continued to contest its consequences.
\nThe next records that could change this conclusion are a governor’s action/chapter record; adopted GovOps rules and actual designation decisions; and actual procurement, insurance, litigation or customer terms that condition participation on recognized verification. An application or decision naming Fathom would resolve prospective interest versus achieved status. A contract requiring an IVO audit would test indirect compulsion. No such record is established in this packet. Return to the whole-case review before opening those branches.
", "path": "assurance-wave-2026-09-09/fathom-bill.md", "sourceIds": [ "w9-assure-s-sb813-status", "w9-assure-s-sb813-enrolled", "w9-assure-s-sb813-judiciary", "w9-assure-s-sb813-author", "w9-assure-s-sb813-floor" ], "markdown": "Research cutoff: 9 September 2026. Bounded instrument review following the whole-case review in forest-reviews.md. No canonical case changes. All obligations below describe the enrolled bill if enacted, not obligations already established by this bill.\n\n## Instrument-level result\n\nThe advertised instrument is California SB 813 (McNerney), 2025–26 session. The official record shows passage by both houses on 30 August 2026 and enrollment on 1 September. As inspected on 9 September, its status is “Active Bill – Passed”; no governor approval or chapter entry appears. The senator’s 30 August release says it proceeds to the governor. The evidence supports passed/enrolled, pending enactment in the inspected record, rather than signed law. Official status (see source references) history (see source references) author’s passage announcement (see source references)\n\nThe enrolled text creates a Government Operations Agency (GovOps) framework to qualify independent verification organizations (IVOs). It expressly does **not** require a developer, deployer or operator to engage an IVO or obtain a covered audit as a condition of developing, deploying or operating AI in California (§8898.4(a)(3)). A qualified evaluator is not a state endorsement of an AI model or system. The bill neither appoints Fathom nor makes an IVO a universal permission gate. Enrolled text, §§8898–8898.4 (see source references)\n\n## Who would decide, and what would become obligatory?\n\n| Owner or participant | Enrolled power or obligation | Limit and exact locator |\n| --- | --- | --- |\n| GovOps | By 1 January 2028, establish application requirements, qualification criteria and procedures for suspending or terminating IVO designation. Publish requirements and criteria; review and revise them. | §§8898.1(a)–(d), (f). The bill supplies considerations rather than a named vendor or fixed numerical certification benchmark. Actual implementing criteria remain future records. |\n| GovOps selecting evaluators | Consider risk-assessment expertise, metrics and methods, technical staffing, financial conflicts and operational independence. Applicants supply qualifications, proposed methods and supporting documentation. | §8898.1(a), (c). An AI auditor may be a person, partnership, academic institution, nonprofit or corporation (§8898 definitions); the designation route is not limited to nonprofits. |\n| GovOps supervising evaluators | Develop suspension/termination procedures considering standards failures, material misrepresentation, conflicts, missing documentation, integrity/objectivity/competence and cybersecurity lapses. | §8898.1(b). This concerns evaluator designation. The final text does not reproduce the old model-certification/decertification system. |\n| Assessed party and evaluator | Reasonable market-rate payment is allowed. Payment or its amount cannot depend on the assessment outcome. The evaluator must maintain operational and management independence, including freedom from assessed-party control over conclusions and recommendations. | §8898.1(c)(2)(C)–(D). This permits an assessed-party-paid market subject to independence requirements; it does not establish an executed contract or actual fee schedule. |\n| GovOps and advisory participants | Consult relevant stakeholders as appropriate; convene working groups including engineers from competing AI companies and AI-safety experts; report findings to the Legislature. | §8898.2. Advisory participation is not a private veto or an appointment of any named organization. |\n| Designated IVO | Beginning no sooner than 12 months after designation, report annually on methods/standards, relevant governance or funding changes and application changes. Retain unredacted reports for five years. | §8898.3. Specified redactions protect trade secrets, security and other legally protected material, with justification where possible. Public reporting will not necessarily disclose all underlying material. |\n| Developer, deployer or operator | May choose whether to use this verification route. | §8898.4(a)(3). The bill also does not require an IVO to perform covered compliance audits to register (§8898.4(a)(4)). This conclusion concerns SB 813, not duties imposed by other laws. |\n| Court considering an audit | An audit performed according to an identified standard may be relevant but is not conclusive in a harm proceeding. | §8898.4(b). No affirmative defense or rebuttable presumption survives. Section 8898.4(a)(1) disclaims liability arising solely from failure to comply with a standard under this chapter. |\n\nAlternative compliance material matters: §8898.1(g)–(h) calls for practicable alignment with existing auditing standards and permits substantially similar reports, assessments, audits or assurance work when the agency’s requirements are met. That is not automatic interstate reciprocity or acceptance of every existing certificate. The complete enrolled text contains no bespoke appeal, reconsideration or hearing procedure. This does not establish that ordinary administrative or judicial remedies are unavailable; the future agency procedures and any applicable general law would need separate examination.\n\n## Sponsorship, prospective benefit and consequential revisions\n\nThe senator’s 26 March 2025 announcement identifies Fathom as sponsoring through its advocacy arm Meridian. Fathom’s 31 August 2026 release claims that it sponsored, developed and championed the IVO approach. The latter is an organization’s account of its contribution, not independent proof that it wrote each clause or controlled revisions. The exact legal relationship between Fathom and Meridian was not resolved in this lane. Author’s original announcement (see source references) Fathom’s passage account (see source references)\n\nThe most revealing interest is documented in an official source: the Senate Judiciary analysis for the 29 April 2025 hearing identifies Fathom as sponsor and reports that it hoped to become a model review organization (MRO), p.1. On p.9 it attributes to the sponsor the assessment that only a handful of entities were qualified and that Fathom considered itself among them. This establishes a historical prospective-beneficiary interest. It does not establish a later designation, award, monopoly or successful attempt to control an assessment. Judiciary analysis, pp.1 and 9 (see source references)\n\n| Version or decision | Material rule or change | Evidentiary boundary |\n| --- | --- | --- |\n| 26 March 2025, official version 20250SB81398AMD | Attorney General designation of MROs, renewable every three years; model/application certification could furnish an affirmative defense to specified personal-injury or property-damage liability, except intentional misconduct. | Historical text, particularly old §8898.4. This stronger legal incentive is absent from the enrolled bill. |\n| 29 April hearing; 1 May 2025 version 20250SB81397AMD | Judiciary p.16 records author agreement to replace the affirmative defense with a rebuttable presumption. The 1 May text actually makes that replacement: certification could support presumed reasonable care, overcome by contrary admissible evidence. | Proposal/acceptance and subsequent text are separately supported. Deleted and added markup was checked, rather than treating both as operative. |\n| 5 January 2026, official version 20250SB81396AMD | Recasts the institutional structure and deletes the rebuttable presumption. | Header/digest and the relevant deletion markup were read; this was not a complete audit of every January provision. |\n| 1 July 2026 committee hearing on 16 June text; 2 July version reflected in August Appropriations | Intermediate seven-member commission, two levels of safety standards and dependencies involving AB 1405 and potentially AB 1709. Committee pp.14–15 report Encode’s request to distribute appointment authority and the author’s agreement to give appointments to the Speaker and Senate Rules. | A concrete instance of accepted input, but the commission and these appointment rights do not survive in the enrolled framework. Neither referenced Assembly bill was separately investigated. |\n| 26 August amendment; enrolled 1 September 2026 | GovOps qualification framework, express voluntary scope, audit relevant but not conclusive; no commission, two-tier scheme, affirmative defense or rebuttable presumption. | Complete enrolled operative text read. Intermediate amendment dates for each individual deletion were not separately established. |\n\nOfficial historical texts are available through the bill text page’s Version selector (see source references) An ordinary URL parameter did not change the selected version and was not used as historical evidence. Normal read-only form selection returned the verified historical headers and amendment markup.\n\n## Contested choices and fiscal limits\n\nThe final Senate floor analysis, dated 30 August and based on the 26 August text, records Anthropic, Encode and Fathom in support and the Business Software Alliance (BSA) opposed. Anthropic supports evaluator designation while explicitly advocating additional work to require independently evaluated developer safeguards. That additional requirement is advocacy beyond this bill. BSA objects to premature implementation, evaluator competence safeguards, international alignment and the possibility that undefined standards in litigation could make nominally voluntary audits a practical expectation. The last point is a documented prediction of indirect pressure, not evidence that insurers, procurers or courts have already made certification mandatory. Official analysis index, Senate Floor 30 August 2026, analysisId 405892, pp.5–7 (see source references)\n\nThe Assembly floor analysis dated 27 August calls Fathom the sponsor and reproduces its strategic case for competitive advantage and a potential national de facto standard; that passage expressly concerns a previous version. It must not be silently restated as an operative legal requirement. Its fiscal comments say **Unknown**. Same official index, Assembly Floor 27 August 2026, analysisId 405661, pp.2–3 (see source references)\n\nA material fiscal trap remains in the official packet. The 5 August Appropriations hearing analysis, based on the 2 July commission version, estimates costs in the tens of millions for that framework. The final Senate floor document still reproduces fiscal discussion of superseded AG/MRO, rebuttable-presumption and commission provisions. Those are not reliable estimates of the final GovOps bill. Report the latest Assembly estimate as unknown; no enacted appropriation, operational budget or actual fee revenue is established here. Same index, Assembly Appropriations entry 3 August, hearing 5 August, analysisId 402751, pp.1–3; Senate Floor 30 August, pp.4–5 (see source references)\n\n## Candidate directional relations\n\nThese are candidate assertions, not edits to the canonical graph.\n\n| Candidate ID | Direction and typed relation | Date/state and limit |\n| --- | --- | --- |\n| assurance-fathom-sponsors-sb813 | Fathom → SB 813: sponsors legislative proposal | Official author announcement, 26 March 2025; confirmed as sponsor in committee discussion. Sponsorship does not establish exclusive drafting or approval authority. |\n| assurance-fathom-prospective-mro | Fathom → proposed MRO designation: expresses interest in becoming eligible operator | Official Judiciary report of sponsor position, 29 April 2025, pp.1, 9. Historical aspiration, not appointment. |\n| assurance-sb813-govops-designation | Enrolled SB 813 → GovOps: would authorize evaluator qualification and suspension procedures | Enrolled 1 September 2026; enactment not established. Final §§8898.1–8898.2. |\n| assurance-govops-ivo-conditions | GovOps → applicant/designated IVO: would impose qualification, independence and reporting conditions | Proposed statutory delegation; detailed implementing standards not yet established in this source packet. |\n| assurance-assessed-party-ivo-payment | Assessed party → IVO: may pay reasonable market-rate assessment fees, without outcome contingency | Enrolled §8898.1(c)(2)(C). Permitted transaction class, not observed payment. |\n| assurance-encode-intermediate-appointments | Encode → author’s intermediate commission amendment: requests shared appointment authority, reported accepted | July committee analysis pp.14–15; superseded institution, no surviving current appointment right. |\n| assurance-bsa-voluntary-pressure-objection | BSA → SB 813: objects that litigation treatment could create practical pressure for voluntary audits | Final Senate floor pp.6–7, reported concern rather than demonstrated effect. |\n| assurance-sb813-voluntary-operation | Enrolled SB 813 → AI developer/deployer/operator: preserves operation without this verification route | §8898.4(a)(3), proposed law. Does not negate other statutory or contractual requirements. |\n\n## Source custody and reading scope\n\nRecords were retrieved from their original publishers on 9 September 2026. The following hashes describe returned bodies inspected in memory; original bodies were **not retained as separate local files** in this lane. The official analysis PDFs are delivered by the index’s ordinary public form selectors, so the stable citation is the index plus dated entry, analysisId and PDF page. Hashes are corroborating retrieval identifiers, not a claim of locally preserved originals. No access restriction was bypassed.\n\n| Record | Scope actually read | SHA-256 where recorded |\n| --- | --- | --- |\n| Enrolled official text, billTextClient.xhtml, published 1 September 2026 21:00 | Header/digest and complete §§8898–8898.4 | 82a81fb20e59151762055d3c7461c3fbf5048ef7e1d92a58d4403b11a98bb646 |\n| Official status and history pages | Full bill-specific status/history, through enrollment | Not recorded |\n| Author 26 March and 30 August announcements; Fathom 31 August account | Complete substantive bill announcement bodies | Not recorded |\n| Senate Judiciary PDF linked above | pp.1–3, 8–11, 14–17; decisive pp.1, 9, 16 | Not recorded |\n| Official analysis index | Dated entry list and public form selectors | f04803891c081237681924f75aefdaf2192d5e98062f7f94fb47b482c95300f9 |\n| Senate Floor, analysisId 405892, 30 August 2026 | All seven PDF pages | c3ac5b1617aef0a3770742619d2d1f1b31f01856aa3c7e778ed4a44fc5cd8414 |\n| Assembly Floor, analysisId 405661, 27 August 2026 | All three PDF pages | 5e1efe9e6b56496c8df63dda2337a382f87ef22a98b540c09d518039025fdd9f |\n| Assembly Appropriations, analysisId 402751, index 3 August/hearing 5 August | All three PDF pages; header identifies 2 July version | 0571f9c79fdda04b6566173d2b5764ba2c4fc78a94dc587cfd62c9a30d2453c9 |\n| Assembly Privacy and Consumer Protection, analysisId 401958, index 30 June/hearing 1 July | PDF pp.1, 10–23; header identifies 16 June version and proposed amendments. Earlier output for pp.2–9 was truncated and is not claimed read. | 6bb665073294fd314946d0d407cc30bfe51b618135e9c57fe9579755c0e942f3 |\n| 26 March 2025 text, selector 20250SB81398AMD | Header/digest and relevant designation/liability provisions | 46962fbb523a016c058f4a3f77271c877698b422af840a284735b8f987ced7bc |\n| 1 May 2025 text, selector 20250SB81397AMD | Header/digest and liability revision, including deletion/addition markup | 5cf00166a54d2d8267478b9844cb6c1ce908aa057d8dbd45917d2e3ca43cee3d |\n| 5 January 2026 text, selector 20250SB81396AMD | Header/digest and relevant provisions/deleted presumption markup | 1566fc7d12befdbc0789355b0025c5c39f5291592458eea2fff933d94f116a2f |\n\n## Stopping point and discriminating missing records\n\nThis instrument supports a recognized verification market with government selection and independence rules, alongside a documented historical sponsor interest in becoming a recognized operator. It does not establish compulsory private certification, a current Fathom designation or enacted state authority. An ordinary policy-development explanation remains strong: legislators revised a sponsored proposal, removed stronger liability protections and preserved a voluntary route while opponents continued to contest its consequences.\n\nThe next records that could change this conclusion are a governor’s action/chapter record; adopted GovOps rules and actual designation decisions; and actual procurement, insurance, litigation or customer terms that condition participation on recognized verification. An application or decision naming Fathom would resolve prospective interest versus achieved status. A contract requiring an IVO audit would test indirect compulsion. No such record is established in this packet. Return to the whole-case review before opening those branches.\n", "file": "research/assurance-wave-2026-09-09-fathom-bill.md" }, { "id": "assurance-wave-2026-09-09-g7-minors", "title": "G7 minors instruments: assurance commitments, privacy limits and the missing compulsory gate", "description": "Instrument-level finding for forest review. Observed 9 September 2026. Read the current forest-reviews.md before acquisition. Only this note was written; no canonical graph, automation, other branch or publication was changed.", "html": "Instrument-level finding for forest review. Observed 9 September 2026. Read the current forest-reviews.md before acquisition. Only this note was written; no canonical graph, automation, other branch or publication was changed.
\nThe 2026 G7 texts politically endorse age assurance, including for conversational AI, but the inspected instruments do not themselves create a universal identity requirement, a named certification monopoly or an enforceable market-entry condition. A concrete compulsory gate appears in 5Rights' response as a demand for future policy: certification before affected products reach the market. That demand is not an adopted provision of these G7 texts.
\nThere is also a direct institutional counterweight to broad identification: the G7 privacy authorities subsequently reference the ministers' agreement and oppose using age assurance for identification, tracking, profiling or monitoring. This is expressed guidance, not evidence that every implementation complies. The most discriminating next record is an actual implementation instrument that assigns the access decision and specifies its legal basis, methods, override and redress.
\nG1: Ministers' Common Set of Principles, English, three pages. Principle1 calls for reliable, proportionate, rights-respecting, privacy-preserving and interoperable age assurance, using minimally intrusive feasible means. It connects assurance both to age-appropriate experiences and to restrictions under applicable law; it permits parental consent where needed. Principles3/6 likewise refer to current legal obligations. Section4 leaves jurisdictions their own implementation approaches and asks providers to develop practical measures with stakeholders. Principle7 promotes shared evidence and impartial risk assessment. It names no verification vendor, certification scheme or technical standard. It does not specify passports, biometrics, real names, exact birth dates, age-estimation methods or a universal check at every visit. “Anonymity” is not explicit. The source uses the umbrella term age assurance; its French translation's use of vérification does not supply an otherwise absent identity-document requirement. [Official English PDF](http[local research file] PDF1 Principle1; PDF2 Principles3–4; PDF3 Principles6–7/§4. [Official French translation](http[local research file]
\nG2: Leaders' Call, 17 June, three pages. Its opening defines the protected population as under18; that is not an across-the-board minimum access age. PDF2 bullet1 asks providers to apply age assurance consistently with national circumstances and applicable frameworks. Bullet2 specifically calls for default protections for conversational AI, including parental tools and age assurance. Bullet4 uses mandatory wording about detecting/removing CSAM and criminal NCII, while grounding prohibitions in national legal frameworks. PDF3 asks ministers to meet regularly and assess progress by year-end. The call supports common assessment methods without naming their authors/operators. No sanction schedule, compliance adjudicator, certification condition, prescribed identifier, authentication protocol or implementation commencement date appears. English original hosted by Germany's federal government.
\nThe legal boundary follows from these operative provisions: the statements articulate joint policy and expectations, while domestic law supplies any compulsory access restriction. Treating the word “must” inside this leaders' call as a newly enacted transnational offence would bypass the missing legal mechanism. Conversely, calling the texts merely empty rhetoric would miss their explicit direction to providers, research coordination and ministerial review. This inquiry establishes neither a newly enforceable G7-wide gate nor the absence of gates already created elsewhere.
\nG3: G7 privacy-authorities statement, adopted 25–26 June in Paris. Paragraph3 explicitly references the ministers' agreement. Paragraphs2/4 favor context-specific use and warn against indiscriminate application. Paragraph7(2) says assurance should not enable identification, tracking, profiling or monitoring; results should not be repurposed. Paragraph7(3) limits collection/use/retention; other provisions address proportionality, security, transparency and design. Paragraph6 names ISO27566 alongside EDPB, FTC/COPPA, Canadian OPC and a joint international statement. It takes note of these resources, rather than requiring ISO certification or endorsing a supplier. Paragraph7(8) commits to further exchange and regulator engagement. No individual enforcement decision or audited deployment is established here. This is a separate regulators' instrument, not a clause silently added to G1/G2. [OPC-hosted original](http[local research file] all numbered paragraphs; page modified8July2026.
\nG4: UNICEF's 1 June statement says it supplied ministers with global evidence before the meetings. It calls for publication and delivery of an agreed implementation plan with timelines/accountability. This is an attributable claim of input, not the submitted evidence package, a government acknowledgment of particular wording, or the plan itself. Its appeal to convert principles into company obligations is also informative about the implementation gap. No inspected original allocates a UNICEF approval/veto right. [UNICEF original](http[local research file] paragraph beginning “With over100million,” and deputy executive director's comment.
\nG5: 5Rights' 17 June response welcomes the call but criticizes its reliance on digital literacy, parental controls and content removal. Leanda Barrington-Leach demands premarket safety certification for digital products/services likely to be used by or affect children. That is a concrete proposed market-access condition, but the response neither identifies a selected certifier nor demonstrates adoption. Its assertion of industry influence is advocacy interpretation; no company submission or negotiated edit is supplied to establish that causal attribution. [5Rights original](http[local research file] International Affairs Team, paragraph preceding Barrington-Leach's quotation and the quotation itself.
\nNo named NGO appears in the operative G1/G2 texts. UNICEF's claimed evidence provision is stronger than shared vocabulary, but weaker than clause-level uptake. 5Rights' reaction supplies an explicit example of divergence between an advocate's desired control mechanism and the actual outcome. Do not create an adopted-certification or industry-authored-clause edge from either account.
\n29 May political agreement versus17June document export: the contemporaneous UK DSIT announcement says ministers agreed the shared principles at that day's Paris meeting. It places possible UK under16 restrictions in a then-closed domestic consultation, not inside the G7 agreement. [UK original](http[local research file] published29May, opening paragraphs. UNICEF's1June statement independently dates its response before the summit. G3 explicitly dates the ministerial declaration's adoption to29May.
\nThe currently served G1 English PDF has17June creation/modification metadata; its French counterpart is explicitly marked translation17June. These are current official-hosted copies of the ministerial principles, not demonstrated archival bytes from29May. The English PDF's Word-author field says “LEPESQUEUX Marine”; G2's says “Warda OUZEMMAL.” Such fields may identify an export account or editor and do not prove substantive authorship or approval authority. No personal identity/author-control node should be created from them. Exact draft changes between May and June remain untested.
\nG2 is the later leaders' call, which welcomes the ministers' principles. It must not be merged with them as a single document. The French translation and German-hosted English G2 are counterparts, not independent corroboration of who drafted it. G3 is subsequent explicit reference/qualification by a different group of officials. These three document identities matter more than the shared G7 label.
\nIDs are provisional. Reuse existing organization IDs; no canonical merge performed.
\n| Candidate | \nRelation | \nEvidence state / action stage | \nImportant boundary | \n
|---|---|---|---|
| assurance-g7-e1 | \nG7 digital ministers → G1: adopt shared principles | \nG1 plus contemporaneous UK account; documented statement / adopted29May | \nCurrent PDF bytes exported17June | \n
| assurance-g7-e2 | \nG7 leaders → G1: welcome ministerial principles | \nG2 PDF2/3; documented / adopted policy statement17June | \nNo new domestic statute | \n
| assurance-g7-e3 | \nG7 leaders → conversational-AI providers: call for age assurance/default protections | \nG2 PDF2 second bullet; documented / recommendation | \nNo chosen method, vendor or compulsory universal identity gate | \n
| assurance-g7-e4 | \nG7 leaders → digital ministers: request continuing meetings/year-end assessment | \nG2 PDF3; documented / announced review obligation within political process | \nNo evidence review has occurred | \n
| assurance-g7-e5 | \nG7 privacy authorities → G1: expressly reference/qualify assurance implementation | \nG3 §§3–7; documented / adopted guidance25–26June | \nNot an enforcement order or proven technical safeguard | \n
| assurance-g7-e6 | \nG7 privacy authorities → ISO27566: note framework | \nG3 §6; documented / reference | \nNot mandatory certification or vendor accreditation | \n
| assurance-g7-e7 | \nUNICEF → digital ministers: reports providing advance evidence | \nG4; reported / claimed completed input | \nOriginal package, recipient receipt and clause uptake absent | \n
| assurance-g7-e8 | \n5Rights → governments: demands premarket safety certification | \nG5; documented advocacy / proposed condition | \nAbsent from inspected adopted text; operator unspecified | \n
The decisive instrument is the implementation plan UNICEF calls agreed, together with any adopted country-specific rule or platform implementation notice giving it effect. The test is concrete: covered service; trigger; required evidence; decision-maker; discretion to refuse or override; legal source; privacy restrictions; appeal and enforcement. A copy of the plan, approval minutes and task assignments would reveal whether this remains coordinated encouragement or becomes a compulsory condition. This note does not infer that a plan must contain all those elements or that it has already been enacted.
\nFor provenance, requestable holder records would be UNICEF's transmitted evidence and recipient acknowledgment, negotiating drafts, and the French presidency's/digital ministers' plan and review papers. No outreach occurred. Public searches for the specific implementation plan and NGO input returned UNICEF's claim and 5Rights' reaction, but no plan or attributable adopted clause. Existing independent domestic policy histories remain a plausible explanation for convergent language.
\nAcquisition stops at this instrument-level comparison. Newly surfaced Franco-German digital-majority/European blueprint material, commercial webinars and national legal implementation leads were not followed; a new branch awaits the root's forest assessment. Closed UN/UNESCO routes were not retried.
\nAll links are public primary publisher/government records, read9September2026. G1English, G1French and G2English were fetched by ordinary HTTPGET and parsed with pypdf; all three pages of G1English and G2English were read. G1French was read to check counterpart terminology/date and substantive provisions. G2French, G3 and G4/G5 substantive bodies were read in the web reader; UK opening/digital-policy paragraphs were read. Whole-text term checks supplemented, but did not replace, reading. No page-image interpretation is asserted. Only this note was retained; hashes below identify response bytes inspected in memory and are not claims of saved original-file custody.
\n| Record | \nHTTP200 bytes | \nSHA-256 | \n
|---|---|---|
| G1English | \n66352 | \n51bcee69d5c81c5b2f158d615f9e46df10c729a24e7febaaf895fa6c27e04837 | \n
| G1French | \n274724 | \nfe9359a6ce62ed632928391a73d75ecbd769411203b8c14a6c2e0291ac00bde1 | \n
| G2English | \n280787 | \nf2d7945c33bf7d071fab8e7c7d121580a0feeae615663efe1fad8f96301110b8 | \n
| G2French | \n182453 | \n8cbd67ec475162fc4e1672050027dc7ac590cbaeb2c40f9f4a06e71b771f53a0 | \n
| G3OPC | \n24336 | \nb2cbc1ad8ec83200c432f648d3c2fe6df69a31a75cb606ec78c13911b7ec392c | \n
| G4UNICEF | \n204863 | \n641e56d9998ef8fd7f452ebf97d05ddc0b2c2c6aa3d9a5a1244171e942f08f13 | \n
| G5FiveRights | \n189556 | \nb752297563c8579c318caabacc5b14fd57ac615fc2554697c4c3757f2251c3fb | \n
| UKDSIT | \n78902 | \n3a0352a7169f7c697dc69d4b3b8e1fd4d12cc60e042e2702dab008a479e9457e | \n
Observed 2026-09-09. The required whole-case review in forest-reviews.md was read before this pass. Only this file was written; no canonical case changes. Research stops here pending the root's whole-case reassessment.
The documentary mechanism is a jointly considered research competition with differentiated award conditions. It gives the supporting parties a way to choose research and attach conditions to their awards. It does not establish regulatory authority, a deployed admission gate, paid partner shares, or a common body controlling all AI agents. The clearest new instrument is the call-specific FAQ's optional Google.org funding route; public announcement wording alone conceals that distinction.
\nGoogle DeepMind's own announcement, dated 11 June 2026, describes a worldwide technical-research fund of up to US$10 million alongside Schmidt Sciences, ARIA and CAIF, supported by Google.org. It expects awardees in autumn 2026. This is an advertised ceiling, not evidence of an executed US$10 million grant or money received. The page discloses neither partner amounts nor payment dates. [DeepMind announcement](http[local research file]
\nCAIF's announcement likewise uses the ceiling. Its article displays 11 June 2026, while the site's listing showed 16 July 2026; retain the conflict rather than invent a second launch. CAIF's foundation page marked applications closed when inspected. Its footer identifies charity 1201294 and company 13485176. Neither page identifies the joint fund as a separate legal entity. [CAIF announcement](http[local research file] [foundation page](http[local research file]
\nThe Schmidt-branded application portal and Schmidt contact establish an administrative front door. They do not establish Schmidt as sole grantor. Named brands, particularly Google.org and Google DeepMind, should not be silently substituted for an award's contracting legal entity. No executed award or contribution agreement was acquired. [Actual application portal](http[local research file]
\nThe portal offers Tier 1 awards up to US$300,000 and Tier 2 awards of US$300,000–1 million for 1–2 years. Its selection criteria cover agenda fit, scientific rigour, impact, philanthropic fit, feasibility, expertise and cost. Joint supporting parties consider applications; selected funders may add limitations or guidance, agreed in final award documents. Lobbying, legislative influence and political activity are excluded. Schmidt-funded indirect costs are capped at 10%.
\nThe portal table gives 9 August 2026, 11:59pm AoE, consistent with its UI's 10 August, 07:59am EDT. The announcements instead say 8 August AoE. No acquired change history explains the one-day difference. Opening is shown as 11 June. This preserves observed versions without asserting why the deadline changed. [Portal, Timeline, Funding Levels, Eligibility and Selection Criteria](http[local research file]
\nThe FAQ is linked directly from the application portal. These are public descriptions of intended conditions, not an acquired executed agreement.
\n| Route / issue | \nTerms disclosed | \n
|---|---|
| Eligibility | \nCharitable/scientific/educational recipients; commercial participants cannot lead. Independent researchers need an eligible host or fiscal sponsor before funding. Schmidt uses NGOsource for non-US equivalency. | \n
| Schmidt funding | \nAnnual gifts/grants depend on progress and reporting; subsequent years are not assured. IP normally stays with awardees unless the agreement says otherwise. | \n
| Google.org option | \nApplicants may opt out and remain eligible for Schmidt, ARIA and CAIF consideration. Google.org-funded code/inventions require Apache 2.0; other specified IP requires CC-BY 4.0, within three months. Remaining IP rights go into the public domain; patent applications are prohibited. Data must be freely searchable, subject to privacy law. Funded projects must comply with Google AI Principles. | \n
| Limits | \nConditional renewal is not a demonstrated termination power over already-paid funds. Neither this FAQ nor the inspected call identifies reviewers, recusal rules, voting shares, publication-approval procedures or executed allocations. | \n
Locators: export lines 9–22 eligibility; 55–75 funding and international administration; 101–103 IP; 105–121 Google funding conditions. The FAQ includes some general Schmidt wording referring to approximately three years; the specific call says 1–2 years, which governs the duration reported here. [Call FAQ](http[local research file] [public text export inspected](http[local research file]
\nThe linked Google principles currently address innovation, responsible development/deployment and collaboration, including human oversight, safety testing, privacy and intellectual property. The call FAQ does not identify a dated edition, change-control rule, dispute process or a party's exclusive interpretation right. Reading this webpage does not supply those missing contractual terms. [Google AI Principles, substantive principles and implementation sections](http[local research file]
\nSchmidt's general FAQ identifies Schmidt Sciences LLC, a 501(c)(3) nonprofit subsidiary of The Eric and Wendy Schmidt Fund for Strategic Innovation. The displayed EIN 46-3460261 belongs to that parent; it must not be assigned to the LLC. General overhead policy permits none without written agreement; where allowed, the normal ceiling is 10%. Thus the call's cap is not an automatic entitlement. Its general policy also discusses protected enterprise AI use during proposal processing; this is not proof that AI evaluated this competition. These are organization-wide policies, not acquired award clauses. [Schmidt general FAQ, organization, indirect-cost and AI-use questions](http[local research file]
\nNo separate ARIA announcement or award agreement assigning its contribution was acquired in this pass. ARIA's separate Scaling Trust programme amounts must not be treated as its share of this joint call. Nor does listing Google DeepMind as a joint supporting party establish that it supplies cash separately from Google.org.
\nSchmidt explicitly credits DeepMind's Distributional AGI Safety work, ARIA's Scaling Trust thesis and CAIF's Multi-Agent Risks from Advanced AI report as inspirations. The agenda includes identity/authentication, admission controls, proof of agent/human credentials, platform vetting and revocation as research topics. It prioritizes strengthening existing infrastructure partly because incumbent protocols may become entrenched. These are concrete agenda choices relevant to the assurance inquiry.
\nCounterevidence matters: the programme excludes commercial product development, purely nontechnical policy work, and solutions that depend on one privileged principal controlling every agent. It funds research on coordination among independently controlled agents; an identity research topic is not a universal identity requirement. No acquired record selects an operational gatekeeper or compels deployment. [Schmidt research agenda, introductory inspirations, Agent Infrastructure and Out of Scope](http[local research file]
\n| Source | \nActual scope inspected / state | \n
|---|---|
| CAIF foundation | \nScoped current-call block and footer; web reader. No whole-site knowledge claimed. | \n
| CAIF announcement | \nFull 72-line web rendering, including displayed date, amounts, named parties and timetable. | \n
| DeepMind announcement | \nFull substantive announcement in 160-line web rendering; navigation not used as evidence. | \n
| Schmidt portal | \nAll substantive public call text through selection criteria in a 268-line rendering; second copy of the body is duplicate page output. No private application submitted or applicant account entered. | \n
| Call-specific FAQ | \nEntire public text export read, 14,895 bytes, HTTP 200, text/plain UTF-8. Export SHA-256: aa91f1a4a06b83602f38639e5711e1d3412f1ea947f5371699fddc43d7a16c94. Public document UI yielded a JavaScript shell; ordinary public export succeeded as the single substitute. Bytes were read in memory and are not retained as an original file. No publication/revision date visible. Line locators refer to this exact export. | \n
| Schmidt agenda | \nSubstantive programme introduction, agenda, research topics and exclusions. The alternate /scaling-ai-safety-for-a-multi-agent-world/ page substantially repeats the agenda; it is not an independent witness. | \n
| Schmidt general FAQ | \nOrganization identity, indirect-cost and AI-use sections. Generic policy scope retained. | \n
| Google AI Principles | \nSubstantive principles and implementation sections, web lines 124–169. Current linked text only; no historical incorporated version inferred. | \n
Web renderings are inspected source views, not locally preserved originals; no original-file hashes are invented for them. Public webinar links were identified but their recordings were not watched. Closed CAIF accounts and other prior blocked routes were not reopened.
\nSuitable relations are: the named parties jointly consider applications under this call; Schmidt hosts the application front door; Google.org offers a conditional funding route; applicants can decline that route without leaving the competition; the agenda attributes inspiration to three named research programmes/documents. Each belongs to the research-funding layer. Do not convert these into paid amounts, veto rights, government endorsement of findings, operational authority or donor instructions over a particular result.
\nThe next decisive records are the joint selection/conflicts protocol, the contribution or administration agreement identifying contracting entities and committed shares, and an award-specific agreement plus disbursement record. Schmidt's programme administrator is the likely first custodian of the selection process; each actual award's grantor and recipient would hold its conditions and receipts. These are record targets only: no outreach, further branch or acquisition follows this pass before the whole-case reassessment.
", "path": "assurance-wave-2026-09-09/multiagent-fund.md", "sourceIds": [ "w9-assure-s-fund-deepmind", "w9-assure-s-fund-caif", "w9-assure-s-fund-portal", "w9-assure-s-fund-faq", "w9-assure-s-fund-agenda", "w9-assure-s-schmidt-faq", "w9-assure-s-google-principles" ], "markdown": "Observed 2026-09-09. The required whole-case review in `forest-reviews.md` was read before this pass. Only this file was written; no canonical case changes. Research stops here pending the root's whole-case reassessment.\n\nThe documentary mechanism is a jointly considered research competition with differentiated award conditions. It gives the supporting parties a way to choose research and attach conditions to their awards. It does not establish regulatory authority, a deployed admission gate, paid partner shares, or a common body controlling all AI agents. The clearest new instrument is the call-specific FAQ's optional Google.org funding route; public announcement wording alone conceals that distinction.\n\n## Amount, parties, administration and timing\n\nGoogle DeepMind's own announcement, dated 11 June 2026, describes a worldwide technical-research fund of **up to US$10 million** alongside Schmidt Sciences, ARIA and CAIF, supported by Google.org. It expects awardees in autumn 2026. This is an advertised ceiling, not evidence of an executed US$10 million grant or money received. The page discloses neither partner amounts nor payment dates. DeepMind announcement (see source references)\n\nCAIF's announcement likewise uses the ceiling. Its article displays 11 June 2026, while the site's listing showed 16 July 2026; retain the conflict rather than invent a second launch. CAIF's foundation page marked applications closed when inspected. Its footer identifies charity 1201294 and company 13485176. Neither page identifies the joint fund as a separate legal entity. CAIF announcement (see source references) foundation page (see source references)\n\nThe Schmidt-branded application portal and Schmidt contact establish an administrative front door. They do not establish Schmidt as sole grantor. Named brands, particularly Google.org and Google DeepMind, should not be silently substituted for an award's contracting legal entity. No executed award or contribution agreement was acquired. Actual application portal (see source references)\n\n## Public call and selection\n\nThe portal offers Tier 1 awards up to US$300,000 and Tier 2 awards of US$300,000–1 million for 1–2 years. Its selection criteria cover agenda fit, scientific rigour, impact, philanthropic fit, feasibility, expertise and cost. Joint supporting parties consider applications; selected funders may add limitations or guidance, agreed in final award documents. Lobbying, legislative influence and political activity are excluded. Schmidt-funded indirect costs are capped at 10%.\n\nThe portal table gives **9 August 2026, 11:59pm AoE**, consistent with its UI's **10 August, 07:59am EDT**. The announcements instead say 8 August AoE. No acquired change history explains the one-day difference. Opening is shown as 11 June. This preserves observed versions without asserting why the deadline changed. Portal, Timeline, Funding Levels, Eligibility and Selection Criteria (see source references)\n\n## Call-linked FAQ: differentiated award conditions\n\nThe FAQ is linked directly from the application portal. These are public descriptions of intended conditions, not an acquired executed agreement.\n\n| Route / issue | Terms disclosed |\n| --- | --- |\n| Eligibility | Charitable/scientific/educational recipients; commercial participants cannot lead. Independent researchers need an eligible host or fiscal sponsor before funding. Schmidt uses NGOsource for non-US equivalency. |\n| Schmidt funding | Annual gifts/grants depend on progress and reporting; subsequent years are not assured. IP normally stays with awardees unless the agreement says otherwise. |\n| Google.org option | Applicants may opt out and remain eligible for Schmidt, ARIA and CAIF consideration. Google.org-funded code/inventions require Apache 2.0; other specified IP requires CC-BY 4.0, within three months. Remaining IP rights go into the public domain; patent applications are prohibited. Data must be freely searchable, subject to privacy law. Funded projects must comply with Google AI Principles. |\n| Limits | Conditional renewal is not a demonstrated termination power over already-paid funds. Neither this FAQ nor the inspected call identifies reviewers, recusal rules, voting shares, publication-approval procedures or executed allocations. |\n\nLocators: export lines 9–22 eligibility; 55–75 funding and international administration; 101–103 IP; 105–121 Google funding conditions. The FAQ includes some general Schmidt wording referring to approximately three years; the specific call says 1–2 years, which governs the duration reported here. Call FAQ (see source references) public text export inspected (see source references)\n\nThe linked Google principles currently address innovation, responsible development/deployment and collaboration, including human oversight, safety testing, privacy and intellectual property. The call FAQ does not identify a dated edition, change-control rule, dispute process or a party's exclusive interpretation right. Reading this webpage does not supply those missing contractual terms. Google AI Principles, substantive principles and implementation sections (see source references)\n\n## Generic policies kept separate\n\nSchmidt's general FAQ identifies **Schmidt Sciences LLC**, a 501(c)(3) nonprofit subsidiary of **The Eric and Wendy Schmidt Fund for Strategic Innovation**. The displayed EIN 46-3460261 belongs to that parent; it must not be assigned to the LLC. General overhead policy permits none without written agreement; where allowed, the normal ceiling is 10%. Thus the call's cap is not an automatic entitlement. Its general policy also discusses protected enterprise AI use during proposal processing; this is not proof that AI evaluated this competition. These are organization-wide policies, not acquired award clauses. Schmidt general FAQ, organization, indirect-cost and AI-use questions (see source references)\n\nNo separate ARIA announcement or award agreement assigning its contribution was acquired in this pass. ARIA's separate Scaling Trust programme amounts must not be treated as its share of this joint call. Nor does listing Google DeepMind as a joint supporting party establish that it supplies cash separately from Google.org.\n\n## What the agenda can and cannot show\n\nSchmidt explicitly credits DeepMind's Distributional AGI Safety work, ARIA's Scaling Trust thesis and CAIF's Multi-Agent Risks from Advanced AI report as inspirations. The agenda includes identity/authentication, admission controls, proof of agent/human credentials, platform vetting and revocation as research topics. It prioritizes strengthening existing infrastructure partly because incumbent protocols may become entrenched. These are concrete agenda choices relevant to the assurance inquiry.\n\nCounterevidence matters: the programme excludes commercial product development, purely nontechnical policy work, and solutions that depend on one privileged principal controlling every agent. It funds research on coordination among independently controlled agents; an identity research topic is not a universal identity requirement. No acquired record selects an operational gatekeeper or compels deployment. Schmidt research agenda, introductory inspirations, Agent Infrastructure and Out of Scope (see source references)\n\n## Source custody and reading scope\n\n| Source | Actual scope inspected / state |\n| --- | --- |\n| CAIF foundation | Scoped current-call block and footer; web reader. No whole-site knowledge claimed. |\n| CAIF announcement | Full 72-line web rendering, including displayed date, amounts, named parties and timetable. |\n| DeepMind announcement | Full substantive announcement in 160-line web rendering; navigation not used as evidence. |\n| Schmidt portal | All substantive public call text through selection criteria in a 268-line rendering; second copy of the body is duplicate page output. No private application submitted or applicant account entered. |\n| Call-specific FAQ | Entire public text export read, 14,895 bytes, HTTP 200, text/plain UTF-8. Export SHA-256: `aa91f1a4a06b83602f38639e5711e1d3412f1ea947f5371699fddc43d7a16c94`. Public document UI yielded a JavaScript shell; ordinary public export succeeded as the single substitute. Bytes were read in memory and are **not retained as an original file**. No publication/revision date visible. Line locators refer to this exact export. |\n| Schmidt agenda | Substantive programme introduction, agenda, research topics and exclusions. The alternate `/scaling-ai-safety-for-a-multi-agent-world/` page substantially repeats the agenda; it is not an independent witness. |\n| Schmidt general FAQ | Organization identity, indirect-cost and AI-use sections. Generic policy scope retained. |\n| Google AI Principles | Substantive principles and implementation sections, web lines 124–169. Current linked text only; no historical incorporated version inferred. |\n\nWeb renderings are inspected source views, not locally preserved originals; no original-file hashes are invented for them. Public webinar links were identified but their recordings were not watched. Closed CAIF accounts and other prior blocked routes were not reopened.\n\n## Narrow integration and stopping point\n\nSuitable relations are: the named parties **jointly consider applications** under this call; Schmidt **hosts the application front door**; Google.org **offers a conditional funding route**; applicants **can decline that route without leaving the competition**; the agenda **attributes inspiration to three named research programmes/documents**. Each belongs to the research-funding layer. Do not convert these into paid amounts, veto rights, government endorsement of findings, operational authority or donor instructions over a particular result.\n\nThe next decisive records are the **joint selection/conflicts protocol**, the **contribution or administration agreement identifying contracting entities and committed shares**, and an **award-specific agreement plus disbursement record**. Schmidt's programme administrator is the likely first custodian of the selection process; each actual award's grantor and recipient would hold its conditions and receipts. These are record targets only: no outreach, further branch or acquisition follows this pass before the whole-case reassessment.\n", "file": "research/assurance-wave-2026-09-09-multiagent-fund.md" }, { "id": "assurance-wave-2026-09-09-root-findings", "title": "Assurance markets: implementation and incentives", "description": "Observed 9 September 2026. Scope: Fathom public strategy, workshop reports and evaluation-industry interviews. The California instrument is separately assessed in fathom-bill.md.", "html": "Observed 9 September 2026. Scope: Fathom public strategy, workshop reports and evaluation-industry interviews. The California instrument is separately assessed in fathom-bill.md.
\nFathom explicitly seeks legislation to create demand for an independent AI-verification market. Its model combines public outcome requirements, accredited private evaluation and legal/commercial benefits for passing products. Its philanthropy-only funding description is self-report, not a completed donor audit. Source: http[local research file] (substantive approach read).
\nThe 16 March Paris workshop account describes a February 2026 IASEAI event and Gillian Hadfield keynote. Its published full report contains a simulation, not an executed regulatory arrangement. PDF pages 9-11 identify unresolved revocation/appeal rights, transparency versus proprietary methods, and incentives when developers pay evaluators. Participants explicitly compare this to credit-rating failures: limited verifier competition can make revocation difficult; government, evaluator and developer can shift responsibility between themselves. The report also records participants who thought the structure improved responses to harm. These are organizer-reported simulation findings, not evidence of an actual cover-up, crisis operation or captured regulator. Sources: http[local research file] and http[local research file] . Landing page substantive body read; PDF targeted pages 9-11 plus scoped extraction of other sections. Original retained with hash in originals-manifest.json; extraction contains spaced lettering.
\nAshby2026 connects voluntary verification to insurance pricing, coverage and liability, explicitly describing these as incentives for uptake. Its children workshop proposes an independent AI-rating institution with access to proprietary developer data; it acknowledges unresolved age-verification/First Amendment tensions. Its separate fraud workshop advocates identity verification and cross-sector data exchange at high-risk interactions. These are workshop proposals, not established institutions or a single operational identity system. Read PDF pages 8-10 and relevant extracted sections. Sources: http[local research file] and http[local research file] . Original retained, hash in manifest. Do not infer participants endorsed every proposal or reproduce unsourced factual claims inside the report.
\nThe 21 July evaluator article interviews Rumman Chowdhury of Humane Intelligence PBC, Shea Brown of BABL AI and Patrick Sullivan of A-LIGN. It identifies test design, access rights, commercial relationships and comparability as unresolved practical issues, advocating professional standards, accreditation, independent oversight and liability insurance. Interviewed aspirations are not adopted industry rules. The PBC must not be silently merged with the similarly named nonprofit on a partnership roster. Source: http[local research file] (full substantive article read; landing http[local research file] is same source family).
\nThere is a concrete economic mechanism worth investigating: a formally optional certificate can become commercially difficult to avoid if insurers, procurement officers or litigants rely on it. This is a supported design intention and an open implementation question. Actual policy endorsements, procurement clauses, insurer schedules and declined-coverage decisions would establish whether it operates, by whose standards and with what contestability. We have not acquired those instruments in this run. Public-interest conviction, market-building and an institution's desire for a role can coexist. None alone demonstrates illicit control.
", "path": "assurance-wave-2026-09-09/root-findings.md", "sourceIds": [ "w9-assure-s-fathom-approach", "w9-assure-s-paris-report", "w9-assure-s-ashby-report" ], "markdown": "Observed 9 September 2026. Scope: Fathom public strategy, workshop reports and evaluation-industry interviews. The California instrument is separately assessed in fathom-bill.md.\n\nFathom explicitly seeks legislation to create demand for an independent AI-verification market. Its model combines public outcome requirements, accredited private evaluation and legal/commercial benefits for passing products. Its philanthropy-only funding description is self-report, not a completed donor audit. Source: http[local research file] (substantive approach read).\n\nThe 16 March Paris workshop account describes a February 2026 IASEAI event and Gillian Hadfield keynote. Its published full report contains a simulation, not an executed regulatory arrangement. PDF pages 9-11 identify unresolved revocation/appeal rights, transparency versus proprietary methods, and incentives when developers pay evaluators. Participants explicitly compare this to credit-rating failures: limited verifier competition can make revocation difficult; government, evaluator and developer can shift responsibility between themselves. The report also records participants who thought the structure improved responses to harm. These are organizer-reported simulation findings, not evidence of an actual cover-up, crisis operation or captured regulator. Sources: http[local research file] and http[local research file] . Landing page substantive body read; PDF targeted pages 9-11 plus scoped extraction of other sections. Original retained with hash in originals-manifest.json; extraction contains spaced lettering.\n\nAshby2026 connects voluntary verification to insurance pricing, coverage and liability, explicitly describing these as incentives for uptake. Its children workshop proposes an independent AI-rating institution with access to proprietary developer data; it acknowledges unresolved age-verification/First Amendment tensions. Its separate fraud workshop advocates identity verification and cross-sector data exchange at high-risk interactions. These are workshop proposals, not established institutions or a single operational identity system. Read PDF pages 8-10 and relevant extracted sections. Sources: http[local research file] and http[local research file] . Original retained, hash in manifest. Do not infer participants endorsed every proposal or reproduce unsourced factual claims inside the report.\n\nThe 21 July evaluator article interviews Rumman Chowdhury of Humane Intelligence PBC, Shea Brown of BABL AI and Patrick Sullivan of A-LIGN. It identifies test design, access rights, commercial relationships and comparability as unresolved practical issues, advocating professional standards, accreditation, independent oversight and liability insurance. Interviewed aspirations are not adopted industry rules. The PBC must not be silently merged with the similarly named nonprofit on a partnership roster. Source: http[local research file] (full substantive article read; landing http[local research file] is same source family).\n\n## Whole-case implication\n\nThere is a concrete economic mechanism worth investigating: a formally optional certificate can become commercially difficult to avoid if insurers, procurement officers or litigants rely on it. This is a supported design intention and an open implementation question. Actual policy endorsements, procurement clauses, insurer schedules and declined-coverage decisions would establish whether it operates, by whose standards and with what contestability. We have not acquired those instruments in this run. Public-interest conviction, market-building and an institution's desire for a role can coexist. None alone demonstrates illicit control.\n", "file": "research/assurance-wave-2026-09-09-root-findings.md" }, { "id": "australia-decision-wave-2026-09-08-encryption-findings", "title": "Encryption-guidance origins: funding, review access and the missing wording record", "description": "The new evidence identifies a funded research route and a real toolkit consultation process. It does not identify the person or organisation that supplied eSafety's encrypted-environment recommendations. This lane made no canonical case edits, outreach or public submissions.", "html": "The new evidence identifies a funded research route and a real toolkit consultation process. It does not identify the person or organisation that supplied eSafety's encrypted-environment recommendations. This lane made no canonical case edits, outreach or public submissions.
\nThe toolkit labelled v 4 March 2026's printed24/PDF13 places footnote29 immediately after its homomorphic-encryption item. The full list has seven items: client-side scanning, homomorphic encryption, risk-based E2EE deployment, conditional encryption limits, interventions, signal-based detection and limiting viral spread. All are framed as approaches industry should consider. Footnote34 elsewhere supports SaferAI-related investigative triage and cannot provide ancestry for this list.
The footnote29 URL points to the 61-page December2024 report Guiding principles for addressing technology-facilitated child sexual exploitation and abuse, by Anri van der Spuy, Sabine Witting, Patrick Burton, Emma Day, Sonia Livingstone and Kim R. Sylwander. The report cover and p61 citation identify a Digital Futures for Children centre publication; the LSE repository calls it a technical report. The toolkit citation instead includes Security and Safety, vol.4. That is a bibliographic discrepancy, not yet proof of an incorrectly linked document.
\nThe report's actual relevant content is more discriminating than a keyword absence. P34 discusses offenders moving to end-to-end encrypted messaging. The toolkit accurately points to that page separately in footnote26, printed22/PDF12. P28 warns that protective interventions can enable surveillance harmful to children's rights; pp50-51 call for proportionate assessment across rights, including privacy, and consultation with technologists and other experts. No homomorphic-encryption recommendation or technical support was identified in the scoped reading or exact-phrase search. This establishes an unresolved support gap for that specific technical claim. It does not establish that the report never discusses encryption, that eSafety's claim is technically false, or that the whole list originates with the report.
\nThe first prefix search for encrypt returned no hits, but full-word encryption found a bibliography entry and reading p34 found encrypted messaging. Prefix absence was therefore not used as evidence. The UNICEF2020 encryption paper cited at reportp57 was not read or promoted into a new origin claim.
Reportp5 identifies the research as part of Toolbox for Measuring Online Child Sexual Exploitation and Abuse, funded through REPHRAIN's fourth Strategic Funding Call. It identifies UKRI as REPHRAIN's funder. P8 says Amy Orben of Cambridge led the project and this report is work package3. The current DFC project page independently documents the publisher's account of project funding and participating institutions; it shares project ancestry with the report and is not an independent audit of the money.
\nThe report's back cover separately acknowledges 5Rights Foundation funding of the DFC centre, describing DFC as a joint LSE/5Rights research centre. Centre support is not evidence of a report-specific earmark. No amount, disbursement, grant conditions, manuscript-approval right or publication veto was acquired.
\nP5 names Mariesa Nicholas, Australian e-Safety Commissioner's Office, among four independent peer reviewers, alongside Hazel Bitaña, Michael Salter and Fabio Senne. This establishes named review participation and contemporaneous institutional affiliation. It does not establish current employment, the substance of her comments, corporate endorsement, or authorship of the toolkit's encryption sentences. The six cover authors remain named authors without allocating particular passages to individuals.
\nThe official eSafety FOI25139 / LOG118 document set contains a May2025 industry consultation chain. The consequential readable record is Document19, PDF29-31, which preserves the 5May invitation and a 16May response from an @x.com correspondent. The invitation says the toolkit had been developed since the December2024 Safety by Design CSEA workshop, using workshop insights. Its draft core-content list already includes end-to-end encryption and AI. That establishes the subject's presence in a May2025 draft outline, not the four particular recommendations or a December2024 version of those sentences.
\nThe email identifies AttachmentA, draft core content, and AttachmentB, practical tools, seeks feedback with particular attention to B, offers marked-up written edits or a 20May online discussion, and announces another opportunity for final-draft redlines before a then-anticipated September launch. These are offered process steps, not proof every step occurred. The sender signs as eSafety's Industry Insights and Enablement Team.
\nX's 16May reply says it has no additional feedback and looks forward to publication. This establishes draft review access and that particular response. It does not show X supplied wording, approved the final2026 text or had no other communications. The draft attachment bodies and any comment-disposition records were not acquired. Doc13's first page was incompletely exposed by the reader, so this result relies on the repeated invitation in Doc19 rather than pretending the whole original email was read there. Redacted names were not reconstructed.
\nThe best next record is now concrete: May2025 AttachmentA and AttachmentB, the final-draft redlines, comment dispositions, footnote29 source entry and publication clearance, held by eSafety's toolkit team. These would identify when the encrypted-environment language entered the draft, which participant suggested or amended it, and whether footnote29 was an editing error, indirect principle citation, or intended to point somewhere else. A participant-specific insertion and accepted comment would establish textual influence; a pre-existing paragraph with an editorial citation error would support ordinary drafting/citation failure instead.
\nFor the money route, the complementary record is the REPHRAIN fourth-call award and work-package budget. That can distinguish ordinary research funding from deliverable-specific approval rights. For the reviewer route, DFC's invitation, comments and author response can distinguish access from substantive changes. Neither inquiry can be replaced by general biographies or shared affiliations.
\nRead/capture occurred 8September2026 America/Chicago, after midnight9SeptemberUTC. The original researchonline PDF GET returned403; no retry. The public eprints cached reader supplied scoped report text. The original FOI PDF GET timed out after20seconds; no retry. Two publisher pages were retained as original HTTP200 bodies. The three PDF reading extracts have derivative hashes in encryption-capture-manifest.json and are not represented as original PDFs.
\nThe 5Rights launch is dated4December2024 and links a /126219/1/ file; eSafety links /126219/3/. The /1/ reader failed and was not retried. The numeric path components are recorded as different file locators, not authenticated version numbers; no byte or revision comparison is claimed. Repository metadata records deposit3December2024 and published-version status. Screenshot calls did not yield an image available for inspection here, so footnote placement and the title comparison are grounded in parsed original-page text plus repository metadata, with no visual-verification claim.
The candidate is structurally separate from case-r8. Context nodes retain existing IDs. New records use w5-enc. The toolkit label is retained literally; the exact publication day is not independently established. The fresh toolkit source is another scoped observation of the same official document, not independent corroboration of its technical assertions.
Source URLs: [Report](http[local research file] [LSE metadata](http[local research file] [DFC project](http[local research file] [5Rights launch](http[local research file] [eSafety toolkit](http[local research file] [released FOI correspondence](http[local research file]
", "path": "australia-decision-wave-2026-09-08/encryption-findings.md", "sourceIds": [ "w5-enc-s-paper", "w5-enc-s-record", "w5-enc-s-dfc", "w5-enc-s-launch" ], "markdown": "The new evidence identifies a funded research route and a real toolkit consultation process. It does not identify the person or organisation that supplied eSafety's encrypted-environment recommendations. This lane made no canonical case edits, outreach or public submissions.\n\n## What the cited report supports\n\nThe toolkit labelled `v 4 March 2026`'s printed24/PDF13 places footnote29 immediately after its homomorphic-encryption item. The full list has seven items: client-side scanning, homomorphic encryption, risk-based E2EE deployment, conditional encryption limits, interventions, signal-based detection and limiting viral spread. All are framed as approaches industry should consider. Footnote34 elsewhere supports SaferAI-related investigative triage and cannot provide ancestry for this list.\n\nThe footnote29 URL points to the 61-page December2024 report **Guiding principles for addressing technology-facilitated child sexual exploitation and abuse**, by Anri van der Spuy, Sabine Witting, Patrick Burton, Emma Day, Sonia Livingstone and Kim R. Sylwander. The report cover and p61 citation identify a Digital Futures for Children centre publication; the LSE repository calls it a technical report. The toolkit citation instead includes *Security and Safety, vol.4*. That is a bibliographic discrepancy, not yet proof of an incorrectly linked document.\n\nThe report's actual relevant content is more discriminating than a keyword absence. P34 discusses offenders moving to end-to-end encrypted messaging. The toolkit accurately points to that page separately in footnote26, printed22/PDF12. P28 warns that protective interventions can enable surveillance harmful to children's rights; pp50-51 call for proportionate assessment across rights, including privacy, and consultation with technologists and other experts. No homomorphic-encryption recommendation or technical support was identified in the scoped reading or exact-phrase search. This establishes an unresolved support gap for that specific technical claim. It does not establish that the report never discusses encryption, that eSafety's claim is technically false, or that the whole list originates with the report.\n\nThe first prefix search for `encrypt` returned no hits, but full-word `encryption` found a bibliography entry and reading p34 found encrypted messaging. Prefix absence was therefore not used as evidence. The UNICEF2020 encryption paper cited at reportp57 was not read or promoted into a new origin claim.\n\n## Who funded and contributed\n\nReportp5 identifies the research as part of **Toolbox for Measuring Online Child Sexual Exploitation and Abuse**, funded through **REPHRAIN's fourth Strategic Funding Call**. It identifies UKRI as REPHRAIN's funder. P8 says Amy Orben of Cambridge led the project and this report is work package3. The current DFC project page independently documents the publisher's account of project funding and participating institutions; it shares project ancestry with the report and is not an independent audit of the money.\n\nThe report's back cover separately acknowledges **5Rights Foundation funding of the DFC centre**, describing DFC as a joint LSE/5Rights research centre. Centre support is not evidence of a report-specific earmark. No amount, disbursement, grant conditions, manuscript-approval right or publication veto was acquired.\n\nP5 names **Mariesa Nicholas, Australian e-Safety Commissioner's Office**, among four independent peer reviewers, alongside Hazel Bitaña, Michael Salter and Fabio Senne. This establishes named review participation and contemporaneous institutional affiliation. It does not establish current employment, the substance of her comments, corporate endorsement, or authorship of the toolkit's encryption sentences. The six cover authors remain named authors without allocating particular passages to individuals.\n\n## Released correspondence supplies a specific drafting route\n\nThe official **eSafety FOI25139 / LOG118** document set contains a May2025 industry consultation chain. The consequential readable record is **Document19, PDF29-31**, which preserves the 5May invitation and a 16May response from an @x.com correspondent. The invitation says the toolkit had been developed since the December2024 Safety by Design CSEA workshop, using workshop insights. Its draft core-content list already includes end-to-end encryption and AI. That establishes the subject's presence in a May2025 draft outline, not the four particular recommendations or a December2024 version of those sentences.\n\nThe email identifies **AttachmentA, draft core content**, and **AttachmentB, practical tools**, seeks feedback with particular attention to B, offers marked-up written edits or a 20May online discussion, and announces another opportunity for final-draft redlines before a then-anticipated September launch. These are offered process steps, not proof every step occurred. The sender signs as eSafety's Industry Insights and Enablement Team.\n\nX's 16May reply says it has no additional feedback and looks forward to publication. This establishes draft review access and that particular response. It does not show X supplied wording, approved the final2026 text or had no other communications. The draft attachment bodies and any comment-disposition records were not acquired. Doc13's first page was incompletely exposed by the reader, so this result relies on the repeated invitation in Doc19 rather than pretending the whole original email was read there. Redacted names were not reconstructed.\n\n## The separating observation\n\nThe best next record is now concrete: **May2025 AttachmentA and AttachmentB, the final-draft redlines, comment dispositions, footnote29 source entry and publication clearance**, held by eSafety's toolkit team. These would identify when the encrypted-environment language entered the draft, which participant suggested or amended it, and whether footnote29 was an editing error, indirect principle citation, or intended to point somewhere else. A participant-specific insertion and accepted comment would establish textual influence; a pre-existing paragraph with an editorial citation error would support ordinary drafting/citation failure instead.\n\nFor the money route, the complementary record is the REPHRAIN fourth-call award and work-package budget. That can distinguish ordinary research funding from deliverable-specific approval rights. For the reviewer route, DFC's invitation, comments and author response can distinguish access from substantive changes. Neither inquiry can be replaced by general biographies or shared affiliations.\n\n## Provenance and limits\n\nRead/capture occurred 8September2026 America/Chicago, after midnight9SeptemberUTC. The original researchonline PDF GET returned403; no retry. The public eprints cached reader supplied scoped report text. The original FOI PDF GET timed out after20seconds; no retry. Two publisher pages were retained as original HTTP200 bodies. The three PDF reading extracts have derivative hashes in **encryption-capture-manifest.json** and are not represented as original PDFs.\n\nThe 5Rights launch is dated4December2024 and links a `/126219/1/` file; eSafety links `/126219/3/`. The `/1/` reader failed and was not retried. The numeric path components are recorded as different file locators, not authenticated version numbers; no byte or revision comparison is claimed. Repository metadata records deposit3December2024 and published-version status. Screenshot calls did not yield an image available for inspection here, so footnote placement and the title comparison are grounded in parsed original-page text plus repository metadata, with no visual-verification claim.\n\nThe candidate is structurally separate from case-r8. Context nodes retain existing IDs. New records use `w5-enc`. The toolkit label is retained literally; the exact publication day is not independently established. The fresh toolkit source is another scoped observation of the same official document, not independent corroboration of its technical assertions.\n\nSource URLs: Report (see source references) LSE metadata (see source references) DFC project (see source references) 5Rights launch (see source references) eSafety toolkit (see source references) released FOI correspondence (see source references)\n", "file": "research/australia-decision-wave-2026-09-08-encryption-findings.md" }, { "id": "australia-decision-wave-2026-09-08-funding-findings", "title": "Research funding and the monitoring proposal", "description": "The new mechanism is resource support plus named advisory access to a published recommendation. Childlight Searchlight2025 Study C (PDF31–32, printed23–24) credits Westpac funding and names a committee including Westpac's Mark Dickson and ICMEC Australia's Danielle Kelly. It recommends broader financial monitoring and cross-sector sharing. This is a publishe", "html": "The new mechanism is resource support plus named advisory access to a published recommendation. Childlight Searchlight2025 Study C (PDF31–32, printed23–24) credits Westpac funding and names a committee including Westpac's Mark Dickson and ICMEC Australia's Danielle Kelly. It recommends broader financial monitoring and cross-sector sharing. This is a published recommendation, not a demonstrated detection system or bank decision. The report acknowledges panel recruitment, self-report and legal-category limitations; no base-rate calibration or operational validation was recovered in the inspected material.
\nWestpac's 2021 recipients page lists the UNSW Human Rights Institute OCSE survey project in a $150,000-per-award research category, over up to two years. AUD is context-inferred. The 2023 JSS report (p56) thanks Westpac and names the researchers and contributions; the 2025 report links that earlier Australian report as related work. These are distinct outputs. No signed award schedule connects the full 2021 amount to Study C, and none allocates it from the ICMEC $25m award. We therefore keep a separate project node and do not draw an allocated-funding edge between award and later study.
\nThe study also identifies support for UNSW's Childlight hub through an Edinburgh collaboration agreement under Human Dignity Foundation core grant INT21-01. Global acknowledgements thank JSS for funding two studies. Payment flow, amounts and any double-counting remain unresolved; the case records the disclosed layers separately without summing them. Searchlight's programme wording is Safer Families, Safer Communities, whereas Westpac's page uses Safer Children, Safer Communities; retain the wording difference rather than silently rewriting the source.
\nThe most useful next records are the executed award output schedule, committee terms/minutes and author response-to-review table. They would test whether advisers shaped the monitoring recommendation, whether the funder had clearance rights, and what outputs the award actually bought. Sources name ordinary public-health/financial-crime research roles as well as an influence opportunity. Neither funding nor advisory listing settles bad faith, censorship motive or implementation.
\nCustody: funding-capture.json lists original URLs, byte sizes and SHA-256 values. Reading scope is recorded in each candidate source. PDF32 funding attribution was visually checked against a rendered original. No holder contact or acquisition payment occurred. Westpac FY22 working-group membership search result was read as discovery context only; no new canonical claim is based on that duplicate/general lane.
", "path": "australia-decision-wave-2026-09-08/funding-findings.md", "sourceIds": [ "w5-fund-s-childlight", "w5-fund-s-jss2023", "w5-fund-s-westpac-grants" ], "markdown": "The new mechanism is resource support plus named advisory access to a published recommendation. Childlight Searchlight2025 Study C (PDF31–32, printed23–24) credits Westpac funding and names a committee including Westpac's Mark Dickson and ICMEC Australia's Danielle Kelly. It recommends broader financial monitoring and cross-sector sharing. This is a published recommendation, not a demonstrated detection system or bank decision. The report acknowledges panel recruitment, self-report and legal-category limitations; no base-rate calibration or operational validation was recovered in the inspected material.\n\nWestpac's 2021 recipients page lists the UNSW Human Rights Institute OCSE survey project in a $150,000-per-award research category, over up to two years. AUD is context-inferred. The 2023 JSS report (p56) thanks Westpac and names the researchers and contributions; the 2025 report links that earlier Australian report as related work. These are distinct outputs. No signed award schedule connects the full 2021 amount to Study C, and none allocates it from the ICMEC $25m award. We therefore keep a separate project node and do not draw an allocated-funding edge between award and later study.\n\nThe study also identifies support for UNSW's Childlight hub through an Edinburgh collaboration agreement under Human Dignity Foundation core grant INT21-01. Global acknowledgements thank JSS for funding two studies. Payment flow, amounts and any double-counting remain unresolved; the case records the disclosed layers separately without summing them. Searchlight's programme wording is Safer Families, Safer Communities, whereas Westpac's page uses Safer Children, Safer Communities; retain the wording difference rather than silently rewriting the source.\n\nThe most useful next records are the executed award output schedule, committee terms/minutes and author response-to-review table. They would test whether advisers shaped the monitoring recommendation, whether the funder had clearance rights, and what outputs the award actually bought. Sources name ordinary public-health/financial-crime research roles as well as an influence opportunity. Neither funding nor advisory listing settles bad faith, censorship motive or implementation.\n\nCustody: funding-capture.json lists original URLs, byte sizes and SHA-256 values. Reading scope is recorded in each candidate source. PDF32 funding attribution was visually checked against a rendered original. No holder contact or acquisition payment occurred. Westpac FY22 working-group membership search result was read as discovery context only; no new canonical claim is based on that duplicate/general lane.\n", "file": "research/australia-decision-wave-2026-09-08-funding-findings.md" }, { "id": "australia-decision-wave-2026-09-08-governance-findings", "title": "Australian governance: the board has published decision rights; the July 2023 votes remain unidentified", "description": "The new records locate a concrete approval chain inside ICMEC Australia. Its published Board Charter, adopted and effective 6 December 2023, gives its board approval over strategic direction and budgets, major capital expenditure and partnerships, and material funding submissions, new programs and external reporting. Management develops and implements plans ", "html": "The new records locate a concrete approval chain inside ICMEC Australia. Its published Board Charter, adopted and effective 6 December 2023, gives its board approval over strategic direction and budgets, major capital expenditure and partnerships, and material funding submissions, new programs and external reporting. Management develops and implements plans within board-set limits; matters exceeding those limits return for approval. This is a documented governance mechanism in a period before the November 2024 SaferAI paper. It does not prove that any particular paper, grant or product decision reached the board, or how anyone voted. [Board Charter, PDF3–4/printed2–3, clauses2.1–2.3; PDF6/printed5,4.2; PDF7/printed6,5.4](http[local research file]
\nThe new route was the live homepage's actual Governance link, /icmec-australia-governance/. This is a different URL from the previous failed /governance/ guess. That independently captured page linked the current constitution, charter and policies. Its public WordPress API, discovered from the homepage's rel=api.w.org link, disclosed earlier constitution attachments. No ACNC route, proxy or mirror was used.
\nBoth a document headed ICMEC Australia Constitution – July 2023 and the current March 2025 text exempt ICMEC US from the usual membership-renewal requirement in clause15.2. Each specifically defines the US corporation by EIN22-3630133. This is affirmative evidence of a continuing named place in the constitutional scheme. It is not the membership register: the exemption alone cannot establish that the US body actually remained a member on any chosen date, that it exercised a vote, or that it retained control. The audit's loss-of-sole-membership/control finding on6July2023 remains compatible with a retained, non-controlling membership. [July2023 publication, PDF6 clause15.2 and PDF23 clause73](http[local research file] [March2025 publication, PDF6 clause15.2 and PDF20 clause73](http[local research file]
\nIn both texts, the board admits applicants, the secretary enters them in the member register, and each member has one vote. Members elect directors and may remove them; the board may fill vacancies under the specified conditions. The board manages the company and records delegations. The register, admission decisions, representative appointments and election resolutions therefore reveal actual governance power; website lists of partner or coalition members do not. Neither inspected appointment scheme gives a named donor a reserved appointment or veto. An undisclosed grant agreement or separate instrument could contain other rights. July2023 locators: PDF5–6 clauses9–15; PDF13 clause33; PDF14–16 clauses39–45; PDF19 clauses58–60. March2025 counterparts: PDF5–6,12–15,17–18.
\n| Original | \nPublisher metadata and observed contents | \nWhat it can establish | \n
|---|---|---|
| ICMEC-AU-Constitution-2023.pdf, media602 | \nUploaded10September2024; PDF created27July2023; headers July2023;24 pages | \nThe July-labeled text and its rights, including15.2. Creation/upload timestamps are not an adoption resolution. | \n
| FOR-WEBSITE-ICMEC-AU-Proposed-Constitution-27-July-2023.pdf, media603 | \nUploaded10September2024; PDF created29September2023; title says Proposed/27July2023; cover says last updated July2022;30 pages | \nSame renewal exception at PDF7; unresolved internal version labels. Do not treat it as executed on27July or6July. | \n
| ICMEC-AU-Constitution.pdf, media609 | \nUploaded11September2024; metadata calls it a reformatted proposed27July2023 copy;29 pages | \nReorganized numbering: same exception3.6.2 at PDF8; no separately established adoption date. | \n
| ICMEC-AU-Constitution-1.pdf, media1410 | \nUploaded13May2025; PDF created12May2025; title says Jul2023 updated Mar2025;21 pages | \nCurrently linked publisher constitution, with March2025 headers; not proof of its exact amendment vote or of every earlier clause. | \n
These are related publisher versions, not four independent confirmations. The media API also lists a second May2025 file (media1409), which was discovered but not fetched. No earlier-membership names or resolution establishing who replaced or joined the sole member were recovered. The decisive date gap is the adopted text and resolutions effective6July2023, not whether a file with 2023 in its name exists.
\nThe current board page lists Jodie Arthur, Westpac General Manager Financial Crime and Group MLRO, as an Australian non-executive director; Jeremy Moller is listed as board chair. This refreshes the prior report-period role observation. It supplies a direct institutional entry point to the board approval process, while leaving appointment basis, representational authority and participation in particular decisions unresolved. A director's external employer does not inherit the board's powers. [Current board page, Jodie Arthur and Jeremy Moller sections](http[local research file]
\nThe published Conflict of Interest Policy places the register of interests with the company secretary. It calls for disclosures, management decisions and their voting results to be recorded. The current constitution separately requires material-conflict disclosure and ordinarily excludes an interested director from discussion/voting, subject to explicit exceptions. These are published rules, not proof of compliance or recusal in a specific Westpac decision. [Policy, PDF4–5,5.2–6.2](http[local research file] current constitution PDF16,50.1–50.4.
\nThe already retained US2023 audit, PDF3/printed2, bears GRF CPAs & Advisors, grfcpa.com, and says the Australian component auditor's statements/reports for the period ended6July2023 were furnished to it. That makes GRF a specifically evidenced recipient of the underlying report. It does not name the Australian auditor. The US audit remains a USD consolidated disclosure; no new cash transfer or Australian expense allocation is inferred.
\nThe Australian secretary is the constitutionally specified holder of the membership register and minutes, including admission, member and director resolutions (July text PDF19,58–60; March text PDF17–18). The governance page supplies info@icmec.org.au for policy questions. The already captured FY2024–25 Impact Report, PDF18/printed35, names Griselda Afonso, Senior Manager Accounts and Office Manager, a report-period finance-route identifier. The PDF author metadata also names her, but metadata does not prove legal authorship, secretary status, audit responsibility or personal custody of a particular record. No named secretary, Australian component auditor or personal email was established.
\nThe strongest rival to donor-directed policy is an Australian board exercising its own judgment, with a bank executive contributing sector expertise, a legacy non-controlling US relationship, and conflicts handled under its published procedures. A donor-direction explanation predicts specific nomination or reserved rights, funding conditions, instructions, draft clearance, or documented participation in linked approvals. The public rules locate where those mechanisms could operate; they do not show them being exercised.
\nThe highest-value bundle is the membership-register extract showing names and start/end dates around6July2023, the adopted constitution and amendment/admission/representative/election resolutions, and the Australian component financial statements/report supplied to GRF for that period. For the donor-to-program chain, pair the executed Westpac grant/extension with the relevant board approval papers, delegated-authority schedule, interests declarations and recusal/vote minutes. Inspect institutional roles and dates; unnecessary personal addresses need not be collected.
\nThe component report and standalone Australian financials were not obtained. Public exact-entity/identifier/record-family searches and the publisher's media search for audit returned no relevant standalone report; financial returned a symposium item. These bounded metadata results do not establish absence from the website or absence of accounts. Broad search incidentally returned ACNC titles; none was opened, fetched or used as financial evidence. The governance page's linked June2026 Non-Conduit Policy PDF returned404 once; no retry or inferred contents. The web reader's cache miss for the new governance URL was followed by a successful direct publisher GET, whose original is retained; no restricted surface was involved.
\nRead: July2023 publication's clauses9–15,33,39–45,58–60,73, with related context on the cited pages; March2025 corresponding provisions plus39.1,47,49–50; the complete seven-page charter; conflict policy5.2–6.2 and surrounding policy; current board role panels; API metadata and version-label passages of the two proposed/reformatted copies. The renewal clauses in July2023/March2025, the charter's program/management approvals, and the GRF header/component paragraph were visually checked. Other extracted text is retained but is not declared wholly reviewed. Originals, derived text and hashes are in governance-capture-manifest.json. No outreach, payment, publication or canonical-case mutation occurred.
", "path": "australia-decision-wave-2026-09-08/governance-findings.md", "sourceIds": [ "w5-gov-s-page", "w5-gov-s-constitution2025", "w5-gov-s-charter", "w5-gov-s-conflicts", "w5-gov-s-board", "w5-gov-s-media-constitution", "w5-gov-s-constitution2023", "w5-gov-s-constitution-reformatted", "w5-gov-s-constitution-proposed", "w5-gov-s-media-audit", "w5-gov-s-media-financial", "w5-gov-s-nonconduit", "w5-gov-s-constitution2025-alt" ], "markdown": "The new records locate a concrete approval chain inside ICMEC Australia. Its published Board Charter, adopted and effective **6 December 2023**, gives its board approval over strategic direction and budgets, major capital expenditure and partnerships, and material funding submissions, new programs and external reporting. Management develops and implements plans within board-set limits; matters exceeding those limits return for approval. This is a documented governance mechanism in a period before the November 2024 SaferAI paper. It does not prove that any particular paper, grant or product decision reached the board, or how anyone voted. Board Charter, PDF3–4/printed2–3, clauses2.1–2.3; PDF6/printed5,4.2; PDF7/printed6,5.4 (see source references)\n\nThe new route was the live homepage's actual Governance link, **/icmec-australia-governance/**. This is a different URL from the previous failed /governance/ guess. That independently captured page linked the current constitution, charter and policies. Its public WordPress API, discovered from the homepage's rel=api.w.org link, disclosed earlier constitution attachments. No ACNC route, proxy or mirror was used.\n\n## The former parent retains a named constitutional exception\n\nBoth a document headed **ICMEC Australia Constitution – July 2023** and the current **March 2025** text exempt ICMEC US from the usual membership-renewal requirement in clause15.2. Each specifically defines the US corporation by EIN22-3630133. This is affirmative evidence of a continuing named place in the constitutional scheme. It is not the membership register: the exemption alone cannot establish that the US body actually remained a member on any chosen date, that it exercised a vote, or that it retained control. The audit's loss-of-sole-membership/control finding on6July2023 remains compatible with a retained, non-controlling membership. July2023 publication, PDF6 clause15.2 and PDF23 clause73 (see source references) March2025 publication, PDF6 clause15.2 and PDF20 clause73 (see source references)\n\nIn both texts, the board admits applicants, the secretary enters them in the member register, and each member has one vote. Members elect directors and may remove them; the board may fill vacancies under the specified conditions. The board manages the company and records delegations. The register, admission decisions, representative appointments and election resolutions therefore reveal actual governance power; website lists of partner or coalition members do not. Neither inspected appointment scheme gives a named donor a reserved appointment or veto. An undisclosed grant agreement or separate instrument could contain other rights. July2023 locators: PDF5–6 clauses9–15; PDF13 clause33; PDF14–16 clauses39–45; PDF19 clauses58–60. March2025 counterparts: PDF5–6,12–15,17–18.\n\n## Version custody keeps the date question honest\n\n| Original | Publisher metadata and observed contents | What it can establish |\n| --- | --- | --- |\n| ICMEC-AU-Constitution-2023.pdf, media602 | Uploaded10September2024; PDF created27July2023; headers July2023;24 pages | The July-labeled text and its rights, including15.2. Creation/upload timestamps are not an adoption resolution. |\n| FOR-WEBSITE-ICMEC-AU-Proposed-Constitution-27-July-2023.pdf, media603 | Uploaded10September2024; PDF created29September2023; title says Proposed/27July2023; cover says last updated July2022;30 pages | Same renewal exception at PDF7; unresolved internal version labels. Do not treat it as executed on27July or6July. |\n| ICMEC-AU-Constitution.pdf, media609 | Uploaded11September2024; metadata calls it a reformatted proposed27July2023 copy;29 pages | Reorganized numbering: same exception3.6.2 at PDF8; no separately established adoption date. |\n| ICMEC-AU-Constitution-1.pdf, media1410 | Uploaded13May2025; PDF created12May2025; title says Jul2023 updated Mar2025;21 pages | Currently linked publisher constitution, with March2025 headers; not proof of its exact amendment vote or of every earlier clause. |\n\nThese are related publisher versions, not four independent confirmations. The media API also lists a second May2025 file (media1409), which was discovered but not fetched. No earlier-membership names or resolution establishing who replaced or joined the sole member were recovered. The decisive date gap is **the adopted text and resolutions effective6July2023**, not whether a file with 2023 in its name exists.\n\n## A donor employee sits at the documented decision interface\n\nThe current board page lists **Jodie Arthur**, Westpac General Manager Financial Crime and Group MLRO, as an Australian non-executive director; **Jeremy Moller** is listed as board chair. This refreshes the prior report-period role observation. It supplies a direct institutional entry point to the board approval process, while leaving appointment basis, representational authority and participation in particular decisions unresolved. A director's external employer does not inherit the board's powers. Current board page, Jodie Arthur and Jeremy Moller sections (see source references)\n\nThe published Conflict of Interest Policy places the register of interests with the company secretary. It calls for disclosures, management decisions and their voting results to be recorded. The current constitution separately requires material-conflict disclosure and ordinarily excludes an interested director from discussion/voting, subject to explicit exceptions. These are published rules, not proof of compliance or recusal in a specific Westpac decision. Policy, PDF4–5,5.2–6.2 (see source references) current constitution PDF16,50.1–50.4.\n\n## The auditor and record holders are more precise\n\nThe already retained US2023 audit, PDF3/printed2, bears **GRF CPAs & Advisors**, grfcpa.com, and says the Australian component auditor's statements/reports for the period ended6July2023 were furnished to it. That makes GRF a specifically evidenced recipient of the underlying report. It does not name the Australian auditor. The US audit remains a USD consolidated disclosure; no new cash transfer or Australian expense allocation is inferred.\n\nThe Australian secretary is the constitutionally specified holder of the membership register and minutes, including admission, member and director resolutions (July text PDF19,58–60; March text PDF17–18). The governance page supplies info@icmec.org.au for policy questions. The already captured FY2024–25 Impact Report, PDF18/printed35, names **Griselda Afonso, Senior Manager Accounts and Office Manager**, a report-period finance-route identifier. The PDF author metadata also names her, but metadata does not prove legal authorship, secretary status, audit responsibility or personal custody of a particular record. No named secretary, Australian component auditor or personal email was established.\n\n## Rival and discriminating record\n\nThe strongest rival to donor-directed policy is an Australian board exercising its own judgment, with a bank executive contributing sector expertise, a legacy non-controlling US relationship, and conflicts handled under its published procedures. A donor-direction explanation predicts specific nomination or reserved rights, funding conditions, instructions, draft clearance, or documented participation in linked approvals. The public rules locate where those mechanisms could operate; they do not show them being exercised.\n\nThe highest-value bundle is the membership-register extract showing names and start/end dates around6July2023, the adopted constitution and amendment/admission/representative/election resolutions, and the Australian component financial statements/report supplied to GRF for that period. For the donor-to-program chain, pair the executed Westpac grant/extension with the relevant board approval papers, delegated-authority schedule, interests declarations and recusal/vote minutes. Inspect institutional roles and dates; unnecessary personal addresses need not be collected.\n\n## Bounded access and read scope\n\nThe component report and standalone Australian financials were not obtained. Public exact-entity/identifier/record-family searches and the publisher's media search for audit returned no relevant standalone report; financial returned a symposium item. These bounded metadata results do not establish absence from the website or absence of accounts. Broad search incidentally returned ACNC titles; none was opened, fetched or used as financial evidence. The governance page's linked June2026 Non-Conduit Policy PDF returned404 once; no retry or inferred contents. The web reader's cache miss for the new governance URL was followed by a successful direct publisher GET, whose original is retained; no restricted surface was involved.\n\nRead: July2023 publication's clauses9–15,33,39–45,58–60,73, with related context on the cited pages; March2025 corresponding provisions plus39.1,47,49–50; the complete seven-page charter; conflict policy5.2–6.2 and surrounding policy; current board role panels; API metadata and version-label passages of the two proposed/reformatted copies. The renewal clauses in July2023/March2025, the charter's program/management approvals, and the GRF header/component paragraph were visually checked. Other extracted text is retained but is not declared wholly reviewed. Originals, derived text and hashes are in governance-capture-manifest.json. No outreach, payment, publication or canonical-case mutation occurred.\n", "file": "research/australia-decision-wave-2026-09-08-governance-findings.md" }, { "id": "australia-seams-wave-2026-09-08-award-findings", "title": "Award instruments: bounded public-source return", "description": "Prepared 8 September 2026 (America/Chicago); retrieval timestamps fall on 9 September UTC. Candidate-only additions against case-r9. Acquisition frozen. No canonical edits, outreach or private access.", "html": "Prepared 8 September 2026 (America/Chicago); retrieval timestamps fall on 9 September UTC. Candidate-only additions against case-r9. Acquisition frozen. No canonical edits, outreach or private access.
\nThe public record now connects the fourth REPHRAIN call to the Cambridge-led Toolbox award and exposes an actual, separately credited Work Package 2 output. It does not expose the Toolbox's award value, payment schedule, final acceptance, publication-rights clause or split between work packages.
\nFor the Westpac-supported survey, the referenced OSF registration is now readable in both original and revised form. It expressly records registration after analysis, and therefore cannot establish an analysis plan fixed before data collection. The authors' methods account explains a scope expansion and recruitment arrangement, but supplies neither the grant agreement nor a per-output budget. This improves the money map by narrowing the unresolved instrument, rather than assigning existing large amounts to later publications.
\nThe centre announced on 18 January 2024 that four projects had received grants following competition; the Toolbox is one of them. UKRI's EP/V011189/1 record identifies a Cambridge collaboration under call four with a 2023 start year, and says its outputs feed the centre's research resources. The March2025 date on the current project page is a page date; it must not be substituted for an award start.
\nThe same UKRI record links the WP2 reliability-analysis page. That artifact is dated 1 October2024 and credits Matthew Richards, Sakshi Ghai and Amy Orben. Its rationale names an evaluation report and shareable analytical code, alongside intended updating with Cambridge/LSE/UNICEF teams through2025. This is a real published work product; it is separate from the retained December2024 DFC WorkPackage3 report. I read its identity and rationale, not its scientific validity or executable code.
\nUKRI's GBP6,972,599, funded September2020-August2024, belongs to the parent centre award. It supplies no basis for a Toolbox or DFC allocation. Two similarly titled UKRI outcome rows also differ on whether material was supplied to others; the public linked artifact prevents treating that metadata difference as evidence of concealed research. The rows are not two awards.
\nWhat would change interpretation: the Cambridge/Bristol award letter and cost schedule, the collaboration or subaward agreement, work-package budget, and actual output-acceptance/publication clauses. REPHRAIN/Bristol and Cambridge are the identifiable institutional holders. Announced funding and integration of outputs are established; donor editing, approval or veto is not.
\nChildlight2025 StudyC, PDF32/printed24, links http[local research file] and identifies HC220317**. An ordinary public API request recovered the registration and its linked response history.
\nThe registration timestamp is 10 November2023,16:41:26 UTC. Both original and revised answers to q8 / Existing Data say “Registration following analysis of the data”. The generic template name is OSF Preregistration; the actual selected answer governs this classification.
\nThe original response record 654e5d376caf8e05d96f32fe leaves measured variables and statistical models as TBC. Revision 65703b40788e5c03f74629f4, submitted 6 December2023, fills those sections, including internet banking, cryptocurrency and privacy-software measures. The current registration endpoint contains these revised responses; they must not be backdated to10 November. The revised model description cautions against causal inference from cross-sectional data.
\nNeither response version supplies grant amount, payment schedule, a funding-role statement or publication rights. Post-analysis registration is a declared evidentiary limitation, not proof of misconduct. It also cannot show that a bank selected the later financial-monitoring recommendations.
\nWhitten, Salter and Woodlock's 29 July2025 CrimRxiv methodology preprint, release1, reports an original design for3,000Australian men, Westpac funding, and advisory review. It says CloudResearch subsequently recruited/administered through PrimePanels after an earlier contractor declined the subject matter. This is the authors' account, not vendor correspondence or an executed service contract.
\nThe authors report expansion to UK/US samples after ethical and stakeholder approval at no extra cost, with fieldwork in November-December2022. Neither approving individuals nor the cost baseline is specified. The statement does not establish that later analyses, publication or institutional support cost nothing.
\nStudyC's4,918-person survey and protocol/ethics identifiers match this lineage. Another2025 chapter, Swipe wrong, links the same registration and funding acknowledgment (PDF43/printed35). Shared data and acknowledgments are not separate awards. No2021$150,000-to-2025StudyC budget identity, ICMEC$25m earmark, or HDF allocation is established. Survey-team/UNSW/Westpac grant and amendment records are the next useful records.
\nAll original successful responses and failed-route receipts are recorded in award-captures.jsonl. award-scoped-extracts.json contains derivative text locators and selected JSON records; derivative extracts are not mislabeled as originals. Existing Childlight and Westpac originals remain in the prior wave and were not redownloaded. The candidate references their stable source IDs.
\nREPHRAIN: Launch of Strategic Funding Call 4 Projects
\nhttp[local research file]
\nLocator: Main announcement dated 18 January 2024; project list
\nRead scope: Read entire main announcement; navigation and unrelated related posts excluded.
\nOriginal: australia-seams-wave-2026-09-08/award-rephrain-launch.html
\nSHA-256: c48a9ffbfde54c48bf9e454a627bc6e9d6f748e02078ce4b51c19a153f3ca037
\nUKRI Gateway to Research: REPHRAIN, EP/V011189/1
\nhttp[local research file]
\nLocator: Funding fields; collaboration row REPHRAIN Strategic Funding call 4 - University of Cambridge; two Toolbox research database/model rows. Derived text lines 33-48,1042-1070,1770-1784.
\nRead scope: Scoped read of funding header, Cambridge call-4 row and two Toolbox outcome rows. Did not review the centre-wide publication/partner roster or all claimed impacts.
\nOriginal: australia-seams-wave-2026-09-08/award-gtr-rephrain.html
\nSHA-256: 7bb844ac1adcb644ed797373823ce94d58aeb225d2a188eb1ea077fa4cfd1188
\nREPHRAIN Work Package 2, Project 1: Reliability Analyses
\nhttp[local research file]
\nLocator: Title/byline/date; section 0.1 Rationale, Objective, Method, Output. Derived text lines1-66.
\nRead scope: Read title/byline/date and section0.1; code/result body captured but not scientifically audited or executed.
\nOriginal: australia-seams-wave-2026-09-08/award-toolbox-wp2.html
\nSHA-256: 4566e1e82db81cd551981cd149953ba7a8f8233ed31b6682682f7720ecefc5a8
\nOSF registration j7xh2: The nature of online offending against children
\nhttp[local research file]
\nLocator: Public API http[local research file] ; data.attributes.date_registered, registered_meta, registration_responses (especially q8, q15.question, q17.question), relationships.schema_responses.
\nRead scope: Read all registration attributes and responses; relationship metadata used only to locate response history and schema.
\nOriginal: australia-seams-wave-2026-09-08/award-osf-registration.json
\nSHA-256: 5a506a840c1c8a44184e88b58f96bcfce2fbb56f9dfcacf90f4926b9a4bdbe75
\nOSF j7xh2 response revision history
\nhttp[local research file]
\nLocator: data ids654e5d376caf8e05d96f32fe (original) and65703b40788e5c03f74629f4 (revision); attributes.date_created/date_submitted/revision_responses/is_original_response.
\nRead scope: Read both response records, metadata and complete original/revised answers.
\nOriginal: australia-seams-wave-2026-09-08/award-osf-responses.json
\nSHA-256: 1111174d503cf066b771d173a070d0e46dac2609ffc16fd0d438d12f1d81d8c1
\nOSF Preregistration schema version3
\nhttp[local research file]
\nLocator: data.attributes.schema.pages: Sampling Plan / q8 Existing Data; q15 Measures; q17 Statistical models.
\nRead scope: Scoped read of the named schema questions and q8 option definitions; remaining template retained, not applied as instructions.
\nOriginal: australia-seams-wave-2026-09-08/award-osf-schema.json
\nSHA-256: e5549f4d090ca36f2c9265e2893623bd016e9d90f56b7e0b453a8bc7c9363c07
\nWhitten, Salter and Woodlock: study methodology, CrimRxiv release1
\nhttp[local research file]
\nLocator: Published29 July2025; DOI10.21428/cb6ab371.d5f32f4f; Methodology/participant recruitment paragraphs; publisher text lines57-100. Download http[local research file]
\nRead scope: Read web introduction/methodology/measures and limitations; funding/role/disclosure searches; retained publisher-linked plain-text export. Entire 60KB export retained but full results/reference list not audited.
\nOriginal: australia-seams-wave-2026-09-08/award-methodology-publisher.txt
\nSHA-256: 5dfc4d6aa9becdc8ef7d1f4e070f06f07826978dfd2f56517b4b8b86c3585b23
\nThe direct OSF web-reader route failed and its normal HTML route yielded a JavaScript shell; the public, unauthenticated registration API and linked response history succeeded. This is a completed substitute, not a protocol access gap.
\nThe CrimRxiv article's ordinary GET returned403; the web reader exposed its text and publisher download links. The linked publisher plain-text export returned200 and is retained with its own hash. Failed PDF/plain-text web-parser requests were not treated as missing originals after that download succeeded.
\nThe original long UKRI search URL failed in the web parser; an ordinary public search and the parent-grant page succeeded. No further search retry is needed.
\nThe current REPHRAIN funding page concerns a different2025call. A targeted search found the old fourth-call announcement and a secondary ResearchConnect summary, but no executed project terms. An institutional RISCS archive substitute failed; REPHRAIN's public search endpoint returned401. Those routes were closed without attempting authentication or transplanting another call's terms. Secondary call ceilings were not promoted into project amounts.
\nA PMC article route presented a browser challenge; it was not bypassed. The already readable methodology paper was sufficient for the bounded implementation point. No no-role-by-funder declaration was acquired. Absence of one in the read protocol/methods passages does not prove none exists elsewhere.
\nThe handoff contains4newnodes,5newedges and2updates to existing edges. It creates no new money-flow edge and changes no existing amount. The source families remain separate from independent corroboration: OSF template/current/history are one platform record; the author methods text and Childlight report share research ancestry; UKRI outcomes contain researcher reports rather than an independent acceptance audit.
", "path": "australia-seams-wave-2026-09-08/award-findings.md", "sourceIds": [ "w6-award-s-launch", "w6-award-s-gtr", "w6-award-s-wp2", "w6-award-s-registration", "w6-award-s-revisions", "w6-award-s-schema", "w6-award-s-methods" ], "markdown": "Prepared 8 September 2026 (America/Chicago); retrieval timestamps fall on 9 September UTC. Candidate-only additions against case-r9. Acquisition frozen. No canonical edits, outreach or private access.\n\n## Result\n\nThe public record now connects the fourth REPHRAIN call to the Cambridge-led Toolbox award and exposes an actual, separately credited Work Package 2 output. It does **not** expose the Toolbox's award value, payment schedule, final acceptance, publication-rights clause or split between work packages.\n\nFor the Westpac-supported survey, the referenced OSF registration is now readable in both original and revised form. It expressly records registration **after analysis**, and therefore cannot establish an analysis plan fixed before data collection. The authors' methods account explains a scope expansion and recruitment arrangement, but supplies neither the grant agreement nor a per-output budget. This improves the money map by narrowing the unresolved instrument, rather than assigning existing large amounts to later publications.\n\n## 1. REPHRAIN: award confirmation, output identity, missing instrument\n\nThe centre announced on **18 January 2024** that four projects had received grants following competition; the Toolbox is one of them. UKRI's EP/V011189/1 record identifies a Cambridge collaboration under call four with a **2023 start year**, and says its outputs feed the centre's research resources. The March2025 date on the current project page is a page date; it must not be substituted for an award start.\n\nThe same UKRI record links the WP2 reliability-analysis page. That artifact is dated **1 October2024** and credits **Matthew Richards, Sakshi Ghai and Amy Orben**. Its rationale names an evaluation report and shareable analytical code, alongside intended updating with Cambridge/LSE/UNICEF teams through2025. This is a real published work product; it is separate from the retained December2024 DFC WorkPackage3 report. I read its identity and rationale, not its scientific validity or executable code.\n\nUKRI's **GBP6,972,599**, funded September2020-August2024, belongs to the parent centre award. It supplies no basis for a Toolbox or DFC allocation. Two similarly titled UKRI outcome rows also differ on whether material was supplied to others; the public linked artifact prevents treating that metadata difference as evidence of concealed research. The rows are not two awards.\n\nWhat would change interpretation: the Cambridge/Bristol award letter and cost schedule, the collaboration or subaward agreement, work-package budget, and actual output-acceptance/publication clauses. REPHRAIN/Bristol and Cambridge are the identifiable institutional holders. Announced funding and integration of outputs are established; donor editing, approval or veto is not.\n\n## 2. OSF: exact registration chronology, not prospective planning\n\nChildlight2025 StudyC, PDF32/printed24, links **http[local research file] and identifies **HC220317**. An ordinary public API request recovered the registration and its linked response history.\n\nThe registration timestamp is **10 November2023,16:41:26 UTC**. Both original and revised answers to **q8 / Existing Data** say **“Registration following analysis of the data”**. The generic template name is OSF Preregistration; the actual selected answer governs this classification.\n\nThe original response record **654e5d376caf8e05d96f32fe** leaves measured variables and statistical models as TBC. Revision **65703b40788e5c03f74629f4**, submitted **6 December2023**, fills those sections, including internet banking, cryptocurrency and privacy-software measures. The current registration endpoint contains these revised responses; they must not be backdated to10 November. The revised model description cautions against causal inference from cross-sectional data.\n\nNeither response version supplies grant amount, payment schedule, a funding-role statement or publication rights. Post-analysis registration is a declared evidentiary limitation, not proof of misconduct. It also cannot show that a bank selected the later financial-monitoring recommendations.\n\n## 3. Survey implementation and later output budgets\n\nWhitten, Salter and Woodlock's **29 July2025 CrimRxiv methodology preprint**, release1, reports an original design for3,000Australian men, Westpac funding, and advisory review. It says CloudResearch subsequently recruited/administered through PrimePanels after an earlier contractor declined the subject matter. This is the authors' account, not vendor correspondence or an executed service contract.\n\nThe authors report expansion to UK/US samples after ethical and stakeholder approval **at no extra cost**, with fieldwork in November-December2022. Neither approving individuals nor the cost baseline is specified. The statement does not establish that later analyses, publication or institutional support cost nothing.\n\nStudyC's4,918-person survey and protocol/ethics identifiers match this lineage. Another2025 chapter, Swipe wrong, links the same registration and funding acknowledgment (PDF43/printed35). Shared data and acknowledgments are not separate awards. No2021$150,000-to-2025StudyC budget identity, ICMEC$25m earmark, or HDF allocation is established. Survey-team/UNSW/Westpac grant and amendment records are the next useful records.\n\n## Acquired evidence and exact read scope\n\nAll original successful responses and failed-route receipts are recorded in **award-captures.jsonl**. **award-scoped-extracts.json** contains derivative text locators and selected JSON records; derivative extracts are not mislabeled as originals. Existing Childlight and Westpac originals remain in the prior wave and were not redownloaded. The candidate references their stable source IDs.\n\n### w6-award-s-launch\n\nREPHRAIN: Launch of Strategic Funding Call 4 Projects\n\nhttp[local research file]\n\nLocator: Main announcement dated 18 January 2024; project list\n\nRead scope: Read entire main announcement; navigation and unrelated related posts excluded.\n\nOriginal: australia-seams-wave-2026-09-08/award-rephrain-launch.html\n\nSHA-256: c48a9ffbfde54c48bf9e454a627bc6e9d6f748e02078ce4b51c19a153f3ca037\n\n### w6-award-s-gtr\n\nUKRI Gateway to Research: REPHRAIN, EP/V011189/1\n\nhttp[local research file]\n\nLocator: Funding fields; collaboration row REPHRAIN Strategic Funding call 4 - University of Cambridge; two Toolbox research database/model rows. Derived text lines 33-48,1042-1070,1770-1784.\n\nRead scope: Scoped read of funding header, Cambridge call-4 row and two Toolbox outcome rows. Did not review the centre-wide publication/partner roster or all claimed impacts.\n\nOriginal: australia-seams-wave-2026-09-08/award-gtr-rephrain.html\n\nSHA-256: 7bb844ac1adcb644ed797373823ce94d58aeb225d2a188eb1ea077fa4cfd1188\n\n### w6-award-s-wp2\n\nREPHRAIN Work Package 2, Project 1: Reliability Analyses\n\nhttp[local research file]\n\nLocator: Title/byline/date; section 0.1 Rationale, Objective, Method, Output. Derived text lines1-66.\n\nRead scope: Read title/byline/date and section0.1; code/result body captured but not scientifically audited or executed.\n\nOriginal: australia-seams-wave-2026-09-08/award-toolbox-wp2.html\n\nSHA-256: 4566e1e82db81cd551981cd149953ba7a8f8233ed31b6682682f7720ecefc5a8\n\n### w6-award-s-registration\n\nOSF registration j7xh2: The nature of online offending against children\n\nhttp[local research file]\n\nLocator: Public API http[local research file] ; data.attributes.date_registered, registered_meta, registration_responses (especially q8, q15.question, q17.question), relationships.schema_responses.\n\nRead scope: Read all registration attributes and responses; relationship metadata used only to locate response history and schema.\n\nOriginal: australia-seams-wave-2026-09-08/award-osf-registration.json\n\nSHA-256: 5a506a840c1c8a44184e88b58f96bcfce2fbb56f9dfcacf90f4926b9a4bdbe75\n\n### w6-award-s-revisions\n\nOSF j7xh2 response revision history\n\nhttp[local research file]\n\nLocator: data ids654e5d376caf8e05d96f32fe (original) and65703b40788e5c03f74629f4 (revision); attributes.date_created/date_submitted/revision_responses/is_original_response.\n\nRead scope: Read both response records, metadata and complete original/revised answers.\n\nOriginal: australia-seams-wave-2026-09-08/award-osf-responses.json\n\nSHA-256: 1111174d503cf066b771d173a070d0e46dac2609ffc16fd0d438d12f1d81d8c1\n\n### w6-award-s-schema\n\nOSF Preregistration schema version3\n\nhttp[local research file]\n\nLocator: data.attributes.schema.pages: Sampling Plan / q8 Existing Data; q15 Measures; q17 Statistical models.\n\nRead scope: Scoped read of the named schema questions and q8 option definitions; remaining template retained, not applied as instructions.\n\nOriginal: australia-seams-wave-2026-09-08/award-osf-schema.json\n\nSHA-256: e5549f4d090ca36f2c9265e2893623bd016e9d90f56b7e0b453a8bc7c9363c07\n\n### w6-award-s-methods\n\nWhitten, Salter and Woodlock: study methodology, CrimRxiv release1\n\nhttp[local research file]\n\nLocator: Published29 July2025; DOI10.21428/cb6ab371.d5f32f4f; Methodology/participant recruitment paragraphs; publisher text lines57-100. Download http[local research file]\n\nRead scope: Read web introduction/methodology/measures and limitations; funding/role/disclosure searches; retained publisher-linked plain-text export. Entire 60KB export retained but full results/reference list not audited.\n\nOriginal: australia-seams-wave-2026-09-08/award-methodology-publisher.txt\n\nSHA-256: 5dfc4d6aa9becdc8ef7d1f4e070f06f07826978dfd2f56517b4b8b86c3585b23\n\n## Closed routes and negative frontier\n\nThe direct OSF web-reader route failed and its normal HTML route yielded a JavaScript shell; the public, unauthenticated registration API and linked response history succeeded. This is a completed substitute, not a protocol access gap.\n\nThe CrimRxiv article's ordinary GET returned403; the web reader exposed its text and publisher download links. The linked publisher plain-text export returned200 and is retained with its own hash. Failed PDF/plain-text web-parser requests were not treated as missing originals after that download succeeded.\n\nThe original long UKRI search URL failed in the web parser; an ordinary public search and the parent-grant page succeeded. No further search retry is needed.\n\nThe current REPHRAIN funding page concerns a different2025call. A targeted search found the old fourth-call announcement and a secondary ResearchConnect summary, but no executed project terms. An institutional RISCS archive substitute failed; REPHRAIN's public search endpoint returned401. Those routes were closed without attempting authentication or transplanting another call's terms. Secondary call ceilings were not promoted into project amounts.\n\nA PMC article route presented a browser challenge; it was not bypassed. The already readable methodology paper was sufficient for the bounded implementation point. No no-role-by-funder declaration was acquired. Absence of one in the read protocol/methods passages does not prove none exists elsewhere.\n\nThe handoff contains4newnodes,5newedges and2updates to existing edges. It creates **no new money-flow edge** and changes no existing amount. The source families remain separate from independent corroboration: OSF template/current/history are one platform record; the author methods text and Childlight report share research ancestry; UKRI outcomes contain researcher reports rather than an independent acceptance audit.\n", "file": "research/australia-seams-wave-2026-09-08-award-findings.md" }, { "id": "australia-seams-wave-2026-09-08-draft-findings", "title": "draft findings", "description": "The new public seam is Common Sense Media's dated review of a later eSafety draft. Its three-page comments, dated 31 October 2025, say the organisation participated in the December 2024 workshop and was answering a request to review Designing for Safety: Preventing Grievous Harm Before It Happens. Page 2 requests clearer age-assurance error rates, fallback p", "html": "The new public seam is Common Sense Media's dated review of a later eSafety draft. Its three-page comments, dated 31 October 2025, say the organisation participated in the December 2024 workshop and was answering a request to review Designing for Safety: Preventing Grievous Harm Before It Happens. Page 2 requests clearer age-assurance error rates, fallback procedures and data-minimisation practices, citing a New York proposed-rulemaking model. Page 3 supports safety education in engineering/company leadership and proactive detection. This is a specific contribution to trace; no recipient acknowledgment, accepted edit, payment or encryption-list authorship is established. All three pages were read, and the PDF's outgoing links were inspected for a draft link; none was identified. “End-to-end reporting flows” on page 2 is not E2EE guidance. [Original comments](http[local research file]
\nThe [publisher's Public Filings and Letters page](http[local research file] links the exact comments under both 2025 and International. Both the PDF and the listing HTML were captured by ordinary public GET with HTTP200. The PDF is 134,371 bytes, SHA-256 41c0f4bd80c977d2b8647edee6efd14ba2c5cd7ec9c216fd7b154ed7b5fbede2. The listing is 103,581 bytes, SHA-256 2319a52809690ceb8f8da2e3a667a4249c238846f621dfcb70a1d84e0b5fe46c. They are the same publisher family, not independent confirmation of receipt. The date printed on the letter is not proof of its public-upload date. The listing reading scope was the two relevant links only.
The [eSafety disclosure log](http[local research file] as retrieved through the search index, identifies Log118 as FOI25139, listed 5 September 2025: non-legal eSafety/X correspondence from 1 January to 23 June 2025; eighteen documents partly released and four wholly exempt. It lists redactions under sections22,37(2)(b),47E(d),47F, but provides no allocation to particular records in the inspected entry. The listing date is not treated as the decision date. Its general statement says documents are published for twelve months and older records remain available on request. Direct page-open returned InternalError; one ordinary GET timed out after twenty seconds. Only the indexed entry/retention statement was preserved, not original HTML. Those four unidentified exempt records must not be described as draft attachments. No full decision notice or document schedule was obtained. This is a release-summary observation, not legal advice on the validity of exemptions.
\nExact May draft filenames were recovered from search-indexed Document13, the 5 May 2025,8:08PM email header in the [existing LOG118 PDF](http[local research file] They are:
\nAttachment A - DRAFT CSEA SbD Toolkit - Industry feedback - May 2025.docx
Attachment B - DRAFT CSEA SbD Toolkit - Tools only - Industry feedback - May 2025.docx
The main web-reader still omitted Document13's first page, PDF20; the exact filenames are therefore specifically an indexed-primary-document observation. Existing Document19, PDF30–31, remains the read evidence for the content outline and markup invitation. Neither attached DOCX was acquired, and the outline does not establish any recommendation's wording. The changed URL query parameter is not evidence of a changed PDF version. Preserve source ID w5-enc-s-foi and append this narrower observation; do not duplicate the source or invent an original PDF hash.
Bounded searches covered the exact AttachmentA/B filenames, the distinctive draft title, FOI25139/LOG118 plus decision terms, and a Right to Know domain query. Exact filename results returned the official email record, not a downloadable draft. The genuinely new public path was the Common Sense Media comments and their publisher listing. No later redline, disposition table, decision schedule or underlying October draft emerged within that search boundary. This is not an exhaustive absence finding.
\nThe timing matters: October2025 lies outside FOI25139's January–June2025 window. The new comments demonstrate a later review opportunity but cannot establish when particular final text appeared. Three obtainable record families are now precisely identified: the two named May DOCX files with returned edits; FOI25139's decision and schedule explaining attachment treatment; and the October review invitation, circulated draft, receipt and comment-disposition records. The separating observation is a dated version/redline with an attributed change and editorial disposition. A refusal, an already-existing passage or a rejected suggestion would narrow the influence claim.
\nStopping condition met: named records sharpened, release scope and unresolved exemptions recorded, and one new contributor-owned public path acquired and read. No unchanged GET retry of the toolkit or original FOI PDF, no outreach, submissions, payment or canonical edits. Toolkit technical-reference tracing belongs to the root lane.
\nLocal custody: draft-csm-comments.pdf, draft-csm-filings.html and draft-capture-manifest.json retain originals and hashes. draft-foi-filenames-index-extract.json and draft-disclosure-index-extract.json are expressly derivative captures. draft-candidate.json proposes four new nodes, seven edges, three new sources, one stable-source observation update and three leads. Context nodes are carried only for validation and must not replace canonical context. No hypothesis of funding or donor control is added.
The new public record is ICMEC US's signed 2023 Form 990, recovered from the live US annual-financial index. It supplies a specific inter-entity transaction category and a tax-return characterization that need reconciliation. It does not reveal who joined the Australian company, what resolution took effect on 6 July 2023, or donor reserved rights. Those membership questions have reached a bounded public-record frontier.
\nThe return covers calendar 2023; its officer signature is dated 6 December 2024 (PDF 1). This is the publisher-hosted signed public copy; no IRS receipt or acceptance was independently read. Schedule R Part IV (PDF 50), Australian row, lists ICMEC as direct controlling entity, C CORP as tax classification, 100% ownership and controlled-entity Yes. Its reported shares of income and end-of-year assets are USD 182,463 and USD 13,922,178. These are the filing's entries; they are not new proof of Australian legal ownership or of control continuing after 6 July.
\nSchedule R Part V (PDF 51), Australian row, reports USD 868,571 under transaction code L and valuation method FMV. Code L covers services, membership or fundraising solicitations for a related organization. Loan categories D and E are marked No. The amount therefore must not be called an Australian loan. It also cannot yet be treated as a cash payment, service fee received, grant, donor-directed expenditure or price for a specific project.
\nThe 2023 IRS instructions are the relevant interpretive record. Their PDF 6, Part V line 2, columns (c)–(d), defines the amount using fair market value of what the filer provided or received, whichever was higher. The return alone does not identify the underlying service, cash direction or agreement. The Australian and Singapore rows must remain separate. The return's USD 13,922,178 share of assets is not directly comparable to the audit's USD 13,519,626 net assets removed from consolidation; they differ in gross/net basis and their precise period alignment remains unresolved.
\nThe IRS's 2023 Schedule R instructions PDF 1, Relationships, require disclosure for relationships existing at any time during the tax year. Australia's presence on Schedule R is therefore compatible with the US audit's loss of control on 6 July 2023. Presence alone creates no contradiction.
\nThere is a narrower unresolved issue in the 100% entry. Instructions PDF 5, Part IV column (h), use the end of the related organization's tax year ending with or within the US filer's year; column (g) has the corresponding year-end asset rule. A different Australian tax year or a filing-basis explanation could matter. We must not invent one. Targeted reuse of the already captured July-2023 constitution, clause 66 (PDF 21), and March-2025 constitution, clause 66 (PDF 19), shows a default 1 January–31 December financial year, changeable by directors' resolution. These are reused sources, not new constitution findings, and still do not establish exact adoption on 6 July or absence of a year-change resolution.
\nThe strongest rival to an actual continued-control reading is a tax-reporting basis or timing difference, including the related entity's applicable period and completion method; stale or mistaken completion also remains possible. Neither can be selected from the return alone. The discriminating record is the Schedule R supporting workpaper identifying the Australian period, ownership/control basis and transaction L ledger, read with the executed membership-change instruments and Australian component statements. Do not allege a false return, continuing control, or a resolved audit contradiction.
\nDirect Australian publisher WordPress media searches for resolution, agm and membership, plus pages/posts searches for annual general, each returned HTTP 200 and an empty array. These are five bounded metadata queries at one publisher, not five independent negative confirmations and not proof the documents do not exist. Targeted external search results did not produce an actual admission/adoption instrument. The independent US annual-financial route produced the return, but the inspected Form 990 Part VI and Schedule O concern US governance and do not supply the Australian membership chronology.
\nPark further public acquisition of the membership branch until an actual member-register extract, admission/resignation/amendment resolution, signed adopted constitution, or equivalent transition instrument becomes publicly available through an authorized independent publisher. The Australian company secretary and board records remain the direct record holders; the US former-parent legal archive is a counterpart route. The return names Shawn Valentine as its books-and-records contact (PDF 6 line 20), a filing-period custodian selector rather than a verified current role. The component statements/report already acknowledged by the US auditor remain a concrete missing record, not a speculative document class.
\nNo ACNC route was accessed or retried. No paid search, outreach or policy-404 retry was performed. No fresh donor reserved-rights instrument was obtained. The absence of one from these routes does not establish that none exists.
\nOriginal bytes and SHA-256 hashes are in membership-capture-manifest.json. The US return's ancestry is the actual link on the live ICMEC US annual-financial index, not an ACNC source or mirror. The 2023 IRS instructions came directly from irs.gov. Australian metadata queries came directly from the publisher's established public API.
\nUS return: inspected PDF 1 and 6, PDF 45–48 Schedule O, and PDF 49–52 Schedule R; relevant PDF 1, 6, 48, 49, 50, 51 and 52 rendered and read, with PDF 47 also retained. Plain extraction was used because layout extraction split words into letters. Critical entries are on the original page 50 and 51 renders. IRS instructions: PDF 1 Relationships, PDF 2 control definitions, PDF 5 Part IV columns (d)–(i), and PDF 6 Part V line 2 read; no entire-document reading claim.
\nExact reused sources: w4-fin-s-fs2023 (audit findings from prior wave); w5-gov-s-constitution2023 clause 66/PDF 21 and w5-gov-s-constitution2025 clause 66/PDF 19 newly read for this interpretation. The current IRS Schedule R instruction webpage applies to 2024 and later; it was excluded as authority for the 2023 return, and the archived 2023 instructions were used instead.
", "path": "australia-seams-wave-2026-09-08/membership-findings.md", "sourceIds": [ "w6-member-s-financial-index", "w6-member-s-us9902023", "w6-member-s-irs2023", "w6-member-s-media-resolution", "w6-member-s-media-agm", "w6-member-s-media-membership", "w6-member-s-pages-agm", "w6-member-s-posts-agm" ], "markdown": "The new public record is ICMEC US's signed 2023 Form 990, recovered from the live US annual-financial index. It supplies a specific inter-entity transaction category and a tax-return characterization that need reconciliation. It does not reveal who joined the Australian company, what resolution took effect on 6 July 2023, or donor reserved rights. Those membership questions have reached a bounded public-record frontier.\n\n## What the new return establishes\n\nThe return covers calendar 2023; its officer signature is dated 6 December 2024 (PDF 1). This is the publisher-hosted signed public copy; no IRS receipt or acceptance was independently read. Schedule R Part IV (PDF 50), Australian row, lists ICMEC as direct controlling entity, C CORP as tax classification, 100% ownership and controlled-entity Yes. Its reported shares of income and end-of-year assets are USD 182,463 and USD 13,922,178. These are the filing's entries; they are not new proof of Australian legal ownership or of control continuing after 6 July.\n\nSchedule R Part V (PDF 51), Australian row, reports USD 868,571 under transaction code L and valuation method FMV. Code L covers services, membership or fundraising solicitations for a related organization. Loan categories D and E are marked No. The amount therefore must not be called an Australian loan. It also cannot yet be treated as a cash payment, service fee received, grant, donor-directed expenditure or price for a specific project.\n\nThe 2023 IRS instructions are the relevant interpretive record. Their PDF 6, Part V line 2, columns (c)–(d), defines the amount using fair market value of what the filer provided or received, whichever was higher. The return alone does not identify the underlying service, cash direction or agreement. The Australian and Singapore rows must remain separate. The return's USD 13,922,178 share of assets is not directly comparable to the audit's USD 13,519,626 net assets removed from consolidation; they differ in gross/net basis and their precise period alignment remains unresolved.\n\n## The period question remains real, but bounded\n\nThe IRS's 2023 Schedule R instructions PDF 1, Relationships, require disclosure for relationships existing at any time during the tax year. Australia's presence on Schedule R is therefore compatible with the US audit's loss of control on 6 July 2023. Presence alone creates no contradiction.\n\nThere is a narrower unresolved issue in the 100% entry. Instructions PDF 5, Part IV column (h), use the end of the related organization's tax year ending with or within the US filer's year; column (g) has the corresponding year-end asset rule. A different Australian tax year or a filing-basis explanation could matter. We must not invent one. Targeted reuse of the already captured July-2023 constitution, clause 66 (PDF 21), and March-2025 constitution, clause 66 (PDF 19), shows a default 1 January–31 December financial year, changeable by directors' resolution. These are reused sources, not new constitution findings, and still do not establish exact adoption on 6 July or absence of a year-change resolution.\n\nThe strongest rival to an actual continued-control reading is a tax-reporting basis or timing difference, including the related entity's applicable period and completion method; stale or mistaken completion also remains possible. Neither can be selected from the return alone. The discriminating record is the Schedule R supporting workpaper identifying the Australian period, ownership/control basis and transaction L ledger, read with the executed membership-change instruments and Australian component statements. Do not allege a false return, continuing control, or a resolved audit contradiction.\n\n## Membership search boundary and reopening condition\n\nDirect Australian publisher WordPress media searches for resolution, agm and membership, plus pages/posts searches for annual general, each returned HTTP 200 and an empty array. These are five bounded metadata queries at one publisher, not five independent negative confirmations and not proof the documents do not exist. Targeted external search results did not produce an actual admission/adoption instrument. The independent US annual-financial route produced the return, but the inspected Form 990 Part VI and Schedule O concern US governance and do not supply the Australian membership chronology.\n\nPark further public acquisition of the membership branch until an actual member-register extract, admission/resignation/amendment resolution, signed adopted constitution, or equivalent transition instrument becomes publicly available through an authorized independent publisher. The Australian company secretary and board records remain the direct record holders; the US former-parent legal archive is a counterpart route. The return names Shawn Valentine as its books-and-records contact (PDF 6 line 20), a filing-period custodian selector rather than a verified current role. The component statements/report already acknowledged by the US auditor remain a concrete missing record, not a speculative document class.\n\nNo ACNC route was accessed or retried. No paid search, outreach or policy-404 retry was performed. No fresh donor reserved-rights instrument was obtained. The absence of one from these routes does not establish that none exists.\n\n## Custody and read scope\n\nOriginal bytes and SHA-256 hashes are in membership-capture-manifest.json. The US return's ancestry is the actual link on the live ICMEC US annual-financial index, not an ACNC source or mirror. The 2023 IRS instructions came directly from irs.gov. Australian metadata queries came directly from the publisher's established public API.\n\nUS return: inspected PDF 1 and 6, PDF 45–48 Schedule O, and PDF 49–52 Schedule R; relevant PDF 1, 6, 48, 49, 50, 51 and 52 rendered and read, with PDF 47 also retained. Plain extraction was used because layout extraction split words into letters. Critical entries are on the original page 50 and 51 renders. IRS instructions: PDF 1 Relationships, PDF 2 control definitions, PDF 5 Part IV columns (d)–(i), and PDF 6 Part V line 2 read; no entire-document reading claim.\n\nExact reused sources: w4-fin-s-fs2023 (audit findings from prior wave); w5-gov-s-constitution2023 clause 66/PDF 21 and w5-gov-s-constitution2025 clause 66/PDF 19 newly read for this interpretation. The current IRS Schedule R instruction webpage applies to 2024 and later; it was excluded as authority for the 2023 return, and the archived 2023 instructions were used instead.\n", "file": "research/australia-seams-wave-2026-09-08-membership-findings.md" }, { "id": "australia-seams-wave-2026-09-08-reference-findings", "title": "The earlier encryption-policy route", "description": "The dated route reaches back before either 2024 report previously traced. A July2022 paper by Ian Levy and Crispin Robinson identifies their posts at the UK National Cyber Security Centre and GCHQ. The European Commission cited it in a public defence of its child-sexual-abuse proposal on 7August2022. eSafety cited it in an encryption position dated 13October", "html": "The dated route reaches back before either 2024 report previously traced. A July2022 paper by Ian Levy and Crispin Robinson identifies their posts at the UK National Cyber Security Centre and GCHQ. The European Commission cited it in a public defence of its child-sexual-abuse proposal on 7August2022. eSafety cited it in an encryption position dated 13October2023. These are actual citations by public authorities, not merely institutional proximity.
\nThe paper advocates technically possible client-side protections, but explicitly disclaims UK government policy or requirements and a complete security analysis (PDF10). Its discussion recognises coercion, database manipulation and mission creep; it reports no evidence for its considered charity-funding capture scenario (PDF41–42). It also points to REPHRAIN's separate Safety Tech Challenge Fund evaluation work (PDF6,56). That is not the later Toolbox award or proof of a common programme. Read cover, selected introduction/executive-summary pages, section6.1 and the opening of section8; no full technical validation claimed. Original arXiv PDF retained.
\nThe Commission post is an archived policy argument dated 7August2022, not a legal-status determination. Its paragraphs naming Levy and Robinson use the paper to support possible detection within encrypted services. The current page attributes authorship to the Directorate-General for Communication; this investigation does not infer a named commissioner from first-person language. Original HTML retained.
\nThe eSafety statement predates the 2024 DFC and SaferAI reports. PDF9–10 discusses client-side checks, secure enclaves, on-device detection and homomorphic techniques. It explicitly describes scalability and security concerns, with Levy/Robinson as the countervailing CSS citation. Its advertising-research references are not validation of an operational abuse-detection system. The date is printed on PDF11. This proves prior consideration; it does not prove how the later toolkit was drafted. Original retrieval timed out; the scoped web-reader observation is retained as a derivative, not a PDF capture.
\nWeProtect's January2024 submission to the Australian draft-standards consultation advocates client-side scanning, homomorphic encryption and secure enclaves while saying the technologies are nascent and require more testing (PDF4). Its disclaimer attributes the position to the Secretariat rather than necessarily to Alliance members (PDF10). Digital Rights Watch's opposing submission is dated 21December2023, despite the February2024 upload path. It disputes feasibility and privacy claims (PDF10–14). These are contrasting formal inputs, not proof either was adopted. Both scoped observations are derivatives after one original-retrieval timeout each.
\nThe 2026 toolkit itself still includes general scale, latency and specialist-tool limitations in its common-barriers discussion (PDF11 / printed20–21). Consequently, we cannot claim all caveats disappeared. The specific homomorphic-encryption citation29 remains unresolved. A homomorphic-learning article in the same journal volume is only a title/topic match: no evidence identifies it as the intended citation, and no Ant Group influence node is earned. The earlier policy route is the stronger finding.
\nThe next discriminating record is an actual reference-manager entry or tracked source substitution for citation29, together with comment dispositions connecting the 2023 statement or 2024 submissions to the later toolkit. Shared terminology alone does not establish that transmission.
", "path": "australia-seams-wave-2026-09-08/reference-findings.md", "sourceIds": [ "w6-ref-s-paper", "w6-ref-s-ec", "w6-ref-s-position", "w6-ref-s-wp", "w6-ref-s-drw" ], "markdown": "The dated route reaches back before either 2024 report previously traced. A July2022 paper by Ian Levy and Crispin Robinson identifies their posts at the UK National Cyber Security Centre and GCHQ. The European Commission cited it in a public defence of its child-sexual-abuse proposal on 7August2022. eSafety cited it in an encryption position dated 13October2023. These are actual citations by public authorities, not merely institutional proximity.\n\nThe paper advocates technically possible client-side protections, but explicitly disclaims UK government policy or requirements and a complete security analysis (PDF10). Its discussion recognises coercion, database manipulation and mission creep; it reports no evidence for its considered charity-funding capture scenario (PDF41–42). It also points to REPHRAIN's separate Safety Tech Challenge Fund evaluation work (PDF6,56). That is not the later Toolbox award or proof of a common programme. Read cover, selected introduction/executive-summary pages, section6.1 and the opening of section8; no full technical validation claimed. Original arXiv PDF retained.\n\nThe Commission post is an archived policy argument dated 7August2022, not a legal-status determination. Its paragraphs naming Levy and Robinson use the paper to support possible detection within encrypted services. The current page attributes authorship to the Directorate-General for Communication; this investigation does not infer a named commissioner from first-person language. Original HTML retained.\n\nThe eSafety statement predates the 2024 DFC and SaferAI reports. PDF9–10 discusses client-side checks, secure enclaves, on-device detection and homomorphic techniques. It explicitly describes scalability and security concerns, with Levy/Robinson as the countervailing CSS citation. Its advertising-research references are not validation of an operational abuse-detection system. The date is printed on PDF11. This proves prior consideration; it does not prove how the later toolkit was drafted. Original retrieval timed out; the scoped web-reader observation is retained as a derivative, not a PDF capture.\n\nWeProtect's January2024 submission to the Australian draft-standards consultation advocates client-side scanning, homomorphic encryption and secure enclaves while saying the technologies are nascent and require more testing (PDF4). Its disclaimer attributes the position to the Secretariat rather than necessarily to Alliance members (PDF10). Digital Rights Watch's opposing submission is dated 21December2023, despite the February2024 upload path. It disputes feasibility and privacy claims (PDF10–14). These are contrasting formal inputs, not proof either was adopted. Both scoped observations are derivatives after one original-retrieval timeout each.\n\nThe 2026 toolkit itself still includes general scale, latency and specialist-tool limitations in its common-barriers discussion (PDF11 / printed20–21). Consequently, we cannot claim all caveats disappeared. The specific homomorphic-encryption citation29 remains unresolved. A homomorphic-learning article in the same journal volume is only a title/topic match: no evidence identifies it as the intended citation, and no Ant Group influence node is earned. The earlier policy route is the stronger finding.\n\nThe next discriminating record is an actual reference-manager entry or tracked source substitution for citation29, together with comment dispositions connecting the 2023 statement or 2024 submissions to the later toolkit. Shared terminology alone does not establish that transmission.\n", "file": "research/australia-seams-wave-2026-09-08-reference-findings.md" }, { "id": "australia-wave-2026-09-08-5rights-findings", "title": "5Rights amendment attachment: the archive route", "description": "The August 5, 2022 proposed-author-amendment attachment was not recovered. This pass tested California State Archives public discovery, rather than repeating the FARA filing index. The Archives describes committee bill files containing amendments and correspondence, and author files that may contain original-sponsor correspondence. Deposited legislator recor", "html": "The August 5, 2022 proposed-author-amendment attachment was not recovered. This pass tested California State Archives public discovery, rather than repeating the FARA filing index. The Archives describes committee bill files containing amendments and correspondence, and author files that may contain original-sponsor correspondence. Deposited legislator records can carry access restrictions. Its public Governor chaptered-bill-file description currently ends at 2018, so it is not proof of custody of the 2022 Newsom file.
\nThe ordinary portal form was inspected and submitted with AB 2273, Portantino and Senate Appropriations. The first ten of 53 broad AB results and first ten of 37 broad committee results were irrelevant historical material; Portantino returned zero. These results are partial portal coverage, not an exhaustive archival holdings search or proof the attachment is absent. External indexed searches for the exact bill and Skaggs produced analyses and later litigation, not the transmitted draft.
\nThe practical next record is an accession/box/folder identification for the 2022 Senate Appropriations AB2273 bill file, followed by the July18-August11 proposed-amendment correspondence. Archives Reference Services can identify likely series but says it does not search for a specified document within a series. No request was sent and no paid copying ordered. CapConn and 5Rights remain likely holders of the actual August5 email; archival custody remains to be confirmed.
\nSources: http[local research file] ; http[local research file] . HTML captures, ordinary-form results and hashes accompany this note. No new actor, money or authorship edge is earned.
", "path": "australia-wave-2026-09-08/5rights-findings.md", "sourceIds": [ "w4-5r-s-archive-resources", "w4-5r-s-reference" ], "markdown": "The August 5, 2022 proposed-author-amendment attachment was not recovered. This pass tested California State Archives public discovery, rather than repeating the FARA filing index. The Archives describes committee bill files containing amendments and correspondence, and author files that may contain original-sponsor correspondence. Deposited legislator records can carry access restrictions. Its public Governor chaptered-bill-file description currently ends at 2018, so it is not proof of custody of the 2022 Newsom file.\n\nThe ordinary portal form was inspected and submitted with AB 2273, Portantino and Senate Appropriations. The first ten of 53 broad AB results and first ten of 37 broad committee results were irrelevant historical material; Portantino returned zero. These results are partial portal coverage, not an exhaustive archival holdings search or proof the attachment is absent. External indexed searches for the exact bill and Skaggs produced analyses and later litigation, not the transmitted draft.\n\nThe practical next record is an accession/box/folder identification for the 2022 Senate Appropriations AB2273 bill file, followed by the July18-August11 proposed-amendment correspondence. Archives Reference Services can identify likely series but says it does not search for a specified document within a series. No request was sent and no paid copying ordered. CapConn and 5Rights remain likely holders of the actual August5 email; archival custody remains to be confirmed.\n\nSources: http[local research file] ; http[local research file] . HTML captures, ordinary-form results and hashes accompany this note. No new actor, money or authorship edge is earned.\n", "file": "research/australia-wave-2026-09-08-5rights-findings.md" }, { "id": "australia-wave-2026-09-08-finance-findings", "title": "Australian financial/legal lane: a dated separation, an extended grant, and the remaining instrument gap", "description": "Read September8,2026 local time. New captures are in this directory, prefixed finance-; original URLs, HTTP outcomes and SHA256 values are in finance-capture-manifest.json. No canonical case or reader edits, outreach, spending or private API access. Registry access stopped at the exact restriction recorded below.", "html": "Read September8,2026 local time. New captures are in this directory, prefixed finance-; original URLs, HTTP outcomes and SHA256 values are in finance-capture-manifest.json. No canonical case or reader edits, outreach, spending or private API access. Registry access stopped at the exact restriction recorded below.
The strongest new records establish that ICMEC US stopped controlling its Australian affiliate on 6July2023, with a substantial donor-restricted resource position removed from the consolidated accounts. Separately, Westpac explicitly identifies the original grant as AUD25million and reports a later two-year extension. Those findings strengthen the history and funding-purpose chain. They still do not supply the Australian grant agreement, permission rights, final term dates, donor-specific ledger allocations or current standalone Australian accounts.
\n[Australian Business Register, ABN86645682828](http[local research file] identifies INTERNATIONAL CENTRE FOR MISSING & EXPLOITED CHILDREN, AUSTRALIA LTD, an Australian public company, active from5November2020, ACN645682828. It records business name Global Centre For Child Protection from3July2026 under the same ABN. That is an additional trading-name selector, not evidence of a newly created charity or changed legal entity. The current register also dates DGR endorsement to22March2024. These are registry classifications, not findings about individual donations.
\nThe [2023 ICMEC consolidated financial statements](http[local research file] Note1, say the US body created the Australian company inNovember2020 and ceased being its sole member on6July2023. Note8, PDFp17/printed16, separately states it no longer controls Australia. “No longer sole member” does not establish complete withdrawal from membership. The former affiliate's late2021 operational-establishment account is compatible with earlier legal incorporation.
\nThe same audit's Note8 removes USD13,519,626 net assets from consolidation. Components include cash4,131,064, investments3,337,000 and grants/contributions receivable6,359,338, plus other assets and liabilities. ExhibitB, PDFp7/printed6, records the deconsolidation entry as +1,408,807 without donor restrictions; −14,928,433 with donor restrictions; −13,519,626 total. These are accounting removals, not a new grant or bank transfer. The receivable is not assigned to Westpac by the note. The entry is not a present Australian solvency assessment.
\nThe auditor's report, PDFp3/printed2, explicitly relies on another auditor's Australian financial statements for the period ended6July2023, with additional conversion procedures. This identifies a specific underlying record that exists; that component report was not obtained here. The [2022 consolidated statements](http[local research file] PDFpp9–10/printed8–9, explain the former consolidation basis and translation of Australian financial statements into US dollars. Do not compare these USD balances directly with AUD25million. Their programme-restricted aggregate is for the then-consolidated group, not a Westpac-specific Australian grant schedule.
\n[Westpac FY21 Modern Slavery Statement](http[local research file] expressly labels the award $25million(AUD). Currency is now directly documented for the original award, upgrading the earlier context-based inference. PDFp21/printed20 describes a three-year grant, managed by ICMEC, for designing, building, deploying and maintaining a platform and analytical tools using cross-industry data. This is the bank's stated grant purpose; the executed instrument's enforceable restrictions remain unavailable. The passage relates the programme to possible remediation following AUSTRAC's2019 allegations, rather than describing a court order requiring this charity payment.
\n[Westpac2024 Modern Slavery Statement](http[local research file] says the initially three-year AUD25million partnership was extended for an additional two years during the reporting year. It does not announce another AUD25million or state a new payment schedule. PDFp17/printed16 also describes a four-year relationship to that point. An extension is therefore documented, but the full contractual chronology is not resolved by simple addition.
\nThe [current programme page](http[local research file] still describes the three-year grant and priorFY20 scoping. It is undated historical programme copy on a currently accessible page. The retained [2024 Safer Children report](http[local research file] describes a six-year commitment/partnership. Preserve each source period and the now-known extension. Possible explanations include scoping versus award clocks, changed delivery periods, later changes, or stale/inconsistent summaries; none is selected as fact. The signed initial agreement and amendments can distinguish them.
\nThe bank's stated grant purpose now explicitly reaches platform construction and maintenance. It does not identify Kablamo, a licence, a hosting charge, a software-capitalisation balance or a SaferAI policy allocation. No identifiable Lighthouse/Kablamo expense line was recovered in the acquired financial statements. The USD20,699 fixed-asset balance at deconsolidation cannot be renamed Lighthouse software. Later Australian operating expenses are outside the US consolidation perimeter.
\nThe strongest rival to funder-directed policy is a restricted technical-capacity grant supporting an operationally independent recipient, with programme development and later advocacy chosen under the recipient's governance. Donor selection and renewal dependence can coexist with that explanation. A competing donor-direction model predicts actual reserved approvals, conditions, intervention correspondence or allocation-linked deliverables. This pass recovered a stated technical purpose and an extension, not those decision rights.
\nPublic search located the exact [ACNC profile](http[local research file] and [AIS2024 record](http[local research file] Their search/web-reader shells establish titles/locators only; the financial contents were not read. Ordinary GETs each timed out after40seconds. One public-browser fallback then returned this exact refusal:
\n\n\nBrowser Use could not complete this action because a browser security check was unavailable. Reason: The admin-enforced policy could not be verified, so access was not granted. Browser use could not verify the admin-enforced policy before accessing http[local research file] This failure may be temporary. The agent may retry after the issue is resolved, but must not bypass browser security controls or use an indirect workaround.
\n
This is a browser admin-policy verification failure, not an automatic approval review rejection or a financial-report absence. All ACNC retrieval stopped; no indirect workaround was attempted. The independently published US audits are substitutes for their particular disclosures, not ACNC or standalone Australian filings. An independently attempted Australian governance URL returned404; its cached search description is not treated as a live constitution. No repeat route was attempted.
\nRead scope: ABN live entity/business-name/tax-status fields; FY21 statement PDFpp5/21; FY24 statement PDFpp17/28; current Westpac ICMEC subsection; US2023 audit PDFpp3,7,9–17 at the organisation, conversion, restriction and deconsolidation passages; US2022 component-audit, consolidation, currency and restriction notes. These are not exhaustive audits of the financial statements or the programmes. Key FY21 currency and US2023 tables were visually checked. Other captured text is available but should not be imported as wholly reviewed.
\nThe highest-value specific financial record is the Australian component financial statements and auditor report for the period ended6July2023, expressly identified by the US audit. The Australian finance/board custodian and the consolidating auditor received or hold that record. It can reveal the Australian grant receivable, restrictions and accounting transition without inventing a cash transfer. Subsequent standalone Australian statements and the publicly lodged AIS attachments, if available after the access restriction clears, can extend the picture into Lighthouse's operating period.
\nThe executed Westpac–ICMEC grant, FY24 extension and any novation associated with the2023 separation should identify legal grantee(s), start/end dates, drawdowns, remaining promises, reporting, permitted costs, clawback, termination and renewal rights. Likely holders are Westpac's programme/legal records custodian and the relevant ICMEC finance/legal archives. A novation is a candidate record, not a document proved to exist merely because control changed.
\nTo join funding to the technology, obtain the Australian project ledger, intangible-asset register, contract and maintenance invoices with donor/project codes. They would distinguish grant-funded development, ordinary operating support and a separately financed service, and permit a legitimate comparison with the vendor's public account. A supplier statement alone cannot allocate an expense.
\nFinally, the membership register, constitutional amendments and member/board resolutions effective6July2023 can specify who obtained or retained membership and appointment rights. The audit proves changed sole-membership/control status; it does not name the replacement members. These records are likely held by the Australian company secretary and the US former-parent archive. No holder requests have been sent.
", "path": "australia-wave-2026-09-08/finance-findings.md", "sourceIds": [ "w4-fin-s-abn", "w4-fin-s-fs2022", "w4-fin-s-fs2023", "w4-fin-s-westpac2021", "w4-fin-s-westpac2024", "w4-fin-s-programme", "w4-fin-s-acnc-profile", "w4-fin-s-acnc-ais2024", "w4-fin-s-governance404" ], "markdown": "Read September8,2026 local time. New captures are in this directory, prefixed `finance-`; original URLs, HTTP outcomes and SHA256 values are in `finance-capture-manifest.json`. No canonical case or reader edits, outreach, spending or private API access. Registry access stopped at the exact restriction recorded below.\n\n## What changed\n\nThe strongest new records establish that ICMEC US stopped controlling its Australian affiliate on **6July2023**, with a substantial donor-restricted resource position removed from the consolidated accounts. Separately, Westpac explicitly identifies the original grant as **AUD25million** and reports a later **two-year extension**. Those findings strengthen the history and funding-purpose chain. They still do not supply the Australian grant agreement, permission rights, final term dates, donor-specific ledger allocations or current standalone Australian accounts.\n\n### Legal identity and the separation\n\nAustralian Business Register, ABN86645682828 (see source references) identifies **INTERNATIONAL CENTRE FOR MISSING & EXPLOITED CHILDREN, AUSTRALIA LTD**, an Australian public company, active from5November2020, ACN645682828. It records business name **Global Centre For Child Protection** from3July2026 under the same ABN. That is an additional trading-name selector, not evidence of a newly created charity or changed legal entity. The current register also dates DGR endorsement to22March2024. These are registry classifications, not findings about individual donations.\n\nThe 2023 ICMEC consolidated financial statements (see source references) Note1, say the US body created the Australian company inNovember2020 and ceased being its **sole member** on6July2023. Note8, PDFp17/printed16, separately states it **no longer controls** Australia. “No longer sole member” does not establish complete withdrawal from membership. The former affiliate's late2021 operational-establishment account is compatible with earlier legal incorporation.\n\nThe same audit's Note8 removes USD13,519,626 net assets from consolidation. Components include cash4,131,064, investments3,337,000 and grants/contributions receivable6,359,338, plus other assets and liabilities. ExhibitB, PDFp7/printed6, records the deconsolidation entry as **+1,408,807 without donor restrictions; −14,928,433 with donor restrictions; −13,519,626 total**. These are accounting removals, not a new grant or bank transfer. The receivable is not assigned to Westpac by the note. The entry is not a present Australian solvency assessment.\n\nThe auditor's report, PDFp3/printed2, explicitly relies on another auditor's Australian financial statements for the period ended6July2023, with additional conversion procedures. This identifies a specific underlying record that exists; that component report was not obtained here. The 2022 consolidated statements (see source references) PDFpp9–10/printed8–9, explain the former consolidation basis and translation of Australian financial statements into US dollars. Do not compare these USD balances directly with AUD25million. Their programme-restricted aggregate is for the then-consolidated group, not a Westpac-specific Australian grant schedule.\n\n### Westpac's own terms history\n\nWestpac FY21 Modern Slavery Statement (see source references) expressly labels the award **$25million(AUD)**. Currency is now directly documented for the original award, upgrading the earlier context-based inference. PDFp21/printed20 describes a three-year grant, managed by ICMEC, for designing, building, deploying and maintaining a platform and analytical tools using cross-industry data. This is the bank's stated grant purpose; the executed instrument's enforceable restrictions remain unavailable. The passage relates the programme to possible remediation following AUSTRAC's2019 allegations, rather than describing a court order requiring this charity payment.\n\nWestpac2024 Modern Slavery Statement (see source references) says the initially three-year AUD25million partnership was extended for **an additional two years** during the reporting year. It does not announce another AUD25million or state a new payment schedule. PDFp17/printed16 also describes a four-year relationship to that point. An extension is therefore documented, but the full contractual chronology is not resolved by simple addition.\n\nThe current programme page (see source references) still describes the three-year grant and priorFY20 scoping. It is undated historical programme copy on a currently accessible page. The retained 2024 Safer Children report (see source references) describes a six-year commitment/partnership. Preserve each source period and the now-known extension. Possible explanations include scoping versus award clocks, changed delivery periods, later changes, or stale/inconsistent summaries; none is selected as fact. The signed initial agreement and amendments can distinguish them.\n\n### Where this does and does not reach operating capacity\n\nThe bank's stated grant purpose now explicitly reaches platform construction and maintenance. It does not identify Kablamo, a licence, a hosting charge, a software-capitalisation balance or a SaferAI policy allocation. No identifiable Lighthouse/Kablamo expense line was recovered in the acquired financial statements. The USD20,699 fixed-asset balance at deconsolidation cannot be renamed Lighthouse software. Later Australian operating expenses are outside the US consolidation perimeter.\n\nThe strongest rival to funder-directed policy is a restricted technical-capacity grant supporting an operationally independent recipient, with programme development and later advocacy chosen under the recipient's governance. Donor selection and renewal dependence can coexist with that explanation. A competing donor-direction model predicts actual reserved approvals, conditions, intervention correspondence or allocation-linked deliverables. This pass recovered a stated technical purpose and an extension, not those decision rights.\n\n## Access receipt and bounded coverage\n\nPublic search located the exact ACNC profile (see source references) and AIS2024 record (see source references) Their search/web-reader shells establish titles/locators only; the financial contents were not read. Ordinary GETs each timed out after40seconds. One public-browser fallback then returned this exact refusal:\n\n> Browser Use could not complete this action because a browser security check was unavailable. Reason: The admin-enforced policy could not be verified, so access was not granted. Browser use could not verify the admin-enforced policy before accessing http[local research file] This failure may be temporary. The agent may retry after the issue is resolved, but must not bypass browser security controls or use an indirect workaround.\n\nThis is a **browser admin-policy verification failure**, not an automatic approval review rejection or a financial-report absence. All ACNC retrieval stopped; no indirect workaround was attempted. The independently published US audits are substitutes for their particular disclosures, not ACNC or standalone Australian filings. An independently attempted Australian governance URL returned404; its cached search description is not treated as a live constitution. No repeat route was attempted.\n\nRead scope: ABN live entity/business-name/tax-status fields; FY21 statement PDFpp5/21; FY24 statement PDFpp17/28; current Westpac ICMEC subsection; US2023 audit PDFpp3,7,9–17 at the organisation, conversion, restriction and deconsolidation passages; US2022 component-audit, consolidation, currency and restriction notes. These are not exhaustive audits of the financial statements or the programmes. Key FY21 currency and US2023 tables were visually checked. Other captured text is available but should not be imported as wholly reviewed.\n\n## The remaining obtainable records\n\nThe highest-value specific financial record is the **Australian component financial statements and auditor report for the period ended6July2023**, expressly identified by the US audit. The Australian finance/board custodian and the consolidating auditor received or hold that record. It can reveal the Australian grant receivable, restrictions and accounting transition without inventing a cash transfer. Subsequent standalone Australian statements and the publicly lodged AIS attachments, if available after the access restriction clears, can extend the picture into Lighthouse's operating period.\n\nThe **executed Westpac–ICMEC grant, FY24 extension and any novation associated with the2023 separation** should identify legal grantee(s), start/end dates, drawdowns, remaining promises, reporting, permitted costs, clawback, termination and renewal rights. Likely holders are Westpac's programme/legal records custodian and the relevant ICMEC finance/legal archives. A novation is a candidate record, not a document proved to exist merely because control changed.\n\nTo join funding to the technology, obtain the Australian **project ledger, intangible-asset register, contract and maintenance invoices** with donor/project codes. They would distinguish grant-funded development, ordinary operating support and a separately financed service, and permit a legitimate comparison with the vendor's public account. A supplier statement alone cannot allocate an expense.\n\nFinally, the **membership register, constitutional amendments and member/board resolutions effective6July2023** can specify who obtained or retained membership and appointment rights. The audit proves changed sole-membership/control status; it does not name the replacement members. These records are likely held by the Australian company secretary and the US former-parent archive. No holder requests have been sent.\n", "file": "research/australia-wave-2026-09-08-finance-findings.md" }, { "id": "australia-wave-2026-09-08-lighthouse-findings", "title": "Lighthouse: the application, its operating support and the missing decision record", "description": "Bounded public-source acquisition, 8 September 2026 local time. The candidate adds seven nodes, eight relations and six sources, reusing the existing Australian entity, Lighthouse and 2024-25 report IDs. No canonical file was edited.", "html": "Bounded public-source acquisition, 8 September 2026 local time. The candidate adds seven nodes, eight relations and six sources, reusing the existing Australian entity, Lighthouse and 2024-25 report IDs. No canonical file was edited.
\nThe new evidence earns a named delivery/support relationship and a dated access restriction. It does not yet earn a Lighthouse-to-account-blocking relationship. A separate documented closure operation is preserved alongside it.
\nThe existing ICMEC Australia impact report, PDF p12 / printed pp22-23, describes a queryable application that combines suspected-activity records with analytics and geographic/network context. It reports dozens of users and just under eight million records processed during 1 July 2024–30 June 2025, including over 70,000 geolocated to Australia. These are records, not unique people, proven offences, alerts sent or bank accounts blocked. The report does not name the feed suppliers. Its proposed university analysis concerns future research, not a demonstrated evaluation of downstream decisions. [Existing original report](http[local research file] scoped local extract: lighthouse-existing-impact-page12.txt.
The current Our Work page describes expansion beyond Lighthouse into additional sovereign, real-time data streams as a plan. That language does not establish deployed expansion or a population-wide identification scheme. Its audience remains organisations and frontline professionals. [Current service description](http[local research file] CSEA data insights and analytics.
\nKablamo's commercial case study reports building Lighthouse, subsequent annually renewed Product Care support, and technical delivery decisions by an unnamed ICMEC product owner. It names AWS infrastructure and NextGIS mapping. No fee, signed instrument or independent acceptance record was recovered. The account and client quotation are one commercial source family; its broad claims of protection outcomes are not an evaluation. The unnamed owner is not reconciled with a named executive or assigned policy/bank-action authority. [Vendor account](http[local research file] The Approach / The Results; lighthouse-kablamo.txt, lines 34-59.
The original 30 October 2024 symposium programme states Lighthouse use was restricted to Australian soil. Its introductory session was limited to specified Australian regulators, financial/CSE professionals, banks/payment remitters and child-protection law enforcement. That is a real published access boundary, not proof of the current licence or of who attended or held accounts. I inspected the original PDF's second page visually: [Programme](http[local research file] lighthouse-symposium2024-page2.png.
The public privacy-policy URL returned HTTP 404 on ordinary acquisition. Search-index text describes a policy updated 3 July 2024, including data-product access and general access/correction/complaint provisions, but the current original was not acquired. This is not treated as a current intelligence-feed licence or proof that an adverse inference is corrected throughout recipient systems. The current footer's Terms of use link is only #. Neither observation proves such rules do not exist. No retention schedule, named provider licence, match-quality assessment, recipient permission matrix or correction-propagation procedure was recovered.
ICMEC's July 2026 No Bank Left Behind paper discusses closure outcomes under Operation Huntsman and cites AUSTRAC. Its Hume and Beyond Bank passages describe collaboration, not a specific Lighthouse alert followed by an account decision. [Paper](http[local research file] pp1-3,5-9.
\nThe paper labels the relevant AUSTRAC reference 2024. That URL now redirects to an official release dated 15 September 2025, which reports more than 3,000 Australian bank accounts linked to sextortion payments closed since June 2022 through the joint ACCCE–AUSTRAC Operation Huntsman. This is an agency-reported cumulative outcome, not an inspected bank ledger. The paper's older 1,800-plus callout and the current page must retain their different dates/versions. The cited historical version was not recovered. [Current AUSTRAC release](http[local research file] date and Operation Huntsman paragraph.
\nNo recovered record joins Lighthouse to those closures. That is a missing attribution, not proof Lighthouse was uninvolved. The candidate therefore creates a separate Huntsman operation and aggregate outcome, with no Lighthouse edge. The paper's closure passage derives from the AUSTRAC publication family; it is not a second independent confirmation.
\nThe highest-value unacquired family is a de-identified, versioned Lighthouse query/alert-to-decision audit trail held by ICMEC and a recipient institution: input provenance and caveats, human review, corroborating evidence, action or no action, and correction. Pair it with supplier licence schedules, access terms, retention rules and correction/withdrawal propagation. No abuse data, account numbers or identifiable case material is needed to test the mechanism.
\nA realistic alternative is an ordinary intelligence-support service: analysts use Lighthouse as context while existing transaction monitoring, other intelligence and institutional duties independently determine investigations, reporting or closures. The acquired records support that interpretation at least as readily as automatic blocking or donor-controlled enforcement. Kablamo's support contract is a separate attainable family for service dependency, access privileges, handover and decision rights.
\nlighthouse-capture-manifest.json records original HTML/PDF bytes, URLs, hashes and failures. The AUSTRAC ordinary-download route timed out; its successful official web-reader extraction is preserved as lighthouse-austrac-web-extract.json, explicitly not raw HTML. Privacy remains an unavailable original; no substantive candidate edge relies on it. The research-partnership page was captured and scoped for Lighthouse output; no such evaluated output was found there, and its other research was not promoted into this lane.
The bounded search included exact Lighthouse/ICMEC queries, public service/privacy material, the builder, the original programme, the current research page and AUSTRAC/AFP-targeted counterpart searches. These searches did not recover the missing institution-level Lighthouse outcome or licence record; they are not an exhaustive proof of absence. Acquisition stops at the specified holder frontier. lighthouse-candidate-check.json confirms candidate IDs, endpoint/source references and absence of baseline collisions; it does not certify the underlying claims.
Selected output: AI and Child Protection: A Collaborative Approach to a Safer Future, the SaferAI for Children Coalition's discussion paper publicised by ICMEC Australia on 26 November 2024. Acquisition/read: 8 September 2026 local time. This lane did not edit the canonical case. The 2025 call to action was read as a bounded comparison, not expanded into a second campaign.
\nThe new result is an evidenced circulation route, with important gaps still visible. A coalition publication proposing greater use of AI for child protection entered both an independent review addressed to Tasmania's Premier and eSafety's later industry guidance. The records show what those recipients used. They do not establish who financed or cleared the original wording, or that either recipient adopted the paper's device-level proposal.
\nThe publisher's dated launch notice describes a collaborative paper meant to inform policy and industry. Original PDF p8 goes beyond general awareness: among possible initiatives to support, it proposes AI detection/reporting inside encrypted environments and detection or blocking of abuse-material creation, storage and distribution at device level. That is a content-screening proposal. It is not a device-age/identity signal architecture, a contract to implement scanning, proof the technique works, or evidence that a government adopted it.
\nThe same paper p15 advocates development, funding and deployment of investigative AI: categorising material, ranking cases and detecting behavioural patterns. Pages16-19 propose legislative change, research support, cross-sector cooperation, public communication and specialised consultation with the coalition. It acknowledges AI benefits as well as harms. The paper's legal/technical/statistical assertions were read as the advocates' claims, not independently reverified for this lane.
\nOriginal: http[local research file] The publisher notice at http[local research file] fixes the displayed launch date. Both originals are retained locally.
\nPage3 says the paper was collaboratively written using member insights. Page4 identifies ICMEC Australia as coalition leader and displays a historical member panel. Visual inspection confirms, among others, AiLECS Lab, University of Tasmania, UNICEF Australia, Australian Childhood Foundation, Bravehearts and the Grace Tame Foundation. These are publisher-presented member identities. They are not a task allocation, author list, sponsor ledger or publication-clearance record. I did not infer the undisplayed members, reconstruct a broad roster or import the January2026 participation rules into this 2024 document.
\nOne useful production lead is embedded in the acquired original: the PDF Author metadata names Mikaela Jago, with Canva as creator/producer. The 2025 call PDF contains the same metadata name. The retained ICMEC Australia 2024-25 impact report, printed p35/PDF18, separately lists Jago as Senior Capacity and Prevention Coordinator. The name is therefore attached both to an institutional role and these document-production records. That still does not establish who drafted a particular passage, whether the field reflects design/export rather than prose, or who approved publication. The candidate deliberately has no “Jago authored/approved paper” edge.
\nNo named paper-specific funding award, sponsor allocation, drafting payment, donor approval clause or sign-off appears in the inspected paper, notice or call. Existing Westpac programme funding cannot fill that gap. The decisive holder record is the paper's version/comment history and publication clearance, paired with its project budget or commissioning agreement if one exists. ICMEC Australia's programme/policy and finance teams are plausible custodians; Jago is a specific production-record lead.
\nThe original Nov2024 paper remains attributed to contemporary ICMEC Australia. This lane does not reinterpret its authorship using a later business-name change or retrospectively treat the international ICMEC organisation as publisher.
\nThe November2025 Woolcott Independent Review, Part A, has a cover identifying Peter Woolcott as independent reviewer and Jeremy Rockliff, Premier of Tasmania, as report addressee. In section6.3.2, printed p131/PDF132, it explicitly cites the coalition paper in notes124-125 and reproduces four proposed AI uses: research, automated analysis, collaboration and professional training. It also calls attention to the reThink chatbot example involving the University of Tasmania. This is substantive citation/use, not an inference from shared vocabulary.
\nThere is a small, concrete institutional path here: University of Tasmania appears in the coalition paper's member panel; that paper discusses its involvement in a trial on p9; the Tasmanian reviewer explicitly highlights the example's local university connection. This does not show that the university authored the paper, commissioned the review's reference or received money because of it.
\nThe review proceeds to recommendations about cross-agency information sharing, privacy-law exceptions and a regulatory roadmap, printed pp132-134. Its own rationale also draws on Commission of Inquiry findings, statutory problems and consultations. We cannot assign those broader recommendations solely to this paper. Its four-item list does not reproduce the device/encrypted-environment screening bullet. The acquired publication is a report and recommendation record; a separate acceptance/implementation record was not acquired.
\nOriginal official substitute: http[local research file] The cover and foreword say November2025; the filename's October31 is not being used as a release/delivery date. The same title/content is readable in the Parliament-hosted tabled-paper copy, but this lane does not claim byte identity between copies without both originals.
\nThe Designing for Safety toolkit, displayed by the web reader as the 4 March2026 version, lists the coalition paper among relevant initiatives. More importantly, printed p28/PDF15 footnote34 cites it specifically for AI classifiers categorising material, prioritising cases and identifying patterns in grooming/perpetrator behaviour. The citation records access on 26 November2025. This supports a direct, bounded claim of use in official guidance.
\nThe toolkit says it followed an eSafety/US Department of Homeland Security workshop in December2024. The paper's appearance in a 2026 toolkit does not show it was used at that earlier workshop. The toolkit also discusses age assurance, identity verification and encrypted-environment measures elsewhere. Those sections have their own context and references. Footnote34 is not a licence to attribute the whole toolkit to SaferAI, or to claim that its 2024 device-screening proposal became law.
\nOfficial source: http[local research file] One direct download timed out; the web reader successfully inspected the original PDF's scoped passages. A derivative reading extract is saved and labelled as such, with its own hash. There is no invented original-PDF hash or local original claim.
\nThe three-page May2025 national call asks the Australian government to prioritise the issue, engage with the coalition, explore investment, strengthen education/legal responses and coordinate internationally. The original p3 “Signed by” panel differs from the 2024 member panel. This supplies a particular endorsement presentation, not proof of private approval permissions or withdrawal by an organisation absent from a later panel. No named ministerial receipt or acceptance appears in this call.
\nThe strongest rival explanation remains ordinary policy circulation: a relevant coalition report is publicly available, and official writers select useful portions while drawing on other material. Direct submission and negotiated drafting are also possible, but their records are missing. A source-selection/transmittal record from the Woolcott review or eSafety would distinguish those routes. General membership, general donor support or an editor's PDF metadata cannot do that work.
\nFive original publications/pages are preserved with SHA-256 values in policy-capture-manifest.json: the current index, historical launch notice, 23-page discussion paper, three-page call and 242-page official review. The last was read only at the precise scope recorded in policy-candidate.json; extracting pages is not a whole-report audit. The eSafety item is a separately labelled web-reader extract. The existing institutional impact-report source was reused by its stable case ID, with a new staff-panel locator.
\nThe initial Parliament PDF acquisition did not yield retained bytes; its exact non-success status was lost when the later eSafety timeout terminated that batch. Rather than manufacture an HTTP code or repeat the route, one alternate official Tasmanian publisher supplied the original. A failed web screenshot was replaced by rendering the already-held paper with installed pypdfium2. No packages, private endpoints, outreach or paid services were used. No further retries are required for this bounded result.
", "path": "australia-wave-2026-09-08/policy-findings.md", "sourceIds": [ "w4-pol-s-about", "w4-pol-s-launch", "w4-pol-s-paper", "w4-pol-s-call", "w4-pol-s-tas-review", "w4-pol-s-esafety-toolkit" ], "markdown": "Selected output: **AI and Child Protection: A Collaborative Approach to a Safer Future**, the SaferAI for Children Coalition's discussion paper publicised by ICMEC Australia on 26 November 2024. Acquisition/read: 8 September 2026 local time. This lane did not edit the canonical case. The 2025 call to action was read as a bounded comparison, not expanded into a second campaign.\n\nThe new result is an evidenced circulation route, with important gaps still visible. A coalition publication proposing greater use of AI for child protection entered both an independent review addressed to Tasmania's Premier and eSafety's later industry guidance. The records show what those recipients used. They do not establish who financed or cleared the original wording, or that either recipient adopted the paper's device-level proposal.\n\n## The concrete intervention\n\nThe publisher's dated launch notice describes a collaborative paper meant to inform policy and industry. Original PDF p8 goes beyond general awareness: among possible initiatives to support, it proposes AI detection/reporting inside encrypted environments and detection or blocking of abuse-material creation, storage and distribution at device level. That is a content-screening proposal. It is not a device-age/identity signal architecture, a contract to implement scanning, proof the technique works, or evidence that a government adopted it.\n\nThe same paper p15 advocates development, funding and deployment of investigative AI: categorising material, ranking cases and detecting behavioural patterns. Pages16-19 propose legislative change, research support, cross-sector cooperation, public communication and specialised consultation with the coalition. It acknowledges AI benefits as well as harms. The paper's legal/technical/statistical assertions were read as the advocates' claims, not independently reverified for this lane.\n\nOriginal: http[local research file] The publisher notice at http[local research file] fixes the displayed launch date. Both originals are retained locally.\n\n## Who contributed, and who approved?\n\nPage3 says the paper was collaboratively written using member insights. Page4 identifies ICMEC Australia as coalition leader and displays a historical member panel. Visual inspection confirms, among others, AiLECS Lab, University of Tasmania, UNICEF Australia, Australian Childhood Foundation, Bravehearts and the Grace Tame Foundation. These are publisher-presented member identities. They are not a task allocation, author list, sponsor ledger or publication-clearance record. I did not infer the undisplayed members, reconstruct a broad roster or import the January2026 participation rules into this 2024 document.\n\nOne useful production lead is embedded in the acquired original: the PDF Author metadata names **Mikaela Jago**, with Canva as creator/producer. The 2025 call PDF contains the same metadata name. The retained ICMEC Australia 2024-25 impact report, printed p35/PDF18, separately lists Jago as Senior Capacity and Prevention Coordinator. The name is therefore attached both to an institutional role and these document-production records. That still does not establish who drafted a particular passage, whether the field reflects design/export rather than prose, or who approved publication. The candidate deliberately has no “Jago authored/approved paper” edge.\n\nNo named paper-specific funding award, sponsor allocation, drafting payment, donor approval clause or sign-off appears in the inspected paper, notice or call. Existing Westpac programme funding cannot fill that gap. The decisive holder record is the paper's version/comment history and publication clearance, paired with its project budget or commissioning agreement if one exists. ICMEC Australia's programme/policy and finance teams are plausible custodians; Jago is a specific production-record lead.\n\nThe original Nov2024 paper remains attributed to contemporary ICMEC Australia. This lane does not reinterpret its authorship using a later business-name change or retrospectively treat the international ICMEC organisation as publisher.\n\n## A public counterpart uses the paper\n\nThe **November2025 Woolcott Independent Review, Part A**, has a cover identifying Peter Woolcott as independent reviewer and Jeremy Rockliff, Premier of Tasmania, as report addressee. In section6.3.2, printed p131/PDF132, it explicitly cites the coalition paper in notes124-125 and reproduces four proposed AI uses: research, automated analysis, collaboration and professional training. It also calls attention to the reThink chatbot example involving the University of Tasmania. This is substantive citation/use, not an inference from shared vocabulary.\n\nThere is a small, concrete institutional path here: University of Tasmania appears in the coalition paper's member panel; that paper discusses its involvement in a trial on p9; the Tasmanian reviewer explicitly highlights the example's local university connection. This does not show that the university authored the paper, commissioned the review's reference or received money because of it.\n\nThe review proceeds to recommendations about cross-agency information sharing, privacy-law exceptions and a regulatory roadmap, printed pp132-134. Its own rationale also draws on Commission of Inquiry findings, statutory problems and consultations. We cannot assign those broader recommendations solely to this paper. Its four-item list does not reproduce the device/encrypted-environment screening bullet. The acquired publication is a report and recommendation record; a separate acceptance/implementation record was not acquired.\n\nOriginal official substitute: http[local research file] The cover and foreword say November2025; the filename's October31 is not being used as a release/delivery date. The same title/content is readable in the Parliament-hosted tabled-paper copy, but this lane does not claim byte identity between copies without both originals.\n\n## eSafety's narrower use\n\nThe **Designing for Safety** toolkit, displayed by the web reader as the 4 March2026 version, lists the coalition paper among relevant initiatives. More importantly, printed p28/PDF15 footnote34 cites it specifically for AI classifiers categorising material, prioritising cases and identifying patterns in grooming/perpetrator behaviour. The citation records access on 26 November2025. This supports a direct, bounded claim of use in official guidance.\n\nThe toolkit says it followed an eSafety/US Department of Homeland Security workshop in December2024. The paper's appearance in a 2026 toolkit does not show it was used at that earlier workshop. The toolkit also discusses age assurance, identity verification and encrypted-environment measures elsewhere. Those sections have their own context and references. Footnote34 is not a licence to attribute the whole toolkit to SaferAI, or to claim that its 2024 device-screening proposal became law.\n\nOfficial source: http[local research file] One direct download timed out; the web reader successfully inspected the original PDF's scoped passages. A derivative reading extract is saved and labelled as such, with its own hash. There is no invented original-PDF hash or local original claim.\n\n## The comparison and the limit\n\nThe three-page May2025 national call asks the Australian government to prioritise the issue, engage with the coalition, explore investment, strengthen education/legal responses and coordinate internationally. The original p3 “Signed by” panel differs from the 2024 member panel. This supplies a particular endorsement presentation, not proof of private approval permissions or withdrawal by an organisation absent from a later panel. No named ministerial receipt or acceptance appears in this call.\n\nThe strongest rival explanation remains ordinary policy circulation: a relevant coalition report is publicly available, and official writers select useful portions while drawing on other material. Direct submission and negotiated drafting are also possible, but their records are missing. A source-selection/transmittal record from the Woolcott review or eSafety would distinguish those routes. General membership, general donor support or an editor's PDF metadata cannot do that work.\n\n## Custody and closed routes\n\nFive original publications/pages are preserved with SHA-256 values in policy-capture-manifest.json: the current index, historical launch notice, 23-page discussion paper, three-page call and 242-page official review. The last was read only at the precise scope recorded in policy-candidate.json; extracting pages is not a whole-report audit. The eSafety item is a separately labelled web-reader extract. The existing institutional impact-report source was reused by its stable case ID, with a new staff-panel locator.\n\nThe initial Parliament PDF acquisition did not yield retained bytes; its exact non-success status was lost when the later eSafety timeout terminated that batch. Rather than manufacture an HTTP code or repeat the route, one alternate official Tasmanian publisher supplied the original. A failed web screenshot was replaced by rendering the already-held paper with installed pypdfium2. No packages, private endpoints, outreach or paid services were used. No further retries are required for this bounded result.\n", "file": "research/australia-wave-2026-09-08-policy-findings.md" }, { "id": "cashapp-ag-resolution-wave-2026-09-09-july-ag-resolution", "title": "July 2026 multistate AG / Block resolution", "description": "Observed 9 September 2026. The July resolution disclosed by Block is identifiable as the 8 July 2026, USD 45 million settlement announced by 46 state attorneys general, co-led by Oregon and Texas. Two state-issued copies establish the substantive agreed terms. Oregon's copy includes the parties' signatures; neither inspected copy establishes judicial entry. ", "html": "Observed 9 September 2026. The July resolution disclosed by Block is identifiable as the 8 July 2026, USD 45 million settlement announced by 46 state attorneys general, co-led by Oregon and Texas. Two state-issued copies establish the substantive agreed terms. Oregon's copy includes the parties' signatures; neither inspected copy establishes judicial entry. The central coordination provision credits redress paid under the January 2025 CFPB order and supplies a state-controlled fallback. It is not an additional USD 75–120 million consumer award to stack on top of the CFPB amount.
\n[Block's Q2 2026 filing](http[local research file] Note 16, printed page 33, says it entered a July 2026 resolution with multiple state AGs concerning complaints and disputes, and calls the amount immaterial. This is an issuer disclosure, not the operative amount or proof of payment. The identification is corroborated by the [California AG announcement](http[local research file] and [Oregon AG announcement](http[local research file] both dated 8 July 2026. Oregon says it co-led the investigation with Texas.
\nThe [California instrument](http[local research file] is a [PROPOSED] Final Judgment and Permanent Injunction, People of California v. Block, Inc. f/k/a Square, Inc., Superior Court, Alameda County. Its case-number field is blank on PDF page 1 and judicial signature/date fields are blank on page 21, visually checked. Its opening and ¶8, pages 2–3, say the parties agreed without factual/liability admission, Block denies wrongdoing, and parties waive appeal. This is an official published settlement text, not a visually verified entered judgment.
\nThe [Oregon instrument](http[local research file] is a Stipulated General Judgment and Money Award, State of Oregon ex rel. Dan Rayfield v. Block, Circuit Court, Multnomah County. PDF page 29 bears Block Chief Legal Officer Chrysty Esperanza's DocuSign signature dated 6 July 2026; page 30 bears Oregon AAG Joseph Ferretti's signature dated 8 July 2026. Page 28's court-approval line is blank; page 31 is a certificate dated 8 July calling the proposed judgment ready for judicial signature. Its case-number field is also blank. Thus signed party assent is observed; court entry, an assigned docket number and final entered-version identity remain unverified. A blank posted copy is not evidence that the court never entered a judgment.
\nBoth instruments' first-page footnote lists the 46 participating states and explains that separate state judgments incorporate substantive terms, with local-rule/law differences. Neither is a single document bearing all 46 AG signatures. Defendant wording includes Block's subsidiaries, affiliates, successors and assigns; no claim is made that each separately signed. California ¶19 / Oregon ¶11(k) define 8 July 2026 as the effective date. Relevant laws include the state's consumer-protection statutes, EFTA and Regulation E. California ¶18 / Oregon ¶11(j) define the resolved pre-effective-date conduct and exclude Afterpay US's BNPL products. Complaint allegations are not admissions or contested judicial findings.
\nLocators below refer to the published California text unless Oregon is specified; PDF and printed page numbers agree. Statements describe the agreed terms, subject to the court-entry limitation above.
\n| Mechanism | \nAgreed requirement / decision right | \nRelationship to the earlier CFPB order | \n
|---|---|---|
| Staffed service and escalation | \n¶¶44–46, pages 8–9: complaint/error/account-access/fraud intake and resolution or escalation; live customer support 24 hours/day, toll-free human phone at least 13.5 hours/day, including 5–6:30 p.m. Pacific, human chat at least 18 hours/day, reasonable hold/callback timelines, clearly displayed contact/hours and a direct AG complaint contact. Oregon ¶¶31–33, pages 9–10. | \nHuman phone minimum is 13.5 hours rather than CFPB ¶108(a)'s 12. Do not describe this as 24-hour human telephone service. The AG contact is a specific external escalation route; it does not make the AG the initial reviewer of every case. | \n
| Error investigation and evidence | \n¶¶47–50, pages 9–10: prompt, thorough, reasonable investigation without extra prerequisites; results within three business days of completion; explanation and access to relied-on documents for no-error/different-error findings; correction within one business day; provisional credit after ten business days if incomplete; finalization under Regulation E; two-year records. No mandatory contact with the P2P transfer recipient or law enforcement, or extra information before starting, and no linked-instrument exclusion. Oregon ¶¶34–37, pages 11–12. | \nLargely parallels the CFPB transaction-error requirements and explicitly adds provisional-credit finalization wording. This document-access requirement concerns a Notice of Error, not every closure appeal. | \n
| Locks, suspensions and deactivations | \n¶51, pages 10–11, says “By June 30, 2026”: reasonably designed procedures for prompt notice, funds-access information and how restrictions may be addressed; prompt review/response and reasonable restoration timelines/efforts, with appropriateness and legal-prohibition qualifications. Oregon ¶38, pages 12–13, repeats this. | \nThis wording does not reproduce CFPB ¶108(c)–(d)'s fixed 24-hour notice and ten-business-day appeal decision requirements. The later settlement does not say it replaces or relaxes that order. June 30 precedes the stipulated July 8 effective date; retain the literal date and seek the final entered text/implementation record before interpreting why. It is not evidence of breach or deliberate backdating. | \n
| Marketing and fraud policy | \n¶¶36–43, pages 7–8: no misleading safety/bank/FDIC statements; discontinue marketing known to increase fraud, direct fraud education, public educational link and takeover-prevention procedures. ¶52, pages 11–12: identify suspected scam accounts, stop their operation and associated persons returning, complaint/trend tracking, good-faith refund policies and warnings. Oregon ¶¶23–30,39. | \nAnti-fraud restrictions and customer recourse are both required. No Lantern/Tech Coalition supplier or mandatory signal-sharing arrangement is specified. | \n
| Corporate oversight | \n¶¶30–35, pages 6–7: documented compliance management, qualified committee, quarterly Board reporting, employee/contractor training and retained training records. ¶¶53–54, page 12: annual committee-approved compliance report, available to state on request. Oregon ¶¶17–22,40–41. | \nDo not import the CFPB order's annual sworn Board-approved submission into this different report: this text requires preparation, committee approval and availability on request. | \n
| Supervision and changes | \n¶¶66,75–76, pages 15–17: modification can be sought through the executive committee and court; court retains enforcement/modification jurisdiction. Written non-material changes, including time extensions, can be agreed through counsel without a court order. Where the State determines a material failure and, in its sole discretion, no health/safety/welfare threat, it will give written notice before legal action. Block then has15businessdays to provide a good-faith written response; that response may describe completed correction or ongoing corrective work with a reasonable timetable. Oregon ¶¶54,63–64, pages 17,19–20. | \nThe notice/response requirement applies in the specified circumstances, not to all enforcement;15businessdays is not necessarily a completed-cure deadline. There is no independently observed monitor appointment or prior regulator approval of every operating policy. | \n
General Section III compliance is due no later than 30 days after the effective date unless otherwise specified (California ¶26; Oregon ¶13). Existing procedures may satisfy the requirements if adequate (¶27 / ¶14), so prospective wording is not an admission that every required practice was absent. Training under California ¶33 / Oregon ¶20 is due within 90 days, then within 60 days of assuming the relevant responsibilities; transaction-error training is annual and before assuming that work (¶34 / ¶21).
\nOnly enumerated obligations expire after five years, including specified committee/reporting, support-hours/contact, account-review and fraud measures (California ¶64; Oregon ¶52). Direct consumer education has a seven-year term (¶65 / ¶53). The texts do not say the whole judgment ends after five years. Written modifications or court entry could affect implementation; none was acquired.
\nCalifornia ¶¶59–60, page 13, and Oregon ¶¶46–48, pages 15–16, provide USD 45,000,000 payable within 30 days to the AGs, allocated at the participating states' discretion. California's specified share is USD 2,941,785.27, allocated under California Business and Professions Code §17206(c); the state's portion and interest are for AG consumer-protection enforcement. Oregon's specified share is USD 3,067,141.53, deposited into its ORS 180.095 account for any lawful purpose. Oregon's money-award table, pages 26–27, separately lists USD 281 filing costs. These are published obligations, not receipts.
\nOregon ¶48 expressly allocates USD 264,153.73 out of the same USD 45 million to the National Association of Attorneys General (NAAG), as reimbursement for certain costs and expenses of the participating states' investigation. It directs wire payment using Oregon AG instructions. This is a specific institutional reimbursement entitlement in the signed settlement text. It is neither a separate additional fine nor evidence that NAAG controlled the legal findings, and no bank/recipient payment record was read.
\nCalifornia ¶¶55–58, pages 12–13, and Oregon ¶¶42–45, pages 14–15, establish consumer redress of USD 75–120 million with three decisive conditions. First, redress payments under CFPB docket 2025-CFPB-0001 satisfy this obligation, with no further redress required under this judgment. Second, if Block has not paid at least USD 75 million in connection with that order by 16 January 2030, it must instead use a materially similar redress plan proposed to the Multistate AG Executive Committee by 18 March 2030, approved by that committee before payment. If payments are underway but unfinished on 16 January 2030, an update is required. Third, within ten days after sending CFPB the ¶122 Redress Report, Block must notify the committee that payments have been completed. Do not recast March 18 as a payment-completion deadline; it is the fallback plan-submission deadline.
\nThe executive committee is precisely the AGs of Oregon, Texas, Colorado, Connecticut, Florida, Indiana, Vermont and Washington (California ¶22; Oregon ¶11(n)). This is an explicit redress-plan approval route, not a general veto over every Cash App decision. The published text does not specify its internal voting rule.
\nThe earlier CFPB packet explains its USD 75 million reserve/floor and possible residue-to-CFPB/Treasury mechanism. This July text uses redress paid “in connection with” that order and requires a materially similar fallback plan. It does not support a fresh claim that all of the floor must already have reached consumers, or that a separate second USD 75–120 million is owed in addition. The signed terms, AG announcements and issuer statement establish an agreed resolution; none of these records proves actual July-settlement payment totals.
\nRelease follows full payment of the state-payment amount (California ¶¶61–63, pages 13–14; Oregon ¶¶49–51, pages 16–17). It broadly covers the state's civil claims for defined pre-effective conduct, preserves enforcement of the judgment and does not create, waive or limit private rights of action. Criminal liability and specified other liabilities remain outside the release. Parties' waiver of appeal is not a waiver by every customer.
\nThe California announcement's final settlement-information paragraph visibly labels and links http[local research file] through a publisher-supplied Proofpoint wrapper. More directly, the Oregon announcement's final paragraph links that exact domain with an ordinary href, alongside the CFPB official case page. Both links are retained in the official HTML. This establishes an official AG referral to the remediation site, correcting the earlier narrow search failure to find one. It still does not establish the administrator's appointment instrument, validate each check, or prove a payout total. This run did not revisit or transact with that site.
Strongest supported account: state AGs agreed an additional public enforcement and payment arrangement that expressly preserves/credits the existing federal redress commitment and assigns fallback-plan approval to a named eight-state committee. NAAG has a disclosed, bounded investigation-expense reimbursement within the state-payment pool. Customer remedies and anti-fraud restrictions coexist, with legally meaningful differences between the state and CFPB formulations.
\nThe strongest rival to treating this as a new centralized restriction mechanism is the text itself: repeated existing consumer-law duties, permission to satisfy terms through adequate existing procedures, a specific federal-redress credit, and no named external signal supplier. The record supports coordination of enforcement and remedy; it does not connect this settlement to Lantern decisions or the Cash App closure count in that separate case.
\nThe next decisive records are the entered California/Oregon judgment and docket entry (court custodians), any written timing amendment explaining or carrying forward the June 30 requirement, actual state/NAAG receipts, and the implemented restriction-review procedure plus committee-approved annual report (Block/state AG custodians). The Oregon signatures establish assent but stop short of observed judicial entry. No new court account, outreach or broader state hunt was undertaken. Acquisition stops at this signed-instrument breakpoint for the forest review.
\ncaptures-ag.json records five retained originals, all ordinary unauthenticated GETs with HTTP 200, hashes and dates: California notice, proposed judgment and complaint; Oregon notice and signed stipulated-judgment packet. The California judgment was read in full, PDF 1–21, with pages 1/21 visually checked. Oregon pages 1–21 and 26–31 were read, with signature/approval pages 28–30 visually checked; pages 22–25 were captured but not substantively read (notice-address continuation). California complaint scope is PDF 1 and 11–13 only: caption, account-access allegations ¶59–63, causes of action/prayer and 8 July 2026 date. The complaint is not an operative order. Both announcements' complete substantive bodies and relevant links were read; unrelated site-navigation content was not treated as substantive reading.
Oregon original SHA-256: 4aa51e7210a316a00ddaf3945765c07d9df02e9ae23b4457eda590086b4b5b9f. California proposed judgment SHA-256: 496f00ab441b042a11bd1742ea07a597d28b8c93052a8d59b5acf84213e45ed1. Extracts and PNGs are reading aids. Q2 filing scope was the July-resolution paragraph, printed 33/web-reader line 1259; no original SEC GET retry, and no full-filing reading claim. No case, atlas or ZIP was edited.
Subsequent enforcement-record finding (9 September 2026): the public descriptions analyzed below do not exhaust Cash App's obligations. The16January2025 CFPB consent order expressly requires restriction explanations and an account-appeal procedure with a10-business-day grant/deny deadline (paragraph108(c)–(d), qualified as stated). See the regulator packet. This narrows the earlier procedural gap; actual compliance, applicable modifications and a Lantern-specific correction/remedy trace remain unestablished. The original source-specific analysis below is preserved.
\nObserved 9 September 2026. Bounded agreement lane; no customer account, support contact, private API, historical contract reconstruction or enforceability opinion. The main instrument is the agreement with Block, Inc. It grants service-access and fund-access discretion separately. Neither contractual power proves why any particular account was closed, nor does closure itself establish loss of ownership of the remaining money.
\nThe current official [Cash App Terms of Service](http[local research file] displays Effective 7 February 2024; Last Updated 26 August 2026. The original HTML and derived text are retained as agreement-cashapp-tos.html / .txt. The opening identifies Block, Inc. as the contracting company and says gray annotations have no legal effect (derived lines 1–10). Section II distinguishes revised terms from terms applicable to a dispute arising before a change (lines 213–219). This is a present public instrument, not proof of the terms accepted by the accounts in the 2025 Lantern example. Earlier search-index text displayed June 2026 dates; the captured original and fresh canonical page control this reading.
The incorporated [Acceptable Use Policy](http[local research file] was captured separately as agreement-cashapp-aup.html / .txt. Its displayed text has no effective/updated date; 9 September is the observation date, not an inferred adoption date. The complete policy was read, emphasizing the clauses below. These instruments are two related company sources, not independent corroboration. Exact URLs, response status, acquisition timestamps, sizes, SHA-256 hashes and scoped reading extent are in captures-agreement.json.
| Mechanism | \nOperative provision and retained locator | \nEvidence boundary | \n
|---|---|---|
| End service access | \nTerms XXIII.9, text 1276–1278: Block can suspend/terminate the account, associated accounts or service access for any reason; reasonable notification steps by email or next attempted app access. | \nReasonable notice steps are not an advance-notice period, detailed-reasons promise or prior hearing. | \n
| AUP enforcement | \nAUP II, text 70–74: reported violations can produce suspension/termination without notice; suspected violations may be investigated. | \nRead alongside the general notice wording, without declaring which clause legally prevails in a specific dispute. | \n
| Delay ordinary cash-out | \nTerms VI.1(e)(iii), text 414: discretion to delay/limit transfers for financial exposure or investigation; also deferral/redirection/restriction for law, subpoena, court order or governmental request. | \nDo not say every restriction requires a court order; the clause includes other stated bases. | \n
| Balance after closure | \nTerms XXIII.11, text 1282–1284: pending transactions settle; withdrawal of custodial funds less applicable fees is conditional/discretionary; a new linked bank may be required; risk and investigation holds are permitted. | \nNo universal release deadline here. The broader gray annotation about returning funds is expressly non-operative. A hold is not proof of confiscation. | \n
| Transaction refusal | \nTerms XXIII.4, text 1247–1252: funding sources and transactions can be restricted; notice has prohibited-transaction exceptions. | \nAvailable authority, not a demonstrated instruction from a particular network/bank. | \n
| Compelled fund action | \nTerms XXIII.3, text 1245–1246: freeze, withholding or remittance in response to specified binding instruments. | \nDistinct from the discretionary risk clauses and ordinary closure. No such order was acquired for the Lantern example. | \n
| Recover liabilities | \nTerms XXIII.1, text 1232–1242: recovery/setoff against account or linked funds for stated liabilities, disputes and other agreement grounds. | \nDo not convert contractual recovery language into a right to take every closed account's balance. | \n
| Custodial access | \nTerms VI.3, text 443–445; XXIII.8, text 1274–1275: separate corporate/customer funds, possible pooling, Block agency, access through Cash App rather than direct withdrawal from custodial banks/brokerages. | \nA documented intermediary access condition; neither partner-bank direction nor sole legal ownership by Block follows. | \n
The AUP has separate restrictions for criminal activity (including child sexual-abuse material), adult-content sales, and businesses believed to create elevated financial risk or breach bank/card-network policies (AUP I, text 35–36, 54–55, 62–63). This is a contractual route by which external policy can matter. It identifies no particular bank, policy version, instruction or causal role in the Lantern closures. Its sanctions subsection separately contemplates indefinite holding in apparent violations; that is not the universal rule for closed accounts (text 50–51).
\nThe terms distinguish the Cash App service contract from a cardholder's agreement with the issuing bank; that bank agreement governs card use/bank relationship in an inconsistency (VIII.1–2, text 620–626). No individual customer's product/card agreement was obtained, and no bank-specific closure authority is asserted here. Custodial placement, card issuance and an instruction to close are different relations.
\nXXIII.19–20 contains a broad dispute definition and a separate pre-filing informal dispute process, including written notice, a personal conference and a 60-day negotiation window (text 1314–1323). This is not a routine support appeal or a promise to restore access. Only the opening and pre-filing paragraphs were substantively read; this packet does not evaluate enforceability, opt-outs or every arbitration condition. Parent's separately retained help article addresses operational closure review. Its ordinary withdrawal instructions should not be treated as overriding the agreement's hold conditions.
\nThe public instruments establish a service-mediated path to money and a broad internal decision space. Competing explanations remain live: fraud/risk protection, compliance with an external obligation, enforcement of Cash App's own acceptable-use choices, or a mistaken decision. The contract alone cannot choose among them. Reinstatement review and balance release also remain different outcomes.
\nThe most discriminating missing records are the actual closure reason/notice, applicable customer/product terms at the event date, any hold rationale and release ledger, and the internal or external instruction that caused the decision. The relevant holders would be Block's account/support/compliance records and, only if a bank or authority actually participated, that actor's corresponding instrument. No inference of current access to such records is made. No public account-level record was acquired for the 18-account Lantern example; no historical backfill, bank-name hunt or additional acquisition is proposed by this packet.
\nAcquisition closed after two successful ordinary public GETs. Full original capture is not a whole-document legal review. No canonical graph, package or research note outside this lane was modified.
", "path": "cashapp-authority-wave-2026-09-09/cashapp-agreement.md", "sourceIds": [ "w23-cash-s-terms", "w23-cash-s-aup" ], "markdown": "Subsequent enforcement-record finding (9 September 2026): the public descriptions analyzed below do not exhaust Cash App's obligations. The16January2025 CFPB consent order expressly requires restriction explanations and an account-appeal procedure with a10-business-day grant/deny deadline (paragraph108(c)–(d), qualified as stated). See [the regulator packet](cashapp-regulator-wave-2026-09-09-cfpb.md). This narrows the earlier procedural gap; actual compliance, applicable modifications and a Lantern-specific correction/remedy trace remain unestablished. The original source-specific analysis below is preserved.\n\nObserved 9 September 2026. Bounded agreement lane; no customer account, support contact, private API, historical contract reconstruction or enforceability opinion. The main instrument is the agreement with **Block, Inc.** It grants service-access and fund-access discretion separately. Neither contractual power proves why any particular account was closed, nor does closure itself establish loss of ownership of the remaining money.\n\n## Instruments and custody\n\nThe current official Cash App Terms of Service (see source references) displays **Effective 7 February 2024; Last Updated 26 August 2026**. The original HTML and derived text are retained as `agreement-cashapp-tos.html` / `.txt`. The opening identifies Block, Inc. as the contracting company and says gray annotations have no legal effect (derived lines 1–10). Section II distinguishes revised terms from terms applicable to a dispute arising before a change (lines 213–219). This is a present public instrument, not proof of the terms accepted by the accounts in the 2025 Lantern example. Earlier search-index text displayed June 2026 dates; the captured original and fresh canonical page control this reading.\n\nThe incorporated Acceptable Use Policy (see source references) was captured separately as `agreement-cashapp-aup.html` / `.txt`. Its displayed text has no effective/updated date; 9 September is the observation date, not an inferred adoption date. The complete policy was read, emphasizing the clauses below. These instruments are two related company sources, not independent corroboration. Exact URLs, response status, acquisition timestamps, sizes, SHA-256 hashes and scoped reading extent are in `captures-agreement.json`.\n\n## Decision rights and limits\n\n| Mechanism | Operative provision and retained locator | Evidence boundary |\n| --- | --- | --- |\n| End service access | Terms XXIII.9, text 1276–1278: Block can suspend/terminate the account, associated accounts or service access for any reason; reasonable notification steps by email or next attempted app access. | Reasonable notice steps are not an advance-notice period, detailed-reasons promise or prior hearing. |\n| AUP enforcement | AUP II, text 70–74: reported violations can produce suspension/termination without notice; suspected violations may be investigated. | Read alongside the general notice wording, without declaring which clause legally prevails in a specific dispute. |\n| Delay ordinary cash-out | Terms VI.1(e)(iii), text 414: discretion to delay/limit transfers for financial exposure or investigation; also deferral/redirection/restriction for law, subpoena, court order or governmental request. | Do not say every restriction requires a court order; the clause includes other stated bases. |\n| Balance after closure | Terms XXIII.11, text 1282–1284: pending transactions settle; withdrawal of custodial funds less applicable fees is conditional/discretionary; a new linked bank may be required; risk and investigation holds are permitted. | No universal release deadline here. The broader gray annotation about returning funds is expressly non-operative. A hold is not proof of confiscation. |\n| Transaction refusal | Terms XXIII.4, text 1247–1252: funding sources and transactions can be restricted; notice has prohibited-transaction exceptions. | Available authority, not a demonstrated instruction from a particular network/bank. |\n| Compelled fund action | Terms XXIII.3, text 1245–1246: freeze, withholding or remittance in response to specified binding instruments. | Distinct from the discretionary risk clauses and ordinary closure. No such order was acquired for the Lantern example. |\n| Recover liabilities | Terms XXIII.1, text 1232–1242: recovery/setoff against account or linked funds for stated liabilities, disputes and other agreement grounds. | Do not convert contractual recovery language into a right to take every closed account's balance. |\n| Custodial access | Terms VI.3, text 443–445; XXIII.8, text 1274–1275: separate corporate/customer funds, possible pooling, Block agency, access through Cash App rather than direct withdrawal from custodial banks/brokerages. | A documented intermediary access condition; neither partner-bank direction nor sole legal ownership by Block follows. |\n\nThe AUP has separate restrictions for criminal activity (including child sexual-abuse material), adult-content sales, and businesses believed to create elevated financial risk or breach bank/card-network policies (AUP I, text 35–36, 54–55, 62–63). This is a contractual route by which external policy can matter. It identifies no particular bank, policy version, instruction or causal role in the Lantern closures. Its sanctions subsection separately contemplates indefinite holding in apparent violations; that is not the universal rule for closed accounts (text 50–51).\n\n## Bank roles and reconsideration\n\nThe terms distinguish the Cash App service contract from a cardholder's agreement with the issuing bank; that bank agreement governs card use/bank relationship in an inconsistency (VIII.1–2, text 620–626). No individual customer's product/card agreement was obtained, and no bank-specific closure authority is asserted here. Custodial placement, card issuance and an instruction to close are different relations.\n\nXXIII.19–20 contains a broad dispute definition and a separate pre-filing informal dispute process, including written notice, a personal conference and a 60-day negotiation window (text 1314–1323). This is not a routine support appeal or a promise to restore access. Only the opening and pre-filing paragraphs were substantively read; this packet does not evaluate enforceability, opt-outs or every arbitration condition. Parent's separately retained help article addresses operational closure review. Its ordinary withdrawal instructions should not be treated as overriding the agreement's hold conditions.\n\n## What this changes, and what remains missing\n\nThe public instruments establish a service-mediated path to money and a broad internal decision space. Competing explanations remain live: fraud/risk protection, compliance with an external obligation, enforcement of Cash App's own acceptable-use choices, or a mistaken decision. The contract alone cannot choose among them. Reinstatement review and balance release also remain different outcomes.\n\nThe most discriminating missing records are the actual closure reason/notice, applicable customer/product terms at the event date, any hold rationale and release ledger, and the internal or external instruction that caused the decision. The relevant holders would be Block's account/support/compliance records and, only if a bank or authority actually participated, that actor's corresponding instrument. No inference of current access to such records is made. No public account-level record was acquired for the 18-account Lantern example; no historical backfill, bank-name hunt or additional acquisition is proposed by this packet.\n\nAcquisition closed after two successful ordinary public GETs. Full original capture is not a whole-document legal review. No canonical graph, package or research note outside this lane was modified.\n", "file": "research/cashapp-authority-wave-2026-09-09-cashapp-agreement.md" }, { "id": "cashapp-authority-wave-2026-09-09-cashapp-help", "title": "Cash App account-closure help: process and unresolved boundary", "description": "Subsequent enforcement-record finding (9 September 2026): the public descriptions analyzed below do not exhaust Cash App's obligations. The16January2025 CFPB consent order expressly requires restriction explanations and an account-appeal procedure with a10-business-day grant/deny deadline (paragraph108(c)–(d), qualified as stated). See the regulator packet(.", "html": "Subsequent enforcement-record finding (9 September 2026): the public descriptions analyzed below do not exhaust Cash App's obligations. The16January2025 CFPB consent order expressly requires restriction explanations and an account-appeal procedure with a10-business-day grant/deny deadline (paragraph108(c)–(d), qualified as stated). See the regulator packet. This narrows the earlier procedural gap; actual compliance, applicable modifications and a Lantern-specific correction/remedy trace remain unestablished. The original source-specific analysis below is preserved.
\nObserved 9 September 2026. Public help article, no stated publication/effective date. This is a provider's account of its support process, not observed performance or a legal conclusion. [Your Account Is Closed](http[local research file]
\nThe article describes a two-stage route: after contact with Support, the closure is first screened for appeal eligibility; an appropriate team then reviews possible re-enablement. The first reviewer is not named. It gives up to 10 days for review and says the customer will be contacted at completion. During the restriction, it says withdrawal and selling stocks or bitcoin remain possible. Pending direct deposits are to be returned to their originator in 3–5 business days. Cashing out may require a new bank account or debit card; normal availability is described as 1–3 business days, with faster options depending on the selected speed.
\nThis establishes a published route to contest closure, not an entitlement for every closure to enter a substantive appeal. No eligibility criteria, evidence disclosure rule, reviewer-independence requirement or outcome statistics appear in this article. Its general withdrawal guidance must be read with the agreement's hold conditions; it does not prove a particular customer could withdraw immediately. Nor does it identify a Lantern-specific appeal or provide a correction-to-reinstatement example.
\nAn ambiguous final sentence says an account opened and linked to a closed Cash App account will also close. The account type is not specified there. Do not interpret it as Cash App closing an external bank account, or add a causal edge without clarification from an authoritative instrument.
\nThe financially consequential distinction is between service restoration and the exit route for remaining money. These may follow different processes. Investigate the contract's hold authority and any actual adjudicated example separately before making a claim about confiscation, indefinite retention or practical remedy.
\nCustody: cashapp-closed-help.html, SHA256 c0dfaf7195b720422d8ea5d16cdaac01b083249cf87f65426b61039047afa28b. The primary web reader returned zero lines; ordinary GET succeeded. Full public article content was read from the HTML's embedded page-data object. No account, support submission, browser execution or nonpublic endpoint was used. captures-help.json records the acquisition. Extracted text is a reading aid, not a separate source.
", "path": "cashapp-authority-wave-2026-09-09/cashapp-help.md", "sourceIds": [ "w23-cash-s-help" ], "markdown": "Subsequent enforcement-record finding (9 September 2026): the public descriptions analyzed below do not exhaust Cash App's obligations. The16January2025 CFPB consent order expressly requires restriction explanations and an account-appeal procedure with a10-business-day grant/deny deadline (paragraph108(c)–(d), qualified as stated). See [the regulator packet](cashapp-regulator-wave-2026-09-09-cfpb.md). This narrows the earlier procedural gap; actual compliance, applicable modifications and a Lantern-specific correction/remedy trace remain unestablished. The original source-specific analysis below is preserved.\n\nObserved 9 September 2026. Public help article, no stated publication/effective date. This is a provider's account of its support process, not observed performance or a legal conclusion. Your Account Is Closed (see source references)\n\nThe article describes a two-stage route: after contact with Support, the closure is first screened for appeal eligibility; an appropriate team then reviews possible re-enablement. The first reviewer is not named. It gives up to 10 days for review and says the customer will be contacted at completion. During the restriction, it says withdrawal and selling stocks or bitcoin remain possible. Pending direct deposits are to be returned to their originator in 3–5 business days. Cashing out may require a new bank account or debit card; normal availability is described as 1–3 business days, with faster options depending on the selected speed.\n\nThis establishes a published route to contest closure, not an entitlement for every closure to enter a substantive appeal. No eligibility criteria, evidence disclosure rule, reviewer-independence requirement or outcome statistics appear in this article. Its general withdrawal guidance must be read with the agreement's hold conditions; it does not prove a particular customer could withdraw immediately. Nor does it identify a Lantern-specific appeal or provide a correction-to-reinstatement example.\n\nAn ambiguous final sentence says an account opened and linked to a closed Cash App account will also close. The account type is not specified there. Do not interpret it as Cash App closing an external bank account, or add a causal edge without clarification from an authoritative instrument.\n\nThe financially consequential distinction is between service restoration and the exit route for remaining money. These may follow different processes. Investigate the contract's hold authority and any actual adjudicated example separately before making a claim about confiscation, indefinite retention or practical remedy.\n\nCustody: cashapp-closed-help.html, SHA256 c0dfaf7195b720422d8ea5d16cdaac01b083249cf87f65426b61039047afa28b. The primary web reader returned zero lines; ordinary GET succeeded. Full public article content was read from the HTML's embedded page-data object. No account, support submission, browser execution or nonpublic endpoint was used. captures-help.json records the acquisition. Extracted text is a reading aid, not a separate source.\n", "file": "research/cashapp-authority-wave-2026-09-09-cashapp-help.md" }, { "id": "cashapp-regulator-wave-2026-09-09-cfpb", "title": "CFPB / Block: account restrictions, error redress and accountable decision-making", "description": "Observed 9 September 2026. The operative instrument is the 16 January 2025 consent order(httplocal research file administrative proceeding 2025-CFPB-0001, against Block, Inc., defined to include successors and assigns. It explicitly reaches account-restriction appeals as well as transaction-error investigation. It supplies a legally required process and a re", "html": "Observed 9 September 2026. The operative instrument is the [16 January 2025 consent order](http[local research file] administrative proceeding 2025-CFPB-0001, against Block, Inc., defined to include successors and assigns. It explicitly reaches account-restriction appeals as well as transaction-error investigation. It supplies a legally required process and a regulator-monitoring route; it does not establish that any particular appeal succeeded, that all customers actually received timely service, or that Lantern caused the investigated conduct.
\nOrder ¶¶1–3, PDF 2–10, establish jurisdiction, effective date on administrative-docket entry and definitions. ¶4, PDF 11, identifies Block as a Delaware corporation operating Cash App. The stipulation.
\nThe retained [official case page](http[local research file] displays “Post Order/Post Judgment”; its last-modified date is 24 January 2025. The retained [official docket](http[local research file] likewise displays post-order status and lists only document 001 (order) and 002 (stipulation), both filed 16 January 2025; last modified 16 January 2025. Two targeted official-domain searches for this docket/order with termination and 2026 amendment terms did not reveal a later CFPB instrument. This is a scoped public-record observation, not certification that no written modification exists. Order ¶¶154–155 permit written changes to non-material requirements for good cause, including extensions/reporting changes. ¶159, PDF 73–74, sets termination at the later of five years from effective date or the relevant qualifying enforcement-action extension, subject to written amendment, suspension, waiver or termination. The initial five-year date is 16 January 2030; the packet does not assert an unchanged end date.
\nThe [official 16 January 2025 notice](http[local research file] is now expressly marked archived, with an outdated-content warning and last-modified date 25 June 2026. That modification stamp is not a new order date. Use the operative order for exact hours, remedies and exceptions.
\nThe order describes failures across several historical periods, not one uniform period applicable to every finding. ¶¶7–18, PDF 12–16: no live telephone customer support until February 2021, misleading contact representations, scams exploiting fake support numbers, and slow/form responses. ¶¶24–27, PDF 18–19: automatic “need more information,” “pushback” and bulk “inquiry” communications delayed or prematurely closed error cases and shifted work back to consumers. These are described templates/macros; the instrument does not establish a generative-AI system or join these practices to Lantern.
\nThe most direct account-access findings are ¶¶34–36, PDF 21. CFPB found missing restriction notifications, generic terms-of-service explanations, failure proactively to publicize the appeal process introduced around 2021, slow reconsideration, weeks to unfreeze even successful appeals, and consumers unaware that withdrawal might be available. ¶¶55–58, PDF 26–28, explain the regulator's unfairness finding: consumers' dependence on reaching the company made loss of access difficult to avoid. This establishes a historically identified failure of recourse, rather than proof that closure always confiscates funds.
\nTransaction-error findings are separate: ¶¶19–33,39–48 and72–105, PDF 16–25,32–40, describe deficient investigations, denied/delayed provisional credits and refunds, missing explanations/document-access notices, misapplied reporting windows, and wrongful substitution of linked-bank chargebacks. The rule-and-order analysis does not depend solely on Cash App's help wording or contract discretion.
\nAll locators below refer to the order; PDF and printed page numbers agree.
\n| Decision or duty | \nActor and operative requirement | \nExact scope / limit | \n
|---|---|---|
| Reachable customer service | \nBlock must maintain service able to intake and appropriately resolve or escalate complaints, error notices, account-access issues and fraud reports, including noncustomers; publish access instructions. ¶108(a)–(b), PDF 44–45. | \nService available 24 hours/day; human voice at least 12 hours/day; human chat at least 18 hours/day. Do not turn the notice's shorthand into 24 hours of human telephone support. | \n
| Restriction notice | \nWithin 24 hours of freezing, suspending, locking, closing or otherwise restricting an account, provide written explanation of the decision and appeal procedure. ¶108(c), PDF 45–46. | \nException is expressly “unless prohibited from doing so by law enforcement.” Except when funds are frozen, also say whether and how funds can be transferred out, if possible. The frozen-funds exception belongs to the transfer-information limb; do not silently extend it to every notice duty. | \n
| Account appeal | \nEstablish/maintain an appeal or change-status procedure, clearly publish instructions on the support primary webpage and in-app; grant or deny within 10 business days from receiving the request, with written explanation if denied. ¶108(d), PDF 46–47. | \nThis is an account-restriction requirement, not merely a transaction-error rule. It does not guarantee reinstatement or establish an independent external adjudicator for each appeal. | \n
| Transaction-error decision and evidence | \nPrompt, thorough, reasonable investigation; report results within 3 business days of completion; explain no-error/different-error findings, notify the right to relied-on documents and promptly supply copies on request. Correct found errors within 1 business day; provisionally credit within 10 business days if investigation incomplete. ¶108(e), PDF 47. | \nApplies to a defined Notice of Error, ¶3(f), PDF 9–10. Do not transfer this express document-access right wholesale to every account-closure appeal. | \n
| Stop obstructing error investigation | \nNo unnecessary-information delaying communications, required police/merchant contact or extra information as investigation prerequisites; no stopping/delaying because further information was not supplied; no sending qualifying linked-account notices away to the issuing bank. ¶107(a)–(g), PDF 41–44. | \nConcerns the specified financial-service/error processes. This is not a prohibition on all lawful verification, fraud control or information requests. | \n
| Fraud controls and account exclusion | \nPolicies must identify accounts suspected of fraudulently inducing transactions, prevent their operation and prevent associated individuals returning; track complaints/trends, provide refunds under Cash App policies and warnings for suspected scam transfers. ¶108(f), PDF 48. | \nThe same order requires anti-fraud restrictions and notice/appeal safeguards. No named signal supplier, mandatory Lantern membership or cross-platform exclusion instruction appears. | \n
| Internal responsibility | \nCommittee with required risk, compliance, business, customer-operations and data-science roles, and three working groups; assessment, escalation, reporting, resources and correction duties. ¶109, PDF 49–52. Board ultimately responsible; executives/Board review plans and submissions; annual Board-approved sworn progress report. ¶¶113–116, PDF 54–56. | \nThese are mandated corporate responsibilities. The order does not identify outside NGOs or vendors as decision-makers. | \n
| Plans and supervisory records | \nCreate/implement Compliance Plan within 120 days; provide it to Bureau on request. ¶¶110–112, PDF 53–54. Keep compliance, redress and complaint records in original electronic form, accessible for historical analysis; produce on request. ¶¶143–145, PDF 67–69. | \nCompliance Plan is not expressly subject to the same prior non-objection procedure as the Redress Plan. Required plans/reports were not obtained here. | \n
| Regulator oversight and complaint route | \nIdentify CFPB complaint channel, remain on Company Portal and comply with response requirements. ¶¶148–149, PDF 70. On written request, additional sworn reports/testimony/documents within 14 days; consenting affiliated interviewees may have counsel. ¶¶150–153, PDF 71. | \nThis supplies monitoring and escalation powers, not proof that CFPB independently adjudicates every restriction appeal. Bureau enforces order; ¶162, PDF 74–75. | \n
The earlier help article's eligibility screen and unspecified “up to 10 days” do not exhaust the operative duty. The order measures the appeal/change-status decision from receipt of the request and expressly uses business days. Establishing an actual violation would require the applicable procedure, request/notice dates and response, plus any relevant written modification; comparing two short public descriptions is insufficient.
\nOrder ¶¶117–125, PDF 56–60, require USD 75 million to be reserved or placed in a segregated deposit account within 10 days; redress to affected consumers is capped at USD 120 million. These are commitments, not observed paid amounts. ¶126–128, PDF 61, separately assess a USD 55 million civil penalty, payable to CFPB within 10 days for its Civil Penalty Fund. The notice's USD 175 million headline is the upper redress figure plus that penalty, not evidence USD 175 million reached consumers. Per ¶160, PDF 74, unspecified day limits are calendar days.
\nA further allocation distinction matters: if completed consumer redress is below USD 75 million, ¶123 requires the shortfall to be sent to CFPB 90 days after the Redress Report. CFPB may use it for further redress and may send impracticable/inappropriate-to-distribute residue to Treasury (¶124). Thus the minimum commitment does not guarantee USD 75 million is directly received by consumers. No return to Block is permitted for funds paid (¶132, PDF 63), and redress cannot be conditioned on customers waiving rights (¶125).
\nAffected categories and formulae are in ¶3(a), PDF 3–8, and ¶118, PDF 56–57. They include noninvestigated and wrongly timed unauthorized-transfer claims, partial-refund cases, omitted provisional credits, and frozen accounts that were later unfrozen. The latter requires a freeze of at least 14 days and balance of at least USD 25, with exclusions for specified law-enforcement/government/statutory grounds and customer-caused verification delay. Its payment is a pro-rata portion of USD 50, not a universal USD 50 per closed account. The category definition refers to transactions as of 1 July 2019; the order's release is separately limited to known pre-effective-date practices (¶157, PDF 73).
\nWithin 120 days Block must submit its Redress Plan for the Enforcement Director's review/non-objection. The Director may require revision, due within 15 days; Block must follow the non-objected-to plan. ¶¶119–120, PDF 57–59. It must pay credit redress and begin non-credit redress within 30 days of non-objection (¶121); completion-report due within 15 days after plan completion (¶122). The report must state recipient/account counts and amounts paid. This packet did not obtain that report, the non-objection, plan-level distribution deadlines or any extension. A required payment is not evidence of receipt or timely completion.
\nDefensible directional relations are CFPB → Block: imposes account-restriction notice/appeal and error-redress obligations; Block Board → compliance: has ultimate responsibility; CFPB Enforcement Director → Redress Plan: reviews/non-objects or requires revision; Block → affected consumers/CFPB: owes specified redress/penalty obligations. Distinguish each from implementation and payment events. The record supplies a material counterweight to an unrestricted-private-discretion narrative: a regulator identified procedural harms and imposed deadlines, reasons, record access for error claims and reporting. It also orders anti-fraud account exclusion, so neither unrestricted access nor unconditional restoration follows.
\nNo mention of Lantern or Tech Coalition was found in the complete extracted order/stipulation or case/notice reading. That does not prove Cash App never uses those systems; it means this instrument supplies no attribution to them. General account restriction, transaction error, redress eligibility and Lantern signal correction remain separate mechanisms unless a later record joins them.
\nThe strongest next discriminator is the Board-approved annual compliance report and implemented account-appeal procedure under ¶¶108(d),110–116, plus any written modification; for distribution, the Redress Plan, non-objection and ¶122 Redress Report. Custodians are Block and CFPB Enforcement. These would show how the legal requirement became a working process, what exceptions were applied and what was actually paid. Their existence is required by the order; this packet does not assert that they are publicly released. Acquisition stops here for the forest assessment.
\nFive originals were obtained by ordinary unauthenticated GET, all HTTP 200, on 9 September 2026 at approximately 10:53 UTC. Metadata, final URLs, byte counts and SHA-256 are in captures-cfpb.json. cfpb-order.pdf SHA-256 8b72ad0538889fd6b4e9d1481806611955fb9a0e4e219bb87f2809471c351ed1; cfpb-stipulation.pdf SHA-256 7634dc07a22f9abfad610af9464c10101a5f897190fa6b9fba2597316f227100.
The order was read in full through the extracted text of PDF 1–75, with PDF 46 visually inspected for the notice/exception/appeal structure. The stipulation was read in full through PDF 1–6, with PDF 6 visually inspected because extraction of the final waivers/signature page was degraded. The full main article/case/docket bodies and displayed metadata were read from retained HTML; navigation/footer inventories were not treated as substantive source reading. Captured originals are sources; extracts and two PNGs are reading aids. No appeal submission, private account, outreach, paid route, canonical edit or further branch was undertaken.
", "path": "cashapp-regulator-wave-2026-09-09/cfpb.md", "sourceIds": [ "w23-cash-s-cfpb-order", "w23-cash-s-stipulation", "w23-cash-s-cfpb-case", "w23-cash-s-cfpb-docket", "w23-cash-s-cfpb-notice" ], "markdown": "Observed 9 September 2026. The operative instrument is the 16 January 2025 consent order (see source references) administrative proceeding **2025-CFPB-0001**, against **Block, Inc.**, defined to include successors and assigns. It explicitly reaches account-restriction appeals as well as transaction-error investigation. It supplies a legally required process and a regulator-monitoring route; it does not establish that any particular appeal succeeded, that all customers actually received timely service, or that Lantern caused the investigated conduct.\n\n## Instrument, consent and currentness\n\nOrder ¶¶1–3, PDF 2–10, establish jurisdiction, effective date on administrative-docket entry and definitions. ¶4, PDF 11, identifies Block as a Delaware corporation operating Cash App. The stipulation.\n\nThe retained official case page (see source references) displays “Post Order/Post Judgment”; its last-modified date is 24 January 2025. The retained official docket (see source references) likewise displays post-order status and lists only document 001 (order) and 002 (stipulation), both filed 16 January 2025; last modified 16 January 2025. Two targeted official-domain searches for this docket/order with termination and 2026 amendment terms did not reveal a later CFPB instrument. This is a scoped public-record observation, **not certification that no written modification exists**. Order ¶¶154–155 permit written changes to non-material requirements for good cause, including extensions/reporting changes. ¶159, PDF 73–74, sets termination at the later of five years from effective date or the relevant qualifying enforcement-action extension, subject to written amendment, suspension, waiver or termination. The initial five-year date is 16 January 2030; the packet does not assert an unchanged end date.\n\nThe official 16 January 2025 notice (see source references) is now expressly marked archived, with an outdated-content warning and last-modified date 25 June 2026. That modification stamp is not a new order date. Use the operative order for exact hours, remedies and exceptions.\n\n## What the Bureau found\n\nThe order describes failures across several historical periods, not one uniform period applicable to every finding. ¶¶7–18, PDF 12–16: no live telephone customer support until February 2021, misleading contact representations, scams exploiting fake support numbers, and slow/form responses. ¶¶24–27, PDF 18–19: automatic “need more information,” “pushback” and bulk “inquiry” communications delayed or prematurely closed error cases and shifted work back to consumers. These are described templates/macros; the instrument does not establish a generative-AI system or join these practices to Lantern.\n\nThe most direct account-access findings are ¶¶34–36, PDF 21. CFPB found missing restriction notifications, generic terms-of-service explanations, failure proactively to publicize the appeal process introduced around 2021, slow reconsideration, weeks to unfreeze even successful appeals, and consumers unaware that withdrawal might be available. ¶¶55–58, PDF 26–28, explain the regulator's unfairness finding: consumers' dependence on reaching the company made loss of access difficult to avoid. This establishes a historically identified failure of recourse, rather than proof that closure always confiscates funds.\n\nTransaction-error findings are separate: ¶¶19–33,39–48 and72–105, PDF 16–25,32–40, describe deficient investigations, denied/delayed provisional credits and refunds, missing explanations/document-access notices, misapplied reporting windows, and wrongful substitution of linked-bank chargebacks. The rule-and-order analysis does not depend solely on Cash App's help wording or contract discretion.\n\n## Who must do what\n\nAll locators below refer to the order; PDF and printed page numbers agree.\n\n| Decision or duty | Actor and operative requirement | Exact scope / limit |\n| --- | --- | --- |\n| Reachable customer service | Block must maintain service able to intake and appropriately resolve or escalate complaints, error notices, account-access issues and fraud reports, including noncustomers; publish access instructions. ¶108(a)–(b), PDF 44–45. | Service available 24 hours/day; human voice at least 12 hours/day; human chat at least 18 hours/day. Do not turn the notice's shorthand into 24 hours of human telephone support. |\n| Restriction notice | Within 24 hours of freezing, suspending, locking, closing or otherwise restricting an account, provide written explanation of the decision and appeal procedure. ¶108(c), PDF 45–46. | Exception is expressly “unless prohibited from doing so by law enforcement.” Except when funds are frozen, also say whether and how funds can be transferred out, if possible. The frozen-funds exception belongs to the transfer-information limb; do not silently extend it to every notice duty. |\n| Account appeal | Establish/maintain an appeal or change-status procedure, clearly publish instructions on the support primary webpage and in-app; grant or deny within 10 business days from receiving the request, with written explanation if denied. ¶108(d), PDF 46–47. | This is an account-restriction requirement, not merely a transaction-error rule. It does not guarantee reinstatement or establish an independent external adjudicator for each appeal. |\n| Transaction-error decision and evidence | Prompt, thorough, reasonable investigation; report results within 3 business days of completion; explain no-error/different-error findings, notify the right to relied-on documents and promptly supply copies on request. Correct found errors within 1 business day; provisionally credit within 10 business days if investigation incomplete. ¶108(e), PDF 47. | Applies to a defined Notice of Error, ¶3(f), PDF 9–10. Do not transfer this express document-access right wholesale to every account-closure appeal. |\n| Stop obstructing error investigation | No unnecessary-information delaying communications, required police/merchant contact or extra information as investigation prerequisites; no stopping/delaying because further information was not supplied; no sending qualifying linked-account notices away to the issuing bank. ¶107(a)–(g), PDF 41–44. | Concerns the specified financial-service/error processes. This is not a prohibition on all lawful verification, fraud control or information requests. |\n| Fraud controls and account exclusion | Policies must identify accounts suspected of fraudulently inducing transactions, prevent their operation and prevent associated individuals returning; track complaints/trends, provide refunds under Cash App policies and warnings for suspected scam transfers. ¶108(f), PDF 48. | The same order requires anti-fraud restrictions **and** notice/appeal safeguards. No named signal supplier, mandatory Lantern membership or cross-platform exclusion instruction appears. |\n| Internal responsibility | Committee with required risk, compliance, business, customer-operations and data-science roles, and three working groups; assessment, escalation, reporting, resources and correction duties. ¶109, PDF 49–52. Board ultimately responsible; executives/Board review plans and submissions; annual Board-approved sworn progress report. ¶¶113–116, PDF 54–56. | These are mandated corporate responsibilities. The order does not identify outside NGOs or vendors as decision-makers. |\n| Plans and supervisory records | Create/implement Compliance Plan within 120 days; provide it to Bureau on request. ¶¶110–112, PDF 53–54. Keep compliance, redress and complaint records in original electronic form, accessible for historical analysis; produce on request. ¶¶143–145, PDF 67–69. | Compliance Plan is not expressly subject to the same prior non-objection procedure as the Redress Plan. Required plans/reports were not obtained here. |\n| Regulator oversight and complaint route | Identify CFPB complaint channel, remain on Company Portal and comply with response requirements. ¶¶148–149, PDF 70. On written request, additional sworn reports/testimony/documents within 14 days; consenting affiliated interviewees may have counsel. ¶¶150–153, PDF 71. | This supplies monitoring and escalation powers, not proof that CFPB independently adjudicates every restriction appeal. Bureau enforces order; ¶162, PDF 74–75. |\n\nThe earlier help article's eligibility screen and unspecified “up to 10 days” do not exhaust the operative duty. The order measures the appeal/change-status decision from receipt of the request and expressly uses business days. Establishing an actual violation would require the applicable procedure, request/notice dates and response, plus any relevant written modification; comparing two short public descriptions is insufficient.\n\n## Required money, allocation and timing\n\nOrder ¶¶117–125, PDF 56–60, require **USD 75 million** to be reserved or placed in a segregated deposit account within 10 days; redress to affected consumers is capped at **USD 120 million**. These are commitments, not observed paid amounts. ¶126–128, PDF 61, separately assess a **USD 55 million civil penalty**, payable to CFPB within 10 days for its Civil Penalty Fund. The notice's USD 175 million headline is the upper redress figure plus that penalty, not evidence USD 175 million reached consumers. Per ¶160, PDF 74, unspecified day limits are calendar days.\n\nA further allocation distinction matters: if completed consumer redress is below USD 75 million, ¶123 requires the shortfall to be sent to CFPB 90 days after the Redress Report. CFPB may use it for further redress and may send impracticable/inappropriate-to-distribute residue to Treasury (¶124). Thus the minimum commitment does not guarantee USD 75 million is directly received by consumers. No return to Block is permitted for funds paid (¶132, PDF 63), and redress cannot be conditioned on customers waiving rights (¶125).\n\nAffected categories and formulae are in ¶3(a), PDF 3–8, and ¶118, PDF 56–57. They include noninvestigated and wrongly timed unauthorized-transfer claims, partial-refund cases, omitted provisional credits, and frozen accounts that were later unfrozen. The latter requires a freeze of at least 14 days and balance of at least USD 25, with exclusions for specified law-enforcement/government/statutory grounds and customer-caused verification delay. Its payment is a pro-rata portion of USD 50, not a universal USD 50 per closed account. The category definition refers to transactions as of 1 July 2019; the order's release is separately limited to known pre-effective-date practices (¶157, PDF 73).\n\nWithin 120 days Block must submit its Redress Plan for the Enforcement Director's review/non-objection. The Director may require revision, due within 15 days; Block must follow the non-objected-to plan. ¶¶119–120, PDF 57–59. It must pay credit redress and begin non-credit redress within 30 days of non-objection (¶121); completion-report due within 15 days after plan completion (¶122). The report must state recipient/account counts and amounts paid. This packet did not obtain that report, the non-objection, plan-level distribution deadlines or any extension. A required payment is not evidence of receipt or timely completion.\n\n## Integration and next discriminating record\n\nDefensible directional relations are **CFPB → Block: imposes account-restriction notice/appeal and error-redress obligations**; **Block Board → compliance: has ultimate responsibility**; **CFPB Enforcement Director → Redress Plan: reviews/non-objects or requires revision**; **Block → affected consumers/CFPB: owes specified redress/penalty obligations**. Distinguish each from implementation and payment events. The record supplies a material counterweight to an unrestricted-private-discretion narrative: a regulator identified procedural harms and imposed deadlines, reasons, record access for error claims and reporting. It also orders anti-fraud account exclusion, so neither unrestricted access nor unconditional restoration follows.\n\nNo mention of Lantern or Tech Coalition was found in the complete extracted order/stipulation or case/notice reading. That does not prove Cash App never uses those systems; it means this instrument supplies no attribution to them. General account restriction, transaction error, redress eligibility and Lantern signal correction remain separate mechanisms unless a later record joins them.\n\nThe strongest next discriminator is the **Board-approved annual compliance report and implemented account-appeal procedure under ¶¶108(d),110–116**, plus any written modification; for distribution, the **Redress Plan, non-objection and ¶122 Redress Report**. Custodians are Block and CFPB Enforcement. These would show how the legal requirement became a working process, what exceptions were applied and what was actually paid. Their existence is required by the order; this packet does not assert that they are publicly released. Acquisition stops here for the forest assessment.\n\n## Custody and reading scope\n\nFive originals were obtained by ordinary unauthenticated GET, all HTTP 200, on 9 September 2026 at approximately 10:53 UTC. Metadata, final URLs, byte counts and SHA-256 are in `captures-cfpb.json`. `cfpb-order.pdf` SHA-256 `8b72ad0538889fd6b4e9d1481806611955fb9a0e4e219bb87f2809471c351ed1`; `cfpb-stipulation.pdf` SHA-256 `7634dc07a22f9abfad610af9464c10101a5f897190fa6b9fba2597316f227100`.\n\nThe order was read in full through the extracted text of PDF 1–75, with PDF 46 visually inspected for the notice/exception/appeal structure. The stipulation was read in full through PDF 1–6, with PDF 6 visually inspected because extraction of the final waivers/signature page was degraded. The full main article/case/docket bodies and displayed metadata were read from retained HTML; navigation/footer inventories were not treated as substantive source reading. Captured originals are sources; extracts and two PNGs are reading aids. No appeal submission, private account, outreach, paid route, canonical edit or further branch was undertaken.\n", "file": "research/cashapp-regulator-wave-2026-09-09-cfpb.md" }, { "id": "cashapp-regulator-wave-2026-09-09-issuer-disclosures", "title": "Block disclosures: paid, promised and still unresolved", "description": "Observed9September2026. This lane tests execution evidence against the regulator instruments. It uses official SEC filings and Block's contemporaneous response, with a separately labelled remediation-site lead whose official referral was established in the subsequent July AG run. Exact source URLs, filing dates and scoped reads are in root-reader-sources.jso", "html": "Observed9September2026. This lane tests execution evidence against the regulator instruments. It uses official SEC filings and Block's contemporaneous response, with a separately labelled remediation-site lead whose official referral was established in the subsequent July AG run. Exact source URLs, filing dates and scoped reads are in root-reader-sources.json. No underlying bank-transfer receipt, consumer account or internal compliance report was obtained.
\nBlock's [2024 Form10-K](http[local research file] filed24February2025, states in its litigation note (printed144) that the CFPB civil penalty of $55million was paid in January2025. This is a company-reported completed payment. The next paragraph describes an $80million state-regulator settlement with payments expected by/about February2025. Expectation is not payment confirmation. The redress amount in the filing is agreement language, not a completed consumer-distribution figure; the operative order's reserve, cap and shortfall mechanics are separately examined in cfpb.md.
\nThe later [2025 Form10-K](http[local research file] litigation note (printed142–143) does not provide an acquired settlement-specific receipt or completed redress total. That scoped omission is not evidence of nonpayment. A further first-quarter2025 filing was only opened and keyword-searched; it supplies no additional substantive payment finding in this pass.
\nA website identifying itself with docket2025-CFPB-0001 reports that checks began mailing on8June2026. Its public home was read. Subsequent source-provenance update,9September2026: the retained official Oregon AG8July announcement directly links that exact domain; California also links it through a publisher-supplied Proofpoint wrapper. Root independently extracted both hrefs. Official AG referral is therefore established, correcting the earlier scoped search failure. The site claim remains distinct from an administrator-appointment instrument, verified individual receipt or complete payout total; none of those was obtained. See July AG packet and its retained announcement originals. No personal information was entered and no contact was made.
\nThe [second-quarter2026 Form10-Q](http[local research file] filed5August2026, reports a separate July2026 multistate attorney-general resolution concerning customer complaints/disputes, without giving an amount in the inspected paragraph (printed33). It also says SEC staff concluded their investigation in March2026 without intending to recommend enforcement. DOJ discussions continued; Block recorded a $526million estimated loss. This is an accrual, not an assessed or paid fine or completed settlement. These are issuer accounts; the underlying agency or court instruments were not acquired in this lane. Their disclosed changes show why a2025 enforcement list cannot be treated as the current state.
\nThe July attorney-general disposition is the most relevant next primary-record lead: it may add or revise customer-remedy obligations. The DOJ accrual and shareholder cases are distinct matters. They do not establish a common chain of command, a Lantern role, or that every allegation was sustained. No new branch was acquired after locating these facts in the selected filing.
\nIn its [16January2025 response](http[local research file] Block disputed the CFPB's characterization, described the issues as historical, and asserted improvements to support, security and detection. That preserves the respondent's account; it does not substitute for verification of outcomes or compliance with the resulting order.
\nOfficial SEC submission metadata was captured successfully to identify filings. Ordinary GETs of the2025K,2026Q2 and Block-response pages returned403; these routes are closed. Their saved HTML files are explicitly error responses, not retained originals. The primary web reader worked and provided the scoped substantive passages above. Investor filing-list pages rendered only navigation, so the successful public SEC metadata route supplied exact filing addresses. No challenge bypass or authenticated route was used.
\nRoot additionally read retained CFPB order PDF44–47,50–55,74–75. These provide actual restriction-notice/appeal requirements, board/executive reporting responsibilities and enforcement/termination limits. The whole order and stipulation are covered by Carver's separate packet and Hubble's independent review, not by a claimed full root read.
", "path": "cashapp-regulator-wave-2026-09-09/issuer-disclosures.md", "sourceIds": [ "w23-cash-s-issuer2024", "w23-cash-s-issuer2025", "w23-cash-s-issuer2026q2", "w23-cash-s-response", "w23-cash-s-unverified-redress-site", "w23-cash-s-ag-referral" ], "markdown": "Observed9September2026. This lane tests execution evidence against the regulator instruments. It uses official SEC filings and Block's contemporaneous response, with a separately labelled remediation-site lead whose official referral was established in the subsequent July AG run. Exact source URLs, filing dates and scoped reads are in root-reader-sources.json. No underlying bank-transfer receipt, consumer account or internal compliance report was obtained.\n\n## Payment evidence\n\nBlock's 2024 Form10-K (see source references) filed24February2025, states in its litigation note (printed144) that the CFPB civil penalty of $55million was paid in January2025. This is a company-reported completed payment. The next paragraph describes an $80million state-regulator settlement with payments expected by/about February2025. Expectation is not payment confirmation. The redress amount in the filing is agreement language, not a completed consumer-distribution figure; the operative order's reserve, cap and shortfall mechanics are separately examined in cfpb.md.\n\nThe later 2025 Form10-K (see source references) litigation note (printed142–143) does not provide an acquired settlement-specific receipt or completed redress total. That scoped omission is not evidence of nonpayment. A further first-quarter2025 filing was only opened and keyword-searched; it supplies no additional substantive payment finding in this pass.\n\nA website identifying itself with docket2025-CFPB-0001 reports that checks began mailing on8June2026. Its public home was read. **Subsequent source-provenance update,9September2026:** the retained official Oregon AG8July announcement directly links that exact domain; California also links it through a publisher-supplied Proofpoint wrapper. Root independently extracted both hrefs. Official AG referral is therefore established, correcting the earlier scoped search failure. The site claim remains distinct from an administrator-appointment instrument, verified individual receipt or complete payout total; none of those was obtained. See [July AG packet](cashapp-ag-resolution-wave-2026-09-09-july-ag-resolution.md) and its retained announcement originals. No personal information was entered and no contact was made.\n\n## Current disclosure changes\n\nThe second-quarter2026 Form10-Q (see source references) filed5August2026, reports a separate July2026 multistate attorney-general resolution concerning customer complaints/disputes, without giving an amount in the inspected paragraph (printed33). It also says SEC staff concluded their investigation in March2026 without intending to recommend enforcement. DOJ discussions continued; Block recorded a $526million estimated loss. This is an accrual, not an assessed or paid fine or completed settlement. These are issuer accounts; the underlying agency or court instruments were not acquired in this lane. Their disclosed changes show why a2025 enforcement list cannot be treated as the current state.\n\nThe July attorney-general disposition is the most relevant next primary-record lead: it may add or revise customer-remedy obligations. The DOJ accrual and shareholder cases are distinct matters. They do not establish a common chain of command, a Lantern role, or that every allegation was sustained. No new branch was acquired after locating these facts in the selected filing.\n\n## Respondent's position and limits\n\nIn its 16January2025 response (see source references) Block disputed the CFPB's characterization, described the issues as historical, and asserted improvements to support, security and detection. That preserves the respondent's account; it does not substitute for verification of outcomes or compliance with the resulting order.\n\nOfficial SEC submission metadata was captured successfully to identify filings. Ordinary GETs of the2025K,2026Q2 and Block-response pages returned403; these routes are closed. Their saved HTML files are explicitly error responses, not retained originals. The primary web reader worked and provided the scoped substantive passages above. Investor filing-list pages rendered only navigation, so the successful public SEC metadata route supplied exact filing addresses. No challenge bypass or authenticated route was used.\n\nRoot additionally read retained CFPB order PDF44–47,50–55,74–75. These provide actual restriction-notice/appeal requirements, board/executive reporting responsibilities and enforcement/termination limits. The whole order and stipulation are covered by Carver's separate packet and Hubble's independent review, not by a claimed full root read.\n", "file": "research/cashapp-regulator-wave-2026-09-09-issuer-disclosures.md" }, { "id": "cashapp-regulator-wave-2026-09-09-state-controls", "title": "Block/Cash App multistate BSA/AML order: decision rights and public limits", "description": "Bounded regulator lane, observed 9 September 2026. Acquisition closed after the original operative text, one fuller official copy, its current state landing page and the coordinator's explanatory release. No outreach, paid/private access, additional state case or NYDFS acquisition. Root's Block disclosures and the separate CFPB lane are not incorporated here", "html": "Bounded regulator lane, observed 9 September 2026. Acquisition closed after the original operative text, one fuller official copy, its current state landing page and the coordinator's explanatory release. No outreach, paid/private access, additional state case or NYDFS acquisition. Root's Block disclosures and the separate CFPB lane are not incorporated here.
\nThis is an actual consent order binding Block, Inc., rather than a voluntary industry recommendation. It establishes a state-to-company oversight mechanism: state Executive Committee receives reports/workpapers and can question the consultant; Block's board approves corrective plans and carries ultimate compliance responsibility; an independent consultant assesses and validates the work. The public instrument does not reveal the consultant's identity, executed engagement, specific implemented controls, payment receipts or completed remediation. It establishes no Lantern causal link and no bank-owner instruction.
\nThe [CSBS announcement](http[local research file] dated 15 January 2025, links a [10-page order body](http[local research file] That PDF ends by saying signature pages follow, but contains no signature pages or appendices. It is retained as state-csbs-order-body.pdf; its operative pages 1–10 were read, with emphasis on recitals, II governance, III payment, IV enforcement, V release and VI duration/private rights.
The [Maine enforcement entry](http[local research file] dated 16 January 2025, reproduces the order and links a [22-page fuller official copy](http[local research file] retained as state-maine-order-full.pdf. Maine's HTML operative sections and signature/appendix text were read; the PDF's signatures/appendices were extracted, and PDF pages 11, 19, 20 and 21 were rendered locally and visually inspected. The text layer of PDF11 is empty although the signed page is visible. Maine's HTML is a derivative presentation of the same instrument, not independent corroboration; use the PDF for signature/table ambiguities. The different copies are not represented as byte-identical.
Originals, URLs, retrieval timestamps, hashes, reading scopes and derivative ancestry are in captures-state.json. All four public GETs succeeded. The web screenshot route returned cache-miss errors; the already retained PDF was rendered locally with pypdfium2 instead, without another network attempt.
The order names Block, Inc., formerly Square, Inc., NMLS 942933, as the company operating Cash App. The other parties are participating states' money-transmission regulatory agencies. CSBS and MTRA supply coordination protocols, and the states work through the Multi-State MSB Examination Taskforce (MMET); they are not additional company respondents or substituted state enforcement authorities. Jack Dorsey signs for Block as Block Head and Chairperson (Maine PDF11), not as a separately named personal respondent.
\nThe recitals describe an examination begun around 15 May 2023, covering 1 January 2021–31 March 2023, conducted by California, Colorado, Kentucky, Massachusetts, New Hampshire, Ohio, Texas and Washington. A joint Report of Examination (ROE) issued 6 November 2023 identified AML-program adequacy findings. Block's approximately 6 December 2023 response described new and planned controls. The actual ROE and response are not attached in the retained copies. Block neither admits nor denies wrongdoing or violations (CSBS PDF1–2; Maine HTML recitals).
\nCSBS's release describes due-diligence, identity-verification, suspicious-activity reporting and high-risk-account requirements, and says regulators found noncompliance with certain requirements. Those explanatory categories must not be silently turned into an enumerated, proved list of specific Block failures: the public order refers to the underlying ROE without setting those particulars out. Likewise, the broad list of practices covered by the release in V.C is a release definition, not a separate finding that every listed practice failed.
\nThere are three different regulator groupings. The examination had the eight jurisdictions above; the public release credits seven enforcement leads (Arkansas, California, Massachusetts, Florida, Maine, Texas and Washington); the initial Executive Committee has five (Arkansas, California, Massachusetts, Texas and Washington). Do not merge these rosters into a single approval body.
\n| Actor / relation suitable for the map | \nExact scope and locus | \n
|---|---|
| Participating state regulators → Block: enforce order | \nJurisdiction under I and IV; each state retains enforcement/examination authority. Before enforcement under the agreement, the state gives the Executive Committee and Block written allegations and an opportunity to respond (IV.C, PDF7). This does not bind every other government agency. | \n
| State Executive Committee → compliance process: receives and questions | \nII.A and II.C.3, PDF2–3: receives consultant/company communications and can obtain all Board Reports. II.D.4/6, PDF5: simultaneous consultant reports to Block and committee; committee can obtain workpapers and other work product. IV.D, PDF7: participating regulators may obtain information held by committee/MMET. | \n
| Block → Compliance Management Committee: appointment | \nII.B, PDF3: company representatives across relevant divisions, quarterly meetings/minutes. Continues until the later of receipt of validation report or eight quarters from effective date; only then may Block dissolve it at its discretion. No present dissolution or meeting record obtained. | \n
| Block board → compliance/CAP: ultimate oversight and formal approval | \nII.C, PDF3; II.D.2, PDF4: board or authorized committee formally approves the Corrective Action Plan (CAP). CAP and updates are sent to the state committee. The text does not separately state that the state committee must formally approve each CAP. | \n
| Independent consultant → AML program: assessment/priority/validation | \nII.D.1–3, PDF3–5: assess adequacy under applicable BSA requirements and 31 CFR 1022.210, prioritize corrective measures, validate completed measures and give progress reports. Block must provide relevant personnel, vendor, facility and record access (II.D.7, PDF5–6). This is review/validation authority, not proof of power to close individual customer accounts. | \n
| Block → replacement consultant: selection/engagement; state committee receives notice | \nII.D.8, PDF6: notice of replacement need within ten days of awareness, notice of selection within ten days, engagement within thirty days of selection notice; qualified independent unaffiliated third party. No named consultant, initial selection procedure or express state appointment-veto provision was located in this operative text. Do not infer veto merely from the word independent. | \n
| State Executive Committee → reporting deadlines: extension discretion | \nII.D.9, PDF6: reasonable-ground request by Block, good-cause grant by committee. CAP duration extensions have a separate consultant-review mechanism in II.D.2(c), PDF5. | \n
Effective date is 15 January 2025 (VI.A, PDF8). Assessment report: within nine months of that date, subject to reporting extensions. CAP submission: within 90 calendar days after the assessment report is issued, with immediate implementation; consultant priority must be appropriately incorporated. CAP completion: 12 months, with the consultant documenting the basis/need for an extension (II.D.2(c)). The CSBS release describes this as twelve months after report filing; the operative paragraph itself states the period without an equally explicit start phrase. Do not turn that summary into proof of an actual filing date or expired obligation.
\nOther specified clocks include consultant progress reports at least quarterly and within ten days of request/quarter end; questions and responses have separate 20/15-day or 45/15-day periods. CAP completion triggers independent validation and a report; continued deficiencies trigger an updated CAP within 60 days of the validation report. Board progress reports are due within thirty days of each company committee meeting. No actual report, validation, extension or schedule performance is established here.
\nUSD80,000,000 is the settlement obligation, composed of USD79,075,000 administrative penalty plus USD925,000 administrative costs (III.A, PDF6). Each payment clock is twenty calendar days after receipt of payment instructions, not simply twenty days after signing. Appendix C (Maine PDF22) allocates costs: Arkansas150,000; California250,000; Massachusetts150,000; Maine75,000; Texas150,000; Washington150,000. These are state administrative-cost allocations, not the consultant's fees or consumer redress. Unpaid sums may be claimed against relevant surety bonds under III.B. The covered-conduct release depends on both effectiveness and full payment (V.D, PDF8). An obligation or release condition is not a payment receipt.
\nThe 15 January announcement reports 48 state regulators. Maine's fuller copy visibly contains a Colorado signature dated 16 January 2025 (PDF19), and Appendix A enumerates 49 jurisdictions including DC (PDF20). Appendix B contains 48 allocation rows, includes Colorado and omits Wyoming, although Wyoming appears in Appendix A and has a signature block in PDF17. Its stated penalty total remains USD79,075,000 (PDF21). Independent numerical review confirms that the48 listed values sum exactly to that total; the open question is the allocation basis for the differing party list, not an arithmetic shortfall. These are visible record differences, not OCR artifacts. The Maine HTML likewise includes Colorado and lacks a Wyoming penalty row.
\nA later signature, publication assembly or allocation adjustment could explain parts of the difference. No corrected allocation schedule, alternative-payment election or reconciliation notice was obtained, so do not assert a particular explanation, missing money or that Wyoming was not a party. Preserve the total obligation as stated and leave per-state payment completeness unestablished.
\nThe live Maine entry still presents its 16 January 2025 date, this agreement text and attached full copy. No linked completion certification, amendment, extension, consultant appointment or payment acknowledgment is displayed in the inspected entry. The CSBS release remains dated 15 January 2025. This narrow present inspection is not an exhaustive docket search and cannot establish current compliance or noncompliance. VI.C/I requires written changes by the parties (or other stated cessation conditions); VI.J creates no private enforcement right under this order. The agreement preserves Block's later hearing/appeal rights in a compliance disagreement (recitals, PDF2).
\nThe next discriminating records are the executed consultant engagement/selection notice, assessment and validation reports, board-approved CAP with state receipt and any extensions, payment instructions/receipts and any corrected Appendix B. Block, the consultant and the state Executive Committee are the directly identified holders for the compliance records; the recipient regulators hold payment records. The unreleased ROE is the record needed to specify original deficiencies. Their absence from this packet is not evidence that none exist or that the required work was not performed. No subsequent branch is opened pending the forest assessment.
", "path": "cashapp-regulator-wave-2026-09-09/state-controls.md", "sourceIds": [ "w23-cash-s-state-body", "w23-cash-s-state-full", "w23-cash-s-state-page", "w23-cash-s-state-release" ], "markdown": "Bounded regulator lane, observed 9 September 2026. Acquisition closed after the original operative text, one fuller official copy, its current state landing page and the coordinator's explanatory release. No outreach, paid/private access, additional state case or NYDFS acquisition. Root's Block disclosures and the separate CFPB lane are not incorporated here.\n\n## Result\n\nThis is an actual consent order binding Block, Inc., rather than a voluntary industry recommendation. It establishes a state-to-company oversight mechanism: state Executive Committee receives reports/workpapers and can question the consultant; Block's board approves corrective plans and carries ultimate compliance responsibility; an independent consultant assesses and validates the work. The public instrument does not reveal the consultant's identity, executed engagement, specific implemented controls, payment receipts or completed remediation. It establishes no Lantern causal link and no bank-owner instruction.\n\n## Sources, versions and scoped reading\n\nThe CSBS announcement (see source references) dated **15 January 2025**, links a 10-page order body (see source references) That PDF ends by saying signature pages follow, but contains no signature pages or appendices. It is retained as `state-csbs-order-body.pdf`; its operative pages 1–10 were read, with emphasis on recitals, II governance, III payment, IV enforcement, V release and VI duration/private rights.\n\nThe Maine enforcement entry (see source references) dated **16 January 2025**, reproduces the order and links a 22-page fuller official copy (see source references) retained as `state-maine-order-full.pdf`. Maine's HTML operative sections and signature/appendix text were read; the PDF's signatures/appendices were extracted, and PDF pages **11, 19, 20 and 21** were rendered locally and visually inspected. The text layer of PDF11 is empty although the signed page is visible. Maine's HTML is a derivative presentation of the same instrument, not independent corroboration; use the PDF for signature/table ambiguities. The different copies are not represented as byte-identical.\n\nOriginals, URLs, retrieval timestamps, hashes, reading scopes and derivative ancestry are in `captures-state.json`. All four public GETs succeeded. The web screenshot route returned cache-miss errors; the already retained PDF was rendered locally with pypdfium2 instead, without another network attempt.\n\n## Parties, examination and allegation boundaries\n\nThe order names **Block, Inc., formerly Square, Inc., NMLS 942933**, as the company operating Cash App. The other parties are participating states' money-transmission regulatory agencies. CSBS and MTRA supply coordination protocols, and the states work through the Multi-State MSB Examination Taskforce (MMET); they are not additional company respondents or substituted state enforcement authorities. Jack Dorsey signs for Block as Block Head and Chairperson (Maine PDF11), not as a separately named personal respondent.\n\nThe recitals describe an examination begun around **15 May 2023**, covering **1 January 2021–31 March 2023**, conducted by California, Colorado, Kentucky, Massachusetts, New Hampshire, Ohio, Texas and Washington. A joint Report of Examination (ROE) issued **6 November 2023** identified AML-program adequacy findings. Block's approximately **6 December 2023** response described new and planned controls. The actual ROE and response are not attached in the retained copies. Block neither admits nor denies wrongdoing or violations (CSBS PDF1–2; Maine HTML recitals).\n\nCSBS's release describes due-diligence, identity-verification, suspicious-activity reporting and high-risk-account requirements, and says regulators found noncompliance with certain requirements. Those explanatory categories must not be silently turned into an enumerated, proved list of specific Block failures: the public order refers to the underlying ROE without setting those particulars out. Likewise, the broad list of practices covered by the release in V.C is a release definition, not a separate finding that every listed practice failed.\n\nThere are three different regulator groupings. The examination had the eight jurisdictions above; the public release credits seven enforcement leads (Arkansas, California, Massachusetts, Florida, Maine, Texas and Washington); the initial **Executive Committee has five** (Arkansas, California, Massachusetts, Texas and Washington). Do not merge these rosters into a single approval body.\n\n## Operative rights\n\n| Actor / relation suitable for the map | Exact scope and locus |\n| --- | --- |\n| Participating state regulators → Block: enforce order | Jurisdiction under I and IV; each state retains enforcement/examination authority. Before enforcement under the agreement, the state gives the Executive Committee and Block written allegations and an opportunity to respond (IV.C, PDF7). This does not bind every other government agency. |\n| State Executive Committee → compliance process: receives and questions | II.A and II.C.3, PDF2–3: receives consultant/company communications and can obtain all Board Reports. II.D.4/6, PDF5: simultaneous consultant reports to Block and committee; committee can obtain workpapers and other work product. IV.D, PDF7: participating regulators may obtain information held by committee/MMET. |\n| Block → Compliance Management Committee: appointment | II.B, PDF3: company representatives across relevant divisions, quarterly meetings/minutes. Continues until the later of receipt of validation report or eight quarters from effective date; only then may Block dissolve it at its discretion. No present dissolution or meeting record obtained. |\n| Block board → compliance/CAP: ultimate oversight and formal approval | II.C, PDF3; II.D.2, PDF4: board or authorized committee formally approves the Corrective Action Plan (CAP). CAP and updates are sent to the state committee. The text does not separately state that the state committee must formally approve each CAP. |\n| Independent consultant → AML program: assessment/priority/validation | II.D.1–3, PDF3–5: assess adequacy under applicable BSA requirements and 31 CFR 1022.210, prioritize corrective measures, validate completed measures and give progress reports. Block must provide relevant personnel, vendor, facility and record access (II.D.7, PDF5–6). This is review/validation authority, not proof of power to close individual customer accounts. |\n| Block → replacement consultant: selection/engagement; state committee receives notice | II.D.8, PDF6: notice of replacement need within ten days of awareness, notice of selection within ten days, engagement within thirty days of selection notice; qualified independent unaffiliated third party. No named consultant, initial selection procedure or express state appointment-veto provision was located in this operative text. Do not infer veto merely from the word independent. |\n| State Executive Committee → reporting deadlines: extension discretion | II.D.9, PDF6: reasonable-ground request by Block, good-cause grant by committee. CAP duration extensions have a separate consultant-review mechanism in II.D.2(c), PDF5. |\n\n## Clocks and money states\n\nEffective date is **15 January 2025** (VI.A, PDF8). Assessment report: within **nine months** of that date, subject to reporting extensions. CAP submission: within **90 calendar days after the assessment report is issued**, with immediate implementation; consultant priority must be appropriately incorporated. CAP completion: **12 months**, with the consultant documenting the basis/need for an extension (II.D.2(c)). The CSBS release describes this as twelve months after report filing; the operative paragraph itself states the period without an equally explicit start phrase. Do not turn that summary into proof of an actual filing date or expired obligation.\n\nOther specified clocks include consultant progress reports at least quarterly and within ten days of request/quarter end; questions and responses have separate 20/15-day or 45/15-day periods. CAP completion triggers independent validation and a report; continued deficiencies trigger an updated CAP within 60 days of the validation report. Board progress reports are due within thirty days of each company committee meeting. No actual report, validation, extension or schedule performance is established here.\n\n**USD80,000,000 is the settlement obligation**, composed of **USD79,075,000 administrative penalty** plus **USD925,000 administrative costs** (III.A, PDF6). Each payment clock is **twenty calendar days after receipt of payment instructions**, not simply twenty days after signing. Appendix C (Maine PDF22) allocates costs: Arkansas150,000; California250,000; Massachusetts150,000; Maine75,000; Texas150,000; Washington150,000. These are state administrative-cost allocations, not the consultant's fees or consumer redress. Unpaid sums may be claimed against relevant surety bonds under III.B. The covered-conduct release depends on both effectiveness and full payment (V.D, PDF8). An obligation or release condition is not a payment receipt.\n\n## Visible 48/49 allocation/version question\n\nThe 15 January announcement reports **48 state regulators**. Maine's fuller copy visibly contains a **Colorado signature dated 16 January 2025** (PDF19), and Appendix A enumerates **49 jurisdictions including DC** (PDF20). Appendix B contains **48 allocation rows**, includes Colorado and omits Wyoming, although Wyoming appears in Appendix A and has a signature block in PDF17. Its stated penalty total remains USD79,075,000 (PDF21). Independent numerical review confirms that the48 listed values sum exactly to that total; the open question is the allocation basis for the differing party list, not an arithmetic shortfall. These are visible record differences, not OCR artifacts. The Maine HTML likewise includes Colorado and lacks a Wyoming penalty row.\n\nA later signature, publication assembly or allocation adjustment could explain parts of the difference. No corrected allocation schedule, alternative-payment election or reconciliation notice was obtained, so do not assert a particular explanation, missing money or that Wyoming was not a party. Preserve the total obligation as stated and leave per-state payment completeness unestablished.\n\n## Current public status and stopping point\n\nThe live Maine entry still presents its 16 January 2025 date, this agreement text and attached full copy. No linked completion certification, amendment, extension, consultant appointment or payment acknowledgment is displayed in the inspected entry. The CSBS release remains dated 15 January 2025. This narrow present inspection is not an exhaustive docket search and cannot establish current compliance or noncompliance. VI.C/I requires written changes by the parties (or other stated cessation conditions); VI.J creates no private enforcement right under this order. The agreement preserves Block's later hearing/appeal rights in a compliance disagreement (recitals, PDF2).\n\nThe next discriminating records are the executed consultant engagement/selection notice, assessment and validation reports, board-approved CAP with state receipt and any extensions, payment instructions/receipts and any corrected Appendix B. Block, the consultant and the state Executive Committee are the directly identified holders for the compliance records; the recipient regulators hold payment records. The unreleased ROE is the record needed to specify original deficiencies. Their absence from this packet is not evidence that none exist or that the required work was not performed. No subsequent branch is opened pending the forest assessment.\n", "file": "research/cashapp-regulator-wave-2026-09-09-state-controls.md" }, { "id": "christchurch-governance-wave-2026-09-09-decision-process", "title": "Christchurch Call: decision process and advisory money", "description": "Root acquisition closed, 9 September 2026. This lane examines published institutional design and advisory rules, not a claim that any participant controls the others. Government transition and actual Foundation constitution belong to the parallel packets.", "html": "Root acquisition closed, 9 September 2026. This lane examines published institutional design and advisory rules, not a claim that any participant controls the others. Government transition and actual Foundation constitution belong to the parallel packets.
\n[Next steps for a resilient Christchurch Call](http[local research file] 2023 Summit meeting paper: root read the complete main text, PDF pages 3–8; contents/cover and annex text also inspected. Annex geometry was not visually verified because screenshot retrieval failed. Reader access succeeded; original GET returned a 404 HTML page, retained as an error response.
\nThe proposed change addressed limitations of a secretariat embedded in France and New Zealand: public-finance requirements complicated outside funding, global hiring and secondments. A separate legal entity would receive wider support and employ staff. The proposal divided corporate/financial oversight by a board from work-programme direction by Call Leaders; founding documents would determine appointments while preserving the founders' leadership. It also considered applying the model to related digital issues in the context of terrorist and violent extremist content. These were proposed arrangements, not the acquired constitution or proof that donor money purchased policy authority.
\n[Join the Call](http[local research file] root read the full substantive page, reader lines 14–53. It describes introductory engagement, written intent, an information response cleared by the applicant, confidential circulation and a community discussion, followed by a final decision by Call Leaders and a coordinated announcement. Leaders are described broadly: government, corporate, civil-society and partner-organisation leaders. The page does not disclose an electorate, quorum, voting formula, veto or rejection-appeal rule. The Secretariat convenes; participation alone is not final approval authority. Undated legacy-site reader snapshot; copyright 2026 is not a rule-adoption date. Ordinary GET returned 403, so current operation has not been independently confirmed.
\n[Current priorities page](http[local research file] attributes work to Call commitments and community-leader direction. Indexed substantive body read; direct reader failed and GET returned 403. This supports stated directional continuity, not an exact decision procedure.
\n[CCAN Terms of Reference](http[local research file] root read all 15 pages through the primary reader. Footer says last updated December 2025; the metadata title mentions 2026 and is not an adoption date. [Governance index](http[local research file] links these terms. Both original GETs returned 406, retained as errors.
\nSections 3–7 put admissions, statements and internal decisions with a Core Committee of at most 25. It elects co-chairs. Modified consensus permits at most two blockers with two-thirds participation; a deadline exception permits a two-thirds vote. Internal meetings ordinarily exclude Foundation staff. Section 8 requests consultation and explanations for rejected feedback, not a general Foundation-policy veto. Section 9 reports Foundation funding; operational funds are held by a member organisation under an agreed memorandum. The Committee approves applications and fund-use rules. Section 10 provides Foundation review/feedback on amended terms, without an express approval veto. Section 3 requires government/company-funding disclosures above its threshold but keeps them within eligibility review. These rules create specific independence safeguards alongside dependence; actual compliance, funding amounts, fundholder identity and signed memorandum remain unverified.
\n[Foundation page](http[local research file] root read the full indexed substantive body. It describes convening and programme delivery, names Paul Ash as CEO, Andy George as deputy and Kristen Cambell as chief partnerships officer, and distinguishes patron/founder roles. Those are not a verified trustee roster or appointment instrument. Reader cache miss and one GET403 closed the original route.
\n[Funding page](http[local research file] root read indexed substantive body. It separates financial, ideas, delivery and community partnerships and asserts independence principles. It acknowledges Pivotal support via the Pivotal Leaders Fund, claims NGOsource equivalency determination and a European donation partner, but the available text omits funder-logo identities and the partner's name. No determination instrument, donation agreement or full linked principles was acquired. GET403 and reader failure closed this route.
\nThe strongest supported structure is several interacting authorities: community leaders select priorities and entrants under published descriptions; a separate nonprofit supplies capacity; an advisory committee controls its own participation and speech while relying on financing and a fundholder. This is more precise than treating a supporter list as a board or a funding logo as an instruction right.
\nThe governing Foundation deed, trustee appointments, funded-programme terms and CCAN fundholder memorandum are the most revealing missing instruments. Public rules make the latter a reasonable document-existence expectation, not proof of an executed agreement. No new branch is opened before the combined forest.
\nOriginal-content acquisition failed for every root URL; root-captures.json retains statuses, hashes and filenames of error responses. These are not original article/PDF bytes. Source reading is primary reader/index text only. No contacts, forms, accounts, grants or donations submitted.
", "path": "christchurch-governance-wave-2026-09-09/decision-process.md", "sourceIds": [ "w27-roost-s-proposal2023", "w27-roost-s-join-call", "w27-roost-s-ccan-tor", "w27-roost-s-ccf-foundation", "w27-roost-s-ccf-funding", "w27-roost-s-call-priorities" ], "markdown": "Root acquisition closed, 9 September 2026. This lane examines published institutional design and advisory rules, not a claim that any participant controls the others. Government transition and actual Foundation constitution belong to the parallel packets.\n\n## 2023 proposal: independent capacity, continuing direction\n\nNext steps for a resilient Christchurch Call (see source references) 2023 Summit meeting paper: root read the complete main text, PDF pages 3–8; contents/cover and annex text also inspected. Annex geometry was not visually verified because screenshot retrieval failed. Reader access succeeded; original GET returned a 404 HTML page, retained as an error response.\n\nThe proposed change addressed limitations of a secretariat embedded in France and New Zealand: public-finance requirements complicated outside funding, global hiring and secondments. A separate legal entity would receive wider support and employ staff. The proposal divided corporate/financial oversight by a board from work-programme direction by Call Leaders; founding documents would determine appointments while preserving the founders' leadership. It also considered applying the model to related digital issues in the context of terrorist and violent extremist content. These were proposed arrangements, not the acquired constitution or proof that donor money purchased policy authority.\n\n## Entry to the wider Call\n\nJoin the Call (see source references) root read the full substantive page, reader lines 14–53. It describes introductory engagement, written intent, an information response cleared by the applicant, confidential circulation and a community discussion, followed by a final decision by Call Leaders and a coordinated announcement. Leaders are described broadly: government, corporate, civil-society and partner-organisation leaders. The page does not disclose an electorate, quorum, voting formula, veto or rejection-appeal rule. The Secretariat convenes; participation alone is not final approval authority. Undated legacy-site reader snapshot; copyright 2026 is not a rule-adoption date. Ordinary GET returned 403, so current operation has not been independently confirmed.\n\nCurrent priorities page (see source references) attributes work to Call commitments and community-leader direction. Indexed substantive body read; direct reader failed and GET returned 403. This supports stated directional continuity, not an exact decision procedure.\n\n## CCAN has its own rules and financial intermediary\n\nCCAN Terms of Reference (see source references) root read all 15 pages through the primary reader. Footer says last updated December 2025; the metadata title mentions 2026 and is not an adoption date. Governance index (see source references) links these terms. Both original GETs returned 406, retained as errors.\n\nSections 3–7 put admissions, statements and internal decisions with a Core Committee of at most 25. It elects co-chairs. Modified consensus permits at most two blockers with two-thirds participation; a deadline exception permits a two-thirds vote. Internal meetings ordinarily exclude Foundation staff. Section 8 requests consultation and explanations for rejected feedback, not a general Foundation-policy veto. Section 9 reports Foundation funding; operational funds are held by a member organisation under an agreed memorandum. The Committee approves applications and fund-use rules. Section 10 provides Foundation review/feedback on amended terms, without an express approval veto. Section 3 requires government/company-funding disclosures above its threshold but keeps them within eligibility review. These rules create specific independence safeguards alongside dependence; actual compliance, funding amounts, fundholder identity and signed memorandum remain unverified.\n\n## Current Foundation self-description and limits\n\nFoundation page (see source references) root read the full indexed substantive body. It describes convening and programme delivery, names Paul Ash as CEO, Andy George as deputy and Kristen Cambell as chief partnerships officer, and distinguishes patron/founder roles. Those are not a verified trustee roster or appointment instrument. Reader cache miss and one GET403 closed the original route.\n\nFunding page (see source references) root read indexed substantive body. It separates financial, ideas, delivery and community partnerships and asserts independence principles. It acknowledges Pivotal support via the Pivotal Leaders Fund, claims NGOsource equivalency determination and a European donation partner, but the available text omits funder-logo identities and the partner's name. No determination instrument, donation agreement or full linked principles was acquired. GET403 and reader failure closed this route.\n\n## Inference and next evidence\n\nThe strongest supported structure is several interacting authorities: community leaders select priorities and entrants under published descriptions; a separate nonprofit supplies capacity; an advisory committee controls its own participation and speech while relying on financing and a fundholder. This is more precise than treating a supporter list as a board or a funding logo as an instruction right.\n\nThe governing Foundation deed, trustee appointments, funded-programme terms and CCAN fundholder memorandum are the most revealing missing instruments. Public rules make the latter a reasonable document-existence expectation, not proof of an executed agreement. No new branch is opened before the combined forest.\n\nOriginal-content acquisition failed for every root URL; root-captures.json retains statuses, hashes and filenames of error responses. These are not original article/PDF bytes. Source reading is primary reader/index text only. No contacts, forms, accounts, grants or donations submitted.\n", "file": "research/christchurch-governance-wave-2026-09-09-decision-process.md" }, { "id": "christchurch-governance-wave-2026-09-09-legal-governance", "title": "Christchurch Call Foundation — legal governance", "description": "Observed and acquisition closed 9 September 2026. This lane distinguishes a constituted support organisation from the wider voluntary Call. It found a primary transition announcement and a detailed proposed division of responsibilities, but did not acquire the foundation's adopted deed/constitution or a verified board register. It therefore does not establis", "html": "Observed and acquisition closed 9 September 2026. This lane distinguishes a constituted support organisation from the wider voluntary Call. It found a primary transition announcement and a detailed proposed division of responsibilities, but did not acquire the foundation's adopted deed/constitution or a verified board register. It therefore does not establish actual appointment quotas, voting rights, quorum, vetoes, delegation limits or amendment powers.
\nThe [joint French–New Zealand announcement of 14 May 2024](http[local research file] announces a new non-governmental organisation to coordinate the Call's work. It ties the change to the November 2023 Leaders' Summit, places headquarters in Wellington and schedules Secretariat operation under the foundation from 1 July 2024. It describes key decisions about direction as remaining with Leaders from all sectors. The announcement says funding pledges had been attracted; it does not disclose donor voting rights or certify receipt. Ardern's stated Patron role is voluntary/unpaid dialogue, outreach and advocacy, rather than an identified trustee appointment. These are announced institutional arrangements, not the signed governing instrument.
\nRead scope: complete substantive announcement and Notes from Linnaeus's retained original transition-launch-france.html, excluding unrelated current-news cards. SHA-256 8b5a17c3a57b171a33640c6a9aaa35edb1be3eb4dddc57c1d9d8e58e02881c43. Original custody remains shared in this wave; it was read locally rather than downloaded again. The NZ mirror's HTTP 200 response was a tiny non-substantive response and was not used as an original body.
The [2023 meeting paper, Next steps for a resilient Christchurch Call](http[local research file] PDF5/printed3, is explicitly an initial proposal. It envisages an independent public-good entity, potentially a trust or nonprofit. A future founding document would define board remit and appointments while maintaining French/New Zealand leadership. The board would supervise corporate/financial functions and performance aligned with Leaders' policy directions; the Call work programme would remain with Leaders. It also says the existing Call community has no corporate/legal structure. This is a proposed separation between entity governance and initiative direction, not proof those precise clauses were adopted. Main text PDF3–8 was read through the public reader; the Annex A diagram was available as extracted labels only and was not visually verified. The June 2024 URL folder is not a deed adoption date. Root's failed original GET is already closed; no original PDF hash is claimed.
\nThe current [foundation page](http[local research file] read as indexed primary text, identifies Paul Ash as Chief Executive, Andy George as Deputy Chief Executive and Kristen Cambell as Chief Partnerships Officer. It describes Ardern and Macron as providing strategic guidance and global advocacy under its patrons heading. Their individual labels differ: Ardern is explicitly foundation Patron; Macron's individual label is Call Co-Founder. The text separates community participation, Call commitments and foundation delivery. None of these titles alone establishes a trustee/director seat, appointment power, legal veto or control over participant governments/platforms. The indexed body was read through its leadership/programme introduction; the ordinary original GET was a challenge response, not substantive HTML. No original-body hash is claimed.
\nThe exact name/identifier lead is The Christchurch Call Foundation, CC62350. [CharityData NZ](http[local research file] supplies that pointer and describes a trust registered 27 May 2024. In this lane those legal-form/date fields remain secondary and unverified against the official entity row. Its automated rating, financial dashboard and abbreviated officer list are not evidence of board composition or control. In particular, a CEO and Patron appearing as officers would not by itself make them the only trustees.
\nThe public [Charities Register](http[local research file] explains that registered-charity information can be searched publicly. The aggregator's direct link to ViewCharity?accountId=CC62350 produced a reader cache miss; ordinary GET of that link and of the official search page both returned HTTP 403. That link's use of CC62350 as accountId was not independently validated. These failures do not establish absence of a registration or deed.
Charities Services separately publishes an [open-data interface for research](http[local research file] Its reader-visible documentation permits filtered Organisation queries and names CharityRegistrationNumber as a field. One unauthenticated query for CC62350 through the documented OData host also returned HTTP 403. No account, browser challenge, alternate-host workaround or access-control bypass was attempted. The register and OData routes are closed for this run. The documentation explains public access possibilities; it does not supply the missing entity record.
\nNo adopted trust deed, constitution, board appointment instrument, current complete trustee roster, delegation instrument, reserved-matters schedule or donor agreement was obtained. Searches for the foundation name with deed, constitution, trustees, board and governing-document terms supplied no such primary instrument within this pass. The July 2024 essay proposing community participation in board appointments was treated as advocacy, not as an adopted rule. Similarly, the 2023 plan does not prove that donors elect trustees, that France/New Zealand presently have reserved seats, or that foundation trustees can compel a participating platform's moderation decision.
\nThe next discriminator is the foundation's adopted governing document and amendments, paired with the current trustee/board appointment record. Charities Services or the foundation is the likely custodian. Those records would resolve appointment/removal rights, financial approvals, conflicts, delegations and any formal relationship to Call Leaders or funders. A funding receipt alone would not answer those questions.
\nlegal-captures.json preserves the failed response bytes, their hashes and precise status; it also identifies the shared successful French original and the reader/index-only scopes. Error-response hashes are not hashes of a deed, registration row or foundation webpage body. Acquisition stops at this documented gap for the parent's forest assessment.
9 September 2026. Bounded transition/funding acquisition closed. This note separates the 2024 support-function handoff, the continuing Call, the new Foundation, and later project funding. Legal-entity governance and community decision processes belong to the parallel lanes. No outreach, account, payment or private record access; no canonical changes.
\nThe 14 May 2024 joint NZ/France announcement says support functions would move to a new secretariat funded by the Foundation, operating from 1 July 2024, while New Zealand taxpayer funding for the Call would end on 30 June. Donor pledges had been attracted, but no amounts or received instalments are stated. The Foundation would coordinate and support the wider community; leaders from all sectors would continue making key directional decisions. Ardern's special-envoy term would end, with her continuing as unpaid patron for dialogue and advocacy. The announcement does not assign her a foundation voting or appointment power.
\n[Joint official statement, French presidency publication](http[local research file] Complete substantive body and notes read. The NZ publication is the same joint-source family, not independent corroboration; its ordinary GET returned a challenge shell and was closed.
\nMFAT subsequently reports the transition as completed, with New Zealand government funding ending in June 2024 and support functions financed by community/philanthropic pledges. This is a retrospective departmental performance account, not proof that every pledge had been paid. [MFAT Annual Report 2023–24](http[local research file] PDF66/printed66, complete Outcome1 paragraph context read. The report's publication date was not separately established; its reporting period is distinct from the September2026 retrieval.
\nThe defensible change is government-funded coordination becoming Foundation-funded coordination, while governments remain in the wider initiative. These documents do not prove full NZ or French withdrawal, French funding termination, or an absence of continued diplomatic/in-kind work. No signed handover deed, staff secondment/novation schedule, asset inventory, IP assignment, transferred bank balance or government reserved appointment/approval instrument was acquired. Moving support functions does not by itself establish a transfer of ownership of assets or employment contracts.
\n[MFAT OIA29223, dated4April2024](http[local research file] both pages read; AppendixA onPDF2 visually inspected. The response identifies costs incurred by MFAT only, in NZD, from1July2018 through29February2024.
\n| Financial year | \nReported actual expenditure NZD | \n
|---|---|
| 2018/19 | \n87,936 | \n
| 2019/20 | \n530,486 | \n
| 2020/21 | \n856,130 | \n
| 2021/22 | \n357,172 | \n
| 2022/23 | \n638,646 | \n
| 2023/24 YTD to29February2024 | \n412,853 | \n
The last year is incomplete and subject to later review. These are not the Call's complete costs across all departments/countries, not a Foundation grant, and not payment vouchers. No cumulative influence budget is constructed. The response was listed for publication on31May2024 in MFAT's2024 OIA index; the letter itself is4April.
\n[Packard's CCF grant card](http[local research file] records 2024, $350,000, Term36, for general support, in Just Societies / Democracy, Rights and Governance. Complete substantive card read and original retained. The card does not explicitly state the term's unit or currency, exact award/payment dates, expenditure allocation or donor appointment/clearance rights. Preserve its displayed fields; do not turn the likely36-month term into verified wording without a further source. The US donor's dollar context is consistent with USD, but the currency is not separately printed in this card.
\nCarnegie's [grant database, page13](http[local research file] records 12December2024, $500,000, 12months, Special Opportunities Fund, to support CCF's launch. The full CCF entry was read through the primary public reader, lines152–194; ordinary GET403 was retained only as an error response and the route closed. The description emphasizes launch as an independent body but dates the spin-out to2023. That donor characterization does not override the joint governments' July2024 operational handoff. The discrepancy could reflect preparation versus operation, or imprecise donor wording; this pass does not resolve it.
\nCarnegie's [10March2026 board-grants announcement](http[local research file] says its trustees approved grants at the 5March2026 meeting, including $2,000,000 over24months for CCF to counter terrorist/violent-extremist content online. Read title/date, approval introduction and CCF entry through the primary reader, lines30–34 and67–68. This earns donor board approval, not payment or a right to run Foundation decisions. Original GET403 was closed and retained as error only. Dollar amounts are displayed by a US donor; the brief entry does not separately label currency.
\nThe two Carnegie records are distinct dated grants; do not treat the second as an amendment or sum them with Packard/Canada into current annual resources. None supplies cash receipt, remaining balance, renewal conditions, reporting schedule, a content veto or a named trustee's individual grant decision. The wider donor-list/funding-principles inquiry is a parallel lane, not silently supplied by these grant cards.
\n[Public Safety Canada contribution26007](http[local research file] reference088-2025-2026-Q3-00044. The complete original detail record identifies legal recipient The Christchurch Call Foundation, agreement value CAD1,369,064, term 23October2025–31March2028, under the Community Resilience Fund.
\nIts expected results are three workstreams: piloting interventions for TVEC risks in new technology sectors; partnering with ROOST to support open-source safety tools, particularly for smaller companies; and using the multistakeholder community to encourage uptake of the resulting interventions/tools/practices. These are funded project purposes and expected results, not proof the outputs have been delivered or accepted. The disclosure does not allocate the award among workstreams or specify a payment/approval schedule.
\nThis directly distinguishes ended NZ secretariat funding from later Canadian project support to the Foundation. It is neither a Canadian payment to TAT nor CCIAO financing. The legal-recipient record supports the Foundation bridge without assigning Canada operational authority over ROOST or a particular classification. It does not establish a government appointment or policy-veto right.
\n| From → to | \nRelation and state | \nSupported boundary | \n
|---|---|---|
| NZ/France-led support functions → Foundation secretariat | \nAnnounced14May2024 for1July; later MFAT-reported completed transition | \nFunction handoff; staff/assets/IP transfer terms not acquired. | \n
| NZ government → prior Call support funding | \nAnnounced end30June2024; MFAT retrospective confirmation | \nNot full governmental withdrawal or a French funding-stop record. | \n
| Packard → CCF | \n2024 general-support donor grant record | \nAmount/Term as displayed; payment and governance rights unknown. | \n
| Carnegie → CCF launch | \n12December2024 grant,$500000,12months | \nLaunch support; donor2023spin-out characterization not authoritative handoff date. | \n
| Carnegie trustees → CCF support grant | \n5March2026 donor board approval,$2m,24months | \nApproval is distinct from disbursement and recipient governance. | \n
| Public Safety Canada → CCF agreement26007 → three expected workstreams | \nDisclosed contribution,2025–28,includingROOSTtool partnership | \nNo payment, completed delivery, acceptance, suballocation or content mandate established. | \n
The high-value missing instrument is the actual2024 transition/handover agreement and schedules, joined to the Foundation's constituent instrument and any donor agreements that reserve rights. That could identify employer changes, assigned assets, retained governmental seats, approvals, obligations and termination remedies. Financial statements or grant-payment records would separately distinguish pledges/awards from receipts. No such document was requested from a person or private portal.
\nThe evidence is consistent with a more broadly financed secretariat retaining recipient discretion, while government participants continue to shape direction through the community and fund specified projects. It does not establish donor command, total governmental exit, or a purely private financing base. Acquisition stops here for the combined forest review.
\ntransition-captures.json records eight ordinary GET responses with exact URLs/times/hashes. Five contain substantive primary originals: French joint statement, MFAT annual report, MFAT OIA response, Packard grant card and Canadian contribution record. Three contain only access/challenge responses: NZ announcement shell and two Carnegie403s. All retained-byte hashes were checked. Reader-only Carnegie passages remain expressly distinguished from original captures; the two pages were not retried after the one public-reader substitute.
The annual-report download has259pages, but only PDF66 is the finding source. A keyword-location scan and incidental PDF60 excerpt do not count as a full report read. The OIA table required visual inspection because PDF2 lacked extracted text; transition-oia29223-page2.png is a derivative of the retained original. Its values were read visually. No handover staff, assets, contracts or financial-note schedule was found in these selected instruments.
Thirteen bounded queries covered the NZ/France handoff, official disclosures and the parent-supplied later grant leads: site:beehive.govt.nz "Christchurch Call Foundation" funding; site:mfat.govt.nz "Christchurch Call Foundation" funding; site:elysee.fr "Christchurch Call Foundation"; site:diplomatie.gouv.fr "Christchurch" "fondation" financement; site:mfat.govt.nz "Christchurch Call" "transition" funding Foundation; site:dpmc.govt.nz "Christchurch Call" Foundation; site:mfat.govt.nz "Christchurch Call" "Foundation" "OIA"; site:beehive.govt.nz "Christchurch Call" "million" 2024; "Christchurch Call Foundation" (transfer OR assets OR secondment OR funding) site:mfat.govt.nz; "Christchurch Call Foundation" site:treasury.govt.nz; "Christchurch Call Foundation" "France" "funding" government; site:carnegie.org "Christchurch Call Foundation" "2026"; site:search.open.canada.ca "26007" "Christchurch".
OIA29223's exact link came from the official2024 OIA index, whose matching row was read; other disclosures were not harvested. DPMC2022/2023 background papers, later ministerial references and the2025 recipient press release appeared as discovery only. Existing closed Foundation/funding website routes were not retried. The French presidency's independently published joint announcement provided the available original after the NZ challenge shell; this was not bypass of a private resource. The scope supplies no executed funding conditions or retained government appointment powers; their non-location in this finite pass is not evidence of nonexistence.
", "path": "christchurch-governance-wave-2026-09-09/transition-funding.md", "sourceIds": [ "w27-roost-s-transition-france", "w27-roost-s-mfat-transition", "w27-roost-s-mfat-costs", "w27-roost-s-packard2024", "w27-roost-s-carnegie2024", "w27-roost-s-carnegie2026", "w27-roost-s-canada26007" ], "markdown": "9 September 2026. Bounded transition/funding acquisition closed. This note separates the 2024 support-function handoff, the continuing Call, the new Foundation, and later project funding. Legal-entity governance and community decision processes belong to the parallel lanes. No outreach, account, payment or private record access; no canonical changes.\n\n## What the handoff actually establishes\n\nThe **14 May 2024 joint NZ/France announcement** says support functions would move to a new secretariat funded by the Foundation, operating from **1 July 2024**, while New Zealand taxpayer funding for the Call would end on **30 June**. Donor pledges had been attracted, but no amounts or received instalments are stated. The Foundation would coordinate and support the wider community; leaders from all sectors would continue making key directional decisions. Ardern's special-envoy term would end, with her continuing as unpaid patron for dialogue and advocacy. The announcement does not assign her a foundation voting or appointment power.\n\nJoint official statement, French presidency publication (see source references) Complete substantive body and notes read. The NZ publication is the same joint-source family, not independent corroboration; its ordinary GET returned a challenge shell and was closed.\n\n**MFAT subsequently reports the transition as completed**, with New Zealand government funding ending in June 2024 and support functions financed by community/philanthropic pledges. This is a retrospective departmental performance account, not proof that every pledge had been paid. MFAT Annual Report 2023–24 (see source references) **PDF66/printed66**, complete Outcome1 paragraph context read. The report's publication date was not separately established; its reporting period is distinct from the September2026 retrieval.\n\nThe defensible change is government-funded coordination becoming Foundation-funded coordination, while governments remain in the wider initiative. These documents do not prove full NZ or French withdrawal, French funding termination, or an absence of continued diplomatic/in-kind work. No signed handover deed, staff secondment/novation schedule, asset inventory, IP assignment, transferred bank balance or government reserved appointment/approval instrument was acquired. Moving support functions does not by itself establish a transfer of ownership of assets or employment contracts.\n\n## The government baseline is operating expenditure, not foundation capitalization\n\nMFAT OIA29223, dated4April2024 (see source references) both pages read; AppendixA onPDF2 visually inspected. The response identifies costs incurred **by MFAT only**, in **NZD**, from1July2018 through29February2024.\n\n| Financial year | Reported actual expenditure NZD |\n| --- | --- |\n| 2018/19 | 87,936 |\n| 2019/20 | 530,486 |\n| 2020/21 | 856,130 |\n| 2021/22 | 357,172 |\n| 2022/23 | 638,646 |\n| 2023/24 YTD to29February2024 | 412,853 |\n\nThe last year is incomplete and subject to later review. These are not the Call's complete costs across all departments/countries, not a Foundation grant, and not payment vouchers. No cumulative influence budget is constructed. The response was listed for publication on31May2024 in MFAT's2024 OIA index; the letter itself is4April.\n\n## Actual named donor support after the restructuring\n\nPackard's CCF grant card (see source references) records **2024**, **$350,000**, **Term36**, for **general support**, in Just Societies / Democracy, Rights and Governance. Complete substantive card read and original retained. The card does not explicitly state the term's unit or currency, exact award/payment dates, expenditure allocation or donor appointment/clearance rights. Preserve its displayed fields; do not turn the likely36-month term into verified wording without a further source. The US donor's dollar context is consistent with USD, but the currency is not separately printed in this card.\n\nCarnegie's grant database, page13 (see source references) records **12December2024**, **$500,000**, **12months**, Special Opportunities Fund, to support CCF's launch. The full CCF entry was read through the primary public reader, lines152–194; ordinary GET403 was retained only as an error response and the route closed. The description emphasizes launch as an independent body but dates the spin-out to2023. That donor characterization does not override the joint governments' July2024 operational handoff. The discrepancy could reflect preparation versus operation, or imprecise donor wording; this pass does not resolve it.\n\nCarnegie's 10March2026 board-grants announcement (see source references) says its trustees approved grants at the **5March2026** meeting, including **$2,000,000 over24months** for CCF to counter terrorist/violent-extremist content online. Read title/date, approval introduction and CCF entry through the primary reader, lines30–34 and67–68. This earns **donor board approval**, not payment or a right to run Foundation decisions. Original GET403 was closed and retained as error only. Dollar amounts are displayed by a US donor; the brief entry does not separately label currency.\n\nThe two Carnegie records are distinct dated grants; do not treat the second as an amendment or sum them with Packard/Canada into current annual resources. None supplies cash receipt, remaining balance, renewal conditions, reporting schedule, a content veto or a named trustee's individual grant decision. The wider donor-list/funding-principles inquiry is a parallel lane, not silently supplied by these grant cards.\n\n## Later public project funding survives the NZ support-function exit\n\nPublic Safety Canada contribution26007 (see source references) reference088-2025-2026-Q3-00044. The complete original detail record identifies legal recipient **The Christchurch Call Foundation**, agreement value **CAD1,369,064**, term **23October2025–31March2028**, under the Community Resilience Fund.\n\nIts expected results are three workstreams: piloting interventions for TVEC risks in new technology sectors; partnering with **ROOST** to support open-source safety tools, particularly for smaller companies; and using the multistakeholder community to encourage uptake of the resulting interventions/tools/practices. These are **funded project purposes and expected results**, not proof the outputs have been delivered or accepted. The disclosure does not allocate the award among workstreams or specify a payment/approval schedule.\n\nThis directly distinguishes ended NZ secretariat funding from later Canadian project support to the Foundation. It is neither a Canadian payment to TAT nor CCIAO financing. The legal-recipient record supports the Foundation bridge without assigning Canada operational authority over ROOST or a particular classification. It does not establish a government appointment or policy-veto right.\n\n## Candidate structure and unresolved decision rights\n\n| From → to | Relation and state | Supported boundary |\n| --- | --- | --- |\n| NZ/France-led support functions → Foundation secretariat | Announced14May2024 for1July; later MFAT-reported completed transition | Function handoff; staff/assets/IP transfer terms not acquired. |\n| NZ government → prior Call support funding | Announced end30June2024; MFAT retrospective confirmation | Not full governmental withdrawal or a French funding-stop record. |\n| Packard → CCF | 2024 general-support donor grant record | Amount/Term as displayed; payment and governance rights unknown. |\n| Carnegie → CCF launch | 12December2024 grant,$500000,12months | Launch support; donor2023spin-out characterization not authoritative handoff date. |\n| Carnegie trustees → CCF support grant | 5March2026 donor board approval,$2m,24months | Approval is distinct from disbursement and recipient governance. |\n| Public Safety Canada → CCF agreement26007 → three expected workstreams | Disclosed contribution,2025–28,includingROOSTtool partnership | No payment, completed delivery, acceptance, suballocation or content mandate established. |\n\nThe high-value missing instrument is the actual2024 transition/handover agreement and schedules, joined to the Foundation's constituent instrument and any donor agreements that reserve rights. That could identify employer changes, assigned assets, retained governmental seats, approvals, obligations and termination remedies. Financial statements or grant-payment records would separately distinguish pledges/awards from receipts. No such document was requested from a person or private portal.\n\nThe evidence is consistent with a more broadly financed secretariat retaining recipient discretion, while government participants continue to shape direction through the community and fund specified projects. It does not establish donor command, total governmental exit, or a purely private financing base. Acquisition stops here for the combined forest review.\n\n## Custody, source scopes and finite search\n\n`transition-captures.json` records eight ordinary GET responses with exact URLs/times/hashes. Five contain substantive primary originals: French joint statement, MFAT annual report, MFAT OIA response, Packard grant card and Canadian contribution record. Three contain only access/challenge responses: NZ announcement shell and two Carnegie403s. All retained-byte hashes were checked. Reader-only Carnegie passages remain expressly distinguished from original captures; the two pages were not retried after the one public-reader substitute.\n\nThe annual-report download has259pages, but only PDF66 is the finding source. A keyword-location scan and incidental PDF60 excerpt do not count as a full report read. The OIA table required visual inspection because PDF2 lacked extracted text; `transition-oia29223-page2.png` is a derivative of the retained original. Its values were read visually. No handover staff, assets, contracts or financial-note schedule was found in these selected instruments.\n\nThirteen bounded queries covered the NZ/France handoff, official disclosures and the parent-supplied later grant leads: `site:beehive.govt.nz \"Christchurch Call Foundation\" funding`; `site:mfat.govt.nz \"Christchurch Call Foundation\" funding`; `site:elysee.fr \"Christchurch Call Foundation\"`; `site:diplomatie.gouv.fr \"Christchurch\" \"fondation\" financement`; `site:mfat.govt.nz \"Christchurch Call\" \"transition\" funding Foundation`; `site:dpmc.govt.nz \"Christchurch Call\" Foundation`; `site:mfat.govt.nz \"Christchurch Call\" \"Foundation\" \"OIA\"`; `site:beehive.govt.nz \"Christchurch Call\" \"million\" 2024`; `\"Christchurch Call Foundation\" (transfer OR assets OR secondment OR funding) site:mfat.govt.nz`; `\"Christchurch Call Foundation\" site:treasury.govt.nz`; `\"Christchurch Call Foundation\" \"France\" \"funding\" government`; `site:carnegie.org \"Christchurch Call Foundation\" \"2026\"`; `site:search.open.canada.ca \"26007\" \"Christchurch\"`.\n\nOIA29223's exact link came from the official2024 OIA index, whose matching row was read; other disclosures were not harvested. DPMC2022/2023 background papers, later ministerial references and the2025 recipient press release appeared as discovery only. Existing closed Foundation/funding website routes were not retried. The French presidency's independently published joint announcement provided the available original after the NZ challenge shell; this was not bypass of a private resource. The scope supplies no executed funding conditions or retained government appointment powers; their non-location in this finite pass is not evidence of nonexistence.\n", "file": "research/christchurch-governance-wave-2026-09-09-transition-funding.md" }, { "id": "cma-danish-support-wave-2026-09-09-kr-support", "title": "KR Foundation, specified advocacy and CMA's changing funding mix", "description": "9 September 2026. Bounded acquisition closed. A named Danish grant is established, with a programme purpose more specific than general organizational support. Later recipient reporting also changes the temporal picture: foundations dominate the2025funding graphic, whereas major donors dominate2024. These sources establish published purposes, resource relatio", "html": "9 September 2026. Bounded acquisition closed. A named Danish grant is established, with a programme purpose more specific than general organizational support. Later recipient reporting also changes the temporal picture: foundations dominate the2025funding graphic, whereas major donors dominate2024. These sources establish published purposes, resource relationships and internal foundation authority, not a command to undertake the February2025Senate intervention.
\nKR Foundation's2023annual-report grant table, printed46/PDF24 of the retained double-page republication, lists Check My Ads, US, DKK1,991,232 granted in2023, start1May2023, end1May2025. Its project is titled 'Dismantle the disinformation economy: cut climate change and election disinformation off at the source.' Root visually inspected the row, headers and amount. The old [donor-hosted report address](http[local research file] remains indexed with the same table but ordinaryGET returned404. Root obtained the report from an [Americans for Public Trust republication](http[local research file] this is a donor-authored document in third-party custody, not an endorsement of the host's framing or a fresh donor original. SHA301137e051f62930e4be97813bdd0996f81a5fd7d4778007bd329128855613d5.
\nThe live donor website links a 2023grant list.
\nThe records identify actual Danish support consistent with CMA's own declaration. They do not establish that KR was its only Danish supporter or permit allocating the whole grant to calendar2024or the Senate campaign. DKK is retained without a speculative exchange-rate conversion. The2025award begins after the Februaryletters; the earlier award term overlaps them, but its named purpose concerns climate and election disinformation, not CSAM.
\nThe same2023report, printed16–17/PDF9, explicitly arranges supported organizations around regulatory change, pressure on fossil-fuel service providers, and capacity building. Root viewed the actual diagram after extraction produced repeated hidden text. It places CMA, Center for Countering Digital Hate and Stop Funding Heat in the same category concerning accountability of enablers. It places Friends of the Earth/CAAD in a separate capacity-building category. This is the funder's published programme architecture. It proves neither that every listed organization collaborated on one campaign nor that any one directed another. The stated2023geographic priorities include Europe, USA and Australia. Those programme statements, together with the named grant, are stronger evidence of intentional funding around an agenda than mere overlap in a donor list.
\nThe current [How we work page](http[local research file] substantive primary reader, says applications are invitation-only and grants board-approved. OrdinaryGET returned455; that response is not the substantive page. The [currently linked statutes translation](http[local research file] original captured and all five pages read, supplies more exact authority. Section5 assigns distribution, eligibility, scale and timing to the board and permits defined delegation to the chairmanship/executive director. Section7 entitles VILLUM FONDEN to appoint up to two of five-to-seven board members; remaining seats are filled by board election. Election of those remaining members requires a majority with no more than one dissent. This means two dissenting eligible directors can block such an election; it does not prove that the founder-appointed directors coordinated votes or can individually dictate grants. Ordinary decisions use simple majority and a chair casting vote. Section8 specifies joint signatures or delegated power to bind the foundation. This is internal KR authority, not authority over CMA's publications.
\nSections1and4 identify VILLUM FONDEN and descendants of Villum Kann Rasmussen as founders and a DKK10million initial cash capital, with no return of that initial capital to founders. No present grant is traced to a particular upstream transfer, and no wider family, company or state financing claim is made. The original first-board names in the statutes are historical, not today's board. Current bios were read only as context and are not used to retroactively assign the2023grant vote.
\nNo executed CMA award agreement, approved proposal, reporting template, tranche condition, clawback, editorial approval, target list or evaluation was found in the inspected grant lists, reports, statutory text and two focused searches. Those instruments remain unacquired; invitation-only selection is an actual programme-entry gate, not proof of later editorial veto. The linked2023financial report was located and the opening reader through the auditor's introductory scope was inspected; its financial totals and grant accounting have not been adopted here.
\nThe [CMA2025annual report](http[local research file] original SHA570059f9a0e37ddbf5993033765238a4717d089f39b81db3c02ba8632e4193a2, PDF35 visually inspected, explicitly reports foundation grants82.8%, major donors12.3%, individual donors4.9%, four new institutional donors and three multi-year renewals. It also adds UK and Canadian foundation support alongside USA and Denmark. This supersedes any implication that2024's major-donor majority describes the later mix. Absolute income, category definitions and accounting basis still are not given on that page. The shift may combine new funding, reduced large gifts, recognition timing or recategorization; the two graphics alone do not resolve the cause or identify every new funder.
\nScoped report text read atPDF26,29,32and35 adds direct institutional leads: CMA claims a TAG complaint, support for industry KYC and MRC changes; it describes an Ofcom Online Safety Act additional-safety submission seeking tighter monetization controls. Its policy platform names KYC due diligence and mandatory supply-chain transparency alongside privacy-respecting advertising. These are the organization's claims and proposals, not observed enforcement or an established duty for all internet users to identify themselves. The actual Ofcom submission and its proposed subjects, triggers and safeguards would decide more than inference from the initials KYC. The rest of the37-page report has been captured, not comprehensively reviewed; incidental excerpts are not a whole-report audit.
\nSecondary search also surfaced a2025Oak Foundation grant claim; it remains unverified discovery for a later named-funder reconciliation, not an accepted edge here. No unsolicited expansion into donor genealogy occurred.
\nPer-file captureJSON records status, URLs, bytes and hashes. The original2023list SHA4cad141c609ac4de80c101d1bcce82f7fc291e62c0470ef60e77117477404742; original2025list SHA40704711dab007be9129e88e933134a8fc15bf4532cde43800e2c1a968a0e65d; originalstatutes SHA14806448fb046190304f56f5471239aeb3405645d176910d9396040bec8a8fa7. The failed oldreport request is a404HTML response, originally named.pdf by the capture script; its manifest explicitly saysoriginal_pdf=false. It was never parsed as aPDF or treated as missing evidence of suppression. PNGs of the table, programme diagram,2025list and CMA2025chart were actually viewed. Sources denied or missing were closed; no challenges, accounts, visible UI, outreach or ZIP work occurred.
", "path": "cma-danish-support-wave-2026-09-09/kr-support.md", "sourceIds": [ "w31-fa-s-kr2023", "w31-fa-s-kr2023list", "w31-fa-s-kr2025list", "w31-fa-s-krstatutes", "w31-fa-s-cma2025" ], "markdown": "9 September 2026. Bounded acquisition closed. A named Danish grant is established, with a programme purpose more specific than general organizational support. Later recipient reporting also changes the temporal picture: foundations dominate the2025funding graphic, whereas major donors dominate2024. These sources establish published purposes, resource relationships and internal foundation authority, not a command to undertake the February2025Senate intervention.\n\n## The actual grant\n\nKR Foundation's2023annual-report grant table, printed46/PDF24 of the retained double-page republication, lists Check My Ads, US, DKK1,991,232 granted in2023, start1May2023, end1May2025. Its project is titled 'Dismantle the disinformation economy: cut climate change and election disinformation off at the source.' Root visually inspected the row, headers and amount. The old donor-hosted report address (see source references) remains indexed with the same table but ordinaryGET returned404. Root obtained the report from an Americans for Public Trust republication (see source references) this is a donor-authored document in third-party custody, not an endorsement of the host's framing or a fresh donor original. SHA301137e051f62930e4be97813bdd0996f81a5fd7d4778007bd329128855613d5.\n\nThe live donor website links a 2023grant list.\n\nThe records identify actual Danish support consistent with CMA's own declaration. They do not establish that KR was its only Danish supporter or permit allocating the whole grant to calendar2024or the Senate campaign. DKK is retained without a speculative exchange-rate conversion. The2025award begins after the Februaryletters; the earlier award term overlaps them, but its named purpose concerns climate and election disinformation, not CSAM.\n\n## What the funder says it is building\n\nThe same2023report, printed16–17/PDF9, explicitly arranges supported organizations around regulatory change, pressure on fossil-fuel service providers, and capacity building. Root viewed the actual diagram after extraction produced repeated hidden text. It places CMA, Center for Countering Digital Hate and Stop Funding Heat in the same category concerning accountability of enablers. It places Friends of the Earth/CAAD in a separate capacity-building category. This is the funder's published programme architecture. It proves neither that every listed organization collaborated on one campaign nor that any one directed another. The stated2023geographic priorities include Europe, USA and Australia. Those programme statements, together with the named grant, are stronger evidence of intentional funding around an agenda than mere overlap in a donor list.\n\nThe current How we work page (see source references) substantive primary reader, says applications are invitation-only and grants board-approved. OrdinaryGET returned455; that response is not the substantive page. The currently linked statutes translation (see source references) original captured and all five pages read, supplies more exact authority. Section5 assigns distribution, eligibility, scale and timing to the board and permits defined delegation to the chairmanship/executive director. Section7 entitles VILLUM FONDEN to appoint up to two of five-to-seven board members; remaining seats are filled by board election. Election of those remaining members requires a majority with no more than one dissent. This means two dissenting eligible directors can block such an election; it does not prove that the founder-appointed directors coordinated votes or can individually dictate grants. Ordinary decisions use simple majority and a chair casting vote. Section8 specifies joint signatures or delegated power to bind the foundation. This is internal KR authority, not authority over CMA's publications.\n\nSections1and4 identify VILLUM FONDEN and descendants of Villum Kann Rasmussen as founders and a DKK10million initial cash capital, with no return of that initial capital to founders. No present grant is traced to a particular upstream transfer, and no wider family, company or state financing claim is made. The original first-board names in the statutes are historical, not today's board. Current bios were read only as context and are not used to retroactively assign the2023grant vote.\n\nNo executed CMA award agreement, approved proposal, reporting template, tranche condition, clawback, editorial approval, target list or evaluation was found in the inspected grant lists, reports, statutory text and two focused searches. Those instruments remain unacquired; invitation-only selection is an actual programme-entry gate, not proof of later editorial veto. The linked2023financial report was located and the opening reader through the auditor's introductory scope was inspected; its financial totals and grant accounting have not been adopted here.\n\n## A consequential time correction\n\nThe CMA2025annual report (see source references) original SHA570059f9a0e37ddbf5993033765238a4717d089f39b81db3c02ba8632e4193a2, PDF35 visually inspected, explicitly reports foundation grants82.8%, major donors12.3%, individual donors4.9%, four new institutional donors and three multi-year renewals. It also adds UK and Canadian foundation support alongside USA and Denmark. This supersedes any implication that2024's major-donor majority describes the later mix. Absolute income, category definitions and accounting basis still are not given on that page. The shift may combine new funding, reduced large gifts, recognition timing or recategorization; the two graphics alone do not resolve the cause or identify every new funder.\n\nScoped report text read atPDF26,29,32and35 adds direct institutional leads: CMA claims a TAG complaint, support for industry KYC and MRC changes; it describes an Ofcom Online Safety Act additional-safety submission seeking tighter monetization controls. Its policy platform names KYC due diligence and mandatory supply-chain transparency alongside privacy-respecting advertising. These are the organization's claims and proposals, not observed enforcement or an established duty for all internet users to identify themselves. The actual Ofcom submission and its proposed subjects, triggers and safeguards would decide more than inference from the initials KYC. The rest of the37-page report has been captured, not comprehensively reviewed; incidental excerpts are not a whole-report audit.\n\nSecondary search also surfaced a2025Oak Foundation grant claim; it remains unverified discovery for a later named-funder reconciliation, not an accepted edge here. No unsolicited expansion into donor genealogy occurred.\n\n## Custody\n\nPer-file captureJSON records status, URLs, bytes and hashes. The original2023list SHA4cad141c609ac4de80c101d1bcce82f7fc291e62c0470ef60e77117477404742; original2025list SHA40704711dab007be9129e88e933134a8fc15bf4532cde43800e2c1a968a0e65d; originalstatutes SHA14806448fb046190304f56f5471239aeb3405645d176910d9396040bec8a8fa7. The failed oldreport request is a404HTML response, originally named.pdf by the capture script; its manifest explicitly saysoriginal_pdf=false. It was never parsed as aPDF or treated as missing evidence of suppression. PNGs of the table, programme diagram,2025list and CMA2025chart were actually viewed. Sources denied or missing were closed; no challenges, accounts, visible UI, outreach or ZIP work occurred.\n", "file": "research/cma-danish-support-wave-2026-09-09-kr-support.md" }, { "id": "cma-ofcom-monetization-wave-2026-09-09-submission-findings", "title": "CMA's monetization proposal and Ofcom's response", "description": "Closed bounded acquisition: 9 September 2026. This packet concerns Check My Ads Institute's response to Ofcom's Additional Safety Measures consultation, not an enacted CMA rule. The consultation opened 30 June and closed 20 October 2025. Its regulator-hosted response entry is dated 11 December 2025; that is the displayed publication/listing date, not an inde", "html": "Closed bounded acquisition: 9 September 2026. This packet concerns Check My Ads Institute's response to Ofcom's Additional Safety Measures consultation, not an enacted CMA rule. The consultation opened 30 June and closed 20 October 2025. Its regulator-hosted response entry is dated 11 December 2025; that is the displayed publication/listing date, not an independently established submission date.
\nThe submission connects platform safety to access to monetization: verification of creators and advertisers, content clearance, account bans, advertiser disclosure and refunds. Ofcom later explicitly considered CMA's crisis-revenue request but did not recommend specific monetization practices in its crisis measures. The disposition of CMA's other proposals is not established by this packet.
\nAll page numbers below are one-based PDF pages of the 31-page [CMA response](http[local research file] Retained full extracted text: submission.reader.txt. It is a reader derivative, not the PDF original.
\n| Scope | \nProposed authority or consequence | \nLocator | \n
|---|---|---|
| Monetizing creators and advertisers | \nStronger identity/customer checks; visibility into creator payment and location; ability to trace repeat violators. No prescribed identity document or technical verification standard. | \nQ6, PDF2–3; Q33, PDF10; Q56, PDF14–15 | \n
| Content showing indications of illegal-harm risk | \nWithhold monetization until platform systems clear the risk; manually review content indicating possible illegality before monetizing it. Platforms' systems and moderation teams make these proposed decisions; no prior court finding is specified. | \nQ31, PDF9; Q35, PDF10 | \n
| Adult CSAM offenders and repeat illegal-content monetizers | \nBan the specified adult users and prevent return; permanently ban repeat illegal-content monetizers and claw back revenue for advertisers. These combine account-access and monetization proposals, which are different restrictions. | \nQ36, PDF10; Q39, PDF11 | \n
| Ads alongside identified harmful/unsafe content | \nReport the incident to Ofcom and advertisers; refund advertisers without a minimum-value exemption. The categorical invoice example is illustrative. The response sometimes uses harmful/unsafe wording rather than only established illegality. | \nQ11–12, PDF4–6 | \n
| Livestream and embedded-video advertising | \nRestrict children's automated ad monetization; favor direct/reserved advertising for livestreams; consider off-platform harm where embedded video still earns platform/creator revenue. | \nQ7 and Q9, PDF3–4; Q14, PDF6 | \n
| Public-safety crisis | \nSupport crisis demonetization and seek platform revenue disclosure/accounting with refunds where possible. This qualified crisis-refund proposal differs from Q12's broader refund request. | \nQ49–52, PDF12–14 | \n
The identified KYC/KYBC subjects include content creators and advertisers, not necessarily incorporated companies. The response does not prescribe verification for every reader, viewer or ordinary internet user. Nevertheless, identifying a monetizing creator could affect pseudonymity and access to income. Absence of universal user ID does not eliminate that consequence. Q56 also uses a different expansion of KYBC (“Know-Your-Business-Consumer”); the response does not supply a settled definition or enrollment procedure.
\nCMA endorses the package's proportionality but supplies no comprehensive verification-data minimization, retention, identity-error correction, creator appeal or correction-propagation mechanism in the inspected response. It declines Q42–44 on age-assessment appeals and highly effective age assurance; those nonanswers are not affirmative support or opposition. Its appendix reproduces Twitch and TikTok appeal language and YouTube human-review/context exceptions (PDF17–18 and25). Those are attributed platform-policy quotations, not a new CMA-designed procedure or independently verified current platform rules. No underlying policy execution or actual refund is established.
\nOfcom's [crisis-response statement](http[local research file] published 9 June 2026, records CMA's revenue-disclosure/refund request at A1.147, PDF47/printed47, citing CMA response p14. At A1.148, Ofcom recognizes advertising incentives but focuses the measures on illegal content and child harms: a provider may choose proportionate advertising-integrity steps, while the measures do not recommend specific monetization practices. A1.150–151 preserve human oversight, transparency, redress and risk-based tailoring. This is explicit consideration and non-adoption of specified monetization practices in this particular measure, not proof that Ofcom rejected every CMA identity, refund or content-review proposal.
\nThe same statement explains that crisis measures do not change existing moderation/appeal requirements (A1.127–131); expression, privacy and data protection still constrain action, including data minimization and storage limitation (A1.132–138). Section3.45, PDF16, excludes a general intervention against misinformation/disinformation unless it falls within the relevant illegal-content or child-harm scope. These are Ofcom's stated boundaries, not evidence of completed implementation or a guarantee against every erroneous restriction.
\nThe consultation index still displayed a pending-statement label and a 12 May 2026 update when inspected. That aggregate label cannot override the dated June crisis decision. Section1.12 of the June statement expected most remaining decisions in autumn 2026; this packet does not establish their status on 9 September or treat the June forecast as a current completion record.
\nThe reused original CMA annual report, PDF29, confirms CMA's own account of this submission and calls for stricter monetization controls/transparency. Its linked context page was inaccessible: primary reader failed, then one ordinary GET returned 403. That route is closed.
\nThe submission's ordinary GET also returned 403; the working official reader supplied all 31 pages of extracted text, including all 60 questions, appendix and footnotes. Image-only Exhibit A was not visually inspected. No submission PDF original was retained. The Ofcom index and June statement were read through the official reader, with no original bytes acquired. ofcom-index-and-disposition.reader.txt preserves the scoped returns. submission-captures.json records exact reading extents, dates, hashes and access boundaries. Stored reader text is not original-byte custody.
\nThe strongest rival to a binding monetization-control interpretation is visible in the regulator's own response: it considered CMA's fiscal request while leaving the specific advertising response to proportionate provider choice. The useful missing instrument is the final disposition of CMA's creator/advertiser verification and pre-monetization review proposals, together with any adopted operational standard that specifies identity evidence, error correction and appeal. A platform's actual implementation would be a further, separate record. No additional acquisition branch was opened.
", "path": "cma-ofcom-monetization-wave-2026-09-09/submission-findings.md", "sourceIds": [ "w31-fa-s-cmasub" ], "markdown": "Closed bounded acquisition: 9 September 2026. This packet concerns Check My Ads Institute's response to Ofcom's Additional Safety Measures consultation, not an enacted CMA rule. The consultation opened 30 June and closed 20 October 2025. Its regulator-hosted response entry is dated 11 December 2025; that is the displayed publication/listing date, not an independently established submission date.\n\nThe submission connects platform safety to access to monetization: verification of creators and advertisers, content clearance, account bans, advertiser disclosure and refunds. Ofcom later explicitly considered CMA's crisis-revenue request but did not recommend specific monetization practices in its crisis measures. The disposition of CMA's other proposals is not established by this packet.\n\n## Proposed subjects, triggers and decisions\n\nAll page numbers below are one-based PDF pages of the 31-page CMA response (see source references) Retained full extracted text: submission.reader.txt. It is a reader derivative, not the PDF original.\n\n| Scope | Proposed authority or consequence | Locator |\n| --- | --- | --- |\n| Monetizing creators and advertisers | Stronger identity/customer checks; visibility into creator payment and location; ability to trace repeat violators. No prescribed identity document or technical verification standard. | Q6, PDF2–3; Q33, PDF10; Q56, PDF14–15 |\n| Content showing indications of illegal-harm risk | Withhold monetization until platform systems clear the risk; manually review content indicating possible illegality before monetizing it. Platforms' systems and moderation teams make these proposed decisions; no prior court finding is specified. | Q31, PDF9; Q35, PDF10 |\n| Adult CSAM offenders and repeat illegal-content monetizers | Ban the specified adult users and prevent return; permanently ban repeat illegal-content monetizers and claw back revenue for advertisers. These combine account-access and monetization proposals, which are different restrictions. | Q36, PDF10; Q39, PDF11 |\n| Ads alongside identified harmful/unsafe content | Report the incident to Ofcom and advertisers; refund advertisers without a minimum-value exemption. The categorical invoice example is illustrative. The response sometimes uses harmful/unsafe wording rather than only established illegality. | Q11–12, PDF4–6 |\n| Livestream and embedded-video advertising | Restrict children's automated ad monetization; favor direct/reserved advertising for livestreams; consider off-platform harm where embedded video still earns platform/creator revenue. | Q7 and Q9, PDF3–4; Q14, PDF6 |\n| Public-safety crisis | Support crisis demonetization and seek platform revenue disclosure/accounting with refunds where possible. This qualified crisis-refund proposal differs from Q12's broader refund request. | Q49–52, PDF12–14 |\n\nThe identified KYC/KYBC subjects include content creators and advertisers, not necessarily incorporated companies. The response does not prescribe verification for every reader, viewer or ordinary internet user. Nevertheless, identifying a monetizing creator could affect pseudonymity and access to income. Absence of universal user ID does not eliminate that consequence. Q56 also uses a different expansion of KYBC (“Know-Your-Business-Consumer”); the response does not supply a settled definition or enrollment procedure.\n\nCMA endorses the package's proportionality but supplies no comprehensive verification-data minimization, retention, identity-error correction, creator appeal or correction-propagation mechanism in the inspected response. It declines Q42–44 on age-assessment appeals and highly effective age assurance; those nonanswers are not affirmative support or opposition. Its appendix reproduces Twitch and TikTok appeal language and YouTube human-review/context exceptions (PDF17–18 and25). Those are attributed platform-policy quotations, not a new CMA-designed procedure or independently verified current platform rules. No underlying policy execution or actual refund is established.\n\n## The explicit decision, and its limit\n\nOfcom's crisis-response statement (see source references) published **9 June 2026**, records CMA's revenue-disclosure/refund request at **A1.147, PDF47/printed47**, citing CMA response p14. At **A1.148**, Ofcom recognizes advertising incentives but focuses the measures on illegal content and child harms: a provider may choose proportionate advertising-integrity steps, while the measures do not recommend specific monetization practices. **A1.150–151** preserve human oversight, transparency, redress and risk-based tailoring. This is explicit consideration and non-adoption of specified monetization practices in this particular measure, not proof that Ofcom rejected every CMA identity, refund or content-review proposal.\n\nThe same statement explains that crisis measures do not change existing moderation/appeal requirements (A1.127–131); expression, privacy and data protection still constrain action, including data minimization and storage limitation (A1.132–138). Section3.45, PDF16, excludes a general intervention against misinformation/disinformation unless it falls within the relevant illegal-content or child-harm scope. These are Ofcom's stated boundaries, not evidence of completed implementation or a guarantee against every erroneous restriction.\n\nThe consultation index still displayed a pending-statement label and a 12 May 2026 update when inspected. That aggregate label cannot override the dated June crisis decision. Section1.12 of the June statement expected most remaining decisions in autumn 2026; this packet does not establish their status on 9 September or treat the June forecast as a current completion record.\n\n## Custody and remaining discriminator\n\nThe reused original [CMA annual report](https://checkmyads.org/wp-content/uploads/2026/04/2025-annual-report-1.pdf), PDF29, confirms CMA's own account of this submission and calls for stricter monetization controls/transparency. Its linked context page was inaccessible: primary reader failed, then one ordinary GET returned 403. That route is closed.\n\nThe submission's ordinary GET also returned 403; the working official reader supplied all 31 pages of extracted text, including all 60 questions, appendix and footnotes. Image-only Exhibit A was not visually inspected. No submission PDF original was retained. The Ofcom index and June statement were read through the official reader, with no original bytes acquired. ofcom-index-and-disposition.reader.txt preserves the scoped returns. submission-captures.json records exact reading extents, dates, hashes and access boundaries. Stored reader text is not original-byte custody.\n\nThe strongest rival to a binding monetization-control interpretation is visible in the regulator's own response: it considered CMA's fiscal request while leaving the specific advertising response to proportionate provider choice. The useful missing instrument is the final disposition of CMA's creator/advertiser verification and pre-monetization review proposals, together with any adopted operational standard that specifies identity evidence, error correction and appeal. A platform's actual implementation would be a further, separate record. No additional acquisition branch was opened.\n", "file": "research/cma-ofcom-monetization-wave-2026-09-09-submission-findings.md" }, { "id": "cma-resources-wave-2026-09-09-donor-resources", "title": "Donor resources behind CMA capacity: Democracy Fund and Ford", "description": "9 September 2026. Bounded donor-side acquisition closed. The record supports general organizational support predating the February 2025 advocacy, but not an earmark for the Senate letters or donor approval of their wording. This lane does not equate approval, a future-payment balance, a reported payment and the recipient's accounting recognition.", "html": "9 September 2026. Bounded donor-side acquisition closed. The record supports general organizational support predating the February 2025 advocacy, but not an earmark for the Senate letters or donor approval of their wording. This lane does not equate approval, a future-payment balance, a reported payment and the recipient's accounting recognition.
\n| Donor-side record | \nInspected state | \nAmount, period and purpose | \nLimit | \n
|---|---|---|---|
| Democracy Fund, Inc., 2023 Form 990-PF, Part XIV | \nCMA is in both the paid-during-year and approved-for-future-payment schedules | \nPaid row is associated in the reader/index with USD 450,000, general support; calendar 2023. Future-payment row is also general support. | \nPDF text is interleaved across rows. This worker has not visually verified the paid amount's alignment and does not assign the future-payment amount. Do not total the two schedules as independent awards. | \n
| Democracy Fund, Inc., 2024 Form 990-PF, Part XIV | \nCMA appears in paid-during-year continuation | \nReader/index associates USD 250,000 with general support; calendar 2024. | \nNumerical row alignment needs a facsimile check before treating it as a fully verified payment amount. The schedule category itself is a donor report of payment, not an acquired bank transaction or proof of the recipient's use. | \n
| Ford grant 144432 | \nComplete substantive grant-detail record read | \nUSD 200,000 approved October 2022; start October 2022, end September 2023; General Support; Regular grant; Technology and Society programme, United States funding office | \nApproval/term record, not cash disbursement evidence or executed conditions. | \n
| Ford March 2025 award | \nOfficial search-index result only | \nIndexed USD 400,000, General Support, March 2025 | \nExact grant identifier, term and actual detailed record were not obtained. Retain discovery; no approved/paid edge from this lane. | \n
The [2023 Democracy Fund return](http[local research file] is a 63-page public donor copy. Relevant locators are PDF20/printed17, Part XIV grants paid, reader2118–2213; and PDF42/printed39, grants approved for future payment, reader4240–4331. PDF5 identifies calendar2023 and Democracy Fund, Inc., EIN38-3926408. The future-payment entry is a meaningful reconciliation warning: later payments could discharge an earlier commitment, but no grant identifier or instalment schedule here proves that particular linkage. Its value is therefore withheld rather than guessed from nearby interleaved numbers.
\nThe [2024 return](http[local research file] has 56 PDF pages. Relevant locus is PDF17/printed18, Part XIV grants paid, reader1968–2067; PDF1 identifies calendar2024 and the same donor/EIN. The website upload path is not the payment date. A return's reported annual payment cannot identify the day money reached CMA or its allocation to work in February2025. No other recipient, asset holding or donor genealogy was investigated.
\n[Ford grant144432](http[local research file] primary-reader lines124–160, expressly supplies the approval and grant term. Its short public purpose is General Support, not a named CSAM, Senate, TAG, DV or Adalytics project. The page does not expose a reporting schedule, tranche conditions, approval veto, work-product licence or communications review right. Those rights are unacquired, not proved absent.
\nFord's [indexed database result](http[local research file] supplies the later USD400,000/March2025 discovery. Opening that result returned a database heading without the grant rows; the actual More about this grantee link from grant144432 returned a reader cache miss. Finite exact-name/date/amount searches did not resolve a grant-specific identifier. No grant page was guessed, no verification challenge was attempted, and the result is not promoted to a confirmed cash receipt. If later authenticated, March approval would still postdate the February letters and would not itself establish retrospective campaign funding.
\nGeneral support can finance staffing, research and advocacy capacity even without a project earmark. The defensible connection is therefore resource provision to the organization, with broad purpose compatibility. The rival explanation to donor-directed Senate advocacy is that donors supported CMA's existing mission while CMA independently selected tactics. Both can coexist with donor reporting requirements; these public records do not resolve the actual decision rights. No sum across Ford approvals, Democracy Fund payment schedules and CMA annual-report revenue is offered.
\nThe best next record is the actual general-support award agreement and instalment/recognition schedule spanning 2023–2025, matched by grant identifier to the donor's payment record and CMA's books. It would distinguish future commitments from subsequent payments and expose any reporting, approval, clawback or earmark condition. The specific Ford2025 grant identifier and a readable facsimile of the three Democracy Fund rows are narrower prerequisite gaps, not invitations to expand into unrelated funders. Root's recipient-side packet should remain the separate source for CMA income, staffing and reported delivery.
\ndonor-captures.json records source scopes, reader references, retained derivatives and hashes. No substantive original PDF or grant HTML was captured in this lane. Each Democracy Fund ordinary GET returned403 and was closed; the official primary PDF reader succeeded. Ford's two ordinary HTTP200 responses were human-verification shells, retained as transport evidence, not original grant data. No challenge was solved or submitted. The primary reader supplied the historical144432 record. The public index had no rows in the opened reader, and the linked grantee reader returned cache miss; those routes were not retried.
Screenshot calls for Democracy Fund2024 PDF17 and2023 PDF20/42 returned only textual references through this worker's wrapper, not viewable images. They are not visual checks. This is why numerical row alignment remains qualified. The full financial returns, unrelated portfolio entries, signed award terms, donor correspondence and bank records were not audited. No account, form, outreach, paid source, visible UI, case, visual, ZIP, current-work or forest edit occurred. Acquisition is closed for the combined forest assessment.
", "path": "cma-resources-wave-2026-09-09/donor-resources.md", "sourceIds": [ "w31-fa-s-democracy2024", "w31-fa-s-democracy2023", "w31-fa-s-ford2022", "w31-fa-s-ford2025" ], "markdown": "9 September 2026. Bounded donor-side acquisition closed. The record supports general organizational support predating the February 2025 advocacy, but not an earmark for the Senate letters or donor approval of their wording. This lane does not equate approval, a future-payment balance, a reported payment and the recipient's accounting recognition.\n\n## What is established and what still needs alignment\n\n| Donor-side record | Inspected state | Amount, period and purpose | Limit |\n| --- | --- | --- | --- |\n| Democracy Fund, Inc., 2023 Form 990-PF, Part XIV | CMA is in both the paid-during-year and approved-for-future-payment schedules | Paid row is associated in the reader/index with USD 450,000, general support; calendar 2023. Future-payment row is also general support. | PDF text is interleaved across rows. This worker has not visually verified the paid amount's alignment and does not assign the future-payment amount. Do not total the two schedules as independent awards. |\n| Democracy Fund, Inc., 2024 Form 990-PF, Part XIV | CMA appears in paid-during-year continuation | Reader/index associates USD 250,000 with general support; calendar 2024. | Numerical row alignment needs a facsimile check before treating it as a fully verified payment amount. The schedule category itself is a donor report of payment, not an acquired bank transaction or proof of the recipient's use. |\n| Ford grant 144432 | Complete substantive grant-detail record read | USD 200,000 approved October 2022; start October 2022, end September 2023; General Support; Regular grant; Technology and Society programme, United States funding office | Approval/term record, not cash disbursement evidence or executed conditions. |\n| Ford March 2025 award | Official search-index result only | Indexed USD 400,000, General Support, March 2025 | Exact grant identifier, term and actual detailed record were not obtained. Retain discovery; no approved/paid edge from this lane. |\n\nThe 2023 Democracy Fund return (see source references) is a 63-page public donor copy. Relevant locators are PDF20/printed17, Part XIV grants paid, reader2118–2213; and PDF42/printed39, grants approved for future payment, reader4240–4331. PDF5 identifies calendar2023 and Democracy Fund, Inc., EIN38-3926408. The future-payment entry is a meaningful reconciliation warning: later payments could discharge an earlier commitment, but no grant identifier or instalment schedule here proves that particular linkage. Its value is therefore withheld rather than guessed from nearby interleaved numbers.\n\nThe 2024 return (see source references) has 56 PDF pages. Relevant locus is PDF17/printed18, Part XIV grants paid, reader1968–2067; PDF1 identifies calendar2024 and the same donor/EIN. The website upload path is not the payment date. A return's reported annual payment cannot identify the day money reached CMA or its allocation to work in February2025. No other recipient, asset holding or donor genealogy was investigated.\n\nFord grant144432 (see source references) primary-reader lines124–160, expressly supplies the approval and grant term. Its short public purpose is General Support, not a named CSAM, Senate, TAG, DV or Adalytics project. The page does not expose a reporting schedule, tranche conditions, approval veto, work-product licence or communications review right. Those rights are unacquired, not proved absent.\n\nFord's indexed database result (see source references) supplies the later USD400,000/March2025 discovery. Opening that result returned a database heading without the grant rows; the actual More about this grantee link from grant144432 returned a reader cache miss. Finite exact-name/date/amount searches did not resolve a grant-specific identifier. No grant page was guessed, no verification challenge was attempted, and the result is not promoted to a confirmed cash receipt. If later authenticated, March approval would still postdate the February letters and would not itself establish retrospective campaign funding.\n\n## Influence interpretation and the discriminating record\n\nGeneral support can finance staffing, research and advocacy capacity even without a project earmark. The defensible connection is therefore resource provision to the organization, with broad purpose compatibility. The rival explanation to donor-directed Senate advocacy is that donors supported CMA's existing mission while CMA independently selected tactics. Both can coexist with donor reporting requirements; these public records do not resolve the actual decision rights. No sum across Ford approvals, Democracy Fund payment schedules and CMA annual-report revenue is offered.\n\nThe best next record is the actual general-support award agreement and instalment/recognition schedule spanning 2023–2025, matched by grant identifier to the donor's payment record and CMA's books. It would distinguish future commitments from subsequent payments and expose any reporting, approval, clawback or earmark condition. The specific Ford2025 grant identifier and a readable facsimile of the three Democracy Fund rows are narrower prerequisite gaps, not invitations to expand into unrelated funders. Root's recipient-side packet should remain the separate source for CMA income, staffing and reported delivery.\n\n## Acquisition and custody\n\n`donor-captures.json` records source scopes, reader references, retained derivatives and hashes. No substantive original PDF or grant HTML was captured in this lane. Each Democracy Fund ordinary GET returned403 and was closed; the official primary PDF reader succeeded. Ford's two ordinary HTTP200 responses were human-verification shells, retained as transport evidence, not original grant data. No challenge was solved or submitted. The primary reader supplied the historical144432 record. The public index had no rows in the opened reader, and the linked grantee reader returned cache miss; those routes were not retried.\n\nScreenshot calls for Democracy Fund2024 PDF17 and2023 PDF20/42 returned only textual references through this worker's wrapper, not viewable images. They are not visual checks. This is why numerical row alignment remains qualified. The full financial returns, unrelated portfolio entries, signed award terms, donor correspondence and bank records were not audited. No account, form, outreach, paid source, visible UI, case, visual, ZIP, current-work or forest edit occurred. Acquisition is closed for the combined forest assessment.\n", "file": "research/cma-resources-wave-2026-09-09-donor-resources.md" }, { "id": "cma-resources-wave-2026-09-09-recipient-resources", "title": "Recipient resources and the decision-rights gap", "description": "9 September 2026. Bounded recipient acquisition closed. Root read the 26-page annual report text in full and visually inspected PDF24 and25 of cma-annual2024-250226.pdf. The latter supplies actual chart alignment and complete professional names. These are CMA's own retrospective statements, not an independent impact or financial audit.", "html": "9 September 2026. Bounded recipient acquisition closed. Root read the retained 26-page annual report text in full and visually inspected PDF24 and25 of cma-annual2024-250226.pdf. The latter supplies actual chart alignment and complete professional names. These are CMA's own retrospective statements, not an independent impact or financial audit.
\nThe [2024 annual report](http[local research file] PDF24, presents a funding pie with major donors59, foundation grants30 and individual donors11. The numbers sum to100 and the graphic represents shares; the page does not expressly print a percent sign, dollar denominator, accounting basis or donor-category definitions. Its major-donor slice is the largest. A separate individual-giving chart divides Checkmates60 and one-time small donors40. Do not confuse the charts or assign60 percent of total funding to monthly donors. It reports490 recurring donors, two new major donors, one new institutional donor, and foundation support in the USA and Denmark. It names neither the major donors nor the Danish foundation on this page. No individual concentration, absolute gift amount or donor condition follows from those shares.
\nPDF25 names founders Claire Atkin and Nandini Jammi as board co-chairs, alongside Zach Edwards, Sabrina Hersi Issa, David Carroll and Dr. Joan Donovan. The team includes Arielle Garcia as COO and Sarah K. Wiley as director of policy. This is a historical organizational roster, not proof of current offices, independent directors under tax definitions, donor appointment rights or a voting agreement. PDF25 is visibly readable; the extracted text truncates several names and titles, so the visual takes precedence.
\nRead as a whole, the report describes resources supporting research, communication, industry engagement and public policy together. PDF13 says its trial resources supported inquiries and publisher litigation and informed investors; PDF20 claims briefings, appearances, drafting and coalition work; PDF21 claims its recommendations entered UN information-integrity principles. Those are explicit claims about the intended pathways of influence. They do not authenticate each causal success or demonstrate that funders picked those tactics. PDF16 also attacks the reliability and intrusiveness of advertising data, while PDF17 advocates advertiser visibility and placement choices. Its stated programme therefore cannot simply be equated with support for pervasive personal identification.
\nThe other retained annual-report original has identical extracted PDF24 content. No canonical-version or whole-document equivalence determination is made from the two filenames. Both originals remain preserved. No contracts, donor letters, internal board minutes, bank transactions or complete audited accounts were obtained.
\n[ProPublica's IRS-derived profile](http[local research file] read at FY2024 and2023, reports EIN87-1895699, 2024 revenueUSD1,461,222, expensesUSD2,055,146 and net assetsUSD572,856, versus2023 revenueUSD2,159,795, expensesUSD1,722,154 and net assetsUSD1,166,780. It reports2024 contributionsUSD1,457,471. These figures are accepted only as that provider's extracted filing data, pending inspection of the actual return. They suggest a materially different accounting picture from a simple growth narrative, but cannot establish a liquidity crisis, cash runway, today's finances or a motive for February2025 advocacy. Net assets are not equivalent to cash. A zero executive-compensation summary category is not evidence that officers were unpaid; the same profile lists officer compensation.
\nThe linked full-filing reader delivered a reconstruction shell, not the actual form. Its XML reader returned cache miss; one ordinary GET returned403 and was closed. The response bytes are explicitly a security page, not XML. Philanthropy.org's reader failed; its indexed summary independently repeats some figures, but its ordinary GET returned403 and was closed. CauseIQ's direct reader was robots-denied and was not retried. No provider was counted as an independent underlying financial return merely because it republished the same IRS data. No full return, public ScheduleB, donor-advised grant identity or financial-statement audit was acquired. The indexed Fidelity figure remains discovery and does not identify an underlying donor. Further acquisition on these routes is closed in this run.
\nRoot read donor-resources.md and independently read Ford grant144432's primary-reader substantive record. Ford approvedUSD200,000 of general support inOctober2022 forOctober2022–September2023, through Technology and Society. It is approval and term evidence, not cash payment or a2025 campaign earmark. The Democracy Fund paid/future-payment schedule distinctions are useful, but the worker's numerical row-alignment caveat remains in force: no new root facsimile check occurred. Neither later cash payments nor future commitments may be added to the Ford approval as a single year's recipient revenue. Ford's indexedMarch2025 USD400,000 remains discovery only.
\nThe resource provision is established more strongly than its conditions. General support can sustain personnel and strategic capacity even without a project earmark; it cannot, on these records, identify the person choosing a legislative target or approving wording. The report's major-donor majority makes unidentified gift conditions at least as relevant as the named foundations. Executed awards, instalment schedules, restrictions and board delegation rules would distinguish selection of a mission from direction of a tactic.
\nAnnual originals and derivatives are in recipient-captures.json; hashes remain73b946f4338a68819eee15930641ea2b7efe0e2e394455534a54d1fb264ec748 and e1868d673b15a793de5316db99fd9e6d2c9f09d2dc2c34ef3f6069cd984f7619. Actual visual derivatives: annual2024-p24.png and annual2024-p25.png. Root used the already saved pypdf text; an attempted pdftotext command was unavailable and abandoned, with no claimed output. Session67515 finished exit0 and supplied page lengths plus both PDF24 texts. recipient-return-capture.json and philanthropy-recipient-capture.json label blocked transport. ProPublica primary profile reader turn1552view0; reconstructed-form shell turn1553view0; Ford primary reader turn1556view0. No visible UI, outreach or packaging occurred.
", "path": "cma-resources-wave-2026-09-09/recipient-resources.md", "sourceIds": [ "w31-fa-s-cma2024" ], "markdown": "9 September 2026. Bounded recipient acquisition closed. Root read the retained 26-page annual report text in full and visually inspected PDF24 and25 of cma-annual2024-250226.pdf. The latter supplies actual chart alignment and complete professional names. These are CMA's own retrospective statements, not an independent impact or financial audit.\n\n## What the organization discloses\n\nThe 2024 annual report (see source references) PDF24, presents a funding pie with major donors59, foundation grants30 and individual donors11. The numbers sum to100 and the graphic represents shares; the page does not expressly print a percent sign, dollar denominator, accounting basis or donor-category definitions. Its major-donor slice is the largest. A separate individual-giving chart divides Checkmates60 and one-time small donors40. Do not confuse the charts or assign60 percent of total funding to monthly donors. It reports490 recurring donors, two new major donors, one new institutional donor, and foundation support in the USA and Denmark. It names neither the major donors nor the Danish foundation on this page. No individual concentration, absolute gift amount or donor condition follows from those shares.\n\nPDF25 names founders Claire Atkin and Nandini Jammi as board co-chairs, alongside Zach Edwards, Sabrina Hersi Issa, David Carroll and Dr. Joan Donovan. The team includes Arielle Garcia as COO and Sarah K. Wiley as director of policy. This is a historical organizational roster, not proof of current offices, independent directors under tax definitions, donor appointment rights or a voting agreement. PDF25 is visibly readable; the extracted text truncates several names and titles, so the visual takes precedence.\n\nRead as a whole, the report describes resources supporting research, communication, industry engagement and public policy together. PDF13 says its trial resources supported inquiries and publisher litigation and informed investors; PDF20 claims briefings, appearances, drafting and coalition work; PDF21 claims its recommendations entered UN information-integrity principles. Those are explicit claims about the intended pathways of influence. They do not authenticate each causal success or demonstrate that funders picked those tactics. PDF16 also attacks the reliability and intrusiveness of advertising data, while PDF17 advocates advertiser visibility and placement choices. Its stated programme therefore cannot simply be equated with support for pervasive personal identification.\n\nThe other retained annual-report original has identical extracted PDF24 content. No canonical-version or whole-document equivalence determination is made from the two filenames. Both originals remain preserved. No contracts, donor letters, internal board minutes, bank transactions or complete audited accounts were obtained.\n\n## Recipient filing discovery, with its source limit\n\nProPublica's IRS-derived profile (see source references) read at FY2024 and2023, reports EIN87-1895699, 2024 revenueUSD1,461,222, expensesUSD2,055,146 and net assetsUSD572,856, versus2023 revenueUSD2,159,795, expensesUSD1,722,154 and net assetsUSD1,166,780. It reports2024 contributionsUSD1,457,471. These figures are accepted only as that provider's extracted filing data, pending inspection of the actual return. They suggest a materially different accounting picture from a simple growth narrative, but cannot establish a liquidity crisis, cash runway, today's finances or a motive for February2025 advocacy. Net assets are not equivalent to cash. A zero executive-compensation summary category is not evidence that officers were unpaid; the same profile lists officer compensation.\n\nThe linked full-filing reader delivered a reconstruction shell, not the actual form. Its XML reader returned cache miss; one ordinary GET returned403 and was closed. The response bytes are explicitly a security page, not XML. Philanthropy.org's reader failed; its indexed summary independently repeats some figures, but its ordinary GET returned403 and was closed. CauseIQ's direct reader was robots-denied and was not retried. No provider was counted as an independent underlying financial return merely because it republished the same IRS data. No full return, public ScheduleB, donor-advised grant identity or financial-statement audit was acquired. The indexed Fidelity figure remains discovery and does not identify an underlying donor. Further acquisition on these routes is closed in this run.\n\n## Combined donor comparison\n\nRoot read donor-resources.md and independently read Ford grant144432's primary-reader substantive record. Ford approvedUSD200,000 of general support inOctober2022 forOctober2022–September2023, through Technology and Society. It is approval and term evidence, not cash payment or a2025 campaign earmark. The Democracy Fund paid/future-payment schedule distinctions are useful, but the worker's numerical row-alignment caveat remains in force: no new root facsimile check occurred. Neither later cash payments nor future commitments may be added to the Ford approval as a single year's recipient revenue. Ford's indexedMarch2025 USD400,000 remains discovery only.\n\nThe resource provision is established more strongly than its conditions. General support can sustain personnel and strategic capacity even without a project earmark; it cannot, on these records, identify the person choosing a legislative target or approving wording. The report's major-donor majority makes unidentified gift conditions at least as relevant as the named foundations. Executed awards, instalment schedules, restrictions and board delegation rules would distinguish selection of a mission from direction of a tactic.\n\n## Custody and scope\n\nAnnual originals and derivatives are in recipient-captures.json; hashes remain73b946f4338a68819eee15930641ea2b7efe0e2e394455534a54d1fb264ec748 and e1868d673b15a793de5316db99fd9e6d2c9f09d2dc2c34ef3f6069cd984f7619. Actual visual derivatives: annual2024-p24.png and annual2024-p25.png. Root used the already saved pypdf text; an attempted pdftotext command was unavailable and abandoned, with no claimed output. Session67515 finished exit0 and supplied page lengths plus both PDF24 texts. recipient-return-capture.json and philanthropy-recipient-capture.json label blocked transport. ProPublica primary profile reader turn1552view0; reconstructed-form shell turn1553view0; Ford primary reader turn1556view0. No visible UI, outreach or packaging occurred.\n", "file": "research/cma-resources-wave-2026-09-09-recipient-resources.md" }, { "id": "creative-advertising-wave-2026-09-09-list-authority", "title": "Operation Creative: police classification, advertiser choice and list correction", "description": "9 September 2026. Bounded acquisition closed. This packet establishes a dated public procedure and a later procurement specification, with attributed delivery accounts. It does not certify the currently deployed portal or a September 2026 operating policy. No operational list, account, form, target website or gated tender was accessed.", "html": "9 September 2026. Bounded acquisition closed. This packet establishes a dated public procedure and a later procurement specification, with attributed delivery accounts. It does not certify the currently deployed portal or a September 2026 operating policy. No operational list, account, form, target website or gated tender was accessed.
\nThe decisive link is an assessment by police between a rightsholder's evidence and a list distributed for advertising decisions. The strongest acquired safeguard is an express opportunity to challenge police action and immediate removal when evidence shows that the site no longer commits offences. The unresolved step is what happens to decisions already made by recipients when an entry changes. Funding for capacity does not by itself give a contributor the power to approve a listing.
\n[College of Policing FOIA-2021-094](http[local research file] response dated 17 September 2021, contains an embedded Policing Fraud document labelled “Revised 070119 MJB”. Preserve that literal revision identifier; a separate adoption date and continuing application in 2026 were not established. Original retained as list-college-foi2021.pdf; PDF 37–39 / printed 25–27 is the complete Operation Creative case study. PDF 38 was visually checked.
The procedure requires PIPCU to assess investigation quality, lawful evidence and a rightsholder/nominee statement meeting a standard sufficient to support prosecution. Qualifying referrals can be distributed to advertising businesses. The Head of Unit determines tactics case by case. Website operators may challenge police activity; evidence of cessation of offences triggers immediate removal. Police first contact operators, offer a route to legitimate operation through rights associations, and review again if positive contact fails. Payment-service withdrawal is described as a request to the provider or merchant bank. It is not a payment-provider command conferred by this document.
\nThe preceding text cautions against unlawful use of compulsory powers solely for disruption. The case study does not specify prior court adjudication as an inclusion condition. That is a description of this procedure, not a comprehensive opinion about legal powers. It provides no fixed recurring review interval, response deadline, independent merits reviewer or recipient-wide correction instruction. Its cessation wording also does not spell out a separate mistaken-original-inclusion test. The general challenge opportunity should not be erased because those details are absent.
\nThe FOI letter's own 40-working-day review route concerns access to information, not IWL inclusion. Likewise, Ashiko's separately acquired DCI approval rule is not imported here. The stated Creative decision actor is the Head of Unit.
\n[Find a Tender notice 012517-2024](http[local research file] published 17 April 2024 at 14:38, is a City of London Corporation contract notice, not an executed operating manual. PDF 4, section II.2.4, describes the proposed allocation:
\n| Actor | \nSpecified role | \n
|---|---|
| Copyright holders and representatives | \nSubmit domains with supporting evidence | \n
| Police officers as administrators | \nReview and accept or reject submissions | \n
| Investigators | \nSubmit cases and see limited data | \n
| Users | \nView some data without editing it and download IWL domains | \n
Two administrator levels are described as desirable. This supports a police-controlled approval design; it does not prove permissions were implemented, that every applicant is admitted, or that the software supplier decides infringement. The notice provides no new challenge procedure or correction-propagation rule. Its High Court procurement-review section concerns the tender, not website operators' list appeals. The full technical pack expressly requires an NDA; it was not requested.
\nThe complete official-reader notice is retained in list-tender-and-award.reader.txt; original GET returned 403. The related award PDF, [027681-2024](http[local research file] failed through both the reader and ordinary GET. Its search-only supplier/value details are not adopted. A separate 2025 procurement search result is likewise only a lead, not proof of the same contract or deployed system. No failed route was retried. The tender's inconsistent annual-versus-total value wording was inspected but no budget amount is needed for this authority conclusion.
The retained [IP Crime and Enforcement Report 2018/19](http[local research file] PDF 59 / printed 55, supplies an operational account: a referral is followed by a cease-and-desist notice and attempted contact before listing. It dates the IWL's creation to 2014 and reports 1,643 websites recorded by April 2019, alongside a separate 1,995 suspended since creation. Those are different measures.
\nThe PRS for Music contribution, PDF 113 / printed 109, reports its Rights Protection Unit submitting 138 website referrals. This appears in the2018/19 report, but the Operation Creative paragraph does not specify the counting period; nearby2018 dates concern separate work. This establishes PRS's attributed exercised supply of referrals. It does not show that all 138 were accepted or identify the advertisements removed because of each referral. Its surrounding MAPS/search/app achievements are different channels and must not be counted as IWL results.
\nThe retained [2022–2027 counter-infringement strategy](http[local research file] “Internet/e-commerce”, explicitly calls industry intelligence the main referral source and describes sharing the IWL to disrupt advertising revenue. Its 5,008 takedowns since 2014 is attributed to the wider collaborative operation, which includes multiple tactics. It is not a measured causal effect of list inclusion alone.
\n[City of London Police, 6 February 2023](http[local research file] reports White Bullet research finding a 58% decrease in UK advertising on IWL websites between January and December 2022, and an estimated GBP 6 million decrease in advertising revenue from UK traffic over that period. The office presents this as operational success; White Bullet's CEO also speaks as a collaborator. The underlying study, sampling, list-composition changes and counterfactual were not acquired. Preserve this as a police-published, supplier-attributed estimate, not observed cash withheld, seized money or proof that the IWL alone caused the change. The article separately reports 750 subscribers and 682 domains suspended in 2022; subscriber access, listing and suspension remain distinct outcomes.
\nCandidate directional relations for later review are: rightsholder/nominee to PIPCU, evidential referral; PIPCU to referred domain, quality assessment; police administrator to portal submission, proposed accept/reject right; Head of Unit to operational tactics, dated decision authority; PIPCU to advertising recipients, distribution for action; website operator to PIPCU, challenge opportunity; PIPCU to list entry, conditional immediate removal; PRS to PIPCU, reported 138 referrals; White Bullet to police announcement, attributed measurement input. None is a new case addition.
\nThe strongest rival to a simple “industry dictates the list” interpretation is substantive police screening followed by recipients' separate choices. The published rules support that formal distinction; their practical quality cannot be inferred from the rules alone. Conversely, the presence of screening and a challenge promise does not establish that reviews are fast, independent or consistently reflected downstream.
\nA finite next discriminator is a current public IWL operating procedure or redacted inclusion/challenge decision with timestamps, reasons, reviewer, removal notice and recipient notification. The likely holder is PIPCU/City of London Police; relevant product evidence would be an accepted technical specification or change-event protocol. No claim is made that a particular unpublished document exists. This would distinguish effective correction from list-only deletion and establish whether the dated safeguards remain operative.
\nlist-captures.json contains original/derivative hashes, URLs, dates, acquisition outcomes and scoped reads. The College original is SHA256 245cd690a2643290ec29690933b65833b3283fb11687282f9d4083f0357ddf76. Reused 2019 report and 2022 strategy originals remain at their prior paths with hashes e0b6e4b753ece534e35a20b46cf2aeeb4b147588e99576f746f083c621bade97 and 9dcb6d69ad37942b0b4005543be05e9f4561eb854659d149e0f95510067e0654. This run freshly read the previously excluded Creative passages; it does not relabel old captures as new acquisitions.
College scope: PDF 1–3 cover/access review, PDF 13 contents/version, PDF 37–39 case study and adjacent caution; PDF 12 incidental; visual PDF 38. Report scope: complete PDF 59 and 113, incidental PDF 75; the unrelated activities on those pages were not adopted as IWL evidence. Strategy scope: original HTML Internet/e-commerce section through the Creative paragraph. Tender scope: full public 10-page reader, especially II.2.4; no technical pack. Police announcement scope: complete substantive official-reader lines 50–67 including date and closing figures; no underlying White Bullet study. Neither the 159-page College bundle nor either long report was read in full.
\nSearches were confined to the exact operation/list, challenge/review and official portal notice. No newer public review cadence or exercised correction case was located in this finite pass. That is a scoped gap, not proof no procedure or correction exists. Acquisition closed before any supplier-contract, operational-list or wider advertising branch.
", "path": "creative-advertising-wave-2026-09-09/list-authority.md", "sourceIds": [ "w30-class-s-creative-guidance", "w30-class-s-creative-report2019", "w30-class-s-creative-portal", "w30-class-s-creative-metric" ], "markdown": "9 September 2026. Bounded acquisition closed. This packet establishes a dated public procedure and a later procurement specification, with attributed delivery accounts. It does not certify the currently deployed portal or a September 2026 operating policy. No operational list, account, form, target website or gated tender was accessed.\n\nThe decisive link is an assessment by police between a rightsholder's evidence and a list distributed for advertising decisions. The strongest acquired safeguard is an express opportunity to challenge police action and immediate removal when evidence shows that the site no longer commits offences. The unresolved step is what happens to decisions already made by recipients when an entry changes. Funding for capacity does not by itself give a contributor the power to approve a listing.\n\n## The disclosed rule and its limits\n\nCollege of Policing FOIA-2021-094 (see source references) response dated **17 September 2021**, contains an embedded *Policing Fraud* document labelled **“Revised 070119 MJB”**. Preserve that literal revision identifier; a separate adoption date and continuing application in 2026 were not established. Original retained as `list-college-foi2021.pdf`; PDF 37–39 / printed 25–27 is the complete Operation Creative case study. PDF 38 was visually checked.\n\nThe procedure requires PIPCU to assess investigation quality, lawful evidence and a rightsholder/nominee statement meeting a standard sufficient to support prosecution. Qualifying referrals can be distributed to advertising businesses. The Head of Unit determines tactics case by case. Website operators may challenge police activity; evidence of cessation of offences triggers immediate removal. Police first contact operators, offer a route to legitimate operation through rights associations, and review again if positive contact fails. Payment-service withdrawal is described as a request to the provider or merchant bank. It is not a payment-provider command conferred by this document.\n\nThe preceding text cautions against unlawful use of compulsory powers solely for disruption. The case study does not specify prior court adjudication as an inclusion condition. That is a description of this procedure, not a comprehensive opinion about legal powers. It provides no fixed recurring review interval, response deadline, independent merits reviewer or recipient-wide correction instruction. Its cessation wording also does not spell out a separate mistaken-original-inclusion test. The general challenge opportunity should not be erased because those details are absent.\n\nThe FOI letter's own 40-working-day review route concerns access to information, not IWL inclusion. Likewise, Ashiko's separately acquired DCI approval rule is not imported here. The stated Creative decision actor is the Head of Unit.\n\n## Later public portal specification locates admission rights\n\nFind a Tender notice 012517-2024 (see source references) published **17 April 2024 at 14:38**, is a City of London Corporation contract notice, not an executed operating manual. PDF 4, section II.2.4, describes the proposed allocation:\n\n| Actor | Specified role |\n| --- | --- |\n| Copyright holders and representatives | Submit domains with supporting evidence |\n| Police officers as administrators | Review and accept or reject submissions |\n| Investigators | Submit cases and see limited data |\n| Users | View some data without editing it and download IWL domains |\n\nTwo administrator levels are described as desirable. This supports a police-controlled approval design; it does not prove permissions were implemented, that every applicant is admitted, or that the software supplier decides infringement. The notice provides no new challenge procedure or correction-propagation rule. Its High Court procurement-review section concerns the tender, not website operators' list appeals. The full technical pack expressly requires an NDA; it was not requested.\n\nThe complete official-reader notice is retained in `list-tender-and-award.reader.txt`; original GET returned 403. The related award PDF, 027681-2024 (see source references) failed through both the reader and ordinary GET. Its search-only supplier/value details are not adopted. A separate 2025 procurement search result is likewise only a lead, not proof of the same contract or deployed system. No failed route was retried. The tender's inconsistent annual-versus-total value wording was inspected but no budget amount is needed for this authority conclusion.\n\n## Attributed referrals and effects\n\nThe retained IP Crime and Enforcement Report 2018/19 (see source references) PDF 59 / printed 55, supplies an operational account: a referral is followed by a cease-and-desist notice and attempted contact before listing. It dates the IWL's creation to 2014 and reports 1,643 websites recorded by April 2019, alongside a separate 1,995 suspended since creation. Those are different measures.\n\nThe **PRS for Music contribution**, PDF 113 / printed 109, reports its Rights Protection Unit submitting **138 website referrals**. This appears in the2018/19 report, but the Operation Creative paragraph does not specify the counting period; nearby2018 dates concern separate work. This establishes PRS's attributed exercised supply of referrals. It does not show that all 138 were accepted or identify the advertisements removed because of each referral. Its surrounding MAPS/search/app achievements are different channels and must not be counted as IWL results.\n\nThe retained 2022–2027 counter-infringement strategy (see source references) “Internet/e-commerce”, explicitly calls industry intelligence the main referral source and describes sharing the IWL to disrupt advertising revenue. Its **5,008 takedowns since 2014** is attributed to the wider collaborative operation, which includes multiple tactics. It is not a measured causal effect of list inclusion alone.\n\nCity of London Police, 6 February 2023 (see source references) reports White Bullet research finding a **58% decrease in UK advertising on IWL websites between January and December 2022**, and an estimated **GBP 6 million decrease in advertising revenue from UK traffic** over that period. The office presents this as operational success; White Bullet's CEO also speaks as a collaborator. The underlying study, sampling, list-composition changes and counterfactual were not acquired. Preserve this as a police-published, supplier-attributed estimate, not observed cash withheld, seized money or proof that the IWL alone caused the change. The article separately reports 750 subscribers and 682 domains suspended in 2022; subscriber access, listing and suspension remain distinct outcomes.\n\n## What the sequence establishes\n\nCandidate directional relations for later review are: rightsholder/nominee to PIPCU, evidential referral; PIPCU to referred domain, quality assessment; police administrator to portal submission, proposed accept/reject right; Head of Unit to operational tactics, dated decision authority; PIPCU to advertising recipients, distribution for action; website operator to PIPCU, challenge opportunity; PIPCU to list entry, conditional immediate removal; PRS to PIPCU, reported 138 referrals; White Bullet to police announcement, attributed measurement input. None is a new case addition.\n\nThe strongest rival to a simple “industry dictates the list” interpretation is substantive police screening followed by recipients' separate choices. The published rules support that formal distinction; their practical quality cannot be inferred from the rules alone. Conversely, the presence of screening and a challenge promise does not establish that reviews are fast, independent or consistently reflected downstream.\n\nA finite next discriminator is a current public IWL operating procedure or redacted inclusion/challenge decision with timestamps, reasons, reviewer, removal notice and recipient notification. The likely holder is PIPCU/City of London Police; relevant product evidence would be an accepted technical specification or change-event protocol. No claim is made that a particular unpublished document exists. This would distinguish effective correction from list-only deletion and establish whether the dated safeguards remain operative.\n\n## Custody and exact scope\n\n`list-captures.json` contains original/derivative hashes, URLs, dates, acquisition outcomes and scoped reads. The College original is SHA256 `245cd690a2643290ec29690933b65833b3283fb11687282f9d4083f0357ddf76`. Reused 2019 report and 2022 strategy originals remain at their prior paths with hashes `e0b6e4b753ece534e35a20b46cf2aeeb4b147588e99576f746f083c621bade97` and `9dcb6d69ad37942b0b4005543be05e9f4561eb854659d149e0f95510067e0654`. This run freshly read the previously excluded Creative passages; it does not relabel old captures as new acquisitions.\n\nCollege scope: PDF 1–3 cover/access review, PDF 13 contents/version, PDF 37–39 case study and adjacent caution; PDF 12 incidental; visual PDF 38. Report scope: complete PDF 59 and 113, incidental PDF 75; the unrelated activities on those pages were not adopted as IWL evidence. Strategy scope: original HTML Internet/e-commerce section through the Creative paragraph. Tender scope: full public 10-page reader, especially II.2.4; no technical pack. Police announcement scope: complete substantive official-reader lines 50–67 including date and closing figures; no underlying White Bullet study. Neither the 159-page College bundle nor either long report was read in full.\n\nSearches were confined to the exact operation/list, challenge/review and official portal notice. No newer public review cadence or exercised correction case was located in this finite pass. That is a scoped gap, not proof no procedure or correction exists. Acquisition closed before any supplier-contract, operational-list or wider advertising branch.\n", "file": "research/creative-advertising-wave-2026-09-09-list-authority.md" }, { "id": "creative-advertising-wave-2026-09-09-recipient-use", "title": "Advertising recipients: recommended contracts, licence duties and commercial list delivery", "description": "9September2026. Root acquisition CLOSED at the recipient marketplace/rule breakpoint. This packet connects documented sources of pressure and offered integrations without treating every list as a court finding or every available control as executed action.", "html": "9September2026. Root acquisition CLOSED at the recipient marketplace/rule breakpoint. This packet connects documented sources of pressure and offered integrations without treating every list as a court finding or every available control as executed action.
\nThe undated live [IAB IWL guide](http[local research file] says PIPCU established the list with IAB UK, IPA, ISBA and rightsholders, and calls it police-verified. It describes an API, encourages advertisers/agencies/intermediaries to use the list in buying/selling/trading, and explicitly recommends incorporating its data into contracts and processes. Access requires requesting PIPCU authorization and authenticated users. No access was requested. This is industry guidance promoting contract adoption; no executed advertiser contract is supplied, and the word new beside API does not date a2026 launch. Police review and exact rights remain Carver's distinct source lane.
\nThe current official [LCCP16.1.1](http[local research file] heading through its complete three-part condition, applies to all licences. It requires operators to avoid placing digital ads on sites providing unauthorized access to copyrighted content; take all reasonable steps regarding relevant contracted third parties; and ensure contracts enable prompt termination when the licensee reasonably considers that a third party placed such ads. The termination capacity is subject to applicable dispute-resolution provisions. This is a requirement to hold a contractual right, not an unconditional order to exercise it every time a list changes. The condition itself neither names IWL nor makes inability to verify customer identity the trigger.
\nThe regulator's [13June2017 announcement](http[local research file] expressly links OperationCreative to the licence condition introduced the preceding October and encourages operators to obtain IWL access. It reports an87% fall in gambling ads on infringing sites over the preceding12months. The historical start is the regulator's account; current wording was separately read. The percentage is an attributed aggregate outcome, without acquired sample/method/counterfactual or operator-level records. No revenue total follows from an ad-count percentage. This establishes a sector-specific route from regulatory responsibility to private contracts, with the list promoted as an aid, rather than a universal legal force attaching to list membership.
\n[Thinkbox's January2018 parliamentary supplementary evidence ADV0038](http[local research file] read in full, reproduces accounts from anonymous AgenciesA/B/C. A describes combining PIPCU, private research, client fraud teams and internal vetting; B describes contractual approved-placement requirements, manual vetting, real-time URL/app matching and an always-on exclusion list with PIPCU among its sources. These are agency statements mediated by Thinkbox, not named client contracts or independently observed deployments. Do not identify the anonymous firms from the vendors they mention.
\nThe distinction between ad blocking and revenue denial is material. Thinkbox's own paragraphs14–15 question effectiveness and say some verification happens after purchase, allowing payment despite a blank replacement ad. This is an attributed limitation, not independent transaction evidence. It nevertheless defeats an inference that every blocked creative necessarily stops every corresponding payment. AgencyA also reports investigating inconsistencies and updating lists; this is not a worked publisher challenge or recipient-wide correction trace.
\nWhite Bullet's [25June2018 announcement](http[local research file] describes signing the EU advertising/IP memorandum and offers two different data products. Its GlobalPiracyIndex is said to combine court-order sites and infringement-list initiatives worldwide. Its separate IPIIndex is advertised as assessing sites beyond court/enforcement adjudication, using machine learning and specialist criteria. These are vendor claims, not an audited database, proven PIPCU-specific feed or a newly acquired EU agreement. No operational list was opened and marketing accuracy/completeness claims are not adopted.
\nThe vendor's [5May2021 WIPO announcement](http[local research file] calls WhiteBullet a WIPOALERT member and describes official national-list information alongside its own tracked content/risk assessments. This supports an attributed membership/combined-offering account, not a dated executed WIPO licence or item-by-item ingestion proof. Linnaeus separately supplies WIPO's actual public rules; any differences in scope prevail over a broad promotional description.
\nIts 26May2021 Peer39 announcement.
\nThe next proof would be a dated recipient agreement or actual configuration identifying which source lists are selected, the trigger for blocking/cessation, pre-bid versus post-purchase placement, and a correction or reconciliation record. A list-entry reversal alone cannot prove that an advertiser resumed spending. A named vendor or offered integration is a real commercial connection but is weaker than an executed case.
\nDiscovery also returned a [D3 procurement record](http[local research file] identifying FTS037229-2026 / ocds-h6vhtk-0689ed, a reported WhiteBullet advertising-monitoring award ofGBP74520. That amount, award and scope remain a THIRD-PARTY LEAD. The exact official notice reader failed and one ordinary GET returned403; neither original nor primary readable notice was acquired. No signed contract, payment or funding-to-metric claim follows. The Stotles lead route also failed and is closed. This is a named record target for a future new credible source, not permission to keep trying the same failures. No procurement award from this lead is promoted into the accepted graph.
\nTwo original HTMLs are retained: recipient-iab.html SHA25635188162720124f1a7b7113f2ead35ce9633023d72f86c2a2637b67647b778bc, and recipient-peer39-marketplace.html SHA256a5ed833b584f5744cc4fe7aca6898bd5b1c15274f65b504a96b6b38136b32660. Root fully read IAB's four substantive sections from its reader and then a scoped original extract; initial whole-page output truncated and was replaced by that complete section read. Root read Peer39's substantive page through its reader and verified original heading/marketplace description/WhiteBullet card. No images or forms were used.
\nGC condition/news, Thinkbox and the2018WhiteBullet ordinary GETs returned403; error files are not originals. Their successful official/vendor readers supplied the scoped text. The two2021WhiteBullet articles were fully read through their vendor reader, excluding related-post snippets/marketing footer as new research; no repeated GET to the already failing host. Reader-based summaries and locators are in recipient-reader-scopes.json; these are derivatives, not facsimiles. The notice failed both primary routes. Exact status/hash facts are in recipient-captures-initial.json and recipient-peer39-capture.json. Search discovery included unsuccessful named-vendor/contract queries and noisy unrelated acronym results; those were excluded, not evidence of absence. No operational list, account, advertising endpoint, contact, purchase, public release, case edit, ZIP or visible app was used. Acquisition closed for independent review and combined forest.
", "path": "creative-advertising-wave-2026-09-09/recipient-use.md", "sourceIds": [ "w30-class-s-iab-guide", "w30-class-s-gambling-condition", "w30-class-s-gambling2017", "w30-class-s-thinkbox2018", "w30-class-s-whitebullet2018", "w30-class-s-whitebullet-wipo", "w30-class-s-whitebullet-peer", "w30-class-s-peer39" ], "markdown": "9September2026. Root acquisition CLOSED at the recipient marketplace/rule breakpoint. This packet connects documented sources of pressure and offered integrations without treating every list as a court finding or every available control as executed action.\n\n## IAB UK: a recommended contract and workflow route\n\nThe undated live IAB IWL guide (see source references) says PIPCU established the list with IAB UK, IPA, ISBA and rightsholders, and calls it police-verified. It describes an API, encourages advertisers/agencies/intermediaries to use the list in buying/selling/trading, and explicitly recommends incorporating its data into contracts and processes. Access requires requesting PIPCU authorization and authenticated users. No access was requested. This is industry guidance promoting contract adoption; no executed advertiser contract is supplied, and the word new beside API does not date a2026 launch. Police review and exact rights remain Carver's distinct source lane.\n\n## Gambling: a licence obligation with a contract requirement\n\nThe current official LCCP16.1.1 (see source references) heading through its complete three-part condition, applies to all licences. It requires operators to avoid placing digital ads on sites providing unauthorized access to copyrighted content; take all reasonable steps regarding relevant contracted third parties; and ensure contracts enable prompt termination when the licensee reasonably considers that a third party placed such ads. The termination capacity is subject to applicable dispute-resolution provisions. This is a requirement to hold a contractual right, not an unconditional order to exercise it every time a list changes. The condition itself neither names IWL nor makes inability to verify customer identity the trigger.\n\nThe regulator's 13June2017 announcement (see source references) expressly links OperationCreative to the licence condition introduced the preceding October and encourages operators to obtain IWL access. It reports an87% fall in gambling ads on infringing sites over the preceding12months. The historical start is the regulator's account; current wording was separately read. The percentage is an attributed aggregate outcome, without acquired sample/method/counterfactual or operator-level records. No revenue total follows from an ad-count percentage. This establishes a sector-specific route from regulatory responsibility to private contracts, with the list promoted as an aid, rather than a universal legal force attaching to list membership.\n\n## Agency practice: multiple list sources and a payment limitation\n\nThinkbox's January2018 parliamentary supplementary evidence ADV0038 (see source references) read in full, reproduces accounts from anonymous AgenciesA/B/C. A describes combining PIPCU, private research, client fraud teams and internal vetting; B describes contractual approved-placement requirements, manual vetting, real-time URL/app matching and an always-on exclusion list with PIPCU among its sources. These are agency statements mediated by Thinkbox, not named client contracts or independently observed deployments. Do not identify the anonymous firms from the vendors they mention.\n\nThe distinction between ad blocking and revenue denial is material. Thinkbox's own paragraphs14–15 question effectiveness and say some verification happens after purchase, allowing payment despite a blank replacement ad. This is an attributed limitation, not independent transaction evidence. It nevertheless defeats an inference that every blocked creative necessarily stops every corresponding payment. AgencyA also reports investigating inconsistencies and updating lists; this is not a worked publisher challenge or recipient-wide correction trace.\n\n## White Bullet and Peer39: official data alongside proprietary assessment\n\nWhite Bullet's 25June2018 announcement (see source references) describes signing the EU advertising/IP memorandum and offers two different data products. Its GlobalPiracyIndex is said to combine court-order sites and infringement-list initiatives worldwide. Its separate IPIIndex is advertised as assessing sites beyond court/enforcement adjudication, using machine learning and specialist criteria. These are vendor claims, not an audited database, proven PIPCU-specific feed or a newly acquired EU agreement. No operational list was opened and marketing accuracy/completeness claims are not adopted.\n\nThe vendor's 5May2021 WIPO announcement (see source references) calls WhiteBullet a WIPOALERT member and describes official national-list information alongside its own tracked content/risk assessments. This supports an attributed membership/combined-offering account, not a dated executed WIPO licence or item-by-item ingestion proof. Linnaeus separately supplies WIPO's actual public rules; any differences in scope prevail over a broad promotional description.\n\nIts 26May2021 Peer39 announcement.\n\n## Missing discriminating evidence and a closed procurement lead\n\nThe next proof would be a dated recipient agreement or actual configuration identifying which source lists are selected, the trigger for blocking/cessation, pre-bid versus post-purchase placement, and a correction or reconciliation record. A list-entry reversal alone cannot prove that an advertiser resumed spending. A named vendor or offered integration is a real commercial connection but is weaker than an executed case.\n\nDiscovery also returned a D3 procurement record (see source references) identifying FTS037229-2026 / ocds-h6vhtk-0689ed, a reported WhiteBullet advertising-monitoring award ofGBP74520. That amount, award and scope remain a THIRD-PARTY LEAD. The exact official notice reader failed and one ordinary GET returned403; neither original nor primary readable notice was acquired. No signed contract, payment or funding-to-metric claim follows. The Stotles lead route also failed and is closed. This is a named record target for a future new credible source, not permission to keep trying the same failures. No procurement award from this lead is promoted into the accepted graph.\n\n## Custody and actual scope\n\nTwo original HTMLs are retained: recipient-iab.html SHA25635188162720124f1a7b7113f2ead35ce9633023d72f86c2a2637b67647b778bc, and recipient-peer39-marketplace.html SHA256a5ed833b584f5744cc4fe7aca6898bd5b1c15274f65b504a96b6b38136b32660. Root fully read IAB's four substantive sections from its reader and then a scoped original extract; initial whole-page output truncated and was replaced by that complete section read. Root read Peer39's substantive page through its reader and verified original heading/marketplace description/WhiteBullet card. No images or forms were used.\n\nGC condition/news, Thinkbox and the2018WhiteBullet ordinary GETs returned403; error files are not originals. Their successful official/vendor readers supplied the scoped text. The two2021WhiteBullet articles were fully read through their vendor reader, excluding related-post snippets/marketing footer as new research; no repeated GET to the already failing host. Reader-based summaries and locators are in recipient-reader-scopes.json; these are derivatives, not facsimiles. The notice failed both primary routes. Exact status/hash facts are in recipient-captures-initial.json and recipient-peer39-capture.json. Search discovery included unsuccessful named-vendor/contract queries and noisy unrelated acronym results; those were excluded, not evidence of absence. No operational list, account, advertising endpoint, contact, purchase, public release, case edit, ZIP or visible app was used. Acquisition closed for independent review and combined forest.\n", "file": "research/creative-advertising-wave-2026-09-09-recipient-use.md" }, { "id": "creative-advertising-wave-2026-09-09-wipo-distribution", "title": "WIPO ALERT: contribution, distribution and challenge authority", "description": "9 September2026. Bounded acquisition CLOSED. Five official originals retained; no accounts, applications, list searches, list entries, API calls, operational actions or outreach. No case/atlas/ZIP/current-work/forest edits. This packet establishes a distribution mechanism and its published limits; an actual PIPCU IWL feed into WIPO was not established.", "html": "9 September2026. Bounded acquisition CLOSED. Five official originals retained; no accounts, applications, list searches, list entries, API calls, operational actions or outreach. No case/atlas/ZIP/current-work/forest edits. This packet establishes a distribution mechanism and its published limits; an actual PIPCU IWL feed into WIPO was not established.
\nThe [current public landing page](http[local research file] says authorized bodies in member states supply lists of websites/apps assessed under national rules. WIPO distributes the information, does not itself determine infringement, and directs listing concerns to the national agency. Advertisers, agencies and technical providers can apply to use aggregated lists in automated advertising systems. This is WIPO's undated public description as observed today, not a finding that every listing rests on a final judicial judgment.
\nThe [public user FAQ](http[local research file] gives actual gatekeeping choices. WIPO may examine a genuine advertising business before approving access. Users can select particular contributors/countries rather than every list; contributors can refuse particular users, although access to all lists is the stated default. Manual download and API access are described. The user agreement is summarized as confining use to preventing misplaced advertising, with best-endeavours confidentiality. Access is currently described as free; this is not evidence of a subsidy amount or paid contract. Contributors remain responsible for accuracy. A described facility notifies contributors when another list deletes a site present on multiple lists. That does not establish automatic deletion from every list or restoration of advertising. No executed agreement, current versioned contractual schedule, recipient refresh deadline or tested runtime was acquired.
\n[WIPO/ACE/14/9](http[local research file] is prepared by the Secretariat. The current landing links it as its explanation of the original BRIP Database project. Paragraphs9–14 describe contributor Letters of Understanding and user agreements, contributor-controlled lists, advertiser eligibility screening, and retained national review/appeal rights. The inclusion definition is a contributor's reasonable suspicion of deliberate infringement or facilitation, including outside the contributor's territory; WIPO presents itself as a distributor. Paragraph12 describes prompt contributor updates when a national entry is removed. Paragraph20 says advertising users cannot add/remove entries. Paragraph22's automated ad-suppression sequence is the preferred implementation, not proof any named recipient deployed it. Paragraphs25–27 describe the2019 startup/test-data stage, not current adoption counts.
\nThis is the important distinction: a national classification can acquire cross-border commercial reach through a shared feed without becoming a WIPO adjudication. Whether a particular recipient must or chooses to suppress an ad still needs its own applicable instrument. The platform is described as voluntary cooperation; that does not establish that every recipient faces no separate legal or contractual duty.
\n[WIPO's contributor-supplied operating-procedures page](http[local research file] undated and observed9September2026, provides two useful contrasts. Brazil's table describes rightsholder evidence/notice with a five-day comment opportunity, CCP assessment, SFI approval and CNCP upload. An operator may seek review for an incorrect listing or changed conduct; the stated analysis period is60days. Removal depends on the specified authorities' review. The underlying executed cooperation agreement and proof of these safeguards in a case were not acquired. Its separate reference to domestic circulation to payment intermediaries must not be treated as WIPO advertising-user permission to freeze bank accounts.
\nJapan's table identifies CODA, an association formed at the instance of Japanese ministries, using member-company/association data and updates approximately every two months. Its own list is also called the Infringing Website List (IWL). The table supplies criteria but no detailed operator-appeal mechanism. CODA's IWL is not PIPCU's IWL. The presence of this association also prevents simplification of every contributor into a government ministry. Neither table's dated/undated operating summary proves current execution of every step.
\nThe2019 framework paragraph3 describes PIPCU's IWL as a national precedent: rightsholder evidence verified by police, then shared via a portal. The FAQ likewise cites a2017 OperationCreative effectiveness claim as rationale. Neither passage says PIPCU signed a contributor agreement, uploaded its list or supplied an ongoing WIPO feed. The currently posted procedures page has eight country headings and none for theUK, but it is not established as an exhaustive or current roster. Its omission does not prove non-participation.
\nBefore acquisition, reused retained2019 IPO report scoped text (../pipcu-resources-wave-2026-09-09/referral-delivery2019.scoped-derived.txt, OperationCreative section) and the2022 strategy's OperationCreative paragraph in referral-strategy2022.html. They describe IWL distribution to advertisers/brands and industry-intelligence input. They do not identify WIPO as the recipient. This was a narrow reread, not a new audit of those reports or their cumulative outcomes. Root's separate recipient lane and Carver's police lane own named uses and PIPCU process details.
\nThe defensible relation is PIPCU/IWL -> cited precedent for WIPO's coordination project. An operational IWL -> WIPO feed remains OPEN. A signed contributor LoU, official dated admission/upload account or a named recipient's explicit feed provenance would change that result. Co-appearance in a blocklist discussion, a shared acronym, or access to both systems does not.
\nCandidate typed relations, not case additions: WIPO -> advertising users [access approval]; national contributors -> national lists [classification and maintenance]; contributor -> selected user [data-access permission/refusal]; user -> selected national lists [selection]; WIPO -> aggregated distribution service [resource provision]; user -> advertising-placement system [permitted purpose, implemented use requires separate evidence]; affected operator -> national contributor [review route]; national removal -> contributor WIPO update [dated prescribed process]; cross-list removal -> other contributors [published notification capability]. Keep notification, actual propagation and downstream restoration as distinct states.
\nThe most consequential missing public instruments are a current versioned user agreement/LoU, a cross-border correction/removal propagation requirement, and an exercised correction showing what recipients did after an entry changed. No WIPO-wide challenge tribunal or compensation obligation was found in the finite inspected sources; national remedies remain source-specific. This does not claim that no additional contractual instruments exist.
\nExact discovery URL: [WIPO Webcast, Advisory Committee on Enforcement,4June2026 afternoon](http[local research file] Search result title was WIPO Webcast; the URL identifies the session/date. A displayed search extract attributes to a WIPO speaker an update describing WIPO ALERT Pay, a reported pilot with rightsholders/payment providers and movement into development. The full webcast/transcript and underlying report were NOT opened/read, the video was not watched, no original was captured, names/figures/stages were not verified, and no fact or case edge is promoted from it. It is a candidate for the next forest because payment distribution might differ materially from the ad-user agreement. It must not be folded into the present WIPO ALERT rules.
\nAllfivecapturedURLs returned200 on one ordinary publicGET each. Manifest wipo-captures.json records exactURLs, acquisitiontimes, bytecounts, hashes and derivative ancestry. No source-access failure needed recovery in this lane.
\nwipo-index.html — SHA256 39f87737f0633818b2d726aeb24d24081f3883e0d394afad88b37870f1b930dc. Complete substantive public landing text from title through roles/warning/procedures link read via official reader; original DOM inspected for project source href. No database search/API/form/video used. Undated page observed9September2026; latest visible news dates are not publication dates for the page.
wipo-faq.html — SHA256 ba3e517615565ef9f865343b45c10698b00bbc17ae24c1481568392bf1bef5f9. Complete substantive FAQ read, including eligibility, listselection/contributoraccess, useagreement/confidentiality/freeaccess, accuracydisclaimer and crosslistdeletionnotice. Undated public account; not an executed useragreement.
wipo-procedures.html — SHA256 ab8bb5196d50f0d5664438cd7059020565e056f77e16c383e8d831966f7b7e3f. Country heading inventory inspected; complete Brazil and Japan tables read directly from original. Other country tables not systematically reviewed. These are contributor-supplied summaries hostedbyWIPO, not underlying executed agreements or complete currentnationallegalaudit. Undated; historical officeholder details not promoted ascurrent.
wipo-project-index.html — SHA256 a07fe913e2a571051127031b3a6e2b4ed17cbf5f9d56d8f5c7d27ac06ac13d6f. Documentcode/title/publicationdate/preparedbySecretariat/Englishlink verified; navigation not substantive. Publication18June2019.
wipo-project2019.pdf — SHA256 103b26be0b382778579688a170db1078f24d462301e50dfef18b56168827f535. Entire8-page substantive text read from officialreader and directoriginalPDF extraction, including background/PIPCUprecedent, paras6-27, LoUs/UserAgreements, sitesofconcern, appeals/updates, country/list controls and testdata/currentstatus. Embedded UI screenshots/figures not independently visually inspected. Source date18June2019; documented design/reportatthatdate, not evidenceofcurrentruntimeexecution.
Discovery was limited to two initial officialWIPOqueries for rules/agreements, followed by three targeted questions for UK/contributor/appeal joins. Their results led to the current pages and the projectdocument linked byWIPO itself. The 2025EUIPO discussionpaper, Ukraine legislative accounts, WIPO2026report/webcast, secondaryblog and academic results were discoveryonly and did not become findings. No wholesale globalIP survey or underlying national statute/appeal-case acquisition was attempted. Fulltext extraction is not proof of visual inspection or runtime execution. Acquisition closed pending combinedforest.
", "path": "creative-advertising-wave-2026-09-09/wipo-distribution.md", "sourceIds": [ "w30-class-s-wipo-index", "w30-class-s-wipo-faq", "w30-class-s-wipo-procedures", "w30-class-s-wipo2019" ], "markdown": "9 September2026. Bounded acquisition CLOSED. Five official originals retained; no accounts, applications, list searches, list entries, API calls, operational actions or outreach. No case/atlas/ZIP/current-work/forest edits. This packet establishes a distribution mechanism and its published limits; an actual PIPCU IWL feed into WIPO was not established.\n\n## The distribution boundary\n\nThe current public landing page (see source references) says authorized bodies in member states supply lists of websites/apps assessed under national rules. WIPO distributes the information, does not itself determine infringement, and directs listing concerns to the national agency. Advertisers, agencies and technical providers can apply to use aggregated lists in automated advertising systems. This is WIPO's undated public description as observed today, not a finding that every listing rests on a final judicial judgment.\n\nThe public user FAQ (see source references) gives actual gatekeeping choices. WIPO may examine a genuine advertising business before approving access. Users can select particular contributors/countries rather than every list; contributors can refuse particular users, although access to all lists is the stated default. Manual download and API access are described. The user agreement is summarized as confining use to preventing misplaced advertising, with best-endeavours confidentiality. Access is currently described as free; this is not evidence of a subsidy amount or paid contract. Contributors remain responsible for accuracy. A described facility notifies contributors when another list deletes a site present on multiple lists. That does not establish automatic deletion from every list or restoration of advertising. No executed agreement, current versioned contractual schedule, recipient refresh deadline or tested runtime was acquired.\n\n## The dated framework is explicit about suspicion and retained national authority\n\nWIPO/ACE/14/9 (see source references) is prepared by the Secretariat. The current landing links it as its explanation of the original BRIP Database project. Paragraphs9–14 describe contributor Letters of Understanding and user agreements, contributor-controlled lists, advertiser eligibility screening, and retained national review/appeal rights. The inclusion definition is a contributor's reasonable suspicion of deliberate infringement or facilitation, including outside the contributor's territory; WIPO presents itself as a distributor. Paragraph12 describes prompt contributor updates when a national entry is removed. Paragraph20 says advertising users cannot add/remove entries. Paragraph22's automated ad-suppression sequence is the preferred implementation, not proof any named recipient deployed it. Paragraphs25–27 describe the2019 startup/test-data stage, not current adoption counts.\n\nThis is the important distinction: a national classification can acquire cross-border commercial reach through a shared feed without becoming a WIPO adjudication. Whether a particular recipient must or chooses to suppress an ad still needs its own applicable instrument. The platform is described as voluntary cooperation; that does not establish that every recipient faces no separate legal or contractual duty.\n\n## National procedures differ; the acronym is not an identity\n\nWIPO's contributor-supplied operating-procedures page (see source references) undated and observed9September2026, provides two useful contrasts. Brazil's table describes rightsholder evidence/notice with a five-day comment opportunity, CCP assessment, SFI approval and CNCP upload. An operator may seek review for an incorrect listing or changed conduct; the stated analysis period is60days. Removal depends on the specified authorities' review. The underlying executed cooperation agreement and proof of these safeguards in a case were not acquired. Its separate reference to domestic circulation to payment intermediaries must not be treated as WIPO advertising-user permission to freeze bank accounts.\n\nJapan's table identifies CODA, an association formed at the instance of Japanese ministries, using member-company/association data and updates approximately every two months. Its own list is also called the Infringing Website List (IWL). The table supplies criteria but no detailed operator-appeal mechanism. CODA's IWL is not PIPCU's IWL. The presence of this association also prevents simplification of every contributor into a government ministry. Neither table's dated/undated operating summary proves current execution of every step.\n\n## What the UK connection actually establishes\n\nThe2019 framework paragraph3 describes PIPCU's IWL as a national precedent: rightsholder evidence verified by police, then shared via a portal. The FAQ likewise cites a2017 OperationCreative effectiveness claim as rationale. Neither passage says PIPCU signed a contributor agreement, uploaded its list or supplied an ongoing WIPO feed. The currently posted procedures page has eight country headings and none for theUK, but it is not established as an exhaustive or current roster. Its omission does not prove non-participation.\n\nBefore acquisition, reused retained2019 IPO report scoped text (../pipcu-resources-wave-2026-09-09/referral-delivery2019.scoped-derived.txt, OperationCreative section) and the2022 strategy's OperationCreative paragraph in referral-strategy2022.html. They describe IWL distribution to advertisers/brands and industry-intelligence input. They do not identify WIPO as the recipient. This was a narrow reread, not a new audit of those reports or their cumulative outcomes. Root's separate recipient lane and Carver's police lane own named uses and PIPCU process details.\n\nThe defensible relation is PIPCU/IWL -> cited precedent for WIPO's coordination project. An operational IWL -> WIPO feed remains OPEN. A signed contributor LoU, official dated admission/upload account or a named recipient's explicit feed provenance would change that result. Co-appearance in a blocklist discussion, a shared acronym, or access to both systems does not.\n\n## Candidate relationships and unresolved records\n\nCandidate typed relations, not case additions: WIPO -> advertising users [access approval]; national contributors -> national lists [classification and maintenance]; contributor -> selected user [data-access permission/refusal]; user -> selected national lists [selection]; WIPO -> aggregated distribution service [resource provision]; user -> advertising-placement system [permitted purpose, implemented use requires separate evidence]; affected operator -> national contributor [review route]; national removal -> contributor WIPO update [dated prescribed process]; cross-list removal -> other contributors [published notification capability]. Keep notification, actual propagation and downstream restoration as distinct states.\n\nThe most consequential missing public instruments are a current versioned user agreement/LoU, a cross-border correction/removal propagation requirement, and an exercised correction showing what recipients did after an entry changed. No WIPO-wide challenge tribunal or compensation obligation was found in the finite inspected sources; national remedies remain source-specific. This does not claim that no additional contractual instruments exist.\n\n### Parked discovery only: WIPO ALERT Pay\n\nExact discovery URL: WIPO Webcast, Advisory Committee on Enforcement,4June2026 afternoon (see source references) Search result title was WIPO Webcast; the URL identifies the session/date. A displayed search extract attributes to a WIPO speaker an update describing WIPO ALERT Pay, a reported pilot with rightsholders/payment providers and movement into development. The full webcast/transcript and underlying report were NOT opened/read, the video was not watched, no original was captured, names/figures/stages were not verified, and no fact or case edge is promoted from it. It is a candidate for the next forest because payment distribution might differ materially from the ad-user agreement. It must not be folded into the present WIPO ALERT rules.\n\n## Original custody and actual read scope\n\nAllfivecapturedURLs returned200 on one ordinary publicGET each. Manifest wipo-captures.json records exactURLs, acquisitiontimes, bytecounts, hashes and derivative ancestry. No source-access failure needed recovery in this lane.\n\n**wipo-index.html** — SHA256 `39f87737f0633818b2d726aeb24d24081f3883e0d394afad88b37870f1b930dc`. Complete substantive public landing text from title through roles/warning/procedures link read via official reader; original DOM inspected for project source href. No database search/API/form/video used. Undated page observed9September2026; latest visible news dates are not publication dates for the page.\n\n**wipo-faq.html** — SHA256 `ba3e517615565ef9f865343b45c10698b00bbc17ae24c1481568392bf1bef5f9`. Complete substantive FAQ read, including eligibility, listselection/contributoraccess, useagreement/confidentiality/freeaccess, accuracydisclaimer and crosslistdeletionnotice. Undated public account; not an executed useragreement.\n\n**wipo-procedures.html** — SHA256 `ab8bb5196d50f0d5664438cd7059020565e056f77e16c383e8d831966f7b7e3f`. Country heading inventory inspected; complete Brazil and Japan tables read directly from original. Other country tables not systematically reviewed. These are contributor-supplied summaries hostedbyWIPO, not underlying executed agreements or complete currentnationallegalaudit. Undated; historical officeholder details not promoted ascurrent.\n\n**wipo-project-index.html** — SHA256 `a07fe913e2a571051127031b3a6e2b4ed17cbf5f9d56d8f5c7d27ac06ac13d6f`. Documentcode/title/publicationdate/preparedbySecretariat/Englishlink verified; navigation not substantive. Publication18June2019.\n\n**wipo-project2019.pdf** — SHA256 `103b26be0b382778579688a170db1078f24d462301e50dfef18b56168827f535`. Entire8-page substantive text read from officialreader and directoriginalPDF extraction, including background/PIPCUprecedent, paras6-27, LoUs/UserAgreements, sitesofconcern, appeals/updates, country/list controls and testdata/currentstatus. Embedded UI screenshots/figures not independently visually inspected. Source date18June2019; documented design/reportatthatdate, not evidenceofcurrentruntimeexecution.\n\nDiscovery was limited to two initial officialWIPOqueries for rules/agreements, followed by three targeted questions for UK/contributor/appeal joins. Their results led to the current pages and the projectdocument linked byWIPO itself. The 2025EUIPO discussionpaper, Ukraine legislative accounts, WIPO2026report/webcast, secondaryblog and academic results were discoveryonly and did not become findings. No wholesale globalIP survey or underlying national statute/appeal-case acquisition was attempted. Fulltext extraction is not proof of visual inspection or runtime execution. Acquisition closed pending combinedforest.\n", "file": "research/creative-advertising-wave-2026-09-09-wipo-distribution.md" }, { "id": "dff-allocation-governance-wave-2026-09-09-allocation-governance", "title": "DFF allocation governance: a real selection role, incomplete legal delegation", "description": "Acquisition closed on 9 September 2026. Three directly linked DFF records establish a reported transfer of grant-selection decisions and subsequent board adoption. They do not supply the operative delegation, appointment decision or executed grant instrument. The previously accepted August 2026 announcement is reused context, not new independent confirmation", "html": "Acquisition closed on 9 September 2026. Three directly linked DFF records establish a reported transfer of grant-selection decisions and subsequent board adoption. They do not supply the operative delegation, appointment decision or executed grant instrument. The previously accepted August 2026 announcement is reused context, not new independent confirmation.
\nThe [8 October 2024 pilot announcement](http[local research file] by Thomas Vink, invited expressions of interest between 8 and 27 October. The initial process covered only pre-litigation grants; litigation-track funding was to continue under the existing process. The Peer Group would review applications and collectively decide which ones to recommend for support, replacing the prior expert-panel stage and reducing the grants team's decision role. Expansion into strategy, criteria and future member selection was an intention, not an accomplished transfer. Read: official-reader lines 9–12, 36–44; full substantive article read. Original: allocation-pilot2024.html.
The [25 March 2026 retrospective](http[local research file] reports that the first 2025 round involved ten peers choosing three pre-litigation projects. The second involved 17 peers selecting 13 projects across both grant types. It states that DFF's board formally adopted this decision-making process in March 2026, ending the pilot. Strategy, criteria and priority-setting remained subjects for further change. These are DFF's published accounts, not inspected board minutes. Read: lines 9–19 and 65–68. Original: allocation-retrospective2026.html.
The same retrospective describes staff facilitation and provision of previous-grantee information, other assessments and budgets. Participants report weighing legal avenues, prospects, budget justification, impact and protection of affected people, with competing priorities discussed collectively. These are descriptions of practice, not a published binding scoring rubric. No source inspected here identifies who sets the available budget or gives the group power to expand it. Read: lines 24, 28–38 and 43–49; full substantive discussion read.
\n| Stage | \nEvidence and boundary | \n
|---|---|
| Member recruitment and eligibility | \nThe 2024 call sought a mix including existing experts, grantees, applicants and the wider community; litigation expertise was not essential. Lived experience, intersecting issues and regional representation were sought. This describes desired composition, not exhaustive legal eligibility or a right to appointment. | \n
| Appointment and renewal | \nThe current page says DFF ran an open call, received about 40 applications and began with ten members. It does not identify the appointing person/body, selection votes, term length, removal grounds or renewal rule. | \n
| Conflicts and confidentiality | \nThe current page says all members sign Terms of Reference containing confidentiality and conflict policies, and are compensated. The actual terms, conflict definitions, disclosure obligations, recusal procedure and compensation amounts were not acquired. Permitting grantee/applicant participation does not establish self-dealing or failure to recuse. | \n
| Grant selection | \nLater records describe peers making grant decisions across both types. The earlier pilot called their output recommendations. Preserve these stages and dates; neither wording proves legal power to bind DFF to a contract. | \n
| Board and staff | \nThe board is reported to have adopted the model; staff facilitate reviews. These roles do not establish either a retained award-by-award veto or its complete absence. | \n
| Signature, payment and redress | \nThe three records do not disclose grant signatories, post-selection checks, payment authorisation, suspension/recovery rights or an applicant appeal procedure. Do not import Oak's separate grant-letter workflow into DFF's grants. | \n
The [current Community Peer Group page](http[local research file] supports the recruitment, signed-terms and compensation descriptions above (reader lines 5–9). Its roster records affiliations, not corporate reserved seats, and is not treated as an appointment instrument. Its references to a pilot remain visible alongside the linked March 2026 account of adoption; use the dated retrospective for that transition. Original: allocation-peer-group.html.
The evidence now supports a more precise division: DFF established a process, the board reportedly adopted it, and peers reportedly selected projects within an available funding budget. A claim that DFF staff alone chose the June 2026 Liberty grant would erase this stated process. Conversely, saying the peers control the whole institution or that Oak/DFF retained no contractual rights would exceed the records. Public accounts of selection and internal debate do not identify the people who voted on Liberty or reveal its agreement.
\nThe strongest alternative to a claim of complete authority transfer is that a genuine delegated selection process operates inside an institutional strategy, budget and legal framework whose reserved rights are not disclosed here. The exact discriminating records are the signed Peer Group Terms of Reference and incorporated conflict policy, the March 2026 board adoption/delegation resolution, and the applicable DFF grant agreement or execution policy. The first two would clarify appointment, recusal and the force of selections; the last would locate signature and payment authority. None was read in this bounded pass.
\nThe starting page was an actual link in the accepted August 2026 announcement; the two additional records were actual links on that page. All three original HTMLs were retained by one successful ordinary GET each. Full extracted derivatives and exact hashes/read scopes are in allocation-captures.json. No applications, forms, roster biographies, governance repositories or further linked articles were pursued. There was no failed-route recovery, account activity, outreach, graph, forest, current-work, visual or package change. Acquisition is closed for root review.
Bounded acquisition, 9 September 2026. The 8 September 2026 exposure draft is a proposal, not an enacted or operating duty. This note follows the funding/governance forest assessment and reads the existing policy-lineage, operative-authority, Mannifera-selection and McKinnon-governance packets. It records transmission and decisions; it does not attribute legislative control to a funder.
\nThe supported structure has parallel paths: Reset model → HRLC express endorsement, and Reset submission70 → statutory-review reasoning → government commitment tied to that review. The record does not establish HRLC as the intermediary carrying Reset into the reviewer's reasoning. The public decision record supports the overarching duty while modifying or leaving open important components. It does not establish government adoption of the full HRLC/Reset package, NGO authorship of the exposure draft, or funder approval of statutory wording.
\nHRLC's submission dated 5 July 2024 recommends that government adopt Reset.Tech Australia's approach. Its recommendation section says the “government adopts Reset.Tech Australia’s risk-based approach”; the ensuing five pillars cover duty, assessment, mitigation, transparency and accountability. Section 5 identifies Reset's Five Pillars and cites the April 2024 Digital Platform Regulation Green Paper, footnote 23. The transparency request expressly includes vetted researchers and civil society organisations. This is an attributable endorsement and submission, not similarity between two reports. [HRLC submission 155, PDF/printed p5 §2; p10 §5 and fn23; p11 §5.1.4](http[local research file]
\nHRLC also explicitly adopts the World Economic Forum's 2023 Typology of Online Harms as its framework for understanding online harm. This identifies a separate taxonomy-to-submission transmission, not government adoption of that taxonomy. [Submission155, PDF/printed p9 §4.3, footnotes15–18](http[local research file] The external WEF original was not acquired in this run.
\nThe published updated review, internally dated October 2024 and published 4 February 2025, supplies the official reasoning bridge. Its available text is not assumed identical to the original October delivery; the version evidence does not establish that the Reset citation was newly inserted. In §5.5 the reviewer favours a singular overarching duty, identifies Woods/Perrin and Carnegie antecedents, and quotes Reset submission 70 p5 when explaining why content-specific duties can miss harmful service design. Footnotes 37–38 identify that submission. The same section names other supporters; attribution to Reset is evidence of use in the review's reasoning, not exclusive authorship or but-for causation. [Statutory review, printed p53 §5.5](http[local research file]
\nMinister Rowland announced the government's duty-of-care commitment in the Sydney Institute speech on 13 November 2024. The speech explicitly grounds the direction in the recently received Rickard review and international alignment. The associated release is dated 14 November 2024. Neither relevant announcement passage attributes the decision to HRLC, Reset or their funders. These are separate announcement/publication dates, not competing dates for a single document. [Speech, A Digital Duty of Care section](http[local research file] [release](http[local research file]
\nA newly inspected departmental release establishes official awareness more concretely than an NGO's access claim. FOI 25-247 Document 2, briefing QB24-000162, is marked last edited and SES-cleared 26 November 2024. Its media table summarises HRLC's DSA-style proposal and an October article featuring HRLC and Reset endorsing duty of care. The clearance is after the announcement: it establishes inclusion in an official briefing, not that the minister read these articles before deciding. A separate released presentation dated 26 February 2025 describes review recommendations and implementation risks; it is a discussion document, not a decision minute. [FOI bundle, Document 2 PDF pp10–15, especially pp14–15; Document 5 PDF pp32–35](http[local research file]
\nThe 14 April 2026 response is a government decision record, distinct from recommendations or consultation summaries. Its labels preserve a material distinction: support in principle can leave implementation open; note requires further consideration. The following are dispositions of review recommendations, not individually addressed replies to HRLC. [Government response, PDF/printed pp3–9](http[local research file]
\n| Review proposal | \nGovernment disposition | \nExact locator and consequence | \n
|---|---|---|
| Rec4, singular overarching duty | \nSupport | \np4: accepted policy architecture. | \n
| Rec7, automatic highest tier above 10% population | \nNot supported | \np5; p4 explanation favours proportionate application wherever risk exists. | \n
| Rec9, eSafety mandatory codes | \nSupport in principle | \np5; p4 explanation instead describes government-made rules with parliamentary scrutiny. | \n
| Rec13, safeguarded authorised-researcher access | \nSupport in principle | \np6: implementation remained open. | \n
| Rec25, accessible internal complaints | \nSupport | \np8. | \n
| Rec26, digital Ombuds scheme | \nNote | \np9; p7 prioritises internal dispute mechanisms initially. | \n
The September draft adds a consequential boundary to the research request: proposed 205B(1) requires employment by an Australian university and approval under prospective legislative rules. Research additionally requires ethics approval and a prescribed public-interest category. Section 205C delegates scheme design, including privacy, security, confidentiality, conflicts, fees and review. This narrows direct eligibility compared with the NGOs' request; it neither awards an NGO a data-access entitlement nor excludes all university collaboration. The reason for choosing this boundary is not established here. Exposure draft, PDF pp63–65 / printed pp59–61]
\nHRLC's October Rights-First report also sought clear reasons for content/account restrictions, accessible remedies, and review of significant regulatory decisions. Those safeguards belong in the comparison alongside its coercive powers proposal. Rights-First, PDF p12 / printed pp22–23, and PDF p15 / printed pp28–29]
\nThe official May–June 2026 consultation summary reports more than forty meetings and support across several stakeholder categories, including industry, alongside competing priorities. It groups participants instead of assigning the relevant demands to HRLC or Reset. It cannot close the attribution gap or prove that a recorded suggestion was accepted. [Summary, PDF pp1–2](http[local research file]
\nThese are candidate relations for reconciliation with existing entities and source IDs, not canonical edits.
\n| Candidate | \nDirection / typed relation | \nStage and limit | \n
|---|---|---|
| disp-au-e-hrlc-reset | \nHRLC → Reset five-pillar model: expressly recommends adoption | \nSubmitted advocacy; document dated 5 July 2024, submission155 p5. Exact receipt date not established. | \n
| disp-au-e-hrlc-wef-taxonomy | \nHRLC submission155 → WEF2023 harm taxonomy: expressly adopts framework | \nProposal framing, §4.3 p9. No government-adoption edge. | \n
| disp-au-e-review-reset | \nStatutory review → Reset submission70: cites reasoning for overarching duty | \nPublished updated reviewer recommendation, internally dated October2024/published4February2025; available text not assumed identical to initial delivery; §5.5 p53. Reuse existing lineage edge if equivalent. | \n
| disp-au-e-brief-awareness | \nDepartmental briefing → HRLC/Reset public advocacy: records media coverage | \nOfficial awareness by 26 November 2024; FOI Document2 pp14–15. No predecision receipt inference. | \n
| disp-au-e-government-rec4 | \nGovernment → reviewrec4: supports | \nPublished response, 14 April 2026; p4. Policy acceptance, not enactment. | \n
| disp-au-e-government-rec7 | \nGovernment → reviewrec7: does not support | \nSame response p5. Do not relabel as an HRLC-specific rejection. | \n
| disp-au-e-government-rec13 | \nGovernment → reviewrec13: supports in principle | \nSame response p6. No award, provider selection or payment. | \n
| disp-au-e-research-scope | \nExposure draft → researcher eligibility: requires university employment | \nProposed 205B. Comparative narrowing, with motive and drafting author unresolved. | \n
The strongest rival explanation is a government selecting from convergent international models and many contributors through its independent review, rather than executing a single privately directed package. The records sustain a real, attributed NGO-to-review path while also sustaining government choice and narrowing. Funding and commissioned research increase advocacy capacity; the records read here do not turn that capacity into a statutory approval right.
\nThe next discriminating records would be the department's recommendation-to-clause drafting instructions, an attributed consultation response matrix, or the decision/reasons for university-only researcher eligibility and delegated complaint safeguards. Likely holders are the department's online-safety policy area and ministerial office; future explanatory materials or data-access rules could also resolve the scope. No outreach is authorised. Stop here for the whole-case forest assessment.
\nAll access observations are from 9 September 2026. read below refers to the stated scope, never an entire report by implication. Page numbers in this note are one-based; PDF and printed numbers are distinguished where different.
| Local alias | \nPrimary record and read scope | \nAccess / custody | \n
|---|---|---|
| disp-au-s-hrlc155 | \nLinked original above; cover, recommendations p5, §4.3 p9, §§5–5.1.4 pp10–11 and accountability p17 | \nRead through PDF reader (18 pages; reader refs turn886view1, turn912view0/1). One ordinary GET timed out at 30s; no original retained or hash. | \n
| disp-au-s-government-response | \nLinked original above, cover and response definitions, duty pp3–6, complaints pp7–9, enforcement p11 | \nRead through PDF reader (16 pages; turn882view1, turn907view1/2, turn912view2). Publication date independently read on the department's response landing page. One GET timed out at 30s; no original or hash. | \n
| disp-au-s-foi25247 | \nLinked original above, Document2 pp10–15, Document5 pp32–39 | \nRead through PDF reader (41 pages; turn903view1/2, turn907view3). Specific release date not established. One GET timed out at 30s; no original or hash. Redacted portions were not reconstructed. | \n
| Existing statutory-review source | \nOfficial indexed original §5.5 printed53, including footnotes34–38; official landing page publication 4 February 2025 | \nScoped primary text read; not a new full-PDF inspection. Existing ordinary PDF timeout route remained closed. Actual linked DOCX click returned Internal Error; no retry. | \n
| Existing exposure-draft source | \nReused available reader; current-lane spot checks 25H, 26F–G, 205B–C | \nRead, not newly acquired. Existing ordinary PDF timeout route remained closed; no original-byte hash. | \n
| disp-au-s-announcement | \nLinked ministerial speech13Nov2024 and release14Nov2024, relevant duty passages | \nHTML reader inspected; no local original. | \n
| Existing Rights-First source | \nLinked original above, principles3 and5 at PDF12 and15 | \nScoped PDF reader read; no local original acquired in this lane. | \n
| disp-au-s-consult-summary | \nLinked original above, PDF1–2 of3 | \nScoped PDF reader read; not an attributed participant transcript. No original acquired. | \n
No ACNC, OPD403 or closed normal-GET exposure/review routes were retried. No canonical case, forest file, account, submission or public surface was changed. The only lane output is this note; the three fresh capture attempts yielded no retained originals.
", "path": "disposition-wave-2026-09-09/australia-disposition.md", "sourceIds": [ "w13-disp-s-hrlc155", "w13-disp-s-au-response", "w13-disp-s-rowland-speech", "w13-disp-s-rowland-release", "w13-disp-s-foi25247" ], "markdown": "Bounded acquisition, 9 September 2026. The 8 September 2026 exposure draft is a proposal, not an enacted or operating duty. This note follows the funding/governance forest assessment and reads the existing policy-lineage, operative-authority, Mannifera-selection and McKinnon-governance packets. It records transmission and decisions; it does not attribute legislative control to a funder.\n\nThe supported structure has parallel paths: **Reset model → HRLC express endorsement**, and **Reset submission70 → statutory-review reasoning → government commitment tied to that review**. The record does not establish HRLC as the intermediary carrying Reset into the reviewer's reasoning. The public decision record supports the overarching duty while modifying or leaving open important components. It does not establish government adoption of the full HRLC/Reset package, NGO authorship of the exposure draft, or funder approval of statutory wording.\n\n## What actually travelled\n\nHRLC's submission dated 5 July 2024 recommends that government adopt Reset.Tech Australia's approach. Its recommendation section says the “government adopts Reset.Tech Australia’s risk-based approach”; the ensuing five pillars cover duty, assessment, mitigation, transparency and accountability. Section 5 identifies Reset's Five Pillars and cites the April 2024 Digital Platform Regulation Green Paper, footnote 23. The transparency request expressly includes vetted researchers **and civil society organisations**. This is an attributable endorsement and submission, not similarity between two reports. HRLC submission 155, PDF/printed p5 §2; p10 §5 and fn23; p11 §5.1.4 (see source references)\n\nHRLC also explicitly adopts the World Economic Forum's 2023 Typology of Online Harms as its framework for understanding online harm. This identifies a separate taxonomy-to-submission transmission, not government adoption of that taxonomy. Submission155, PDF/printed p9 §4.3, footnotes15–18 (see source references) The external WEF original was not acquired in this run.\n\nThe published updated review, internally dated October 2024 and published 4 February 2025, supplies the official reasoning bridge. Its available text is not assumed identical to the original October delivery; the version evidence does not establish that the Reset citation was newly inserted. In §5.5 the reviewer favours a singular overarching duty, identifies Woods/Perrin and Carnegie antecedents, and quotes Reset submission 70 p5 when explaining why content-specific duties can miss harmful service design. Footnotes 37–38 identify that submission. The same section names other supporters; attribution to Reset is evidence of use in the review's reasoning, not exclusive authorship or but-for causation. Statutory review, printed p53 §5.5 (see source references)\n\nMinister Rowland announced the government's duty-of-care commitment in the Sydney Institute speech on **13 November 2024**. The speech explicitly grounds the direction in the recently received Rickard review and international alignment. The associated release is dated **14 November 2024**. Neither relevant announcement passage attributes the decision to HRLC, Reset or their funders. These are separate announcement/publication dates, not competing dates for a single document. Speech, A Digital Duty of Care section (see source references) release (see source references)\n\nA newly inspected departmental release establishes official awareness more concretely than an NGO's access claim. FOI 25-247 Document 2, briefing QB24-000162, is marked last edited and SES-cleared **26 November 2024**. Its media table summarises HRLC's DSA-style proposal and an October article featuring HRLC and Reset endorsing duty of care. The clearance is after the announcement: it establishes inclusion in an official briefing, not that the minister read these articles before deciding. A separate released presentation dated 26 February 2025 describes review recommendations and implementation risks; it is a discussion document, not a decision minute. FOI bundle, Document 2 PDF pp10–15, especially pp14–15; Document 5 PDF pp32–35 (see source references)\n\n## Decisions and differences\n\nThe **14 April 2026** response is a government decision record, distinct from recommendations or consultation summaries. Its labels preserve a material distinction: support in principle can leave implementation open; note requires further consideration. The following are dispositions of **review recommendations**, not individually addressed replies to HRLC. Government response, PDF/printed pp3–9 (see source references)\n\n| Review proposal | Government disposition | Exact locator and consequence |\n| --- | --- | --- |\n| Rec4, singular overarching duty | Support | p4: accepted policy architecture. |\n| Rec7, automatic highest tier above 10% population | Not supported | p5; p4 explanation favours proportionate application wherever risk exists. |\n| Rec9, eSafety mandatory codes | Support in principle | p5; p4 explanation instead describes government-made rules with parliamentary scrutiny. |\n| Rec13, safeguarded authorised-researcher access | Support in principle | p6: implementation remained open. |\n| Rec25, accessible internal complaints | Support | p8. |\n| Rec26, digital Ombuds scheme | Note | p9; p7 prioritises internal dispute mechanisms initially. |\n\nThe September draft adds a consequential boundary to the research request: proposed **205B(1)** requires employment by an Australian university and approval under prospective legislative rules. Research additionally requires ethics approval and a prescribed public-interest category. Section **205C** delegates scheme design, including privacy, security, confidentiality, conflicts, fees and review. This narrows direct eligibility compared with the NGOs' request; it neither awards an NGO a data-access entitlement nor excludes all university collaboration. The reason for choosing this boundary is not established here. Exposure draft, PDF pp63–65 / printed pp59–61]\n\nHRLC's October Rights-First report also sought clear reasons for content/account restrictions, accessible remedies, and review of significant regulatory decisions. Those safeguards belong in the comparison alongside its coercive powers proposal. Rights-First, PDF p12 / printed pp22–23, and PDF p15 / printed pp28–29]\n\nThe official May–June 2026 consultation summary reports more than forty meetings and support across several stakeholder categories, including industry, alongside competing priorities. It groups participants instead of assigning the relevant demands to HRLC or Reset. It cannot close the attribution gap or prove that a recorded suggestion was accepted. Summary, PDF pp1–2 (see source references)\n\n## Candidate directional edges and limits\n\nThese are candidate relations for reconciliation with existing entities and source IDs, not canonical edits.\n\n| Candidate | Direction / typed relation | Stage and limit |\n| --- | --- | --- |\n| disp-au-e-hrlc-reset | HRLC → Reset five-pillar model: expressly recommends adoption | Submitted advocacy; document dated 5 July 2024, submission155 p5. Exact receipt date not established. |\n| disp-au-e-hrlc-wef-taxonomy | HRLC submission155 → WEF2023 harm taxonomy: expressly adopts framework | Proposal framing, §4.3 p9. No government-adoption edge. |\n| disp-au-e-review-reset | Statutory review → Reset submission70: cites reasoning for overarching duty | Published updated reviewer recommendation, internally dated October2024/published4February2025; available text not assumed identical to initial delivery; §5.5 p53. Reuse existing lineage edge if equivalent. |\n| disp-au-e-brief-awareness | Departmental briefing → HRLC/Reset public advocacy: records media coverage | Official awareness by 26 November 2024; FOI Document2 pp14–15. No predecision receipt inference. |\n| disp-au-e-government-rec4 | Government → reviewrec4: supports | Published response, 14 April 2026; p4. Policy acceptance, not enactment. |\n| disp-au-e-government-rec7 | Government → reviewrec7: does not support | Same response p5. Do not relabel as an HRLC-specific rejection. |\n| disp-au-e-government-rec13 | Government → reviewrec13: supports in principle | Same response p6. No award, provider selection or payment. |\n| disp-au-e-research-scope | Exposure draft → researcher eligibility: requires university employment | Proposed 205B. Comparative narrowing, with motive and drafting author unresolved. |\n\nThe strongest rival explanation is a government selecting from convergent international models and many contributors through its independent review, rather than executing a single privately directed package. The records sustain a real, attributed NGO-to-review path while also sustaining government choice and narrowing. Funding and commissioned research increase advocacy capacity; the records read here do not turn that capacity into a statutory approval right.\n\nThe next discriminating records would be the department's recommendation-to-clause drafting instructions, an attributed consultation response matrix, or the decision/reasons for university-only researcher eligibility and delegated complaint safeguards. Likely holders are the department's online-safety policy area and ministerial office; future explanatory materials or data-access rules could also resolve the scope. No outreach is authorised. Stop here for the whole-case forest assessment.\n\n## Reading and custody ledger\n\nAll access observations are from 9 September 2026. `read` below refers to the stated scope, never an entire report by implication. Page numbers in this note are one-based; PDF and printed numbers are distinguished where different.\n\n| Local alias | Primary record and read scope | Access / custody |\n| --- | --- | --- |\n| disp-au-s-hrlc155 | Linked original above; cover, recommendations p5, §4.3 p9, §§5–5.1.4 pp10–11 and accountability p17 | Read through PDF reader (18 pages; reader refs turn886view1, turn912view0/1). One ordinary GET timed out at 30s; no original retained or hash. |\n| disp-au-s-government-response | Linked original above, cover and response definitions, duty pp3–6, complaints pp7–9, enforcement p11 | Read through PDF reader (16 pages; turn882view1, turn907view1/2, turn912view2). Publication date independently read on the department's response landing page. One GET timed out at 30s; no original or hash. |\n| disp-au-s-foi25247 | Linked original above, Document2 pp10–15, Document5 pp32–39 | Read through PDF reader (41 pages; turn903view1/2, turn907view3). Specific release date not established. One GET timed out at 30s; no original or hash. Redacted portions were not reconstructed. |\n| Existing statutory-review source | Official indexed original §5.5 printed53, including footnotes34–38; official landing page publication 4 February 2025 | Scoped primary text read; not a new full-PDF inspection. Existing ordinary PDF timeout route remained closed. Actual linked DOCX click returned Internal Error; no retry. |\n| Existing exposure-draft source | Reused available reader; current-lane spot checks 25H, 26F–G, 205B–C | Read, not newly acquired. Existing ordinary PDF timeout route remained closed; no original-byte hash. |\n| disp-au-s-announcement | Linked ministerial speech13Nov2024 and release14Nov2024, relevant duty passages | HTML reader inspected; no local original. |\n| Existing Rights-First source | Linked original above, principles3 and5 at PDF12 and15 | Scoped PDF reader read; no local original acquired in this lane. |\n| disp-au-s-consult-summary | Linked original above, PDF1–2 of3 | Scoped PDF reader read; not an attributed participant transcript. No original acquired. |\n\nNo ACNC, OPD403 or closed normal-GET exposure/review routes were retried. No canonical case, forest file, account, submission or public surface was changed. The only lane output is this note; the three fresh capture attempts yielded no retained originals.\n", "file": "research/disposition-wave-2026-09-09-australia-disposition.md" }, { "id": "disposition-wave-2026-09-09-uk-disposition", "title": "UK network submissions: three published dispositions", "description": "Observed 9 September 2026. Bounded acquisition following the latest forest assessment; no canonical case edits. This packet follows the Online Safety Act Network (OSAN, predecessor name of the current Online Safety Network) from submissions into regulator responses. It does not reassess the whole current Online Safety Act regime.", "html": "Observed 9 September 2026. Bounded acquisition following the latest forest assessment; no canonical case edits. This packet follows the Online Safety Act Network (OSAN, predecessor name of the current Online Safety Network) from submissions into regulator responses. It does not reassess the whole current Online Safety Act regime.
\nThe strongest new pattern is selective, attributable uptake with explicit resistance. These records establish that the network's arguments entered the decision process and received substantive answers. They do not establish control of Ofcom, preferential treatment of the network, or a donor-to-regulatory-decision instruction chain. The prior UK operator and articles packets supply organizational context only.
\nSubmission. OSAN's Full response to Ofcom's consultation on protecting people from illegal harms online is internally dated February 2024, with Maeve Walsh as contact (PDF/printed p2). Its pp25–26 argue that the regulatory threshold should be understood against a civil balance of probabilities, allowing consideration of wider contextual evidence. The cyberflashing example refers to widespread harm and limited countervailing interests, expressly contrasting difficulties around suicide offences. It is therefore inaccurate to recast the original as a universal test that any speech without apparent benefit should be illegal. [Ofcom-hosted original response, pp25–26](http[local research file]
\nDisposition, 16 December 2024. Ofcom identifies OSAN and cites those pages in Volume 3: Transparency, trust and other guidance, 2.19 and footnote11, PDF/printed p10. At 2.20–2.21, pp10–11, it declines the proposed departure: the Act requires reasonable grounds to infer all necessary offence elements; the appropriate inference depends on the offence and context. Ofcom also invokes the expression costs of erroneously treating large quantities of lawful activity as illegal. It agrees that this is not the criminal beyond-reasonable-doubt threshold, while denying that its consultation applied that threshold. This is an explicit adverse disposition of the proposed interpretation, not a court determination that OSAN's legal argument was invalid. [Ofcom decision, 2.18–2.21](http[local research file]
\nPersonal contribution boundary. A separate network article by Lorna Woods, dated 15 February 2024, makes this contextual civil-threshold argument under “Burden of Proof in a Civil Regime.” It supports a named public intellectual contribution. The full February response's contact field alone does not prove Walsh wrote every passage. [Woods article](http[local research file]
\nSubmission. The February response's PDF/printed p97 identifies a mismatch between the risk discussion, thumbnails/click-through pathways and the presentation of search obligations, including direct and indirect routes to harmful material. It provides a concrete concern about how the risk assessment and proposed measures fit together. [OSAN response, p97](http[local research file]
\nDisposition, 16 December 2024. Annex 1: Further stakeholder responses, A1.9.9–A1.9.10, PDF/printed p94, attributes the concern to OSAN (footnote486, p97). Ofcom says it amended Register of Risks paragraph24.13, agreed with the wider risks highlighted and reiterated them in 24.10–24.14. It simultaneously confines “search content” to the Act's definition: content in results or reached through interaction with them, excluding subsequent interactions; footnote487 cites s57(2) and (5). This supports an OSAN concern → acknowledged amendment of the risk register edge. It does not establish expansion of statutory liability, a new content-removal mandate or adoption of an unrestricted downstream-link duty. The final register was not independently acquired and compared in this pass; the amendment is expressly reported by its issuing regulator. [Ofcom Annex1, A1.9.9–10](http[local research file]
\nDated network contribution. The full Protection of Children Consultation: Response from Online Safety Act Network names Professor Lorna Woods and Maeve Walsh on PDF p1 and is dated July 2024 on p2. Its Issue8, pp66–69, criticises reliance on age gates without changing underlying service design, discusses under-13 detection and different-age experiences, and seeks clearer effectiveness metrics plus use of information-gathering powers to establish companies' capabilities. It also supports Ofcom not selecting a particular tool. These are differentiated proposals, not an uncomplicated demand for universal identification. [OSAN response2, pp1–2 and66–69](http[local research file]
\nActual decision, 24 April 2025. Ofcom's children's statement Volume3: Assessing the risks of harms to children online, 8.170–8.171, PDF/printed pp36–37, clarifies that declared minimum ages in terms of service and self-declared ages are insufficient evidence for risk assessment. Without highly effective age assurance enforcing a minimum age, providers should conservatively assess the likely presence and ages of younger children. Ofcom calls these minor drafting changes in this statement. This particular decision is guidance on evidential assumptions for risk assessments: it does not, by itself, require every service to set a minimum age or every user to present identity documents. [Ofcom April decision, 8.169–8.171](http[local research file]
\nExplicit attribution, published 15 May 2025. In the Secondary Legislation Scrutiny Committee's 25th Report, paragraph23, Ofcom's answer credits changes reflecting civil-society feedback including OSAN, giving the minimum-age clarification as its example. It also denies assigning greater weight to any stakeholder group. The Committee describes OSAN's formal submission and publishes Ofcom/DSIT responses; it records OSAN's continuing dissatisfaction at 24 and says at 26 it does not adjudicate the competing arguments. This is stronger evidence of recognized contribution than similar wording or a meeting. It is weaker than an exact accepted-edit record or proof OSAN originated the measure. [SLSC report, paragraphs11,23–26](http[local research file] [Report publication date](http[local research file]
\nThe July response's inspected age-assurance section does not contain the exact minimum-age formulation; text searches of both published OSAN response files also returned no exact “minimum age” match. That negative is limited by PDF extraction and terminology. The April statement's nearby stakeholder footnotes do not uniquely identify an OSAN sentence behind this edit. Consequently, retain the edge as Ofcom publicly credited OSAN among contributing civil-society respondents, not Ofcom adopted Woods/Walsh's specified minimum-age clause.
\nThe network is a submitter and advocate in these records. Woods and Walsh are named authors of the July response, not thereby statutory committee members or holders of regulatory approval rights. Targeted Ofcom searches did not establish a formal appointment for either; an unrelated result names Siobhan Walsh, not Maeve Walsh. This pass makes no claim that no formal role exists anywhere.
\nThe November2023 consultation title describes the consultation's opening period, not the response date. The illegal-harms response itself says February2024. Ofcom's current consultation page displays 20 November 2024 alongside the main response download; that is not treated as its submission date. [Ofcom consultation chronology and response listing](http[local research file]
\nThe reported decisions are historical December2024 / April–May2025 decisions. They are not presented as a complete current-law statement on 9 September2026. A later review must distinguish a subsequent change from reversal of these attributed decisions.
\nSuggested precise relations for root, without creating canonical entries here:
\n| Relation | \nRecord-supported verb and limit | \n
|---|---|
| OSAN → February2024 illegal-harms response | \nSubmitted an identified consultation response; exact day not established | \n
| Ofcom → OSAN inference proposal | \nDeclined proposed interpretation in published December2024 response | \n
| Ofcom → OSAN search-scope concern | \nAcknowledged and reported amending risk-register wording; statutory scope retained | \n
| Woods + Walsh → July2024 response | \nNamed co-authors | \n
| Ofcom → OSAN / minimum-age clarification | \nPublicly credited OSAN among contributors in May2025; no exclusive or verbatim authorship proof | \n
| OSAN → SLSC | \nFormal submission considered during scrutiny; this confers no committee membership or adjudicatory power | \n
Retained reader capture: uk-disposition-reader-extracts.json, SHA256 c15aa176a8a012bfe25cb819787c8c6661d30a1e10ce697234cd06d35d1c22cc. It contains the three scoped public web-reader result groups. It is a derivative text capture, not the original PDFs and not their hashes. Context returned by the reader extends beyond the selected paragraphs; no complete-volume audit is claimed. The operative passages inspected are listed above.
Retained original: uk-disposition-woods-icjg-2024.pdf, ordinary public GET200, 2,062,645 bytes, SHA256 4097e729563bd10b42046cc8d5392c2d5a688aef8a11d5ea219a7df18bec530f, nine pages. Source: [network-hosted article PDF](http[local research file] Text extraction did not expose the relevant heading; this PDF has not been visually audited and is not the evidential basis of the article summary. The corresponding public HTML was successfully read through ordinary GET, including the named author/date and civil-threshold section.
Ofcom's full February response was read through the public PDF reader at pp1–2,25–27 and97 plus date searches. July response2 was read at pp1–6 and66–69, with surrounding reader context; response1's cover/initial sections were read and an exact-term search performed. Ofcom's December Volume3 was read at cover and2.18–2.21; Annex1 at cover andA1.9.9–10 with surrounding search-moderation context. April Volume3 was read at cover and8.169–8.179. SLSC report was read at paragraphs11 and17–26 plus publication metadata.
\nAccess gaps: the separate Ofcom “Our approach to developing Codes measures” PDF gave a reader error and ordinary GET403; that acquisition branch is closed and no unseen safe-harbour disposition from it is claimed. Network HTML initially failed through the web reader but ordinary GET succeeded. A direct Parliament HTML capture returned403; its public reader text was accessible and retained. No access controls were bypassed. No original Ofcom PDF hash is supplied. No visual proof of fine layout, handwritten annotation or hidden/redacted text is asserted.
\nThe three decisions satisfy this run's contribution-to-disposition test: an adverse interpretation decision, an expressly reported amendment, and regulator-attributed contribution to a concrete guidance clarification. No further acquisition is needed before the parent's whole-case reassessment.
\nThe most useful next record is Ofcom's consultation-response coding/decision log or drafting version comparison for April2025 paragraph8.171, with the cited OSAN input and approval trail. Custodian: Ofcom's online-safety policy team / consultation records function. It would separate collective feedback attribution from the exact authorship and decisive influence of a particular proposal. For the search issue, the already published final Register24.10–24.14 and its consultation predecessor would resolve the small remaining text-diff question. Neither missing record warrants claiming that influence was absent; neither is needed to invent donor instruction or a veto. No request or outreach was made.
", "path": "disposition-wave-2026-09-09/uk-disposition.md", "sourceIds": [ "w13-disp-s-osan-feb", "w13-disp-s-ofcom-inference", "w13-disp-s-ofcom-search", "w13-disp-s-woods-civil", "w13-disp-s-osan-july", "w13-disp-s-ofcom-child-risk", "w13-disp-s-slsc-report", "w13-disp-s-slsc-date" ], "markdown": "Observed 9 September 2026. Bounded acquisition following the latest forest assessment; no canonical case edits. This packet follows the Online Safety Act Network (OSAN, predecessor name of the current Online Safety Network) from submissions into regulator responses. It does not reassess the whole current Online Safety Act regime.\n\nThe strongest new pattern is **selective, attributable uptake with explicit resistance**. These records establish that the network's arguments entered the decision process and received substantive answers. They do not establish control of Ofcom, preferential treatment of the network, or a donor-to-regulatory-decision instruction chain. The prior UK operator and articles packets supply organizational context only.\n\n## 1. Illegal-content inference: proposed interpretation not accepted\n\n**Submission.** OSAN's *Full response to Ofcom's consultation on protecting people from illegal harms online* is internally dated **February 2024**, with Maeve Walsh as contact (PDF/printed p2). Its pp25–26 argue that the regulatory threshold should be understood against a civil balance of probabilities, allowing consideration of wider contextual evidence. The cyberflashing example refers to widespread harm and limited countervailing interests, expressly contrasting difficulties around suicide offences. It is therefore inaccurate to recast the original as a universal test that any speech without apparent benefit should be illegal. Ofcom-hosted original response, pp25–26 (see source references)\n\n**Disposition, 16 December 2024.** Ofcom identifies OSAN and cites those pages in *Volume 3: Transparency, trust and other guidance*, **2.19 and footnote11, PDF/printed p10**. At **2.20–2.21, pp10–11**, it declines the proposed departure: the Act requires reasonable grounds to infer all necessary offence elements; the appropriate inference depends on the offence and context. Ofcom also invokes the expression costs of erroneously treating large quantities of lawful activity as illegal. It agrees that this is not the criminal beyond-reasonable-doubt threshold, while denying that its consultation applied that threshold. This is an explicit adverse disposition of the proposed interpretation, not a court determination that OSAN's legal argument was invalid. Ofcom decision, 2.18–2.21 (see source references)\n\n**Personal contribution boundary.** A separate network article by **Lorna Woods**, dated **15 February 2024**, makes this contextual civil-threshold argument under “Burden of Proof in a Civil Regime.” It supports a named public intellectual contribution. The full February response's contact field alone does not prove Walsh wrote every passage. Woods article (see source references)\n\n## 2. Search results: accepted register clarification, without expanding statutory scope\n\n**Submission.** The February response's **PDF/printed p97** identifies a mismatch between the risk discussion, thumbnails/click-through pathways and the presentation of search obligations, including direct and indirect routes to harmful material. It provides a concrete concern about how the risk assessment and proposed measures fit together. OSAN response, p97 (see source references)\n\n**Disposition, 16 December 2024.** *Annex 1: Further stakeholder responses*, **A1.9.9–A1.9.10, PDF/printed p94**, attributes the concern to OSAN (footnote486, p97). Ofcom says it amended **Register of Risks paragraph24.13**, agreed with the wider risks highlighted and reiterated them in **24.10–24.14**. It simultaneously confines “search content” to the Act's definition: content in results or reached through interaction with them, excluding subsequent interactions; footnote487 cites **s57(2) and (5)**. This supports an **OSAN concern → acknowledged amendment of the risk register** edge. It does not establish expansion of statutory liability, a new content-removal mandate or adoption of an unrestricted downstream-link duty. The final register was not independently acquired and compared in this pass; the amendment is expressly reported by its issuing regulator. Ofcom Annex1, A1.9.9–10 (see source references)\n\n## 3. Children's minimum ages: regulator credits feedback including OSAN; exact clause authorship remains open\n\n**Dated network contribution.** The full *Protection of Children Consultation: Response from Online Safety Act Network* names **Professor Lorna Woods and Maeve Walsh** on PDF p1 and is dated **July 2024** on p2. Its **Issue8, pp66–69**, criticises reliance on age gates without changing underlying service design, discusses under-13 detection and different-age experiences, and seeks clearer effectiveness metrics plus use of information-gathering powers to establish companies' capabilities. It also supports Ofcom not selecting a particular tool. These are differentiated proposals, not an uncomplicated demand for universal identification. OSAN response2, pp1–2 and66–69 (see source references)\n\n**Actual decision, 24 April 2025.** Ofcom's children's statement *Volume3: Assessing the risks of harms to children online*, **8.170–8.171, PDF/printed pp36–37**, clarifies that declared minimum ages in terms of service and self-declared ages are insufficient evidence for risk assessment. Without highly effective age assurance enforcing a minimum age, providers should conservatively assess the likely presence and ages of younger children. Ofcom calls these minor drafting changes in this statement. This particular decision is guidance on evidential assumptions for risk assessments: it does not, by itself, require every service to set a minimum age or every user to present identity documents. Ofcom April decision, 8.169–8.171 (see source references)\n\n**Explicit attribution, published 15 May 2025.** In the Secondary Legislation Scrutiny Committee's **25th Report, paragraph23**, Ofcom's answer credits changes reflecting civil-society feedback **including OSAN**, giving the minimum-age clarification as its example. It also denies assigning greater weight to any stakeholder group. The Committee describes OSAN's formal submission and publishes Ofcom/DSIT responses; it records OSAN's continuing dissatisfaction at **24** and says at **26** it does not adjudicate the competing arguments. This is stronger evidence of recognized contribution than similar wording or a meeting. It is weaker than an exact accepted-edit record or proof OSAN originated the measure. SLSC report, paragraphs11,23–26 (see source references) Report publication date (see source references)\n\nThe July response's inspected age-assurance section does not contain the exact minimum-age formulation; text searches of both published OSAN response files also returned no exact “minimum age” match. That negative is limited by PDF extraction and terminology. The April statement's nearby stakeholder footnotes do not uniquely identify an OSAN sentence behind this edit. Consequently, retain the edge as **Ofcom publicly credited OSAN among contributing civil-society respondents**, not **Ofcom adopted Woods/Walsh's specified minimum-age clause**.\n\n## Authority and chronology boundaries for integration\n\nThe network is a submitter and advocate in these records. Woods and Walsh are named authors of the July response, not thereby statutory committee members or holders of regulatory approval rights. Targeted Ofcom searches did not establish a formal appointment for either; an unrelated result names **Siobhan Walsh**, not Maeve Walsh. This pass makes no claim that no formal role exists anywhere.\n\nThe November2023 consultation title describes the consultation's opening period, not the response date. The illegal-harms response itself says February2024. Ofcom's current consultation page displays **20 November 2024** alongside the main response download; that is not treated as its submission date. Ofcom consultation chronology and response listing (see source references)\n\nThe reported decisions are historical **December2024 / April–May2025** decisions. They are not presented as a complete current-law statement on 9 September2026. A later review must distinguish a subsequent change from reversal of these attributed decisions.\n\nSuggested precise relations for root, without creating canonical entries here:\n\n| Relation | Record-supported verb and limit |\n| --- | --- |\n| OSAN → February2024 illegal-harms response | Submitted an identified consultation response; exact day not established |\n| Ofcom → OSAN inference proposal | Declined proposed interpretation in published December2024 response |\n| Ofcom → OSAN search-scope concern | Acknowledged and reported amending risk-register wording; statutory scope retained |\n| Woods + Walsh → July2024 response | Named co-authors |\n| Ofcom → OSAN / minimum-age clarification | Publicly credited OSAN among contributors in May2025; no exclusive or verbatim authorship proof |\n| OSAN → SLSC | Formal submission considered during scrutiny; this confers no committee membership or adjudicatory power |\n\n## Source custody and actual reading scope\n\n**Retained reader capture:** `uk-disposition-reader-extracts.json`, SHA256 **c15aa176a8a012bfe25cb819787c8c6661d30a1e10ce697234cd06d35d1c22cc**. It contains the three scoped public web-reader result groups. It is a derivative text capture, **not the original PDFs and not their hashes**. Context returned by the reader extends beyond the selected paragraphs; no complete-volume audit is claimed. The operative passages inspected are listed above.\n\n**Retained original:** `uk-disposition-woods-icjg-2024.pdf`, ordinary public GET200, **2,062,645 bytes**, SHA256 **4097e729563bd10b42046cc8d5392c2d5a688aef8a11d5ea219a7df18bec530f**, nine pages. Source: network-hosted article PDF (see source references) Text extraction did not expose the relevant heading; this PDF has not been visually audited and is not the evidential basis of the article summary. The corresponding public HTML was successfully read through ordinary GET, including the named author/date and civil-threshold section.\n\nOfcom's full February response was read through the public PDF reader at pp1–2,25–27 and97 plus date searches. July response2 was read at pp1–6 and66–69, with surrounding reader context; response1's cover/initial sections were read and an exact-term search performed. Ofcom's December Volume3 was read at cover and2.18–2.21; Annex1 at cover andA1.9.9–10 with surrounding search-moderation context. April Volume3 was read at cover and8.169–8.179. SLSC report was read at paragraphs11 and17–26 plus publication metadata.\n\n**Access gaps:** the separate Ofcom “Our approach to developing Codes measures” PDF gave a reader error and ordinary GET403; that acquisition branch is closed and no unseen safe-harbour disposition from it is claimed. Network HTML initially failed through the web reader but ordinary GET succeeded. A direct Parliament HTML capture returned403; its public reader text was accessible and retained. No access controls were bypassed. No original Ofcom PDF hash is supplied. No visual proof of fine layout, handwritten annotation or hidden/redacted text is asserted.\n\n## Coherent stopping point and next separating record\n\nThe three decisions satisfy this run's contribution-to-disposition test: an adverse interpretation decision, an expressly reported amendment, and regulator-attributed contribution to a concrete guidance clarification. No further acquisition is needed before the parent's whole-case reassessment.\n\nThe most useful next record is **Ofcom's consultation-response coding/decision log or drafting version comparison for April2025 paragraph8.171**, with the cited OSAN input and approval trail. Custodian: Ofcom's online-safety policy team / consultation records function. It would separate collective feedback attribution from the exact authorship and decisive influence of a particular proposal. For the search issue, the already published final Register24.10–24.14 and its consultation predecessor would resolve the small remaining text-diff question. Neither missing record warrants claiming that influence was absent; neither is needed to invent donor instruction or a veto. No request or outreach was made.\n", "file": "research/disposition-wave-2026-09-09-uk-disposition.md" }, { "id": "disposition-wave-2026-09-09-uk-rights-transition", "title": "Reset UK: the control categories have an identifiable historical origin", "description": "Observed 9 September 2026. A bounded comparison of the 2023 incorporation record, the previously read May2025 replacement constitution, and the April2026 Scott change notice. This reduces one ambiguity without pretending to settle the current member register.", "html": "Observed 9 September 2026. A bounded comparison of the 2023 incorporation record, the previously read May2025 replacement constitution, and the April2026 Scott change notice. This reduces one ambiguity without pretending to settle the current member register.
\nThe [incorporation filing](http[local research file] expressly adopts the guarantee-company model articles in full (PDF2). Its memorandum lists Dale Benjamin Scott as the sole subscriber (PDF10); the guarantee statement names him alone (PDF8). The initial PSC record already reports at least75% voting rights and the right to appoint/remove a board majority (PDF7). These were not newly invented in the2025 replacement or2026 change notice. Appointment as a director did not itself make every director a member under the old general model; do not count directors as historical members.
\nCompanies House identifies the [applicable model articles](http[local research file] as applying to companies incorporated after28April2013. Article21 requires application and director approval for membership;17 permits director appointment by ordinary resolution or board decision;4 gives members a special-resolution reserve power. Those rules are compatible with an initial sole-member position despite two directors. They do not prove that no additional member was admitted beforeMay2025. The current official reproduction was read, not a separately certified2023 copy.
\nThe previously acquired7May2025 resolution is signed by Scott as voting member and replaces the constitution. New article22 makes directors the only members, with admission automatic on appointment;24.2 gives each member one vote on written resolutions. Wood's PSC notification begins on the same7May date. Together these are concrete evidence consistent with a membership transition. The new constitution does not repeat a named Scott reserve. Current PSC percentages still require reconciliation; they must not be added together or confidently replaced by a50/50 figure.
\nThe [April2026 PSC04](http[local research file] effective27June2025, changes service-address/residence details. Both pages were read. It supplies no new nature-of-control statement, so treating this filing as fresh substantive reaffirmation of Scott's old voting percentage would overstate it. No personal address or birth detail is reproduced here.
\nThe best supported explanation now is historical sole-subscriber control followed by a constitution that connects membership to the board, with the displayed PSC categories not reconciled by the records at hand. Alternative arrangements remain possible. The latest member register, PSC basis and any separate voting/appointment instrument would decide the remaining question. No approach to the company has been made. This branch closes at the public-record breakpoint; it should no longer be presented as unexplained overlapping ownership.
\nCustody: originals/root-manifest.json retains URLs, bytes and hashes for both PDFs. Root read extracted operative passages in the incorporation filing and visually inspected its memorandum/guarantee pages and both PSC04 pages. The web reader failed on stale redirect/cache; one ordinary GET per observed stable public endpoint succeeded. There was no use of credentials or expired signed tokens. The ten-page incorporation filing incorporates the general model by reference rather than reproducing all its clauses. Further policy acquisition remains with the two independent workers; root does not open another branch before their forest review.
", "path": "disposition-wave-2026-09-09/uk-rights-transition.md", "sourceIds": [ "w12-gov-s-uk-incorporation", "w12-gov-s-uk-psc04", "w12-gov-s-uk-model" ], "markdown": "Observed 9 September 2026. A bounded comparison of the 2023 incorporation record, the previously read May2025 replacement constitution, and the April2026 Scott change notice. This reduces one ambiguity without pretending to settle the current member register.\n\nThe incorporation filing (see source references) expressly adopts the guarantee-company model articles in full (PDF2). Its memorandum lists **Dale Benjamin Scott as the sole subscriber** (PDF10); the guarantee statement names him alone (PDF8). The initial PSC record already reports at least75% voting rights and the right to appoint/remove a board majority (PDF7). These were not newly invented in the2025 replacement or2026 change notice. Appointment as a director did not itself make every director a member under the old general model; do not count directors as historical members.\n\nCompanies House identifies the applicable model articles (see source references) as applying to companies incorporated after28April2013. Article21 requires application and director approval for membership;17 permits director appointment by ordinary resolution or board decision;4 gives members a special-resolution reserve power. Those rules are compatible with an initial sole-member position despite two directors. They do not prove that no additional member was admitted beforeMay2025. The current official reproduction was read, not a separately certified2023 copy.\n\nThe previously acquired7May2025 resolution is signed by Scott as voting member and replaces the constitution. New article22 makes directors the only members, with admission automatic on appointment;24.2 gives each member one vote on written resolutions. Wood's PSC notification begins on the same7May date. Together these are concrete evidence consistent with a membership transition. The new constitution does not repeat a named Scott reserve. Current PSC percentages still require reconciliation; they must not be added together or confidently replaced by a50/50 figure.\n\nThe April2026 PSC04 (see source references) effective27June2025, changes service-address/residence details. Both pages were read. It supplies **no new nature-of-control statement**, so treating this filing as fresh substantive reaffirmation of Scott's old voting percentage would overstate it. No personal address or birth detail is reproduced here.\n\nThe best supported explanation now is historical sole-subscriber control followed by a constitution that connects membership to the board, with the displayed PSC categories not reconciled by the records at hand. Alternative arrangements remain possible. The latest member register, PSC basis and any separate voting/appointment instrument would decide the remaining question. No approach to the company has been made. This branch closes at the public-record breakpoint; it should no longer be presented as unexplained overlapping ownership.\n\nCustody: originals/root-manifest.json retains URLs, bytes and hashes for both PDFs. Root read extracted operative passages in the incorporation filing and visually inspected its memorandum/guarantee pages and both PSC04 pages. The web reader failed on stale redirect/cache; one ordinary GET per observed stable public endpoint succeeded. There was no use of credentials or expired signed tokens. The ten-page incorporation filing incorporates the general model by reference rather than reproducing all its clauses. Further policy acquisition remains with the two independent workers; root does not open another branch before their forest review.\n", "file": "research/disposition-wave-2026-09-09-uk-rights-transition.md" }, { "id": "dsa-access-enforcement-wave-2026-09-09-commitment-mechanism", "title": "AliExpress public-data access: binding duties and an appointed monitor", "description": "Bounded instrument lane closed, 9 September 2026. Two directly associated official PDF originals acquired: the public version of the adopted decision with its commitment annex, and the Commission's monitoring-trustee notice. No live API, researcher application or third-party data was accessed. Root owns the announcements and current-status/fine review.", "html": "Bounded instrument lane closed, 9 September 2026. Two directly associated official PDF originals acquired: the public version of the adopted decision with its commitment annex, and the Commission's monitoring-trustee notice. No live API, researcher application or third-party data was accessed. Root owns the announcements and current-status/fine review.
\n[Commission Decision C(2025) 3982 final, 18 June 2025](http[local research file] makes the annexed commitments binding under DSA Article 71. Its addressees are AliExpress International (Netherlands) B.V. and Alibaba.com Singapore E-commerce Private Limited (Articles 1–3, PDF 50–51 / decision pages 47–48). Do not replace the latter with the similarly named prospective successor mentioned in the procedural history.
\nThe decision closes the covered proceedings except the specified illegal-content risk-assessment/mitigation grievance in case DSA.100117. The researcher-access section describes suspicions, not an adjudicated Article 40(12) infringement; the annex expressly preserves non-admission (PDF 44–47, recitals 156–164; PDF 54 / annex page 3). Binding acceptance is an actual intervention, but it is neither an infringement finding on access nor proof of fulfilled access.
\nThe official 78-page copy states that confidential material was edited or replaced with summaries/omissions. Its metadata was generated in February 2026; that is not the decision date or receipt date. It is a public edited copy, not an unredacted instrument.
\nThe operative instrument is Annex Section B.V, paragraphs 1–7, PDF 64–66 / annex pages 13–15. It concerns publicly accessible data under Article 40(12). It does not award vetted internal-data access under Article 40(4), trusted-flagger status, or access to a named research organisation.
\nAliExpress must maintain an exception to its general data-access restrictions for qualifying researchers, limited to systemic-risk research and applicable researcher terms. It must maintain a portal with dataset information, field explanations, tool instructions and an intake route; machine-readable access and reasonable downloads to researchers' own devices; near-real-time API access for readily available datasets; and gradual expansion informed by frequently requested datasets. The annex's examples include advertisements, stores/products and product/seller penalties. No such dataset was collected in this investigation.
\n| Who performs the step | \nBinding provision and precise timing | \nLimit | \n
|---|---|---|
| Trained AliExpress staff | \nB.V.3(i–ii): follow an internal SOP and assess eligibility within 15 business days. | \nAssessment is not automatic approval; the internal SOP was not acquired. | \n
| Researcher / data warehouse team | \nB.V.3(iii–iv): approved off-the-shelf access follows tool instructions; customised deployment is expected within 7 business days, with an express complex-development exception. | \nDo not turn the expectation into an unconditional seven-day deadline. | \n
| AliExpress QA team | \nB.V.6: monitor API stability, transmitted-data accuracy and median application processing time; conduct quarterly retrieval, reconciliation and API tests. | \nThese are required checks, not acquired test results. | \n
| AliExpress → trustee | \nB.V.7: semiannual metrics/test reports, counts received/processed/approved/rejected, and researcher-feedback summaries. | \nNo actual report, count or recipient list was acquired. | \n
The dedicated feedback email and the trustee's third-party query route are real stated channels. They are not an acquired adjudicatory appeal procedure for a denied research application. Do not import the decision's separate consumer/content-moderation complaint mechanism as that remedy.
\nThe [official one-page trustee notice](http[local research file] states that the Commission approved Ankura Consulting (Europe), Limited on 23 September 2025. It identifies monitoring and reporting to the Commission as the role and permits third-party queries. This establishes a named appointment-approval step. It does not prove that a particular report was delivered, accepted or acted on, nor disclose fees paid.
\nThe authority allocation is unusually explicit. AliExpress proposes the trustee/mandate and pays fees that must not depend on monitoring results; the Commission approves or rejects, may modify the mandate, and may issue orders/instructions to ensure compliance (C.1–5, PDF 74–75). Independence includes a 12-month restriction on related non-monitoring services before and after, and a limit on prior consecutive monitoring service; ordinary Article 37 audit service is distinguished. No separate independence assessment or executed remuneration terms were acquired.
\nThe trustee may recommend compliance measures and must promptly report reasonably concluded failure, without waiting for scheduled reports. It may not make technical/product/business decisions or exercise the Commission's investigative powers (C.12–13, PDF 76–77). Advance notice of potentially adverse material changes expressly does not require trustee approval, and experimental/beta/testing changes are excepted (C.17). Necessary-information access is bounded by its mandate and coordination with the auditor; the Commission resolves disagreement over the information process within 14 calendar days (C.12(iii), C.15–16). Confidentiality, specific permitted audit-document sharing, and restrictions on public statements apply (C.13–14). A privately paid monitor is not thereby a private regulator with platform-management powers.
\nThe Effective Date is receipt of the decision, not automatically 18 June (definitions, PDF 55; Section E, PDF 78). No receipt record was acquired. The initial implementation period runs to 30 September 2025, with best-efforts compliance during rollout; thereafter obligations apply in full (C.10, PDF 75–76). The first trustee report was due within two months after that cycle; subsequent reports cover cycles ending 30 September 2026 and 30 September 2027 (C.12(iv), PDF 76; recitals 167–171, PDF 48). Their existence or delivery is not established here.
\nThe substantive measures run five years from receipt. The trustee mandate runs through its final report on the cycle ending September 2027. C.7–8 allow replacement/discharge and handover; they do not expressly create automatic renewal beyond that reporting period. Section D allows the Commission to waive, modify or substitute commitments if DSA conditions no longer justify undiminished continuation. This is not treated as an exercised extension or a general renewal right.
\nTwo source-text issues remain visible. C.9, PDF 75, prints “31 November 2027,” an impossible date. Recital 167 instead says 30 November, and C.12(iv) expresses the deadline as two months after 30 September. This was visually verified and is not silently normalized. Separately, recital 174 (PDF 49) says internal QA continues through the five-year term; Section D (PDF 78) says all monitoring/reporting obligations lapse when the commitments expire at final-report submission, while other substantive measures continue for five years. The decision's stated continuing-QA expectation and the annex's broader lapse wording should both be retained; this lane does not resolve their legal interpretation.
\nThat issue does not imply all independent oversight ends in 2027: the decision separately points to continuing DSA risk assessment and Article 37 independent audit obligations (recital 174). The specific trustee, internal QA and general DSA audit are different mechanisms.
\nRecital 9 (PDF 6) describes pre-decision changes to AliExpress' transparency tools, including research-access options and downloads. This is an official historical implementation account, not an audit that every annex obligation was satisfied. The trustee approval provides a further concrete implementation step. No actual application decision, data delivery, QA report, trustee report or Commission acceptance of performance was acquired.
\nFor breach, recital 180 expressly preserves reopening under Article 71(2). The decision does not impose an access-related monetary fine. Root's separate current-status reading identifies the 20 July 2026 EUR 550 million announcement as concerning illegal-product risk assessment/mitigation outside the settled grievances; it must not be used as proof of researcher-access breach or trustee failure. That later announcement, its full infringement decision, payment and appeal status are outside this worker's two-original acquisition scope.
\nThe highest-value next record is the first trustee report due in November 2025, together with the underlying B.V.7 researcher-access statistics/feedback and the Commission's response. These would test whether the process opened access, denied it and why, or triggered correction. Known holders are AliExpress, Ankura and the Commission. The internal researcher-assessment SOP and executed trustee mandate/fees are also specific gaps. No missing record is assumed nonexistent and no holder was contacted.
\nOriginals: ali-decision.pdf (78 pages, scoped read) and ali-trustee.pdf (one page, full text and visual read). Decision text scope: PDF 1–6, 44–51, 53–56, 64–66, 74–78. Decision images actually viewed: 65, 66, 74, 75, 77, 78. Other page-level keyword hits were only used to select these sections; a full-document extraction is not a full read. Both original GETs returned PDFs after the reader failed with Unicode decoding errors. No additional acquisition or retry followed. Exact metadata, hashes and ancestry are in ali-captures.json; no graph or packaging edits.
Root announcement/current-status lane closed,9September2026. The Commission's18June2025 announcement (updated20June2025) describes legally binding AliExpress commitments on researcher access to public data: automated access/scraping, a dedicated API, customised datasets on request and information for researchers. It also describes internal monitoring and reports to an independent trustee. This is an account of an exercised Article71 commitment decision, not a particular researcher approval or measured implementation outcome. The worker acquired the actual decision and trustee appointment, which root separately reviewed within the scope below.
\nThe current official platform index identifies AliExpress International (Netherlands) B.V., links the18June2025 decision and trustee record, and lists a20July2026 EUR550million fine. The index's content says updated31August2026 while the footer says7September2026; neither is the date of each legal event. Reading here was limited to AliExpress's actual entry, document links and those date fields. No full platform register audit.
\nThe fine has a distinct basis. The Commission's French representation announcement,20July2026, says it concerns inadequate assessment and mitigation of risks from illegal, dangerous or counterfeit products. It explicitly separates those grievances from the2025 commitments. The current record therefore cannot be presented as a fine for refusing researcher access, evidence of failed trustee oversight, or GDI-driven enforcement. It describes conduct lasting at least untilJune2025; it does not state all later conduct was identical. The requested action plan is due20October2026. The full infringement decision, any appeal and payment of the fine were not acquired. This is the Commission's published announcement, not an independent judicial finding or our reproduced product testing.
\nThe original presscorner URL linked by the index yielded zero reader lines and a34-character title-only derivative after ordinary GET. It is retained as an application shell, not as the press-release body. The same Commission's French representation supplied a substantive original; root read the complete article in French. The French URL's web reader separately returned a non-retryable unsafe-open error; the ordinary GET was already part of the same orchestration and succeeded. No further reader or fetch attempt followed. The public French page is the substantive source; shell bytes add no corroboration.
\nRoot supplemental decision reading, without reacquisition: ali-decision.pdf PDF1,48-51,54,64-66,74-78 text; PDF65 and78 visual. The initial combined text output truncated part ofPDF74; that page was reread separately. Ali-trustee.pdf complete one-page text and visual. All78decision pages are retained but not claimed read. These locators establish the15-business-day request assessment, customised-data7-business-day expectation with complex-development exception, access QA metrics, nominated/paid-but-Commission-approved trustee, notice-not-approval clause, reporting sunset and legal addressees. The appointment record names Ankura Consulting (Europe), Limited, approved23September2025. Its actual mandate, invoices, work plan and reports were not acquired.
\nA notable drafting issue should remain visible. Recital174 (PDF49) expects internal QA to continue through the five-year substantive term, while AnnexD (PDF78) says all monitoring/reporting obligations expire with the commitment period tied to the final trustee report. Article1 (PDF50) makes commitments binding for five years; AnnexD distinguishes the continued substantive measures from the reporting period. The report schedule and rationale limit the trustee mandate through the2027cycle; no finding that all independent oversight ends then follows, since the decision also points to ordinary Article37 audits. AnnexC9's impossible31November2027 date is a textual typo identified by the worker; recital167 andC12(iv) express two months after30September2027. Root did not separately visually inspect the typo page. The approved mandate and actual compliance practice are the useful next instruments, not an invented resolution of that tension.
\nNo public-data access delivery, individual research approval, current API operation or GDI/AI4TRUST recipient relationship was tested. Root acquisition ends here; findings are supplemental outside r31 and the fixed r32 data candidate.
", "path": "dsa-access-enforcement-wave-2026-09-09/root-enforcement-findings.md", "sourceIds": [ "w33-con-s-index", "w33-con-s-announcement", "w33-con-s-fine" ], "markdown": "Root announcement/current-status lane closed,9September2026. The Commission's18June2025 announcement (updated20June2025) describes legally binding AliExpress commitments on researcher access to public data: automated access/scraping, a dedicated API, customised datasets on request and information for researchers. It also describes internal monitoring and reports to an independent trustee. This is an account of an exercised Article71 commitment decision, not a particular researcher approval or measured implementation outcome. The worker acquired the actual decision and trustee appointment, which root separately reviewed within the scope below.\n\nThe current official platform index identifies AliExpress International (Netherlands) B.V., links the18June2025 decision and trustee record, and lists a20July2026 EUR550million fine. The index's content says updated31August2026 while the footer says7September2026; neither is the date of each legal event. Reading here was limited to AliExpress's actual entry, document links and those date fields. No full platform register audit.\n\nThe fine has a distinct basis. The Commission's French representation announcement,20July2026, says it concerns inadequate assessment and mitigation of risks from illegal, dangerous or counterfeit products. It explicitly separates those grievances from the2025 commitments. The current record therefore cannot be presented as a fine for refusing researcher access, evidence of failed trustee oversight, or GDI-driven enforcement. It describes conduct lasting at least untilJune2025; it does not state all later conduct was identical. The requested action plan is due20October2026. The full infringement decision, any appeal and payment of the fine were not acquired. This is the Commission's published announcement, not an independent judicial finding or our reproduced product testing.\n\nThe original presscorner URL linked by the index yielded zero reader lines and a34-character title-only derivative after ordinary GET. It is retained as an application shell, not as the press-release body. The same Commission's French representation supplied a substantive original; root read the complete article in French. The French URL's web reader separately returned a non-retryable unsafe-open error; the ordinary GET was already part of the same orchestration and succeeded. No further reader or fetch attempt followed. The public French page is the substantive source; shell bytes add no corroboration.\n\nRoot supplemental decision reading, without reacquisition: ali-decision.pdf PDF1,48-51,54,64-66,74-78 text; PDF65 and78 visual. The initial combined text output truncated part ofPDF74; that page was reread separately. Ali-trustee.pdf complete one-page text and visual. All78decision pages are retained but not claimed read. These locators establish the15-business-day request assessment, customised-data7-business-day expectation with complex-development exception, access QA metrics, nominated/paid-but-Commission-approved trustee, notice-not-approval clause, reporting sunset and legal addressees. The appointment record names Ankura Consulting (Europe), Limited, approved23September2025. Its actual mandate, invoices, work plan and reports were not acquired.\n\nA notable drafting issue should remain visible. Recital174 (PDF49) expects internal QA to continue through the five-year substantive term, while AnnexD (PDF78) says all monitoring/reporting obligations expire with the commitment period tied to the final trustee report. Article1 (PDF50) makes commitments binding for five years; AnnexD distinguishes the continued substantive measures from the reporting period. The report schedule and rationale limit the trustee mandate through the2027cycle; no finding that all independent oversight ends then follows, since the decision also points to ordinary Article37 audits. AnnexC9's impossible31November2027 date is a textual typo identified by the worker; recital167 andC12(iv) express two months after30September2027. Root did not separately visually inspect the typo page. The approved mandate and actual compliance practice are the useful next instruments, not an invented resolution of that tension.\n\nNo public-data access delivery, individual research approval, current API operation or GDI/AI4TRUST recipient relationship was tested. Root acquisition ends here; findings are supplemental outside r31 and the fixed r32 data candidate.\n", "file": "research/dsa-access-enforcement-wave-2026-09-09-root-enforcement-findings.md" }, { "id": "dsa-application-decisions-wave-2026-09-09-application-findings", "title": "Research applications: reported reasons, incomplete decision records", "description": "Closed 9 September 2026 after the actual GDI → DSA40 Collaboratory link and three substantive public records: the current aggregate tracker, the September 2025 policy paper, and the linked issue tracker. These reveal applicant-reported validity disputes and a current pending AliExpress entry. They do not publish an identifiable AliExpress application/respons", "html": "Closed 9 September 2026 after the actual GDI → DSA40 Collaboratory link and three substantive public records: the current aggregate tracker, the September 2025 policy paper, and the linked issue tracker. These reveal applicant-reported validity disputes and a current pending AliExpress entry. They do not publish an identifiable AliExpress application/response pair or establish whether a particular refusal was justified.
\n[Tracker Insights](http[local research file] displays “last updated: 8. September 2026”. It reports 70 complete applications registered through its voluntary tracker; 51 decided, comprising 24 accepted and 27 rejected. Those are recorded application statuses, not verified data deliveries. “Complete” concerns the tracker's recorded applications and does not independently establish that a platform received a legally sufficient submission.
\nThe platform table lists one AliExpress public-data application as open. It discloses no submission date, applicant, requested fields, response, paragraph, cure or subsequent outcome. No GDI identity is established. Public-data labelling alone is insufficient to reconstruct the actual legal channel used. This current entry cannot be matched to the confidential population in the separate FTI audit ending June 2025 or treated as a contradiction of that audit's report of no approved applications.
\nA narrow arithmetic check of the original HTML table gives 68 listed applications: 17 open, 27 rejected and 24 accepted. The decided totals match the headline; the table leaves two of the headline 70 unaccounted for. The page does not explain the difference. Do not silently add cases, infer missing outcomes or calculate population-wide success rates. Its narrative initially names fewer platforms than the displayed table; the table itself contains the AliExpress row. Original applications-tracker.html; full substantive text applications-tracker.derived.txt; exact table extraction applications-tracker-table0.json.
The same page's Reasons for Rejection section provides an unattributed platform-style excerpt concerning systemic-risk purpose limitation, plus the following category counts. These are the organiser's presentation of submitted information, not a bundle of publicly inspectable original platform letters.
\n| Reported category | \nCount | \nEvidentiary limit | \n
|---|---|---|
| Failure to confine research to the statutory purpose | \n5 | \nMost common reason within the reasons presented; not proof every proposal was outside scope. | \n
| Necessity and proportionality | \n2 | \nThe requested data, platform reasoning and balancing are absent. | \n
| Incompleteness | \n1 | \nMissing fields and any cure request are not identified. | \n
| Access request differed from application | \n1 | \nThe application/request pair is not published. | \n
| Required university affiliation | \n1 | \nThe organiser labels this an invalid reason during early implementation; no individual decision or adjudication is supplied. | \n
| Required EU location | \n2 | \nLikewise an organiser assessment, with no original correspondence or outcome of a challenge. | \n
These counts sum to 12, not all 27 reported refusals. The page does not identify which platforms or legal paragraphs produced each category, whether the categories overlap, or a complete coding frame. It separately says 27 entries provided application text and only 12 explicitly referred to Article 34. Those are different denominators and must not be equated with the 12 refusal-category counts. An explicit Article 34 citation is not by itself a demonstrated legal completeness test.
\nThe tracker mixes public and non-public requests. Its categories cannot all be assigned automatically to Article 40(12), or its non-public rows to completed Article 40(4) DSC proceedings. It reports platform decisions; it does not publish the original channel and chronology needed to classify each one. The displayed average decision delays are sample summaries and were not treated as statutory clocks or evidence of the AliExpress entry's duration.
\nThe actual [September 2025 policy paper](http[local research file] Data Access for Researchers under the Digital Services Act: From Policy to Practice, is Weizenbaum Policy Paper 14. The current publications index dates it 22 September 2025; the PDF cover and imprint give September 2025. It is not a fresh 2026 study of the tracker snapshot.
\nPDF 12–13 / printed 11–12 expressly discuss Article 40(12). Researchers apply through platform-specific forms or email, with provider or surrogate vetting. The authors report varying information requirements, refusals based on EU residence, and lengthy exchanges requiring progressively narrower purposes or data scope. Footnote 24 describes forms ranging from 8 questions for Snap to 55 for Meta, including requests for phone number or date of birth in some forms. These are the paper's dated accounts; this lane neither accessed those forms nor verified current requirements.
\nThe account distinguishes receiving access from receiving reliable data, reporting quality problems even after approval. It cites an earlier 2024 policy paper for the application obstacles rather than reproducing original letters. It proposes Commission guidance and codes of conduct; those are authors' proposed remedies, not an acquired cure procedure or a successful appeal. PDF 14–15 separately discuss privileged Article 40(4) access and data specificity; those legal explanations must not be used to recategorise the preceding Article 40(12) examples.
\nPDF 35 / printed 34 expressly describes the Collaboratory tracker as researcher self-reporting and the project as offering application support. The entry page openly recruits researchers and non-profit organisations to share requests, with the stated aims of reporting success, helping applications and informing regulation. This is a self-selected sample from an advocacy/support network, not a random sample or a census of platform applications. The inspected sources give no response rate or independently sampled comparison population. Original application texts are apparently available for a subset of the submissions, but their submission to the organiser is not public release or independent verification here.
\nPDF 2 says the paper draws on the transnational network built by the Collaboratory, led by some of its authors, and calls it a joint project funded by “Mercator Stiftung” since early 2024. The retained homepage/publications HTML references a funder logo asset named Stiftung-Mercator-Weiss-RGB.png, with empty alt text. The logo image was not separately acquired or visually read. These observations identify the disclosed foundation name; no exact grantor legal suffix, amount, award agreement, payment or reserved approval right was acquired.
The same page describes institutional funding of the Weizenbaum Institute by the German Federal Ministry of Research, Technology and Space and the State of Berlin. That institutional credit is not proof those bodies funded this particular study or determined the tracker's categories. No funder approval, sampling instruction, editorial control or access to applicant records is disclosed in the inspected scopes. The authors' leadership, the project's open recruitment and application-support mission do establish a closer organiser/study relationship than an independent audit of platform decisions. The paper also argues for more resources for researcher intermediaries (PDF 35–36); that is a stated policy position, not an awarded new resource.
\nThe linked [Issue Tracker](http[local research file] says submissions are checked by the organisers and may prompt follow-up. Its undated substantive body lists a past Meta application-submission interruption from December 2024 to April 2025 and a TikTok API data outage in January 2025. Its statement that it knows of no current critical issues is limited to the organisers' knowledge and this issue-list scope. It does not negate the separate application refusals or pending cases. No AliExpress refusal, cure, complaint, restored access or original response is published there. The linked external Meta page was not followed because it would expand beyond the selected three substantive records.
\nThe home and tracker pages each timed out in the ordinary web reader, then succeeded through one ordinary public GET. The linked April 2024 paper landing failed in the reader; its one GET alternative returned HTTP 200 BunkerWeb Bot Detection content at a challenge URL, requiring JavaScript. That is challenge HTML, not the paper. The route was closed without challenge completion, alternate mirror or DOI retry. The exact response, URL and hash remain in custody. No account, survey, form, contact, application or live platform test was used.
\nThe strongest rival to arbitrary obstruction is that at least some requests failed legitimate eligibility, specificity or proportionality requirements. Conversely, a platform's own label of incompleteness does not prove that its demands were necessary or its process offered a reasonable cure. The present aggregates cannot decide between those explanations for AliExpress or any other individual case.
\nThe most useful next record is an applicant-consented, dated and appropriately redacted application plus platform response, follow-up/cure exchange and final outcome, with the exact Article 40 paragraph and dataset request. A link from that pair to the tracker's coding record would allow a real validity assessment. The Collaboratory/applicant would be expected holders for reported submissions and the platform for its decision record; no release or request is assumed. The root's separate audit evidence remains a different period and source population. This lane closes without a finding that GDI submitted the tracked AliExpress request, that access was delivered, or that a refusal violated the DSA.
\nText reading: full substantive home and tracker bodies; current publications index entries/links; policy PDF 1–3, 12–15, 35–36 and 38 (plus the initial reader's introductory/context scope through PDF 12, not a full-PDF read); full substantive Issue Tracker body. Full policy extraction is retained for root review but is not a claim every page was read. No images/charts were visually inspected; the tracker table was read from actual HTML text. Primary original hashes, exact paths and transport/read scopes are in applications-captures.json.
9 September2026. Root read the complete substantive retained tracker page and its HTML table row/rowspan structure, the entry-page body, and the short issue-tracker body. Policy Paper14 (September2025) was text-read at PDF1-5,12-13,35-36 and38 only; no full38page reading or visual audit. The root webreader timed out on the tracker; already acquired worker original supplied the substantive reading. No fresh duplicate GET, account, survey submission or personal request records.
\nThe tracker updates to8September2026 and reports70 complete registered applications,51 decisions,24 acceptances and27 rejections. The HTML table actually sums to68:24 accepted,27 rejected,17 open. Its public/nonpublic categories sum to60 and8 respectively. The homepage reports61 public-data applications and70 overall; neither exactly matches the table. No explanation for the two-record difference was acquired. Treat the denominator as unreconciled; do not silently repair it, infer fraud, or compute a whole-platform acceptance rate. These are voluntarily reported applications, not an inventory of all requests to platforms.
\nIts AliExpress row is one public-data entry marked open. The page supplies no application date, applicant identity, precise statutory paragraph, original text or correspondence for that entry. It cannot be matched to the prior FTI audit period or to GDI. More generally, requested public/nonpublic data is not enough to establish the legal route: Article40(4) can also be used for some public data. The policy paper separately discusses40(12) and40(4); it does not furnish individual tracker routing.
\nThe rejection section offers an unattributed purpose-limitation quotation and category counts. Purpose limitation5, necessity/proportionality2, incomplete1, request/application mismatch1, university-affiliation1, EU-location2 sum to12, not27 total rejections; coverage and possible overlap are unspecified. The organisers label the latter affiliation/location reasons invalid. That is their legal assessment, not an acquired ruling or visible application/response pair. Prior accepted legal text supports distinguishing eligibility criteria from blanket affiliation/location exclusions; it cannot decide an unacquired individual application. Only27 entries supplied application text, and12 explicitly mentioned Article34; neither this denominator nor listed methods/risks describes all70 registrations. Categories can overlap and are not demonstrated decision causes.
\nThe issue page's separate statements concern practical availability and quality. It currently reports no known critical issues, while listing historical Meta submission unavailability and TikTok API missing data. These are organiser-verified issue accounts with no supporting incident logs read here. Lack of a current listed critical issue does not mean every application succeeds or every dataset is complete. Authorisation, technically usable delivery and researcher judgment remain separate steps.
\nPolicyPDF2 attributes this joint Collaboratory project to Mercator Stiftung funding and says some paper authors lead it. The Weizenbaum Institute's separate institutional funding statement names the German federal ministry (BMFTR) and State of Berlin. A project grant and host-institution support are different resources; no amounts, donor conditions or reviewer rights are disclosed in these pages. The title-page author list and current homepage team are different dated roles and should not be collapsed into identical governance.
\nPDF35 explicitly describes researcher self-reporting, application support and emerging intermediaries; it advocates active EU/member-state support for such roles. This is a substantive institutional connection: the same initiative gathers reports, helps applicants, and informs regulatory discussion. It makes the dataset valuable for finding failure cases and clarifying process, while also requiring attention to self-selection and advocacy context. It is neither a regulator's adjudication log nor inherently disqualified by its advocacy. The page identifies other networks as examples, without proving funding, control or case-specific coordination among them.
\nThis run still lacks the desired public application/response pair. The acquired records narrow the validity question to identifiable tests and separate evidentiary layers. The next acquisition must follow the forest review after this packet closes; no expansion to all listed networks is authorized by this note alone.
\nRoot continuation also text-read PDF12-13 (provider or surrogate vetting; dated reported application obstacles, remedies proposed),36 (advocacy for sustained public funding) and38 (imprint and CC BY4.0). The original application/response bundle remains unacquired. Explicit table arithmetic corrected an initial unaccepted root assumption that the homepage public-data count matched the table; actual60public/8nonpublic are preserved above. This was a root synthesis correction, not a change to source data or the worker's correct68/70 finding.
", "path": "dsa-application-decisions-wave-2026-09-09/root-reading.md", "sourceIds": [], "markdown": "9 September2026. Root read the complete substantive retained tracker page and its HTML table row/rowspan structure, the entry-page body, and the short issue-tracker body. Policy Paper14 (September2025) was text-read at PDF1-5,12-13,35-36 and38 only; no full38page reading or visual audit. The root webreader timed out on the tracker; already acquired worker original supplied the substantive reading. No fresh duplicate GET, account, survey submission or personal request records.\n\nThe tracker updates to8September2026 and reports70 complete registered applications,51 decisions,24 acceptances and27 rejections. The HTML table actually sums to68:24 accepted,27 rejected,17 open. Its public/nonpublic categories sum to60 and8 respectively. The homepage reports61 public-data applications and70 overall; neither exactly matches the table. No explanation for the two-record difference was acquired. Treat the denominator as unreconciled; do not silently repair it, infer fraud, or compute a whole-platform acceptance rate. These are voluntarily reported applications, not an inventory of all requests to platforms.\n\nIts AliExpress row is one public-data entry marked open. The page supplies no application date, applicant identity, precise statutory paragraph, original text or correspondence for that entry. It cannot be matched to the prior FTI audit period or to GDI. More generally, requested public/nonpublic data is not enough to establish the legal route: Article40(4) can also be used for some public data. The policy paper separately discusses40(12) and40(4); it does not furnish individual tracker routing.\n\nThe rejection section offers an unattributed purpose-limitation quotation and category counts. Purpose limitation5, necessity/proportionality2, incomplete1, request/application mismatch1, university-affiliation1, EU-location2 sum to12, not27 total rejections; coverage and possible overlap are unspecified. The organisers label the latter affiliation/location reasons invalid. That is their legal assessment, not an acquired ruling or visible application/response pair. Prior accepted legal text supports distinguishing eligibility criteria from blanket affiliation/location exclusions; it cannot decide an unacquired individual application. Only27 entries supplied application text, and12 explicitly mentioned Article34; neither this denominator nor listed methods/risks describes all70 registrations. Categories can overlap and are not demonstrated decision causes.\n\nThe issue page's separate statements concern practical availability and quality. It currently reports no known critical issues, while listing historical Meta submission unavailability and TikTok API missing data. These are organiser-verified issue accounts with no supporting incident logs read here. Lack of a current listed critical issue does not mean every application succeeds or every dataset is complete. Authorisation, technically usable delivery and researcher judgment remain separate steps.\n\nPolicyPDF2 attributes this joint Collaboratory project to Mercator Stiftung funding and says some paper authors lead it. The Weizenbaum Institute's separate institutional funding statement names the German federal ministry (BMFTR) and State of Berlin. A project grant and host-institution support are different resources; no amounts, donor conditions or reviewer rights are disclosed in these pages. The title-page author list and current homepage team are different dated roles and should not be collapsed into identical governance.\n\nPDF35 explicitly describes researcher self-reporting, application support and emerging intermediaries; it advocates active EU/member-state support for such roles. This is a substantive institutional connection: the same initiative gathers reports, helps applicants, and informs regulatory discussion. It makes the dataset valuable for finding failure cases and clarifying process, while also requiring attention to self-selection and advocacy context. It is neither a regulator's adjudication log nor inherently disqualified by its advocacy. The page identifies other networks as examples, without proving funding, control or case-specific coordination among them.\n\nThis run still lacks the desired public application/response pair. The acquired records narrow the validity question to identifiable tests and separate evidentiary layers. The next acquisition must follow the forest review after this packet closes; no expansion to all listed networks is authorized by this note alone.\n\nRoot continuation also text-read PDF12-13 (provider or surrogate vetting; dated reported application obstacles, remedies proposed),36 (advocacy for sustained public funding) and38 (imprint and CC BY4.0). The original application/response bundle remains unacquired. Explicit table arithmetic corrected an initial unaccepted root assumption that the homepage public-data count matched the table; actual60public/8nonpublic are preserved above. This was a root synthesis correction, not a change to source data or the worker's correct68/70 finding.\n", "file": "research/dsa-application-decisions-wave-2026-09-09-root-reading.md" }, { "id": "dsa-research-access-gates-wave-2026-09-09-gates-findings", "title": "DSA research access: the decision-maker is identifiable; a GDI approval is not established", "description": "Bounded official-source check closed on 9 September 2026. Two original HTML responses retained: one substantive Commission FAQ and the public portal's application shell. No login, application, data query or live platform experiment. Root independently owns the adopted Article 40 and delegated-rule reading; this lane reports explanatory guidance and the actua", "html": "Bounded official-source check closed on 9 September 2026. Two original HTML responses retained: one substantive Commission FAQ and the public portal's application shell. No login, application, data query or live platform experiment. Root independently owns the adopted Article 40 and delegated-rule reading; this lane reports explanatory guidance and the actual public-record boundary.
\nThe [Joint Research Centre / European Centre for Algorithmic Transparency FAQ](http[local research file] published 3 July 2025, shows an update of 29 October 2025. It distinguishes public-data access under Article 40(12) from vetted non-public access under 40(4). Both require commercial independence, disclosed research funding, proportionate research and security/privacy safeguards; vetting adds research affiliation and public, free results. Locators: questions on access/conditions, reader lines 33–52.
\nThe DSC where the platform provider is established makes the final decision and issues a reasoned request. The researcher's home DSC may assess and forward an application (90–96). The FAQ describes an 80-working-day response, with explained extension (139–145); this is not a guaranteed data-delivery date. The establishment DSC specifies access conditions case by case, potentially using a secure environment rather than transferring data (130–138). A provider may seek amendments or mediation, but the DSC retains the decision (154–167). Public-data refusal feedback can inform Commission systemic enforcement; the FAQ says it will not opine on individual cases (80–84).
\nThese are described powers and conditions, not an exercised decision concerning GDI. The Commission's explanatory text should be reconciled with root's operative-law reading rather than substituted for it. One question heading says the delegated act was adopted in June 2025, while the body dates adoption to July; no June date is adopted here.
\nThe FAQ's question on other researchers' outcomes says public overviews of reasoned requests and amendments are intended to appear in the [DSA data access portal](http[local research file] (149–150). This is narrower than a promise that every application, rejection or private exchange is public. Root's independent reading of delegated Regulation 2025/2050 Article 11 identifies topic, provider, requested data and access modality as overview fields; researcher and funder names are not mandatory fields. An exact-name search is consequently a particularly weak exclusion test. This legal detail is cross-lane input, not an additional legal-text read by this worker.
\nThe FAQ's actual link opened /home. The official reader returned zero lines. One ordinary GET returned HTTP 200 and 45,880 bytes of an application shell, with only the title in text extraction and no anchor links. No overview rows were inspected. No application index was enumerated, searched through a logged-in interface, exported or reverse-engineered. A retained shell is not evidence that the portal has no records, or that its intended public summaries require login. This transport branch was closed after the two ordinary methods.
No publicly identifiable official approval or reasoned request for GDI or AI4TRUST was acquired. That is an evidence limit, not a finding that neither has applied, qualified or obtained data. The exact indexed-name checks were:
\nsite:data-access.dsa.ec.europa.eu "GDI"; site:data-access.dsa.ec.europa.eu "AI4TRUST"; site:data-access.dsa.ec.europa.eu "Disinformation Index"; "Global Disinformation Index" "vetted" "approved"; site:cnam.ie "AI4TRUST"; site:cnam.ie "Global Disinformation Index"; and site:bundesnetzagentur.de "GDI" "researcher".
These returned no matching official named request/approval in the displayed results. The broader query returned historical Commission policy references and commentary, neither of which is a vetting record. Earlier discovery queries and the search boundary are recorded in the manifest. Search indexing is incomplete; a request may name a principal researcher, research institution or project variant rather than either searched label. This was not a search of every DSC or a complete public register.
\nThe already accepted GDI retrospective describes a 2024 Article 40 application and refusal. It does not supply the original application, a DSC determination or a paragraph-specific legal route. Its date alone must not be recast as an application under the mechanism that the FAQ says became operational on 29 October 2025. Conversely, this run does not invalidate that historical self-report. The exact procedural channel remains a missing fact.
\nHorizon participation and financial support do not constitute a research-access determination. Nor do any previously documented platform data collections establish which legal channel supplied them. The selected sources do not establish trusted-flagger status for GDI or AI4TRUST; no inference between that separate status and researcher access is made. Commercial independence is an application condition requiring assessment, not a finding in this packet that a particular legal entity satisfies or fails it.
\nAn issued reasoned-request overview, matched to this research through provider, topic, data and modality plus corroborating applicant or decision records, would establish a concrete authority step. The public overview need not name the applicant, so a name match alone cannot be the retrieval criterion. The associated vetting decision and specified access conditions would show what was authorized. Actual provider fulfilment or delivery confirmation would be a further step. A grant record or a statement that research used platform data cannot replace this sequence.
\nKnown expected holders are the establishment DSC, the principal researcher and the provider. The portal is the stated publication channel for reasoned-request overviews. For the reported 2024 refusal, the application and platform response remain the direct records needed to distinguish public-data access, another platform scheme and a government-vetted request. No holder was contacted, and no next branch was opened.
\ngates-faq.html and gates-faq.derived.txt: full substantive FAQ reader-read, questions from DSA introduction through stakeholder input, plus publication/update fields (official-reader lines 3–184). Original HTML retained; its links and date/footer fields inspected locally. This is a 29 October 2025 updated explanation captured on 9 September 2026, not a complete current enforcement-case inventory.
gates-portal.html and gates-portal.derived.txt: application-shell response and title only. No substantive public request index read. gates-transport.json preserves actual requests, timestamps, statuses and hashes. gates-captures.json states each source's limits and the exact search coverage. No original decision, reasoned request, refusal correspondence or data-delivery receipt was acquired.
9 September2026. Root legal-text lane closed. Two adopted EU instruments and their expressly incorporated research-organisation definition were acquired from EUR-Lex. These are powers and conditions in enacted text, not decisions about GDI. Both regulation pages show in-force status; this is not a review of every related national measure or later enforcement case.
\nUnder Regulation2022/2065 Article40(4),(8)-(11), the Digital Services Coordinator of the platform provider's establishment grants vetted status for the specific research and issues the reasoned access request. A researcher's home coordinator may perform an initial assessment but does not acquire the final decision. The awarding coordinator can investigate and terminate access if eligibility lapses, after allowing the researcher to respond. Grant participation does not itself confer this status.
\nArticle40(8) requires research affiliation, independence from commercial interests, disclosure of research funding, appropriate security/confidentiality, necessity/proportionality, the specified systemic-risk research purpose and free public results. The incorporated Directive2019/790 Article2(1) defines a scientific-research organisation acting without profit or reinvesting profits, or under a recognised public-interest mission, without preferential results access for an undertaking exercising decisive influence. Its recital12 explains structural control as an exclusion example. This is not an acquired assessment of GDI's eligibility. A licensing business alone does not prove failure of these conditions, nor does government or foundation support prove independence. The stated commercial-independence condition is not a general certification of political neutrality.
\nDelegated Regulation2025/2050 Article8 and recital12 make the application a particularly revealing expected record. The recital calls for funder identity, amount, nature/duration, awarded versus pending status and, where available, funder evaluations. The operative text requires information about research funding and a declaration of commercial independence relevant to the specific project. We have acquired the required categories, not any applicant's declarations or assessor's reasons.
\nThe coordinator also sets technical, legal and organisational access conditions (Article9). A secure environment can constrain which data researchers see and log processing; the law does not always require transfer of raw data. Article10 requires access-start deadline and termination date. Article15 requires providers to report actual provision/termination within three working days and supply needed documentation. It constrains extra data-management or standard-tool restrictions unless specified in the reasoned request, and additional personal-data conditions outside that request. These are case-specific rules, not a general requirement to integrate any nominated detection product. No approved GDI tool, grant-to-API entitlement or current AI4TRUST connection is shown.
\nArticle7 sets80 working days to issue a reasoned request or explain refusal, with reasoned extension. It is neither GDI's requested maximum60-day process nor a guaranteed data-delivery deadline. Under base Article40(5)-(6), providers can seek amendment within15 days on stated access/security/confidentiality grounds and the coordinator decides within15 days. Delegated Articles12-14 specify assessment, optional mediation with judicial rights preserved, and independent experts. Experts must declare no financial/personal ties to provider or applicant, no outcome interest and no conflict; their opinions are recorded in AGORA. This does not prove any expert has been retained in a named case.
\nThe public record is narrower than the application file. Article11 requires an overview of an issued reasoned request and its changes, containing topic, provider, data description and access modalities. It does not require every applicant, rejected application, funding declaration or reviewer record to appear publicly, and does not require researcher/funder names in that overview. Exact-name failure is therefore not a reliable absence test. Authorised access and actual delivery also leave different records: an overview can show the former, Article15 notification the latter.
\nTwo other legal channels must remain distinct. Article40(12) covers publicly accessible platform data for researchers satisfying conditions8(b)-(e), without the paragraph4 project-specific vetted route. Article22 trusted-flagger status separately gives priority processing of illegal-content notices within an expertise area, with reporting, suspension and revocation rules. It does not make each notice a binding removal command. No status under either channel is inferred for GDI.
\nThe reused GDI20April2026 account says Article40 generically when describing2024 platform applications/refusals. Without original applications and replies, this case cannot assign the exact paragraph or treat them as coordinator vetting decisions. Its calls for a60-day process, standard sharing protocol and mandatory approved-tool API compatibility remain requested changes. The adopted law permits meaningful compelled research access under specified conditions; it does not establish that this organisation obtained it or that any classification controls a commercial or state decision.
\nSources and root read scope: dsa2022.html, original2022/2065, complete Articles40 and22 only; access2025.html, original2025/2050, complete Articles1-16 plus selected recitals11,12,14,17,21,24, title/OJ date/status, not whole preamble or Annex; research-definition2019.html, original2019/790, Article2(1) and recital12, other Article2 definitions only contextual. Complete source bytes retained does not mean complete documents read. Selected derivatives and root-transport.json preserve scope and custody. Reused horizon-gdi2026.derived.txt selected account/recommendations read in this run; its original/read history stays in the accepted Horizon packet. No new source independence. Root acquisition ends here.
", "path": "dsa-research-access-gates-wave-2026-09-09/root-law-findings.md", "sourceIds": [ "w33-con-s-dsa", "w33-con-s-delegated", "w33-con-s-definition" ], "markdown": "9 September2026. Root legal-text lane closed. Two adopted EU instruments and their expressly incorporated research-organisation definition were acquired from EUR-Lex. These are powers and conditions in enacted text, not decisions about GDI. Both regulation pages show in-force status; this is not a review of every related national measure or later enforcement case.\n\nUnder Regulation2022/2065 Article40(4),(8)-(11), the Digital Services Coordinator of the platform provider's establishment grants vetted status for the specific research and issues the reasoned access request. A researcher's home coordinator may perform an initial assessment but does not acquire the final decision. The awarding coordinator can investigate and terminate access if eligibility lapses, after allowing the researcher to respond. Grant participation does not itself confer this status.\n\nArticle40(8) requires research affiliation, independence from commercial interests, disclosure of research funding, appropriate security/confidentiality, necessity/proportionality, the specified systemic-risk research purpose and free public results. The incorporated Directive2019/790 Article2(1) defines a scientific-research organisation acting without profit or reinvesting profits, or under a recognised public-interest mission, without preferential results access for an undertaking exercising decisive influence. Its recital12 explains structural control as an exclusion example. This is not an acquired assessment of GDI's eligibility. A licensing business alone does not prove failure of these conditions, nor does government or foundation support prove independence. The stated commercial-independence condition is not a general certification of political neutrality.\n\nDelegated Regulation2025/2050 Article8 and recital12 make the application a particularly revealing expected record. The recital calls for funder identity, amount, nature/duration, awarded versus pending status and, where available, funder evaluations. The operative text requires information about research funding and a declaration of commercial independence relevant to the specific project. We have acquired the required categories, not any applicant's declarations or assessor's reasons.\n\nThe coordinator also sets technical, legal and organisational access conditions (Article9). A secure environment can constrain which data researchers see and log processing; the law does not always require transfer of raw data. Article10 requires access-start deadline and termination date. Article15 requires providers to report actual provision/termination within three working days and supply needed documentation. It constrains extra data-management or standard-tool restrictions unless specified in the reasoned request, and additional personal-data conditions outside that request. These are case-specific rules, not a general requirement to integrate any nominated detection product. No approved GDI tool, grant-to-API entitlement or current AI4TRUST connection is shown.\n\nArticle7 sets80 working days to issue a reasoned request or explain refusal, with reasoned extension. It is neither GDI's requested maximum60-day process nor a guaranteed data-delivery deadline. Under base Article40(5)-(6), providers can seek amendment within15 days on stated access/security/confidentiality grounds and the coordinator decides within15 days. Delegated Articles12-14 specify assessment, optional mediation with judicial rights preserved, and independent experts. Experts must declare no financial/personal ties to provider or applicant, no outcome interest and no conflict; their opinions are recorded in AGORA. This does not prove any expert has been retained in a named case.\n\nThe public record is narrower than the application file. Article11 requires an overview of an issued reasoned request and its changes, containing topic, provider, data description and access modalities. It does not require every applicant, rejected application, funding declaration or reviewer record to appear publicly, and does not require researcher/funder names in that overview. Exact-name failure is therefore not a reliable absence test. Authorised access and actual delivery also leave different records: an overview can show the former, Article15 notification the latter.\n\nTwo other legal channels must remain distinct. Article40(12) covers publicly accessible platform data for researchers satisfying conditions8(b)-(e), without the paragraph4 project-specific vetted route. Article22 trusted-flagger status separately gives priority processing of illegal-content notices within an expertise area, with reporting, suspension and revocation rules. It does not make each notice a binding removal command. No status under either channel is inferred for GDI.\n\nThe reused GDI20April2026 account says Article40 generically when describing2024 platform applications/refusals. Without original applications and replies, this case cannot assign the exact paragraph or treat them as coordinator vetting decisions. Its calls for a60-day process, standard sharing protocol and mandatory approved-tool API compatibility remain requested changes. The adopted law permits meaningful compelled research access under specified conditions; it does not establish that this organisation obtained it or that any classification controls a commercial or state decision.\n\nSources and root read scope: dsa2022.html, original2022/2065, complete Articles40 and22 only; access2025.html, original2025/2050, complete Articles1-16 plus selected recitals11,12,14,17,21,24, title/OJ date/status, not whole preamble or Annex; research-definition2019.html, original2019/790, Article2(1) and recital12, other Article2 definitions only contextual. Complete source bytes retained does not mean complete documents read. Selected derivatives and root-transport.json preserve scope and custody. Reused horizon-gdi2026.derived.txt selected account/recommendations read in this run; its original/read history stays in the accepted Horizon packet. No new source independence. Root acquisition ends here.\n", "file": "research/dsa-research-access-gates-wave-2026-09-09-root-law-findings.md" }, { "id": "duty-care-wave-2026-09-09-draft-access", "title": "Digital Duty of Care: draft access and missing-text claims", "description": "Observed 9 September 2026. The latest entry in forest-reviews.md was read before acquisition. This is a bounded status inquiry, not an assessment of the merits or legal effect of the proposed powers. Only this file is written by this lane.", "html": "Observed 9 September 2026. The latest entry in forest-reviews.md was read before acquisition. This is a bounded status inquiry, not an assessment of the merits or legal effect of the proposed powers. Only this file is written by this lane.
The 8 September exposure draft is publicly available through the department, despite the government's description of the consultation as targeted. Its readable PDF contains six explicit unfinished-outline placeholders. The acquired record does not establish that a more complete bill was given only to selected consultees, that provisions were censored, or why any undisclosed material might have been withheld. Those claims remain unverified. This finding does not resolve whether the proposed powers could restrict speech; the operative-authority lane addresses that separate question.
\nThe department's publication page is dated 8 September 2026. It links a DOCX advertised as 184.44 KB and a PDF as 636.4 KB, and gives a public feedback address, digitaldutyofcare@communications.gov.au, with a deadline of 12pm Tuesday 22 September. No timezone is specified there. The page presents no invitation or login requirement for reading the draft and does not limit its feedback sentence to named recipients. This proves an exposed document and feedback route, not a guarantee that every submission will receive equal consideration. The page links no bill-specific explanatory memorandum, draft rules, separate invitation terms or recipient list. Absence from that page is not proof those documents exist and are withheld. [Department publication page](http[local research file]
The ministerial announcement calls the consultation targeted, specifies digital platforms, industry bodies, civil-society organisations and advocates as groups whose feedback is sought, and says parliamentary introduction is intended this year. These are categories, not a named invitation roster. [8 September announcement, opening and consultation paragraph](http[local research file]
\nThe official press-conference transcript independently confirms the intended public access. In response to a journalist who had not yet seen the draft, Anika Wells said: “So, the exposure draft is now out. You can peruse it online.” Wayne Holdsworth separately said he had read it twice that morning. That is one attributed account of having read the text, not proof of an exclusive advance version. Other journalists and the Prime Minister spoke of people not yet having seen it; the transcript supplies no document-distribution time or exclusive-access rule. It gives no explanation of unfinished outlines, redaction, or withheld draft clauses. [Official transcript, 8 September, Holdsworth opening remarks and Wells answer beginning “So, the exposure draft”](http[local research file]
\nThe exposed PDF has 75 physical pages, including front matter, and 71 numbered text pages. Its cover identifies an exposure draft dated 08/09/2026. The following are explicit “to be drafted” markers in the readable text; each is a simplified-outline section. The marker itself does not identify withheld substantive wording. This classification follows the headings and words actually published, not an assumption about unpublished drafting history.
\n| Proposed provision | \nLocation, physical PDF page / printed page | \nPublished heading | \n
|---|---|---|
| 86A | \n7 / 3 | \nSimplified outline of Part 6A, fake nude material | \n
| 4 | \n27 / 23 | \nSimplified outline of the Act | \n
| 25 | \n29 / 25 | \nSimplified outline of Part 2, digital duty of care | \n
| 166 | \n47 / 43 | \nSimplified outline of Part 11, eSafety Commissioner | \n
| 192A | \n52 / 48 | \nSimplified outline of Part 13 | \n
| 205A | \n63 / 59 | \nSimplified outline of Part 14, research and data access | \n
The acquired draft text does not label these markers redactions. No official explanation beyond their unfinished-drafting label was found in the inspected publication page or press conference. Do not rename them censorship or claim their unseen wording. [Public exposure-draft PDF](http[local research file]
\nThis is not a full visual redaction audit: neither original bytes nor a usable page image was acquired by this lane. Text search can find declared placeholders but cannot exclude every possible graphical redaction, omitted earlier version, or unpublished attachment.
\nThe department also publishes the May 2026 issues paper, with a publication-page date of 26 May, and links a three-page summary of May–June consultation outcomes. These are policy-development materials, not an explanatory memorandum accompanying the September bill. The summary reports more than 40 consultation meetings across described stakeholder categories. It does not supply a September draft-recipient roster. [Issues paper and summary publication page](http[local research file] [consultation-outcomes summary, page 1 introduction](http[local research file]
\nA distinct, explicit non-publication rule appears on the closed 14 November–7 December 2025 public survey page. Under its privacy collection notice, individual completed surveys and participants' personal information are not to be published; third-party personal-information disclosure needs legal authorization or requirement. Aggregated results may be published, with contact and identifying details redacted. The page also reserves a right not to consider responses that create legal exposure or contain abusive, offensive or defamatory material. Those are stated rules for survey responses, not secrecy terms for the 2026 draft. The old page's closed label and registration requirement must not be applied to the September exposure-draft feedback route. [2025 survey page, Participate and Privacy Collection Notice](http[local research file]
\nA bounded official-domain search also surfaced FOI 26-276, Document 6, Attachment C, described as a summary of late-2025 consultation activities and organizational participants. Only the search-returned passage was inspected in this lane; the underlying release was not acquired here. It may identify earlier consultees but cannot establish who received September 2026 draft text. This lead was sent to the genealogy lane without assuming continuity of the roster. [FOI release, Attachment C target](http[local research file]
\nThe ministerial announcement and department publication pages were read in the public web reader. The press-conference transcript was read in full. The current bill was read selectively for its cover, contents, placeholders and their surrounding headings, using the web reader's 75-page rendering. The outcomes summary was read for its introduction and accessible thematic text; it was not used to attribute a named organization's influence. The 2025 survey's participation and privacy terms were read in full.
\nAn ordinary Python GET to the department publication page timed out after 30 seconds before any draft bytes were obtained. The DOCX link yielded an internal error in the web reader. Screenshot attempts returned references without visible image content, with two cache misses. Those are acquisition limitations, not evidence that the government denies public access. The root's separate ordinary PowerShell substitute also timed out; that branch is closed. No local original, original-file hash, DOCX/PDF equivalence check or complete visual examination is claimed. No account, invitation, access control or publication restriction was bypassed.
\nThe precise unresolved record family is the September consultation invitation and distribution record, together with any covering instructions, confidentiality terms, attachment/version list and later draft comparison. The department's Digital Duty of Care policy team is the likely custodian. If an allegedly censored page or different draft is later supplied, compare it against the dated public version before attributing removal or a motive. No request or outreach was made. This lane stops at the verified public-exposure instrument; further acquisition waits for the next whole-case review.
\nThe user clarified a materially different allegation: large technology companies supposedly received less skeletal versions before the public learned the intended detail. The user identifies Isaac Butterfield as the source they heard, but no exact title, URL, timestamp or quotation was supplied or identified in this bounded pass. This is an attributed user lead, not a verified quotation from Butterfield. Unrelated older appearances were not substituted for the missing clip.
\nPublic availability does not resolve selective advance access. The allegation remains OPEN. Three different claims require different records: earlier circulation of the same text needs a dated covering message and identifiable attachment; circulation of a fuller version needs both versions or a reliable recipient account specifying the extra text; private explanation of an otherwise shared draft needs briefing material, notes or an attributed account. Neither public draft access nor participation in an earlier policy meeting proves or disproves those events. The unfinished outlines listed above likewise cannot establish what companies received.
\nA Senate primary-source search excerpt exposes a relevant historical entry in the Orders for the production of documents status table, No. 169—Digital duty of care, printed page 43. The order, moved by Senator Payman and agreed on 4 September 2025, sought minister/office correspondence and briefing/meeting records with eSafety and the department concerning preparation of a legislated duty. The entry records a response on 7 October 2025, identifying 26 responsive documents, claiming public-interest immunity for 20, and supplying the remaining documents with redactions on public-interest grounds. Six is the derived remainder, 26 minus 20, not a separately printed count. The stated immunity reason concerns Cabinet deliberations and prejudice to informed government consideration of ongoing policy matters. The entry records partial compliance. [Senate OPD table, order 169](http[local research file]
\nThis is evidence of a recorded 2025 preparation-document withholding claim and redacted production, distinct from the survey privacy policy above. It does not identify a complete September 2026 draft, establish its delivery to a company, or identify additional clauses seen only by industry. The underlying ministerial response and individual documents were not acquired. Reading scope is the indexed primary entry only: two targeted searches returned the same No.169 passage with its dates, counts, stated reason and printed page. A direct reader open failed; the root also reported a 403 on its direct open. No full-table review, original capture/hash or later compliance-history claim is made.
\nTargeted public searches for current draft circulation, confidentiality, redactions, advance access and relevant industry accounts did not produce a paired version, September invitation/covering letter, or recipient account specifying extra clauses. That limited search result is not evidence the allegation is false. The official press-conference account of Holdsworth reading the text that morning identifies neither a company recipient nor a distinct version.
\nButterfield's official website was inspected for its current video links. The linked YouTube channel returned a cache miss through the web reader. The single browser substitute then denied access because it could not verify the admin-enforced policy for youtube.com. No browser security control was bypassed, no account was used, and no indirect transcript, mirror or alternate YouTube route was pursued after that denial. The exact clip and its underlying source remain unexamined. [Official site used to identify the channel link](http[local research file]
\nThe present outcome is therefore narrower than either dismissal or confirmation: a public September draft; explicit incomplete outlines; a separate indexed parliamentary record of withheld/redacted preparation documents in 2025; and an unresolved allegation of selective access to fuller 2026 text. The next discriminating evidence would be the exact clip/quotation and its cited document, or the recipient covering letter plus attachment/version comparison. No outreach or further acquisition follows. The current run returns to the whole-case assessment with the selective-version claim open.
", "path": "duty-care-wave-2026-09-09/draft-access.md", "sourceIds": [ "w10-duty-s-publication", "w10-duty-s-announcement", "w10-duty-s-opd169" ], "markdown": "Observed 9 September 2026. The latest entry in `forest-reviews.md` was read before acquisition. This is a bounded status inquiry, not an assessment of the merits or legal effect of the proposed powers. Only this file is written by this lane.\n\n**The 8 September exposure draft is publicly available through the department, despite the government's description of the consultation as targeted.** Its readable PDF contains six explicit unfinished-outline placeholders. The acquired record does not establish that a more complete bill was given only to selected consultees, that provisions were censored, or why any undisclosed material might have been withheld. Those claims remain unverified. This finding does not resolve whether the proposed powers could restrict speech; the operative-authority lane addresses that separate question.\n\n## What is publicly accessible now\n\nThe department's publication page is dated 8 September 2026. It links a DOCX advertised as 184.44 KB and a PDF as 636.4 KB, and gives a public feedback address, `digitaldutyofcare@communications.gov.au`, with a deadline of **12pm Tuesday 22 September**. No timezone is specified there. The page presents no invitation or login requirement for reading the draft and does not limit its feedback sentence to named recipients. This proves an exposed document and feedback route, not a guarantee that every submission will receive equal consideration. The page links no bill-specific explanatory memorandum, draft rules, separate invitation terms or recipient list. Absence from that page is not proof those documents exist and are withheld. Department publication page (see source references)\n\nThe ministerial announcement calls the consultation targeted, specifies digital platforms, industry bodies, civil-society organisations and advocates as groups whose feedback is sought, and says parliamentary introduction is intended this year. These are categories, not a named invitation roster. 8 September announcement, opening and consultation paragraph (see source references)\n\nThe official press-conference transcript independently confirms the intended public access. In response to a journalist who had not yet seen the draft, Anika Wells said: **“So, the exposure draft is now out. You can peruse it online.”** Wayne Holdsworth separately said he had read it twice that morning. That is one attributed account of having read the text, not proof of an exclusive advance version. Other journalists and the Prime Minister spoke of people not yet having seen it; the transcript supplies no document-distribution time or exclusive-access rule. It gives no explanation of unfinished outlines, redaction, or withheld draft clauses. Official transcript, 8 September, Holdsworth opening remarks and Wells answer beginning “So, the exposure draft” (see source references)\n\n## What the public draft leaves unfinished\n\nThe exposed PDF has 75 physical pages, including front matter, and 71 numbered text pages. Its cover identifies an exposure draft dated 08/09/2026. The following are explicit **“to be drafted”** markers in the readable text; each is a simplified-outline section. The marker itself does not identify withheld substantive wording. This classification follows the headings and words actually published, not an assumption about unpublished drafting history.\n\n| Proposed provision | Location, physical PDF page / printed page | Published heading |\n| --- | --- | --- |\n| 86A | 7 / 3 | Simplified outline of Part 6A, fake nude material |\n| 4 | 27 / 23 | Simplified outline of the Act |\n| 25 | 29 / 25 | Simplified outline of Part 2, digital duty of care |\n| 166 | 47 / 43 | Simplified outline of Part 11, eSafety Commissioner |\n| 192A | 52 / 48 | Simplified outline of Part 13 |\n| 205A | 63 / 59 | Simplified outline of Part 14, research and data access |\n\nThe acquired draft text does not label these markers redactions. No official explanation beyond their unfinished-drafting label was found in the inspected publication page or press conference. Do not rename them censorship or claim their unseen wording. Public exposure-draft PDF (see source references)\n\nThis is not a full visual redaction audit: neither original bytes nor a usable page image was acquired by this lane. Text search can find declared placeholders but cannot exclude every possible graphical redaction, omitted earlier version, or unpublished attachment.\n\n## Earlier public materials and actual publication restrictions\n\nThe department also publishes the **May 2026 issues paper**, with a publication-page date of 26 May, and links a **three-page summary of May–June consultation outcomes**. These are policy-development materials, not an explanatory memorandum accompanying the September bill. The summary reports more than 40 consultation meetings across described stakeholder categories. It does not supply a September draft-recipient roster. Issues paper and summary publication page (see source references) consultation-outcomes summary, page 1 introduction (see source references)\n\nA distinct, explicit non-publication rule appears on the closed **14 November–7 December 2025 public survey** page. Under its privacy collection notice, individual completed surveys and participants' personal information are not to be published; third-party personal-information disclosure needs legal authorization or requirement. Aggregated results may be published, with contact and identifying details redacted. The page also reserves a right not to consider responses that create legal exposure or contain abusive, offensive or defamatory material. Those are stated rules for survey responses, not secrecy terms for the 2026 draft. The old page's closed label and registration requirement must not be applied to the September exposure-draft feedback route. 2025 survey page, Participate and Privacy Collection Notice (see source references)\n\nA bounded official-domain search also surfaced **FOI 26-276, Document 6, Attachment C**, described as a summary of late-2025 consultation activities and organizational participants. Only the search-returned passage was inspected in this lane; the underlying release was not acquired here. It may identify earlier consultees but cannot establish who received September 2026 draft text. This lead was sent to the genealogy lane without assuming continuity of the roster. FOI release, Attachment C target (see source references)\n\n## Scope, access failures and stopping point\n\nThe ministerial announcement and department publication pages were read in the public web reader. The press-conference transcript was read in full. The current bill was read selectively for its cover, contents, placeholders and their surrounding headings, using the web reader's 75-page rendering. The outcomes summary was read for its introduction and accessible thematic text; it was not used to attribute a named organization's influence. The 2025 survey's participation and privacy terms were read in full.\n\nAn ordinary Python GET to the department publication page timed out after 30 seconds before any draft bytes were obtained. The DOCX link yielded an internal error in the web reader. Screenshot attempts returned references without visible image content, with two cache misses. Those are acquisition limitations, not evidence that the government denies public access. The root's separate ordinary PowerShell substitute also timed out; that branch is closed. No local original, original-file hash, DOCX/PDF equivalence check or complete visual examination is claimed. No account, invitation, access control or publication restriction was bypassed.\n\nThe precise unresolved record family is the **September consultation invitation and distribution record**, together with any **covering instructions, confidentiality terms, attachment/version list and later draft comparison**. The department's Digital Duty of Care policy team is the likely custodian. If an allegedly censored page or different draft is later supplied, compare it against the dated public version before attributing removal or a motive. No request or outreach was made. This lane stops at the verified public-exposure instrument; further acquisition waits for the next whole-case review.\n\n## Follow-up, 9 September 2026: alleged fuller versions supplied to companies\n\nThe user clarified a materially different allegation: large technology companies supposedly received less skeletal versions before the public learned the intended detail. The user identifies **Isaac Butterfield** as the source they heard, but no exact title, URL, timestamp or quotation was supplied or identified in this bounded pass. This is an attributed user lead, not a verified quotation from Butterfield. Unrelated older appearances were not substituted for the missing clip.\n\n**Public availability does not resolve selective advance access. The allegation remains OPEN.** Three different claims require different records: earlier circulation of the same text needs a dated covering message and identifiable attachment; circulation of a fuller version needs both versions or a reliable recipient account specifying the extra text; private explanation of an otherwise shared draft needs briefing material, notes or an attributed account. Neither public draft access nor participation in an earlier policy meeting proves or disproves those events. The unfinished outlines listed above likewise cannot establish what companies received.\n\n### Separate documented withholding in the preparation process\n\nA Senate primary-source search excerpt exposes a relevant historical entry in the **Orders for the production of documents** status table, **No. 169—Digital duty of care**, printed page 43. The order, moved by Senator Payman and agreed on **4 September 2025**, sought minister/office correspondence and briefing/meeting records with eSafety and the department concerning preparation of a legislated duty. The entry records a response on **7 October 2025**, identifying **26** responsive documents, claiming public-interest immunity for **20**, and supplying the remaining documents with redactions on public-interest grounds. **Six is the derived remainder, 26 minus 20**, not a separately printed count. The stated immunity reason concerns Cabinet deliberations and prejudice to informed government consideration of ongoing policy matters. The entry records partial compliance. Senate OPD table, order 169 (see source references)\n\nThis is evidence of a recorded **2025 preparation-document withholding claim and redacted production**, distinct from the survey privacy policy above. It does not identify a complete September 2026 draft, establish its delivery to a company, or identify additional clauses seen only by industry. The underlying ministerial response and individual documents were not acquired. Reading scope is the indexed primary entry only: two targeted searches returned the same No.169 passage with its dates, counts, stated reason and printed page. A direct reader open failed; the root also reported a 403 on its direct open. No full-table review, original capture/hash or later compliance-history claim is made.\n\n### Exact-source and recipient-record boundary\n\nTargeted public searches for current draft circulation, confidentiality, redactions, advance access and relevant industry accounts did not produce a paired version, September invitation/covering letter, or recipient account specifying extra clauses. That limited search result is not evidence the allegation is false. The official press-conference account of Holdsworth reading the text that morning identifies neither a company recipient nor a distinct version.\n\nButterfield's official website was inspected for its current video links. The linked YouTube channel returned a cache miss through the web reader. The single browser substitute then **denied access because it could not verify the admin-enforced policy for youtube.com**. No browser security control was bypassed, no account was used, and no indirect transcript, mirror or alternate YouTube route was pursued after that denial. The exact clip and its underlying source remain unexamined. Official site used to identify the channel link (see source references)\n\nThe present outcome is therefore narrower than either dismissal or confirmation: a public September draft; explicit incomplete outlines; a separate indexed parliamentary record of withheld/redacted preparation documents in 2025; and an unresolved allegation of selective access to fuller 2026 text. The next discriminating evidence would be the exact clip/quotation and its cited document, or the recipient covering letter plus attachment/version comparison. No outreach or further acquisition follows. The current run returns to the whole-case assessment with the selective-version claim open.\n", "file": "research/duty-care-wave-2026-09-09-draft-access.md" }, { "id": "duty-care-wave-2026-09-09-operative-authority", "title": "Australia Digital Duty of Care: operative authority and adult choice", "description": "Instrument reviewed: Online Safety Amendment (Digital Duty of Care) Bill 2026, exposure draft dated 8 September 2026. Research date: 9 September 2026. This is proposed legislation, not an enacted duty. The preceding whole-case assessment in forest-reviews.md was read before this lane. All draft powers below are conditional on enactment and, where specified, ", "html": "Instrument reviewed: Online Safety Amendment (Digital Duty of Care) Bill 2026, exposure draft dated 8 September 2026. Research date: 9 September 2026. This is proposed legislation, not an enacted duty. The preceding whole-case assessment in forest-reviews.md was read before this lane. All draft powers below are conditional on enactment and, where specified, later instruments.
\nPrimary sources: [D — exposure draft, 75 PDF pages/71 numbered pages](http[local research file] [A — Online Safety Act 2021, official compilation C2024C00852, 11 December 2024](http[local research file] The Register’s [latest-version landing page](http[local research file] identifies that compilation as current. Page locators below are PDF page / printed page, not zero-based reader indices.
\nThe draft would create an enforceable, continuing obligation on service providers and persons exercising day-to-day control to manage online risks. It is not confined to removing already unlawful posts. Parliament would specify initial categories, the minister could expand specified categories and require user controls through legislative instruments, and eSafety could specify assessment requirements and direct remedial action after reasonably believing the duty had been breached. This is a concrete delegation over service design, with legal limits and review routes. It is not evidence that those powers have already been exercised.
\nThe adult-choice boundary is narrower than a general right to assume risk. Under proposed s26(7), the digital duty does not require action concerning lawful communications occurring privately solely between consenting adults. It does not expressly exempt all willing adult consumption of public material or an adult’s private interaction with an AI system. Whether particular AI chats meet the exception remains unresolved; an AI is not expressly treated as a consenting adult. Conversely, pornography is listed in the child-specific category, not automatically in the adult/general list. [D, ss25B–25D, PDF30–33/printed26–29; s26(7), PDF36/printed32.]
\n| Decision owner | \nProposed requirement or power | \nClause, trigger and limit | \n
|---|---|---|
| Parliament, through the enacted text | \nA safe environment so far as reasonably practicable; appropriate design-feature management, written risk assessment and effective measures addressing assessments. | \nD ss25H, 26–26A, PDF35–38/printed31–34. Reasonable practicability weighs likelihood, severity, knowledge, available measures, cost and privacy impact. This is a risk-management duty, not an absolute guarantee that no harm occurs. | \n
| Minister | \nExempt a service/class from one or more duty provisions, possibly conditionally, where risk is small or Australian use minimal. | \nD s25A(3)–(5), PDF30/printed26. Must seek and consider eSafety advice under s27B. This supplies a potential small/low-risk-service route, not an automatic small-business exemption. | \n
| Minister | \nAdd serious-harm categories; add child-harm categories; deem further design features to have negative behavioural impacts. | \nD ss25C(2), 25D(2), 25G(2), PDF32–34/printed28–30. Legislative instruments; s27B requires eSafety advice. Adult/general additions require ministerial satisfaction that material/conduct may cause serious harm. | \n
| Minister | \nTreat a service as social media, or not, for the safe-environment provision; require specified services to provide specified user-empowerment tools. | \nD ss25B(2)–(3), 26(4)–(6), PDF31/printed27 and PDF36/printed32. The special s27B advice list does not include these two powers. Tools may control recommendation/design operation; actual tools and services await instruments. | \n
| eSafety Commissioner | \nSet additional assessment detail, methods, standards/metrics and intervals shorter than annual. | \nD s26A(2)(f), (3)–(4), PDF37–38/printed33–34. Legislative instrument. Assess before changes introducing new/additional risk; keep assessments six years and produce within 30 days of request. | \n
| eSafety Commissioner | \nGive a written direction specifying action to prevent future duty failures. | \nD s26D, PDF38–39/printed34–35. Requires reasonable belief of present/past noncompliance; not a legislative instrument. Failure to comply carries a separate civil penalty. ART review is added by Schedule 2 item 58. | \n
| eSafety Commissioner | \nSelect service classes requiring complaint/dispute processes and specify process requirements; direct remediation of noncompliance. | \nD ss26F–26G, PDF39–40/printed35–36. Processes must be equally available to Australians. Detailed entitlements depend on a later instrument; this is not an express universal right to have removed content restored. | \n
| eSafety Commissioner | \nRequire transparency reports; require classes of providers to publish operational safety information. | \nD ss192C–192N, PDF53–58/printed49–54. Reporting notice must satisfy statutory purposes and consider burden; it must state reasons and review rights. Publication requirements may cover moderation, account removal, complaints and risk assessment. | \n
| eSafety Commissioner | \nCompel information/documents, relevant answers and attendance for examination concerning listed investigations or compliance/possible contraventions. | \nD ss198–205, PDF59–63/printed55–59. Reason to believe the person can provide relevant material; the new general compliance limb excludes Part 4A social-media minimum-age provisions. Existing investigation powers are expanded/recast, not created from nothing. | \n
| Minister through legislative rules; prospective eSafety approval decisions if the rules so provide | \nEstablish researcher-approval and data-access schemes, including data classes, requests, fees, conditions and revocations. | \nD ss205B–205F, PDF63–66/printed59–62; A s240 identifies the minister as legislative-rule maker. Compelled access would depend on those rules. University employment, approval under legislative rules, ethics approval and a prescribed public-interest research kind constrain eligibility. | \n
| Approved researchers and eSafety | \nUse fictitious accounts for authorized research/regulatory purposes despite contrary laws or contractual terms. | \nD ss205G–205L, PDF66–68/printed62–64. Listed uses include observing, recording and testing. Engaging other users is limited to what is necessary to prevent account closure. The Commissioner’s additional material-generation power expressly excludes criminal production/dissemination/possession. Good-faith civil immunity is provided. | \n
| eSafety Commissioner; courts and tribunal in their distinct roles | \nIssue warnings/infringement notices; seek statutory enforcement; make specified decisions reviewable. | \nD ss26B–26D, 163A, 220 amendments, PDF38–39/printed34–35, PDF45–46/printed41–42, PDF68/printed64; A ss162–165. Do not present the court-level maximum as an amount the Commissioner can conclusively impose without court process. | \n
Proposed s25A covers internet carriage, social media, relevant electronic services, designated internet services, hosting, search, app distribution, specified equipment-related services, and services permitting AI generation and sharing through the listed service types (PDF29–30/printed25–26). Existing A s13A expressly includes email, instant messaging, SMS/MMS, chat and multiplayer online games. A s14 supplies a broad internet-content-service category subject to exclusions. Accordingly, messaging and games are within the proposed duty through existing definitions; the private-adult exception affects action required, rather than excluding all messaging services. The AI limb is functionally described, not a blanket designation of every offline model or every AI developer. Other categories might independently cover an online AI product.
\nThe general protection in s25B(1)(a) uses s25C’s list: child sexual exploitation/abuse and grooming; encouragement/instruction of sexual or extreme violence; explicit violent threats; intentional serious-harm harassment; encouragement of suicide, self-harm, sadistic exploitation or animal cruelty; specified terrorism support; encouragement/instruction of crime and illicit drug use; abhorrent violent conduct; and ministerially added categories. There is no single overarching requirement that every item must first be independently criminal. Nor is the entire list subject to the existing adult-targeted cyber-abuse test in A s7. These are distinct schemes. [D PDF31–32/printed27–28; A ss5, 7.]
\nThe retained A s5 definition makes serious harm physical or mental-health harm of sufficient seriousness, including serious psychological harm/distress, and excludes mere ordinary emotional responses. This matters to the minister’s expansion threshold and references to serious harm. It does not require re-proving each expressly listed s25C category under A s7’s separate targeted-abuse test.
\nChild-specific s25D additionally names pornography, disordered-eating promotion/instruction, promotion of hostility toward women or gender equality, crime glorification/dangerous stunts/harmful practices, and abuse/harassment/bullying. It includes a residual category capable of seriously harming a child and ministerially determined additions, the latter using a child-harm satisfaction threshold. The minister’s power is not the sole source of that residual coverage. [D PDF32–33/printed28–29.]
\nFor social media, s25B(1)(c) additionally requires the specified negative-impact features not to operate for children under 16. Sections 25F–25G define and deem recommendation, login-dependent, endless-feed, feedback and time-limited features to have those impacts. This is stronger than an optional parental toggle. It is subject to the overarching reasonably-practicable duty and ministerial service determinations. The draft does not itself prescribe a particular adult algorithm opt-in/opt-out interface: later user-tool instruments would supply that detail. [D PDF31–36/printed27–32.]
\nPrivacy appears expressly in s25H’s proportionality assessment. Section 26(7)’s private-consenting-adult exception is an additional substantive boundary. Neither supplies an explicit general encryption exemption or a technical rule forbidding every possible demand affecting encryption. The inspected operative draft does not prescribe universal government-ID collection, a named age-verification vendor or a general obligation to identify all users. Distinguish absence of such a clause from proof that future implementation has no identity or encryption implications.
\nExisting A ss194–195 already permit notices for held end-user identity/contact information on reasonable grounds of relevance to the Act, with compliance limited by capability. They are not newly invented by this draft, and are not a universal account-ID collection requirement. Their reference to the Act’s operation would sit alongside an expanded Act if enacted. Schedule 2 item 63 also adds eSafety to the Privacy Act definition of enforcement body; the detailed interaction with Australian Privacy Principles was not audited here.
\nExisting A s233 preserves the constitutional implied freedom of political communication. This is not a general statutory right to access all lawful adult material. The draft supplies no comparable comprehensive adult-consent override. Privacy is a weighed factor; a preference to accept risk does not, by itself, switch off the service’s whole duty. That leaves a substantive adult-autonomy question without justifying the claim that every lawful adult choice is prohibited.
\nFor transparency reports published by eSafety, s192G expressly excludes protected confidential and personal information (PDF55/printed51). For provider publication, s192M(6) instead points to kinds of information specified in the determination for exclusion (PDF58/printed54). Do not treat the two protections as identical.
\nResearch access has real institutional selection. Section 205B requires employment by an Australian university, approval under legislative rules, ethics approval through the university and a prescribed public-interest kind of research. It does not grant automatic access to every NGO, advocate or company. Section 205C permits later rules about conflicts, privacy/security/confidentiality, commercial use restrictions, cost-limited provider fees, and review/reconsideration of decisions to grant access requests. Those subjects are delegated; the complete safeguards and review entitlements are not already written into this draft. No named organization is appointed or given a present data-access entitlement. [D PDF64–65/printed60–61.]
\nCross-lane comparison: Reset.Tech Australia's 2024 submission 70 requested vetted researcher access including third-sector organizations (printed p26/PDF28). The draft instead expressly requires employment by an Australian university, plus the prospective approval and research conditions above. That is a narrower independent eligibility route than the advocate requested; it does not award Reset a data right. It does not rule out university collaboration or indirect institutional benefit. This proposal-side passage was inspected by the parent researcher and reconciled in the independent review, rather than independently reacquired in this lane. Sources: [Reset submission 70](http[local research file] [policy-lineage.md]([local research file] draft s205B, PDF64/printed60.
\nThe draft adds ART review for duty, complaint-process and data-access remedial directions through s220(12), and for transparency notices through new s220(23). Existing A s220A requires an internal-review scheme for decisions of the kinds in s220. The draft does not spell out a new automatic stay or a general merits appeal against all rulemaking, warnings or researcher approvals. Do not infer no judicial remedy; the detailed judicial-review/stay framework and future research-review rules were not examined. [D PDF68/printed64; A ss220–220A.]
\nMinisterial appointment and conflict/termination rules remain in A ss167–176. A s188 permits binding ministerial directions of a general nature only; its note expressly identifies disallowance and sunsetting exemptions. That is different from a ministerial power to reverse a particular notice. The draft’s new duty instruments are legislative instruments, while individual remedial directions and the annual enforcement-priority statement are expressly non-legislative. No new exemption from parliamentary disallowance is written alongside the duty instruments read here; an instrument-specific application of the general Legislation Act exemptions was not completed. Do not claim every instrument is immune from scrutiny, or that every one receives an affirmative parliamentary vote.
\nUnder new s166C, eSafety must publish annual monitoring/enforcement priorities, but the statement expressly does not restrict its functions or powers. Existing annual reporting continues; the draft adds counts of new directions. Existing good-faith protection from damages in A s222 is distinct from access to review. The draft also resets the s239A independent-review timing by reference to duty commencement; the surviving provision requires a written report tabled in Parliament. [D PDF49–51/printed45–47 and PDF70/printed66; A ss183, 222, 239A.]
\nA consequential change is information compulsion. Existing A ss198–205 tie powers to specified investigations and allow a reasonable excuse and refusal on self-incrimination grounds. The replacement expands scope to compliance/possible contravention, removes that general reasonable-excuse wording, and compels self-incriminating material while providing individual use and derivative-use protection, subject to stated exceptions. It preserves a journalist-source exception. A private examination, adviser and record remain, although the draft makes obtaining the record request-based and timing subject to the Commissioner/delegate’s view. The existing 12-month criminal maximum for noncompliance is retained/recast; the civil penalty stated becomes 1,000 units. [D PDF59–63/printed55–59; A ss198–205.]
\nThe draft states 60,000 penalty units for duty breach and disobeying duty-remediation directions; 6,000 for complaint-process, transparency and data-access failures; and 1,000 for failing a required Australian contact nomination. New s163A specifies 12 units for an individual or 600 for a corporation per alleged contravention in the listed infringement-notice scheme. These are distinct mechanisms; corporate multipliers, continuing contraventions and dollar conversion were not calculated here. Existing A s162 makes eSafety an applicant and names the Federal Court and Federal Circuit and Family Court (Division 2) as relevant civil-penalty courts.
\nNot every clause is a new power. Proposed s177A substantially relocates existing A s28’s general necessary/convenient power. Grant and educational/research functions substantially continue existing A s27, with transitional continuity in Schedule 2 item 65. By contrast, the general duty, expanded compliance-information limb and research-access scheme are material new instruments. Schedule 2 repeals the Basic Online Safety Expectations Part 4; Schedule 3 repeals the online-content industry-code/standard division. This is a replacement architecture, not simply an additional overlay while all earlier instruments necessarily persist. [D PDF41/printed37, PDF50/printed46, PDF72–74/printed68–70; A ss27–28.]
\nSchedule 1 would commence the day after assent; the duty and industry-code repeal schedules would commence after 12 months (D PDF6/printed2). Proposed s86B permits eSafety, on reasonable grounds of predominant design or use for fake-nude generation, to order app-store/search removal of an app, links, advertising, information or access. Section 86C gives 24 hours unless extended, capability-limited compliance and 6,000 units; s220(10B) adds ART review. This is a specific proposed distribution gate, separate from the general duty. The definition scope of fake nude material was not resolved in this lane: no definition appeared in the inserted lists read, but existing and incorporated definitions were not comprehensively traced. This is not an explicitly marked missing definition or evidence of redaction; the definition question remains unresolved in this scoped reading. [D PDF7–9/printed3–5.]
\nOther Schedule 1 amendments reduce the initial platform-complaint waiting period from 48 to 24 hours, permit waivers for inaccessible complaints systems or foreseeable further harm, and address reposts after a prior removal notice. They do not erase the distinct substantive targeted-adult-abuse test. New search-link deletion notices generally require a prior removal notice, an appropriateness judgment and revocation when no longer required or outstanding removals are complied with. These are extensions of existing removal machinery, not a newly invented power to remove any disliked speech. [D PDF9–22/printed5–18; A ss7, 88–90.]
\nThe shared web-reader rendition of the original publisher’s PDF was used. Read: front matter/commencement; Schedule 2 operative definitions, duty, enforcement, transparency, information, research, review and transition provisions; Schedule 3 repeal provisions; targeted Schedule 1 fake-nude, adult-removal, link-deletion, review and disclosure amendments. Detailed adjacent Schedule 1 paragraphs were not all separately examined. For the existing Act, read the cited definitions, powers, enforcement, ministerial oversight, information, review, expression and rulemaking provisions, not the entire Act.
\nNo original PDF bytes were retained by this lane. Initial direct URL web-open returned a non-retryable safe-open error; one ordinary PDF GET timed out after 30 seconds with no body. Reusing the access agent’s already working source reference supplied the necessary text. These were transport errors, not an automatic sandbox-approval rejection; no blocked browser route was retried. Parliament explanatory-page opens and one current Legislation Act route also failed, so this packet limits its disallowance claim as stated above. Neither this lane nor the access lane acquired original draft bytes or a file hash. The access lane retains scoped public-reader evidence and a stated visual gap; this file is not a preserved original.
\nThe instrument breakpoint is reached. The strongest defensible interpretation is expanded preventative service-design regulation with ministerial category-setting, eSafety implementation/enforcement and selected review mechanisms. A rival interpretation focused on ordinary prevention of serious harms remains compatible with the text; motive is not resolved. The decisive autonomy gap is how the minister/regulator would distinguish protected adult choice from preventable harm in concrete instruments and decisions, especially AI interactions. The private-adult exception belongs to the digital duty; it does not purport to disable every separate notice or information power in the Act. Missing discriminating records are settled definitions, actual user-tool/risk-assessment/data-access rules, safeguards and review procedures, and later decisions applying them. No new branch is opened here; return to the whole-case assessment.
", "path": "duty-care-wave-2026-09-09/operative-authority.md", "sourceIds": [ "w10-duty-s-draft", "w10-duty-s-existing-act" ], "markdown": "Instrument reviewed: **Online Safety Amendment (Digital Duty of Care) Bill 2026, exposure draft dated 8 September 2026**. Research date: 9 September 2026. This is proposed legislation, not an enacted duty. The preceding whole-case assessment in forest-reviews.md was read before this lane. All draft powers below are conditional on enactment and, where specified, later instruments.\n\nPrimary sources: D — exposure draft, 75 PDF pages/71 numbered pages (see source references) A — Online Safety Act 2021, official compilation C2024C00852, 11 December 2024 (see source references) The Register’s latest-version landing page (see source references) identifies that compilation as current. Page locators below are **PDF page / printed page**, not zero-based reader indices.\n\n## Principal finding\n\nThe draft would create an enforceable, continuing obligation on service providers and persons exercising day-to-day control to manage online risks. It is not confined to removing already unlawful posts. Parliament would specify initial categories, the minister could expand specified categories and require user controls through legislative instruments, and eSafety could specify assessment requirements and direct remedial action after reasonably believing the duty had been breached. This is a concrete delegation over service design, with legal limits and review routes. It is not evidence that those powers have already been exercised.\n\nThe adult-choice boundary is narrower than a general right to assume risk. Under proposed s26(7), the digital duty does not require action concerning lawful communications occurring privately solely between consenting adults. It does not expressly exempt all willing adult consumption of public material or an adult’s private interaction with an AI system. Whether particular AI chats meet the exception remains unresolved; an AI is not expressly treated as a consenting adult. Conversely, pornography is listed in the child-specific category, not automatically in the adult/general list. [D, ss25B–25D, PDF30–33/printed26–29; s26(7), PDF36/printed32.]\n\n## Who can require what?\n\n| Decision owner | Proposed requirement or power | Clause, trigger and limit |\n| --- | --- | --- |\n| Parliament, through the enacted text | A safe environment so far as reasonably practicable; appropriate design-feature management, written risk assessment and effective measures addressing assessments. | D ss25H, 26–26A, PDF35–38/printed31–34. Reasonable practicability weighs likelihood, severity, knowledge, available measures, cost and privacy impact. This is a risk-management duty, not an absolute guarantee that no harm occurs. |\n| Minister | Exempt a service/class from one or more duty provisions, possibly conditionally, where risk is small or Australian use minimal. | D s25A(3)–(5), PDF30/printed26. Must seek and consider eSafety advice under s27B. This supplies a potential small/low-risk-service route, not an automatic small-business exemption. |\n| Minister | Add serious-harm categories; add child-harm categories; deem further design features to have negative behavioural impacts. | D ss25C(2), 25D(2), 25G(2), PDF32–34/printed28–30. Legislative instruments; s27B requires eSafety advice. Adult/general additions require ministerial satisfaction that material/conduct may cause serious harm. |\n| Minister | Treat a service as social media, or not, for the safe-environment provision; require specified services to provide specified user-empowerment tools. | D ss25B(2)–(3), 26(4)–(6), PDF31/printed27 and PDF36/printed32. The special s27B advice list does not include these two powers. Tools may control recommendation/design operation; actual tools and services await instruments. |\n| eSafety Commissioner | Set additional assessment detail, methods, standards/metrics and intervals shorter than annual. | D s26A(2)(f), (3)–(4), PDF37–38/printed33–34. Legislative instrument. Assess before changes introducing new/additional risk; keep assessments six years and produce within 30 days of request. |\n| eSafety Commissioner | Give a written direction specifying action to prevent future duty failures. | D s26D, PDF38–39/printed34–35. Requires reasonable belief of present/past noncompliance; not a legislative instrument. Failure to comply carries a separate civil penalty. ART review is added by Schedule 2 item 58. |\n| eSafety Commissioner | Select service classes requiring complaint/dispute processes and specify process requirements; direct remediation of noncompliance. | D ss26F–26G, PDF39–40/printed35–36. Processes must be equally available to Australians. Detailed entitlements depend on a later instrument; this is not an express universal right to have removed content restored. |\n| eSafety Commissioner | Require transparency reports; require classes of providers to publish operational safety information. | D ss192C–192N, PDF53–58/printed49–54. Reporting notice must satisfy statutory purposes and consider burden; it must state reasons and review rights. Publication requirements may cover moderation, account removal, complaints and risk assessment. |\n| eSafety Commissioner | Compel information/documents, relevant answers and attendance for examination concerning listed investigations or compliance/possible contraventions. | D ss198–205, PDF59–63/printed55–59. Reason to believe the person can provide relevant material; the new general compliance limb excludes Part 4A social-media minimum-age provisions. Existing investigation powers are expanded/recast, not created from nothing. |\n| Minister through legislative rules; prospective eSafety approval decisions if the rules so provide | Establish researcher-approval and data-access schemes, including data classes, requests, fees, conditions and revocations. | D ss205B–205F, PDF63–66/printed59–62; A s240 identifies the minister as legislative-rule maker. Compelled access would depend on those rules. University employment, approval under legislative rules, ethics approval and a prescribed public-interest research kind constrain eligibility. |\n| Approved researchers and eSafety | Use fictitious accounts for authorized research/regulatory purposes despite contrary laws or contractual terms. | D ss205G–205L, PDF66–68/printed62–64. Listed uses include observing, recording and testing. Engaging other users is limited to what is necessary to prevent account closure. The Commissioner’s additional material-generation power expressly excludes criminal production/dissemination/possession. Good-faith civil immunity is provided. |\n| eSafety Commissioner; courts and tribunal in their distinct roles | Issue warnings/infringement notices; seek statutory enforcement; make specified decisions reviewable. | D ss26B–26D, 163A, 220 amendments, PDF38–39/printed34–35, PDF45–46/printed41–42, PDF68/printed64; A ss162–165. Do not present the court-level maximum as an amount the Commissioner can conclusively impose without court process. |\n\n## Coverage and harm thresholds\n\nProposed s25A covers internet carriage, social media, relevant electronic services, designated internet services, hosting, search, app distribution, specified equipment-related services, and services permitting AI generation **and** sharing through the listed service types (PDF29–30/printed25–26). Existing A s13A expressly includes email, instant messaging, SMS/MMS, chat and multiplayer online games. A s14 supplies a broad internet-content-service category subject to exclusions. Accordingly, messaging and games are within the proposed duty through existing definitions; the private-adult exception affects action required, rather than excluding all messaging services. The AI limb is functionally described, not a blanket designation of every offline model or every AI developer. Other categories might independently cover an online AI product.\n\nThe general protection in s25B(1)(a) uses s25C’s list: child sexual exploitation/abuse and grooming; encouragement/instruction of sexual or extreme violence; explicit violent threats; intentional serious-harm harassment; encouragement of suicide, self-harm, sadistic exploitation or animal cruelty; specified terrorism support; encouragement/instruction of crime and illicit drug use; abhorrent violent conduct; and ministerially added categories. There is no single overarching requirement that every item must first be independently criminal. Nor is the entire list subject to the existing adult-targeted cyber-abuse test in A s7. These are distinct schemes. [D PDF31–32/printed27–28; A ss5, 7.]\n\nThe retained A s5 definition makes serious harm physical or mental-health harm of sufficient seriousness, including serious psychological harm/distress, and excludes mere ordinary emotional responses. This matters to the minister’s expansion threshold and references to serious harm. It does not require re-proving each expressly listed s25C category under A s7’s separate targeted-abuse test.\n\nChild-specific s25D additionally names pornography, disordered-eating promotion/instruction, promotion of hostility toward women or gender equality, crime glorification/dangerous stunts/harmful practices, and abuse/harassment/bullying. It includes a residual category capable of seriously harming a child and ministerially determined additions, the latter using a child-harm satisfaction threshold. The minister’s power is not the sole source of that residual coverage. [D PDF32–33/printed28–29.]\n\nFor social media, s25B(1)(c) additionally requires the specified negative-impact features not to operate for children under 16. Sections 25F–25G define and deem recommendation, login-dependent, endless-feed, feedback and time-limited features to have those impacts. This is stronger than an optional parental toggle. It is subject to the overarching reasonably-practicable duty and ministerial service determinations. The draft does **not** itself prescribe a particular adult algorithm opt-in/opt-out interface: later user-tool instruments would supply that detail. [D PDF31–36/printed27–32.]\n\n## Privacy, identity, expression and researcher boundaries\n\nPrivacy appears expressly in s25H’s proportionality assessment. Section 26(7)’s private-consenting-adult exception is an additional substantive boundary. Neither supplies an explicit general encryption exemption or a technical rule forbidding every possible demand affecting encryption. The inspected operative draft does not prescribe universal government-ID collection, a named age-verification vendor or a general obligation to identify all users. Distinguish absence of such a clause from proof that future implementation has no identity or encryption implications.\n\nExisting A ss194–195 already permit notices for held end-user identity/contact information on reasonable grounds of relevance to the Act, with compliance limited by capability. They are not newly invented by this draft, and are not a universal account-ID collection requirement. Their reference to the Act’s operation would sit alongside an expanded Act if enacted. Schedule 2 item 63 also adds eSafety to the Privacy Act definition of enforcement body; the detailed interaction with Australian Privacy Principles was not audited here.\n\nExisting A s233 preserves the constitutional implied freedom of political communication. This is not a general statutory right to access all lawful adult material. The draft supplies no comparable comprehensive adult-consent override. Privacy is a weighed factor; a preference to accept risk does not, by itself, switch off the service’s whole duty. That leaves a substantive adult-autonomy question without justifying the claim that every lawful adult choice is prohibited.\n\nFor transparency reports published by eSafety, s192G expressly excludes protected confidential and personal information (PDF55/printed51). For provider publication, s192M(6) instead points to kinds of information specified in the determination for exclusion (PDF58/printed54). Do not treat the two protections as identical.\n\nResearch access has real institutional selection. Section 205B requires employment by an Australian university, approval under legislative rules, ethics approval through the university and a prescribed public-interest kind of research. It does not grant automatic access to every NGO, advocate or company. Section 205C permits later rules about conflicts, privacy/security/confidentiality, commercial use restrictions, cost-limited provider fees, and review/reconsideration of decisions to grant access requests. Those subjects are delegated; the complete safeguards and review entitlements are not already written into this draft. No named organization is appointed or given a present data-access entitlement. [D PDF64–65/printed60–61.]\n\nCross-lane comparison: Reset.Tech Australia's 2024 submission 70 requested vetted researcher access including third-sector organizations (printed p26/PDF28). The draft instead expressly requires employment by an Australian university, plus the prospective approval and research conditions above. That is a narrower independent eligibility route than the advocate requested; it does not award Reset a data right. It does not rule out university collaboration or indirect institutional benefit. This proposal-side passage was inspected by the parent researcher and reconciled in the independent review, rather than independently reacquired in this lane. Sources: Reset submission 70 (see source references) [policy-lineage.md]([local research file] draft s205B, PDF64/printed60.\n\n## Supervision, remedies and changes from existing law\n\nThe draft adds ART review for duty, complaint-process and data-access remedial directions through s220(12), and for transparency notices through new s220(23). Existing A s220A requires an internal-review scheme for decisions of the kinds in s220. The draft does not spell out a new automatic stay or a general merits appeal against all rulemaking, warnings or researcher approvals. Do not infer no judicial remedy; the detailed judicial-review/stay framework and future research-review rules were not examined. [D PDF68/printed64; A ss220–220A.]\n\nMinisterial appointment and conflict/termination rules remain in A ss167–176. A s188 permits binding ministerial directions **of a general nature only**; its note expressly identifies disallowance and sunsetting exemptions. That is different from a ministerial power to reverse a particular notice. The draft’s new duty instruments are legislative instruments, while individual remedial directions and the annual enforcement-priority statement are expressly non-legislative. No new exemption from parliamentary disallowance is written alongside the duty instruments read here; an instrument-specific application of the general Legislation Act exemptions was not completed. Do not claim every instrument is immune from scrutiny, or that every one receives an affirmative parliamentary vote.\n\nUnder new s166C, eSafety must publish annual monitoring/enforcement priorities, but the statement expressly does not restrict its functions or powers. Existing annual reporting continues; the draft adds counts of new directions. Existing good-faith protection from damages in A s222 is distinct from access to review. The draft also resets the s239A independent-review timing by reference to duty commencement; the surviving provision requires a written report tabled in Parliament. [D PDF49–51/printed45–47 and PDF70/printed66; A ss183, 222, 239A.]\n\nA consequential change is information compulsion. Existing A ss198–205 tie powers to specified investigations and allow a reasonable excuse and refusal on self-incrimination grounds. The replacement expands scope to compliance/possible contravention, removes that general reasonable-excuse wording, and compels self-incriminating material while providing individual use and derivative-use protection, subject to stated exceptions. It preserves a journalist-source exception. A private examination, adviser and record remain, although the draft makes obtaining the record request-based and timing subject to the Commissioner/delegate’s view. The existing 12-month criminal maximum for noncompliance is retained/recast; the civil penalty stated becomes 1,000 units. [D PDF59–63/printed55–59; A ss198–205.]\n\nThe draft states 60,000 penalty units for duty breach and disobeying duty-remediation directions; 6,000 for complaint-process, transparency and data-access failures; and 1,000 for failing a required Australian contact nomination. New s163A specifies 12 units for an individual or 600 for a corporation per alleged contravention in the listed infringement-notice scheme. These are distinct mechanisms; corporate multipliers, continuing contraventions and dollar conversion were not calculated here. Existing A s162 makes eSafety an applicant and names the Federal Court and Federal Circuit and Family Court (Division 2) as relevant civil-penalty courts.\n\nNot every clause is a new power. Proposed s177A substantially relocates existing A s28’s general necessary/convenient power. Grant and educational/research functions substantially continue existing A s27, with transitional continuity in Schedule 2 item 65. By contrast, the general duty, expanded compliance-information limb and research-access scheme are material new instruments. Schedule 2 repeals the Basic Online Safety Expectations Part 4; Schedule 3 repeals the online-content industry-code/standard division. This is a replacement architecture, not simply an additional overlay while all earlier instruments necessarily persist. [D PDF41/printed37, PDF50/printed46, PDF72–74/printed68–70; A ss27–28.]\n\n## Separate immediate notice changes\n\nSchedule 1 would commence the day after assent; the duty and industry-code repeal schedules would commence after 12 months (D PDF6/printed2). Proposed s86B permits eSafety, on reasonable grounds of predominant design or use for fake-nude generation, to order app-store/search removal of an app, links, advertising, information or access. Section 86C gives 24 hours unless extended, capability-limited compliance and 6,000 units; s220(10B) adds ART review. This is a specific proposed distribution gate, separate from the general duty. The definition scope of fake nude material was not resolved in this lane: no definition appeared in the inserted lists read, but existing and incorporated definitions were not comprehensively traced. This is not an explicitly marked missing definition or evidence of redaction; the definition question remains unresolved in this scoped reading. [D PDF7–9/printed3–5.]\n\nOther Schedule 1 amendments reduce the initial platform-complaint waiting period from 48 to 24 hours, permit waivers for inaccessible complaints systems or foreseeable further harm, and address reposts after a prior removal notice. They do not erase the distinct substantive targeted-adult-abuse test. New search-link deletion notices generally require a prior removal notice, an appropriateness judgment and revocation when no longer required or outstanding removals are complied with. These are extensions of existing removal machinery, not a newly invented power to remove any disliked speech. [D PDF9–22/printed5–18; A ss7, 88–90.]\n\n## Evidence scope and stopping point\n\nThe shared web-reader rendition of the original publisher’s PDF was used. Read: front matter/commencement; Schedule 2 operative definitions, duty, enforcement, transparency, information, research, review and transition provisions; Schedule 3 repeal provisions; targeted Schedule 1 fake-nude, adult-removal, link-deletion, review and disclosure amendments. Detailed adjacent Schedule 1 paragraphs were not all separately examined. For the existing Act, read the cited definitions, powers, enforcement, ministerial oversight, information, review, expression and rulemaking provisions, not the entire Act.\n\nNo original PDF bytes were retained by this lane. Initial direct URL web-open returned a non-retryable safe-open error; one ordinary PDF GET timed out after 30 seconds with no body. Reusing the access agent’s already working source reference supplied the necessary text. These were transport errors, not an automatic sandbox-approval rejection; no blocked browser route was retried. Parliament explanatory-page opens and one current Legislation Act route also failed, so this packet limits its disallowance claim as stated above. Neither this lane nor the access lane acquired original draft bytes or a file hash. The access lane retains scoped public-reader evidence and a stated visual gap; this file is not a preserved original.\n\nThe instrument breakpoint is reached. The strongest defensible interpretation is expanded preventative service-design regulation with ministerial category-setting, eSafety implementation/enforcement and selected review mechanisms. A rival interpretation focused on ordinary prevention of serious harms remains compatible with the text; motive is not resolved. The decisive autonomy gap is how the minister/regulator would distinguish protected adult choice from preventable harm in concrete instruments and decisions, especially AI interactions. The private-adult exception belongs to the digital duty; it does not purport to disable every separate notice or information power in the Act. Missing discriminating records are settled definitions, actual user-tool/risk-assessment/data-access rules, safeguards and review procedures, and later decisions applying them. No new branch is opened here; return to the whole-case assessment.\n", "file": "research/duty-care-wave-2026-09-09-operative-authority.md" }, { "id": "duty-care-wave-2026-09-09-policy-lineage", "title": "Australian duty of care: attributable model transmission", "description": "Observed 9 September 2026. Bounded lineage run after the recorded assurance forest review. Draft legal operation and access are separate companion notes.", "html": "Observed 9 September 2026. Bounded lineage run after the recorded assurance forest review. Draft legal operation and access are separate companion notes.
\nAustralia's official statutory review, section 5.5, printed p53, attributes the overarching duty-of-care conception to Lorna Woods and William Perrin. It cites Carnegie UK's 2019 model bill and its 2021 criticism of the narrower UK scheme, then favors a singular Australian duty. It directly quotes Reset.Tech Australia's submission70 when explaining why content-specific duties would be insufficient. This is recipient-side acknowledgment of models and submitted reasoning, stronger than a shared vocabulary or conference roster. It does not allocate final clause authorship to those advocates. Source: http[local research file] . Reading scope: primary indexed section5.5 text and footnotes34-38; full PDF reader exceeded13MB limit and both ordinary downloads timed out, so no full-report reading or original custody claimed.
\nCarnegie UK's January2022 submission to Australia's House Select Committee says its clerk invited the submission. It proposes platform risk assessment, mitigation and ongoing supervision, with a stronger regulator and cross-border cooperation. Its envisaged interventions include changes to visibility, forwarding and message prompts, while acknowledging expression and proportionality. That is documented historical submission and intended mechanism, not proof each intervention was enacted. Source: http[local research file] . All six pages read in web rendering, especially paragraphs1,9,19-23. The document itself dates January2022 despite misleading search metadata.
\nThe Online Safety Act Network's June2024 submission68 names Maeve Walsh as lead and Lorna Woods as principal adviser. It explicitly promotes Australian alignment with UK/EU regulation to strengthen regulators against multinational firms. Its approach covers design, operation, resourcing and business models, while warning that systems measures and content interventions cannot be completely separated. It also warns against unnecessary overmoderation and stresses privacy-by-design. Source: http[local research file] . Entire ten-page substantive rendering read. This is an advocate's proposal; the government host is not government endorsement of every statement.
\nA further direct working relationship is a jointly authored 22April2024 article by Rys Farthing and Woods advocating a singular Australian duty and discussing the UK divergence. The author disclosure says Reset Australia received organizational support from Reset Global, Luminate and Internet Society Foundation at that time. It does not identify a grant commissioning this article or funding Woods personally. Its claim about technology companies preferring the narrower scheme is author interpretation, not a recovered company instruction. Source: http[local research file] . Full substantive body/bylines/disclosure read. Hosted outlet identity may have changed; preserve date and original authors rather than treating this as a new2026 article.
\nReset's submission70 explicitly names Woods/Perrin's model and covers access/content creation, discovery, user tools and platform tools. Its singular-duty argument appears on printed pp4-5; later printed p26 proposes vetted researcher access including third-sector organizations. That is a potential institutional-benefit design, not an awarded right to Reset or proof of control over platform data. Source: http[local research file] . Scoped passages PDF6-7 and PDF28, not complete31-page reading.
\nReset.Tech Australia's own2024 transparency report describes a duty-of-care roundtable, paper, small advocacy coalition, review submissions and parliamentary evidence. Its finance page lists Reset.Tech(US), Internet Society Foundation, Susan McKinnon Foundation, Boundless Foundation, Manniferra (spelling as printed), an unnamed private donor, and NSW Government support specifically for attendance at the social-media summit. No amounts or award terms appear there. This is recipient-reported funding; it cannot attribute the duty campaign's expenses to any particular donor, treat summit travel support as state commissioning, or establish donor editorial rights. Sources: http[local research file] , printed pp6,37-38 and scoped project passages. Not read cover-to-cover. The2024 funding roster differs from the earlier article's roster; do not conflate periods or infer withdrawal.
\nCurrent naming needs care: the former Online Safety Act Network announced its change to Online Safety Network on1September2026, expressly expanding attention to AI and emerging technologies. It retains focus on online-safety implementation while advocating product-safety regulation for AI chatbots and broader digital regulation. This is explicit organizational scope expansion, not an inferred AI connection from a name. Sources: http[local research file] (full body) and http[local research file] (current mission/leadership; logo carousel is not a funding or control record).
\nThis is a documented transfer and adaptation of regulatory ideas: model authors and UK civil-society coordination, collaboration with Australian advocates, recipient-side citation, then a draft legislative process. It gives the campaign a concrete UK/Australia bridge. It also preserves the substantive disagreement: advocates wanted a broader duty than the eventual UK implementation, and Australia's draft must be assessed on its own clauses. The material does not connect the September2026 draft to ControlAI direction or a common financial command.
\nPrior local kidron-money.md records an Indigo grant to5Rights and William Perrin's trustee/grant-governance role; that is a useful older money lead, not refreshed here. Indigo's current biography returned a redirect page. No new financial edge should rely on that failed refresh. The most discriminating remaining records are the review's drafting correspondence and submission-response matrix, contemporary consultation invitations/version logs, and the grant agreements funding the relevant Australian advocacy work. Public statements alone do not resolve approval rights or earmarking.
\nThe user's autonomy objection supplies a concrete audit question: does each obligation protect a user's ability to choose, or permit another institution to suppress a lawful choice without that user's agreement? The mechanism must be answered clause by clause, keeping stated aims separate from foreseeable effects.
", "path": "duty-care-wave-2026-09-09/policy-lineage.md", "sourceIds": [ "w10-duty-s-review", "w10-duty-s-carnegie2022", "w10-duty-s-network68", "w10-duty-s-joint-article", "w10-duty-s-reset70", "w10-duty-s-network-rename" ], "markdown": "Observed 9 September 2026. Bounded lineage run after the recorded assurance forest review. Draft legal operation and access are separate companion notes.\n\n## A direct intellectual and advocacy connection\n\nAustralia's official statutory review, section 5.5, printed p53, attributes the overarching duty-of-care conception to Lorna Woods and William Perrin. It cites Carnegie UK's 2019 model bill and its 2021 criticism of the narrower UK scheme, then favors a singular Australian duty. It directly quotes Reset.Tech Australia's submission70 when explaining why content-specific duties would be insufficient. This is recipient-side acknowledgment of models and submitted reasoning, stronger than a shared vocabulary or conference roster. It does not allocate final clause authorship to those advocates. Source: http[local research file] . Reading scope: primary indexed section5.5 text and footnotes34-38; full PDF reader exceeded13MB limit and both ordinary downloads timed out, so no full-report reading or original custody claimed.\n\nCarnegie UK's January2022 submission to Australia's House Select Committee says its clerk invited the submission. It proposes platform risk assessment, mitigation and ongoing supervision, with a stronger regulator and cross-border cooperation. Its envisaged interventions include changes to visibility, forwarding and message prompts, while acknowledging expression and proportionality. That is documented historical submission and intended mechanism, not proof each intervention was enacted. Source: http[local research file] . All six pages read in web rendering, especially paragraphs1,9,19-23. The document itself dates January2022 despite misleading search metadata.\n\nThe Online Safety Act Network's June2024 submission68 names Maeve Walsh as lead and Lorna Woods as principal adviser. It explicitly promotes Australian alignment with UK/EU regulation to strengthen regulators against multinational firms. Its approach covers design, operation, resourcing and business models, while warning that systems measures and content interventions cannot be completely separated. It also warns against unnecessary overmoderation and stresses privacy-by-design. Source: http[local research file] . Entire ten-page substantive rendering read. This is an advocate's proposal; the government host is not government endorsement of every statement.\n\nA further direct working relationship is a jointly authored 22April2024 article by Rys Farthing and Woods advocating a singular Australian duty and discussing the UK divergence. The author disclosure says Reset Australia received organizational support from Reset Global, Luminate and Internet Society Foundation at that time. It does not identify a grant commissioning this article or funding Woods personally. Its claim about technology companies preferring the narrower scheme is author interpretation, not a recovered company instruction. Source: http[local research file] . Full substantive body/bylines/disclosure read. Hosted outlet identity may have changed; preserve date and original authors rather than treating this as a new2026 article.\n\nReset's submission70 explicitly names Woods/Perrin's model and covers access/content creation, discovery, user tools and platform tools. Its singular-duty argument appears on printed pp4-5; later printed p26 proposes vetted researcher access including third-sector organizations. That is a potential institutional-benefit design, not an awarded right to Reset or proof of control over platform data. Source: http[local research file] . Scoped passages PDF6-7 and PDF28, not complete31-page reading.\n\n## Resources and organized activity\n\nReset.Tech Australia's own2024 transparency report describes a duty-of-care roundtable, paper, small advocacy coalition, review submissions and parliamentary evidence. Its finance page lists Reset.Tech(US), Internet Society Foundation, Susan McKinnon Foundation, Boundless Foundation, Manniferra (spelling as printed), an unnamed private donor, and NSW Government support specifically for attendance at the social-media summit. No amounts or award terms appear there. This is recipient-reported funding; it cannot attribute the duty campaign's expenses to any particular donor, treat summit travel support as state commissioning, or establish donor editorial rights. Sources: http[local research file] , printed pp6,37-38 and scoped project passages. Not read cover-to-cover. The2024 funding roster differs from the earlier article's roster; do not conflate periods or infer withdrawal.\n\nCurrent naming needs care: the former Online Safety Act Network announced its change to Online Safety Network on1September2026, expressly expanding attention to AI and emerging technologies. It retains focus on online-safety implementation while advocating product-safety regulation for AI chatbots and broader digital regulation. This is explicit organizational scope expansion, not an inferred AI connection from a name. Sources: http[local research file] (full body) and http[local research file] (current mission/leadership; logo carousel is not a funding or control record).\n\n## Implication and next records\n\nThis is a documented transfer and adaptation of regulatory ideas: model authors and UK civil-society coordination, collaboration with Australian advocates, recipient-side citation, then a draft legislative process. It gives the campaign a concrete UK/Australia bridge. It also preserves the substantive disagreement: advocates wanted a broader duty than the eventual UK implementation, and Australia's draft must be assessed on its own clauses. The material does not connect the September2026 draft to ControlAI direction or a common financial command.\n\nPrior local kidron-money.md records an Indigo grant to5Rights and William Perrin's trustee/grant-governance role; that is a useful older money lead, not refreshed here. Indigo's current biography returned a redirect page. No new financial edge should rely on that failed refresh. The most discriminating remaining records are the review's drafting correspondence and submission-response matrix, contemporary consultation invitations/version logs, and the grant agreements funding the relevant Australian advocacy work. Public statements alone do not resolve approval rights or earmarking.\n\nThe user's autonomy objection supplies a concrete audit question: does each obligation protect a user's ability to choose, or permit another institution to suppress a lawful choice without that user's agreement? The mechanism must be answered clause by clause, keeping stated aims separate from foreseeable effects.\n", "file": "research/duty-care-wave-2026-09-09-policy-lineage.md" }, { "id": "dv-policy-pressure-wave-2026-09-09-counterpressure", "title": "DoubleVerify / Check My Ads: preservation demand and the distinct Adalytics complaint", "description": "Acquisition closed 9 September 2026. This packet addresses the published April 2025 preservation letter and a finite public counterpart check. It does not adjudicate the competing technical claims, establish current litigation status, or investigate underlying harmful material.", "html": "Acquisition closed 9 September 2026. This packet addresses the published April 2025 preservation letter and a finite public counterpart check. It does not adjudicate the competing technical claims, establish current litigation status, or investigate underlying harmful material.
\nThe concrete pressure mechanism is a lawyer's notice of anticipated litigation plus a broad preservation-and-confirmation demand. The next observed judicial step belongs to a different recipient: the acquired May 2025 complaint sues Adalytics Research LLC, not Check My Ads Institute. These should be separate relationships in any graph.
\nThe recipient-published, redacted [14 April 2025 letter](http[local research file] is one page on Paul Hastings letterhead, visibly signed Bradley J. Bondi, acting for DoubleVerify Holdings, Inc., and addressed to Arielle Garcia care of Check My Ads Institute. It identifies possible defamation, tortious interference and injurious falsehood concerning the March 28 Adalytics bot report and related public statements. It says Check My Ads should anticipate litigation and asks for email confirmation by noon EDT on 16 April that records will be preserved. The requested scope includes documents, messages and electronic information under its possession, custody or control relating to DoubleVerify, the named reports, other Adalytics posts referencing DV, communications with regulators or government officials about DV, and communications with Adalytics generally. [PDF1, both substantive paragraphs; full page text and image inspected.]
\nThis copy demands preservation and confirmation, not immediate document production, retraction, cessation of publication, damages or a confidentiality undertaking. The lawyer asserts the litigation posture; the letter is not a court order and does not by itself settle the legally enforceable preservation scope. No confirmation response or implementation record was acquired.
\nThe [16 April Check My Ads statement](http[local research file] is credited to Nandini & Claire. It reports receipt on April 14, says it will comply with its legal obligations, objects that the demand is overbroad, and invokes speech and government-communication protections. It stands by its work and says its work will continue. Those are the recipient's positions. The public statement and publication of the letter do not prove that no chilling effect occurred; equally, they do not establish actual silencing, a judicial overbreadth ruling, completed preservation or production. [Official reader lines18–36, entire substantive statement read.]
\nOrdinary GET returned403. The saved HTML is only that error response. The already-readable official-site text is retained as counter-statement.reader.txt; it is not an HTML original. No alternate language, mirror, form or repeated failed GET was used.
DV's [7 April rebuttal](http[local research file] disputes the report's treatment of bot traffic, billing, code attribution and detection. It says impressions can be removed from billable counts after delivery and characterizes the research and Check My Ads amplification as misleading. It also specifically alleges inaccurate representations to federal agencies and Congress. This is a useful rival explanation for DV's action: it publicly presents the dispute as protection against false commercial claims. It is DV's account, not an independently audited result, and predates rather than responds to the April 14 letter. No direct DV public answer to the recipient's April 16 constitutional objection was located in the finite pass. [Read title/date, opening takeaways and body through technical standards discussion; official reader lines2–77, with related passage95–97. Original HTML retained; no linked research, bot logs, billing data or third-party standards independently audited in this lane.]
\nAn [AdExchanger-hosted republication of the stamped complaint](http[local research file] records Document1 filed12 May2025, civil action8:25-cv-01535, US District Court for the District of Maryland. Plaintiffs are DoubleVerify Holdings Inc. and DoubleVerify Inc.; the sole named defendant in this original complaint is Adalytics Research, LLC. Paragraph68 alleges coordination with Check My Ads to amplify statements. That is a plaintiff allegation, not a finding or a claim that Check My Ads was sued in this document. [PDF1 caption and opening; PDF23 paragraphs68–70.]
\nThe requested remedies include compensation/disgorgement, punitive damages, costs and fees, and an order for retraction and corrective advertising. They are prayers for relief, not granted remedies. The signature page lists Paul Hastings counsel, including Bondi, alongside signing counsel Kwame J. Manley. This gives documentary continuity of counsel and subject matter, not proof of a particular preservation response or a case against Check My Ads. [PDF31 prayer A–E; PDF32 signature.]
\nCustody is a publisher's copy of a filed pleading bearing court filing headers, not a fresh court-docket acquisition. Read PDF1,23–24,29–32 only; the complete32-page PDF was retained and extracted, but intervening factual allegations and all counts were not fully read. No answer, amended complaint, dismissal, judgment, settlement, production order or completed corrective publication was acquired. Current case status remains outside this packet.
\nTwo mechanisms are now evidenced: DV counsel to Check My Ads, preservation demand under a stated litigation threat; and DV corporate plaintiffs to Adalytics, filed claims seeking damages and correction. The shared subject matter and counsel do not erase their different addressees or procedural states. A public research-to-lawmakers route was within the letter's requested preservation scope; it was not shown to be prohibited by an exercised court power.
\nThe most discriminating missing records are Check My Ads' actual response/confirmation and any later filed instrument specifically naming it; for the Adalytics branch, the court's subsequent disposition or an executed settlement would distinguish requested correction from achieved correction. A public rebuttal and a demand letter can be both reputation defense and pressure on a critic; these sources alone do not decide the merits or measure downstream chilling.
\nSeven discovery queries were run: exact statement title on Check My Ads; DoubleVerify with April14/2025/preservation; site.doubleverify.com with Check My Ads/legal letter (literal malformed domain operator, not treated as an exhaustive domain search); DoubleVerify/Bradley/Check My Ads; DoubleVerify/Check My Ads/April2025/response; site:doubleverify.com with Adalytics/May12,2025/lawsuit; and site:doubleverify.com with Check My Ads/letter. Read the exact statement and linked letter, the discovered DV response, and scoped portions of the directly surfaced complaint. Secondary news results were discovery context only, not proof of the operative demands or outcome. No broader donor, legal-firm, related-company, or separate-lawsuit branch opened.
Successful original captures: recipient-posted redacted letter, DV response HTML and media-hosted complaint republication. One403 statement-original route closed; official reader substitute retained. Manifest counter-captures.json records hashes and precise scope. No forms, accounts, private communications, customer records, listed-site visits or harmful images accessed. An incidental unrelated local memory read was rejected by automatic approval review and was not retried; it supplied no evidence.
9 September 2026. The public record now identifies DoubleVerify's membership start and contribution band, and describes what membership supplies. It also shows a route for funding members to participate in governance. None of the inspected records gives DV a reserved board seat, authority to dictate an individual content assessment, or a payment tied to a particular removal. Acquisition is closed after four directly relevant current pages and reuse of retained records.
\nThe current [IWF member directory](http[local research file] places DoubleVerify beneath the heading Members contributing over £25,000, with membership since 1 April 2025. I checked the retained HTML card and its nearest preceding H2 heading, as well as the reader. This is IWF's public contribution-band disclosure, not a £25,000 exact amount, an independently reconciled receipt or a dated2025 invoice. No upper bound is assumed. The snapshot does not identify when DV entered that band or which accounting period it covers. Original-derived191 and210–212; reader401 and451–455.
\nIWF's [current pricing page](http[local research file] describes annual fees paid in GBP, calculated by sector and company size, ranging from £5,000+ to £100,000+. It says membership fees directly support front-line work and that the structure is developed with members. This establishes the direction and stated purpose of the membership resource flow: member funds to IWF. It is a current undated framework, not proof of the tariff or terms accepted by DV in2025. The directory band and annual framework should remain separately attributed rather than manufactured into a specific annual payment. Original-derived57–68; reader53–58.
\nThe retained [FY2025 accounts](http[local research file] PDF21 income policy, distinguish subscription amounts receivable for services from donations and grants. PDF29 says deferred income relates to subscriptions and licence fees. Recognition is therefore not interchangeable with cash receipt. The accounts end31March2025, before the directory's stated DV membership start, and no DV-specific transaction was identified in the inspected passages. They cannot establish its subsequent payment. No grant to IWF from DV or outcome-based bounty is shown in this packet.
\nThe [membership overview](http[local research file] expressly describes licensed access to dynamic blocking lists and alerts, with the underlying imagery assessed by IWF analysts. The [benefits page](http[local research file] also offers expertise, direct analyst contact, research/intelligence, networking with decision makers, policy-development participation and membership publicity/logo use. These are meaningful services and relationship benefits. They are public descriptions rather than an executed service commitment or proof that DV used every opportunity. Overview original-derived85–88 and100–110, reader74/84–87; benefits original-derived62–79, reader57–81.
\nThe accepted April/May2025 partnership records separately establish the stated plan to incorporate IWF's URL List into DV's category, followed by a later participant testimonial about integration. Those retained records are reused, not reacquired here. The public NCMEC notification material described in DV's earlier category work is a different input. Membership gives licensed access, not a demonstrated right to redistribute raw IWF lists freely. DV's separately reported free distribution of its own Highly Illicit category should not be treated as proof of the underlying IWF licence's sublicensing or derivative-use terms.
\nThe current membership overview says members may suggest policy ideas and apply to join the Funding Council. It does not promise an automatic council seat. Original-derived107–109; reader86.
\nThe reused FY2025 report, PDF11, adds the structural route: all members have the opportunity to nominate representatives to the Funding Council; that council supports the board with industry advice and elects three industry trustees. The described eleven-person board comprises an independent chair, six other independent trustees, three industry trustees and one co-opted trustee. The board approves appointments through the described processes. These are the report's governance arrangements, not an assertion that the current2026 composition is unchanged. They establish an institutional channel for industry voice; they do not establish that DV nominated a representative, sat on the council or obtained any seat itself.
\nIWF is identified in that report as company03426366 and charity1112398. The report's legal-member/guarantee discussion is not used to infer that the public DV brand entry proves a specific corporate signatory or guarantor. The member agreement, legal entity and applicable constitutional rights remain unacquired at that level.
\nNo inspected membership description provides DV a veto over IWF's assessment of a URL or image, an entitlement to suppress its own adverse assessment, or a right to demand that DV's commercial category criteria become IWF's criteria. Equally, absence of those rights here is not proof that an undisclosed agreement contains none. PDF33 note19's transaction-related disclosure must not be stretched into proof that funders have no policy influence; the council mechanism shows why those propositions differ.
\nThe disclosed exchange is recurring membership support in return for licensed technical resources, expertise, reputation and opportunities for policy/governance participation. It fits ordinary service-support membership and also creates channels through which an industry member can contribute ideas. Money plus participation does not establish control of a particular assessment. No payment-per-blocking theory is supported by the acquired fee model.
\nThe useful next record is DV's executed2025 membership agreement and fee/service schedule, identifying the exact legal signatory, invoiced period/amount, licensed dataset and permitted derivative distribution, plus any assessment-review, council or conflict provisions. An invoice/receipt would resolve payment without resolving governance; council nomination/election records would resolve exercised participation without proving content decisions. No contract request or contact was made. Older other-company agreements, unrelated donors and generic fundraising genealogy were not pursued.
\nmembership-captures.json records four successful original HTML captures and the reused original accounts, with exact reading scopes and hashes. membership-retrieval-scope.json records the two limited searches and closed acquisition boundary. The directory was read for DV's card and governing heading only; other names returned in its body were not inventoried. No images, lists, list entries, harmful-content locations, forms, accounts or visible UI were opened or operated. The source pages may contain navigation to such services; links were not followed. No case, graph, ZIP, current-work or forest was changed.
| Original | \nSHA256 | \n
|---|---|
membership-directory.html | \n857cf31d4077b15c195388c857f2cb0ebb3cb97a900b27827779b989d32b9425 | \n
membership-fees.html | \n0bb92583593a6b52dcb776644804fddfb2e59f5aca4bfa9c23091449db98b807 | \n
membership-benefits.html | \n0c72ba0d86a6f8fe140c752211aa0c65ace4884374ce2957b91fdf29f348a095 | \n
membership-overview.html | \n835152238c6feeecd487821c4194d02ef1722a9be00dd10b7148c0a7ea4b8835 | \n
Existing recipient-iwf2025.pdf | \n0f96cd95f93b3155cb46163dc64fe04604148d8a57e6d9f5e99c5459d465d433 | \n
9 September2026. Root acquisition closed. The contemporaneous record identifies public criticism, advocacy, Senate demands and vendor changes. It does not establish a sole cause, a government enforcement order or a UK-government instruction.
\nThe [Blackburn office release of7February2025](http[local research file] states that Senators Marsha Blackburn and Richard Blumenthal received Adalytics research before public release and sent letters to Google, Amazon, DoubleVerify, Integral Ad Science, MRC and TAG. Three original letters are retained and fully read here; the other recipients are established by this release, not a reading of their letters.
\nThe [DV letter](http[local research file] dated7February and addressed to Mark Zagorski, seeks answers by14February. It asks about remediation, site blocking, advertiser notification, annual use of NCMEC transparency reports, reports made to authorities and revenue connected to the questioned placements. It also criticizes limited page-level reporting. These are requests from senators, not acquired proof of the underlying allegations, a subpoena or a completed sanction. The footnotes identify the2021 outbound NCMEC-notification table, an explicit input reference. Its notice methodology remains distinct from a judicial finding about the whole service.
\nThe [MRC letter](http[local research file] PDF1–4, challenges the accreditor to review vendors and explain possible revocation, audits and fee policies. It cites2018verification guidance; the letter is evidence of the senators' interpretation and demand, not a fresh reading of that standard or an established violation. This adds leverage over the certification on which commercial trust rests, beyond a direct complaint to the vendor.
\nThe [TAG letter](http[local research file] PDF3 question8, makes the scope dispute explicit: it asks whether certified companies may monetize a website engaged in illegal activity if their ads avoid the prohibited content itself. Questions on review/revocation and a14February response deadline appear onPDF2. This is documented pressure toward considering the funding of a whole website, not merely adjacent content. It is not proof that all such sites were subsequently blocked or that the letter changed TAG's rules.
\n[Check My Ads'7February release](http[local research file] claims its advocacy helped the inquiry. Its named policy/operating leaders urge transparency and advertiser pressure on vendors and accreditors. That is an actor's account of its contribution; the acquired Senate release independently establishes the letters and early research access but does not credit CMA with drafting them. The TAG letter's footnote7 separately cites an Arielle Garcia trial-update article; citation is an actual information route, not authorship of the letter. The underlying trial testimony and complete Adalytics technical report were not inspected in this lane, and no harmful destination was visited.
\nDV's retained7February statement acknowledges additional blocking/review in response to the report while disputing the interpretation of its technology and the specific accusation about placements beside illegal imagery. It announces development of a P2P category. Its13February announcement follows the letters and precedes their requested response deadline; the25February URL-reporting announcement overlaps another criticism. The April24 progress claim follows these, and the separate IWF membership source dates membership from1April2025. Chronological and topical convergence support a reactive commercial/political pressure account; they do not isolate which actor caused which decision, prove compliance with the letters, or establish IWF as the originator of the February policy.
\nTAG's original [2025 Impact and Compliance Report](http[local research file] visibly printedPDF20–21, reports investigations of the CSAM-monetization complaints and says none of the investigated companies violated its BSC guidelines. It cites a23-page response by Mike Zaneis to the senators, explains limits of accreditation and rejects a criminal-adjudication role. It also reports NCMEC training for its threat-exchange members. Its [3February2026 release](http[local research file] repeats the no-violations outcome. This is TAG's own reported determination, not independent exoneration or a complete investigation file. The full23-page response, named findings and applicable underlying audit files remain unacquired. The record contradicts a simple claim that Senate pressure automatically produced certification revocations. Additional protective measures can coexist with a finding that existing rules were not breached.
\nA concrete sequence is now supported: disputed research → claimed advocacy assistance and documented Senate letters → explicit demands for broader responsibility and certification scrutiny → vendor-reported safeguards and transparency changes → a later licensed nonprofit-data relationship. Government pressure, NGO advocacy, commercial decisions and technical inputs are separate acts. IWF's current fee band and membership governance are in membership-resources.md; the later legal counterpressure is in the companion packet. No one-way obedient coalition is established.
\nThe highest-value missing records are DV's actual dated reply and product-change decision record, TAG's specifically described23-page reply and case findings, and the correspondence establishing CMA's contribution. They would distinguish chronological proximity from exercised influence. An acquired fee schedule would answer a different financial question. No outreach was attempted.
\nFour originals from the initial acquisition succeeded: Senate HTML and three PDFs. CMA's ordinary GET returned403, was closed without retry, and its separately available primary reader was retained as cma-advocacy.reader.txt (not original HTML). Root read that substantive reader25–88. The Senate letters' complete reader and original-PDF text were read. The Senate release's full substantive reader99–123 was read.
\nTAG summary and release HTML plus29-page original PDF were retained. Its PDF text layer repeats the same brand-safety text on unrelated pages; those repetitions were not treated as multiple passages. Root rendered and visually inspectedPDF17/18, recognized that mismatch, then visually read actualPDF20–21. Only that visible brand-safety section, the summary's due-process item and full substantive release0–27 support the substantive finding; remaining report metrics and compliance lists are outside the audit. The first report-button GET returned a normal200HTML meta redirect, initially mislabelled as a closed report route; one method correction saved the redirect and followed its explicit public target. This was not an access-denial retry; previous transport metadata is preserved.
\nFocused searches for recipient responses returned the existing DV statement, unrelated reports and TAG's later official report. Two exact searches for the23-page reply returned no relevant copy; that document remains a reopening lead. Search absence is not proof that no reply or public copy exists. No new standards, customer contract, live ad placement, operational list, court merits or full report-methodology audit is claimed.
", "path": "dv-policy-pressure-wave-2026-09-09/policy-pressure.md", "sourceIds": [ "w30-class-s-senate-release", "w30-class-s-cma-advocacy", "w30-class-s-senate-dv", "w30-class-s-senate-mrc", "w30-class-s-senate-tag", "w30-class-s-tag-report", "w30-class-s-tag-release" ], "markdown": "9 September2026. Root acquisition closed. The contemporaneous record identifies public criticism, advocacy, Senate demands and vendor changes. It does not establish a sole cause, a government enforcement order or a UK-government instruction.\n\n## An identifiable request and a deadline\n\nThe Blackburn office release of7February2025 (see source references) states that Senators Marsha Blackburn and Richard Blumenthal received Adalytics research before public release and sent letters to Google, Amazon, DoubleVerify, Integral Ad Science, MRC and TAG. Three original letters are retained and fully read here; the other recipients are established by this release, not a reading of their letters.\n\nThe DV letter (see source references) dated7February and addressed to Mark Zagorski, seeks answers by14February. It asks about remediation, site blocking, advertiser notification, annual use of NCMEC transparency reports, reports made to authorities and revenue connected to the questioned placements. It also criticizes limited page-level reporting. These are requests from senators, not acquired proof of the underlying allegations, a subpoena or a completed sanction. The footnotes identify the2021 outbound NCMEC-notification table, an explicit input reference. Its notice methodology remains distinct from a judicial finding about the whole service.\n\nThe MRC letter (see source references) PDF1–4, challenges the accreditor to review vendors and explain possible revocation, audits and fee policies. It cites2018verification guidance; the letter is evidence of the senators' interpretation and demand, not a fresh reading of that standard or an established violation. This adds leverage over the certification on which commercial trust rests, beyond a direct complaint to the vendor.\n\nThe TAG letter (see source references) PDF3 question8, makes the scope dispute explicit: it asks whether certified companies may monetize a website engaged in illegal activity if their ads avoid the prohibited content itself. Questions on review/revocation and a14February response deadline appear onPDF2. This is documented pressure toward considering the funding of a whole website, not merely adjacent content. It is not proof that all such sites were subsequently blocked or that the letter changed TAG's rules.\n\n## Advocacy and vendor response\n\nCheck My Ads'7February release (see source references) claims its advocacy helped the inquiry. Its named policy/operating leaders urge transparency and advertiser pressure on vendors and accreditors. That is an actor's account of its contribution; the acquired Senate release independently establishes the letters and early research access but does not credit CMA with drafting them. The TAG letter's footnote7 separately cites an Arielle Garcia trial-update article; citation is an actual information route, not authorship of the letter. The underlying trial testimony and complete Adalytics technical report were not inspected in this lane, and no harmful destination was visited.\n\nDV's retained7February statement acknowledges additional blocking/review in response to the report while disputing the interpretation of its technology and the specific accusation about placements beside illegal imagery. It announces development of a P2P category. Its13February announcement follows the letters and precedes their requested response deadline; the25February URL-reporting announcement overlaps another criticism. The April24 progress claim follows these, and the separate IWF membership source dates membership from1April2025. Chronological and topical convergence support a reactive commercial/political pressure account; they do not isolate which actor caused which decision, prove compliance with the letters, or establish IWF as the originator of the February policy.\n\n## An accreditor did not simply adopt the accusation\n\nTAG's original 2025 Impact and Compliance Report (see source references) visibly printedPDF20–21, reports investigations of the CSAM-monetization complaints and says none of the investigated companies violated its BSC guidelines. It cites a23-page response by Mike Zaneis to the senators, explains limits of accreditation and rejects a criminal-adjudication role. It also reports NCMEC training for its threat-exchange members. Its 3February2026 release (see source references) repeats the no-violations outcome. This is TAG's own reported determination, not independent exoneration or a complete investigation file. The full23-page response, named findings and applicable underlying audit files remain unacquired. The record contradicts a simple claim that Senate pressure automatically produced certification revocations. Additional protective measures can coexist with a finding that existing rules were not breached.\n\n## What this changes and what remains\n\nA concrete sequence is now supported: disputed research → claimed advocacy assistance and documented Senate letters → explicit demands for broader responsibility and certification scrutiny → vendor-reported safeguards and transparency changes → a later licensed nonprofit-data relationship. Government pressure, NGO advocacy, commercial decisions and technical inputs are separate acts. IWF's current fee band and membership governance are in membership-resources.md; the later legal counterpressure is in the companion packet. No one-way obedient coalition is established.\n\nThe highest-value missing records are DV's actual dated reply and product-change decision record, TAG's specifically described23-page reply and case findings, and the correspondence establishing CMA's contribution. They would distinguish chronological proximity from exercised influence. An acquired fee schedule would answer a different financial question. No outreach was attempted.\n\n## Custody and limits\n\nFour originals from the initial acquisition succeeded: Senate HTML and three PDFs. CMA's ordinary GET returned403, was closed without retry, and its separately available primary reader was retained as cma-advocacy.reader.txt (not original HTML). Root read that substantive reader25–88. The Senate letters' complete reader and original-PDF text were read. The Senate release's full substantive reader99–123 was read.\n\nTAG summary and release HTML plus29-page original PDF were retained. Its PDF text layer repeats the same brand-safety text on unrelated pages; those repetitions were not treated as multiple passages. Root rendered and visually inspectedPDF17/18, recognized that mismatch, then visually read actualPDF20–21. Only that visible brand-safety section, the summary's due-process item and full substantive release0–27 support the substantive finding; remaining report metrics and compliance lists are outside the audit. The first report-button GET returned a normal200HTML meta redirect, initially mislabelled as a closed report route; one method correction saved the redirect and followed its explicit public target. This was not an access-denial retry; previous transport metadata is preserved.\n\nFocused searches for recipient responses returned the existing DV statement, unrelated reports and TAG's later official report. Two exact searches for the23-page reply returned no relevant copy; that document remains a reopening lead. Search absence is not proof that no reply or public copy exists. No new standards, customer contract, live ad placement, operational list, court merits or full report-methodology audit is claimed.\n", "file": "research/dv-policy-pressure-wave-2026-09-09-policy-pressure.md" }, { "id": "egumball-disposition-wave-2026-09-09-later-disposition", "title": "eGumball later disposition: a failed procedural shield, not delisting", "description": "Acquired and read 9 September 2026. Bounded lane closed after five focused discovery queries and full substantive judicial-text reading.", "html": "Acquired and read 9 September 2026. Bounded lane closed after five focused discovery queries and full substantive judicial-text reading.
\nThe source is CaseMine's republication of G062863M.
\nThe complete substantive web-reader body, lines95-388, was read across two overlapping calls and retained in later-opinion.reader.txt, with no gaps in that interval. The derivative preserves the modification and unmodified body separately. No underlying declarations, contract, docket, judicial PDF or sealed exhibits were acquired. Hashes and exact search scope are in later-captures.json. Root independently retained/read its own complete reader copy.
\nThe correction matters: May6 replaces the original page6 sentence with eGumball's assertion that all its processed 2020/2021 card transactions were directly connected to written client contracts (reader124-133, decisive127). The unmodified body at229 says the opposite. This remains eGumball's assertion, not an adjudicated reconciliation. Rehearing was denied, with no change in judgment.
\nMerrick Bank Corporation was the appellant. The court affirmed the denial of Merrick's special motion to strike under Code of Civil Procedure425.16, the anti-SLAPP statute. The underlying action also named Paysafe Services (US), Corp. and Visa; those parties' ultimate liability was not adjudicated here. The earlier 2023 opinion used Paysafe Payment Processing Solutions, LLC: this packet does not silently merge those legal names.
\nThe appeal applied de novo review. The two-step test first requires the defendant to establish claims arising from protected activity; only after that showing does the plaintiff need to demonstrate a probability of prevailing, described as minimal merit (printed p10; reader273-277). The actual affirmance rests on Merrick's failure to establish protected activity under the catchall public-interest provision425.16(e)(4). Conversion was outside the motion after Merrick conceded it did not arise from protected activity (fn3; reader243).
\nThe court accepted a connection to the broad subject of credit-card security, but held that this specific confidential MATCH report did not contribute to public discussion. It considered the audience, purpose and context, not simply whether the topic affected many people (printed pp11-17; reader287-349). It expressly rejected a categorical rule excluding all confidential commercial speech from anti-SLAPP protection (reader347). Thus this is neither a ruling that private risk reports always lack protection nor a final determination that this report was false or unlawful.
\nMerrick had separately argued forum selection, eGumball's contractual consent, truth and common-interest privilege (printed pp7-8; reader241-249). The decision does not establish that those defenses fail on the merits. eGumball's allegations about account restrictions, withheld funds, a fine, declined test cards and lost processing access are presented as complaint allegations (printed pp5-7; reader220-238), not findings of damages, laundering innocence or payment restoration.
\nThe court expressly applied FilmOn.com Inc. v. DoubleVerify Inc. (2019), involving confidential paid website reports tagged for adult content or copyright infringement. In this opinion, that earlier case supplies the legal comparison between commercial classification for advertisers and commercial risk reporting for acquirers (printed pp13-14; reader305-319). The comparison is significant for the investigation's structure: a privately distributed classifier does not obtain this procedural shield merely by invoking a public-interest subject. It establishes no common ownership, data feed or operational coordination between DoubleVerify and MATCH, and does not adjudicate the truth of FilmOn's challenged classifications. The named precedent remains an unacquired next-question candidate; no separate FilmOn inquiry was opened.
\nThe court also described Merrick's typical merchant waiver/hold-harmless clause as a contractual means of preempting MATCH claims (printed p17; reader349-351). Its statement that Merrick had failed to secure such a release from eGumball is explicitly attributed to eGumball. This is evidence of a described contractual protection mechanism and disputed assent in this case, not a holding of universal enforceability or an acquired executed agreement.
\nThe opinion's MATCH access, confidentiality, quarterly and per-inquiry fee descriptions belong to this historical litigation record. Footnote2 expressly uses the Mastercard Rules dated 9 February 2021, which Merrick cited; no amounts are given (printed pp3-5; reader195-215). These passages must not replace the independently acquired August2026 public rules or imply that this case established current prices.
\nThe court granted Merrick's appellate sealing motion without prejudice on its asserted proprietary/business interests. It likewise partly granted eGumball's motion without prejudice, but refused sealing of 28 specified documents because they were not subject to a trial-court sealing motion and no overriding interest had been explained. A trial referee was still considering overlapping sealing questions. The appellate order expressly disclaimed guidance to that proceeding and allowed a later unsealing application if the referee or trial court denied sealing (printed pp18-20; reader360-378).
\nThe disposition required eGumball, within ten days of issuance of the opinion, to file an updated public redacted brief removing redactions that discussed or cited the 28 documents. Footnote5 identifies Meehan preliminary-injunction declaration exhibits1,3,4,16,17,21,23,24,25,26,29,30,47,48,49,51,52,54,55,56,57,58,61,63,64,65,66, plus exhibit11 to his declaration opposing the anti-SLAPP motion and supporting the preliminary injunction. This is an ordered public-access change, not evidence the revised filing was actually made or of what those exhibits contain. The parties bore their own appeal costs.
\nTwo focused follow-up searches for this docket/brief and eGumball/Meehan/May2025 returned opinion republications and counsel publicity, but no revised brief or these exhibits. No account, docket challenge, sealed-document recovery or contact route was used. The precise frontier is the filed updated redacted brief and its filing date, followed by any public trial-court sealing/removal or final merits disposition. The present packet does not establish that such later records are absent.
\nThe earlier G060853 opinion reversed emergency TRO relief; the 2025 appeal affirmed denial of a different anti-SLAPP motion. Those outcomes answer different procedural questions. Neither acquired opinion establishes a final merits judgment, a completed MATCH correction/removal, or restored card processing.
\nThe useful new fact is that Merrick actually sought and failed to obtain this procedural protection for the listing, while the court preserved separate contract questions and ordered a narrower public record. The next records would distinguish a surviving claim from an exercised substantive remedy. No canonical case, existing evidence packet, reader or ZIP was changed.
", "path": "egumball-disposition-wave-2026-09-09/later-disposition.md", "sourceIds": [ "w30-class-s-egumball2025" ], "markdown": "Acquired and read 9 September 2026. Bounded lane closed after five focused discovery queries and full substantive judicial-text reading.\n\n## Source and version custody\n\nThe source is CaseMine's republication of G062863M.\n\nThe complete substantive web-reader body, lines95-388, was read across two overlapping calls and retained in later-opinion.reader.txt, with no gaps in that interval. The derivative preserves the modification and unmodified body separately. No underlying declarations, contract, docket, judicial PDF or sealed exhibits were acquired. Hashes and exact search scope are in later-captures.json. Root independently retained/read its own complete reader copy.\n\n**The correction matters:** May6 replaces the original page6 sentence with eGumball's assertion that all its processed 2020/2021 card transactions were directly connected to written client contracts (reader124-133, decisive127). The unmodified body at229 says the opposite. This remains eGumball's assertion, not an adjudicated reconciliation. Rehearing was denied, with no change in judgment.\n\n## What was actually decided\n\nMerrick Bank Corporation was the appellant. The court affirmed the denial of Merrick's special motion to strike under Code of Civil Procedure425.16, the anti-SLAPP statute. The underlying action also named Paysafe Services (US), Corp. and Visa; those parties' ultimate liability was not adjudicated here. The earlier 2023 opinion used Paysafe Payment Processing Solutions, LLC: this packet does not silently merge those legal names.\n\nThe appeal applied de novo review. The two-step test first requires the defendant to establish claims arising from protected activity; only after that showing does the plaintiff need to demonstrate a probability of prevailing, described as minimal merit (printed p10; reader273-277). The actual affirmance rests on Merrick's failure to establish protected activity under the catchall public-interest provision425.16(e)(4). Conversion was outside the motion after Merrick conceded it did not arise from protected activity (fn3; reader243).\n\nThe court accepted a connection to the broad subject of credit-card security, but held that this specific confidential MATCH report did not contribute to public discussion. It considered the audience, purpose and context, not simply whether the topic affected many people (printed pp11-17; reader287-349). It expressly rejected a categorical rule excluding all confidential commercial speech from anti-SLAPP protection (reader347). Thus this is neither a ruling that private risk reports always lack protection nor a final determination that this report was false or unlawful.\n\nMerrick had separately argued forum selection, eGumball's contractual consent, truth and common-interest privilege (printed pp7-8; reader241-249). The decision does not establish that those defenses fail on the merits. eGumball's allegations about account restrictions, withheld funds, a fine, declined test cards and lost processing access are presented as complaint allegations (printed pp5-7; reader220-238), not findings of damages, laundering innocence or payment restoration.\n\n## Two concrete accountability mechanisms\n\nThe court expressly applied FilmOn.com Inc. v. DoubleVerify Inc. (2019), involving confidential paid website reports tagged for adult content or copyright infringement. In this opinion, that earlier case supplies the legal comparison between commercial classification for advertisers and commercial risk reporting for acquirers (printed pp13-14; reader305-319). The comparison is significant for the investigation's structure: a privately distributed classifier does not obtain this procedural shield merely by invoking a public-interest subject. It establishes no common ownership, data feed or operational coordination between DoubleVerify and MATCH, and does not adjudicate the truth of FilmOn's challenged classifications. The named precedent remains an unacquired next-question candidate; no separate FilmOn inquiry was opened.\n\nThe court also described Merrick's typical merchant waiver/hold-harmless clause as a contractual means of preempting MATCH claims (printed p17; reader349-351). Its statement that Merrick had failed to secure such a release from eGumball is explicitly attributed to eGumball. This is evidence of a described contractual protection mechanism and disputed assent in this case, not a holding of universal enforceability or an acquired executed agreement.\n\nThe opinion's MATCH access, confidentiality, quarterly and per-inquiry fee descriptions belong to this historical litigation record. Footnote2 expressly uses the Mastercard Rules dated 9 February 2021, which Merrick cited; no amounts are given (printed pp3-5; reader195-215). These passages must not replace the independently acquired August2026 public rules or imply that this case established current prices.\n\n## The public-record remedy and exact missing filing\n\nThe court granted Merrick's appellate sealing motion without prejudice on its asserted proprietary/business interests. It likewise partly granted eGumball's motion without prejudice, but refused sealing of 28 specified documents because they were not subject to a trial-court sealing motion and no overriding interest had been explained. A trial referee was still considering overlapping sealing questions. The appellate order expressly disclaimed guidance to that proceeding and allowed a later unsealing application if the referee or trial court denied sealing (printed pp18-20; reader360-378).\n\nThe disposition required eGumball, within ten days of issuance of the opinion, to file an updated public redacted brief removing redactions that discussed or cited the 28 documents. Footnote5 identifies Meehan preliminary-injunction declaration exhibits1,3,4,16,17,21,23,24,25,26,29,30,47,48,49,51,52,54,55,56,57,58,61,63,64,65,66, plus exhibit11 to his declaration opposing the anti-SLAPP motion and supporting the preliminary injunction. This is an ordered public-access change, not evidence the revised filing was actually made or of what those exhibits contain. The parties bore their own appeal costs.\n\nTwo focused follow-up searches for this docket/brief and eGumball/Meehan/May2025 returned opinion republications and counsel publicity, but no revised brief or these exhibits. No account, docket challenge, sealed-document recovery or contact route was used. The precise frontier is the filed updated redacted brief and its filing date, followed by any public trial-court sealing/removal or final merits disposition. The present packet does not establish that such later records are absent.\n\n## Comparison with the retained 2023 appeal\n\nThe earlier G060853 opinion reversed emergency TRO relief; the 2025 appeal affirmed denial of a different anti-SLAPP motion. Those outcomes answer different procedural questions. Neither acquired opinion establishes a final merits judgment, a completed MATCH correction/removal, or restored card processing.\n\nThe useful new fact is that Merrick actually sought and failed to obtain this procedural protection for the listing, while the court preserved separate contract questions and ordered a narrower public record. The next records would distinguish a surviving claim from an exercised substantive remedy. No canonical case, existing evidence packet, reader or ZIP was changed.\n", "file": "research/egumball-disposition-wave-2026-09-09-later-disposition.md" }, { "id": "esafety-service-wave-2026-09-09-legal-clock", "title": "What could the operator's January 2025 legal claim refer to?", "description": "9 September 2026. Bounded official-source legal-clock pass, now closed. This note does not determine this operator's compliance or identify an eSafety assurance.", "html": "9 September 2026. Bounded official-source legal-clock pass, now closed. This note does not determine this operator's compliance or identify an eSafety assurance.
\nThe closest documented match is the Online Safety (Designated Internet Services—Class 1A and Class 1B Material) Industry Standard 2024, which commenced 22 December 2024, rather than January 2025. Its contemporary announcement expressly discusses some generative-AI and nudification services. That makes it a strong candidate for the instrument being loosely described in the email; it does not establish that this was the operator's intended reference. The separate RES standard has the same clock. The other concrete AI-related changes checked here commenced in May and September 2024.
\nThe source assertion is the respondent's 25 December 2024 email, section 6, retained Tilak declaration Exhibit 6 PDF21: it says eSafety communicated specific requirements, the service complied, and was confirmed aligned with upcoming AI legislation effective January 2025. It is an interested participant's account, with no instrument name or enclosed Australian assurance on that page. Local source: ../sf-compliance-wave-2026-09-09/tilak-declaration-june13-2025.pdf; this pass read the existing PDF21 text extract, not the rest of the declaration afresh.
\n| Instrument | \nMaking / registration | \nCommencement | \nRelevance and boundary | \n
|---|---|---|---|
| DIS Class1A/1B Standard2024, F2024L00710 | \nMade19June2024 by Julie Inman Grant; Federal Register registration21June2024 | \n22December2024 | \nClosest substantive match: includes high-impact generative-AI services and model distribution platforms, with duties directed to specified seriously harmful classes. | \n
| RES Class1A/1B Standard2024, F2024L00711 | \nMade19June2024 by Julie Inman Grant; Federal Register registration21June2024 | \n22December2024 | \nCompanion instrument for relevant electronic services. Its date is a cross-check, not a service-classification finding. | \n
| BOSE Amendment Determination2024, F2024L00590 | \nMinister Michelle Rowland; registered30May2024 | \n31May2024 | \nAdds express generative-AI expectations. This is a different legal mechanism from the mandatory standard, and an earlier clock. | \n
| Criminal Code Amendment (Deepfake Sexual Material) Act2024, C2024A00078 | \nAssent and registration2September2024 | \n3September2024 | \nCriminal transmission offences covering altered/created sexual depictions; not a January service-approval regime. | \n
Both standards' section2 says commencement is six months after the later of the day after registration under the Online Safety Act and the day after registration under the Legislation Act. Original covers/footers establish Federal Register registration; eSafety's published guidance expressly confirms the resulting22December date. The contemporary21June announcement says both were registered that day and would come into force six months after registration, following the disallowance period. The22December date therefore comes from the regulator's explicit account as well as the statutory formula, not an assumption that all registration events must be identical.
\nSources: [DIS original](http[local research file] [RES original](http[local research file] covers and PDF5/printed1, sections1–5; [eSafety21June2024 announcement](http[local research file] article body, reader lines212–237; [eSafety regulatory-guidance index](http[local research file] Online Safety Codes and Standards and Basic Online Safety Expectations sections (current page, last-updated26May2026).
\nSection3 identifies section145 of the Online Safety Act2021 as authority. Section5 applies wherever a designated service is supplied from, only insofar as it serves end-users in Australia, and provides a predominant-purpose exception where another applicable code or standard better aligns with the service. The high-impact generative-AI definition in section6 turns on machine-learning generation capacity and synthetic high-impact material; it excludes a service with controls rendering that risk immaterial. For that category, high-impact material uses X18+/RC film or game and Category2-restricted/RC publication classifications. This is a category trigger, not a finding that every generated adult image breaches the standard. Section4's protective object and the actual duties concern Class1A/1B material. Those classes include child sexual exploitation/pro-terror/extreme crime-and-violence material and specified other crime/violence/drug material. (Original PDF5–7 and12/printed1–3,8.)
\nThe June announcement specifically says the standards may cover apps generating pornography or nudifying images without effective controls preventing their application to children. That sentence supplies the concrete service-type connection to the operator's description. It still does not classify or clear any named service.
\nFor a covered high-impact generative-AI DIS, section22(3) requires preventive systems for child-sexual-exploitation/pro-terror outputs; regular model testing/review and prompt mitigation adjustments; differentiation of AI outputs; warnings for Australian users specifically seeking child-abuse images; report/support information for relevant generation terms; and automated detection/action on child-abuse material in training data, prompts and outputs. The notes acknowledge limited model visibility/control and require reliance on available alternatives in those circumstances. PDF33–34/printed29–30, visually checked.
\nThe instrument also requires contractual provisions governing Class1A/1B use and giving suspension, restriction, termination or material-removal rights for child-exploitation/pro-terror misuse, plus appropriate enforcement in the circumstances specified by section13. Section19 requires management, supervision, internal reporting and sufficient skilled personnel. Section20 requires appropriate known-CSAM detection/removal, with express technical/practicability and encryption/systemic-weakness exceptions and alternative-action duties; it does not displace section22. These are scoped readings of sections13(2),13(5),19 and20, not a claim to have audited every applicable clause. Locators: PDF24–25/printed20–21; PDF30–31/printed26–27.
\nTwo timing qualifications matter. Section7's ordinary first risk-assessment deadline can be six months after commencement, but section7(6)(f) expressly exempts high-impact generative-AI DIS from subsections(1) and(4), with a material-change qualification in the following page's note. Do not mechanically assign that initial assessment deadline to every AI service. Section36(3) prevents a compliance-report request under that particular section before the first anniversary of commencement, and permits no more than one in12months; subsection(5) gives two months to respond. Those limits do not mean the standard's substantive duties wait a year or that all other reporting powers are unavailable. Section38 requires compliance-action records to be kept for at least two years after the relevant calendar year's end. PDF17–18,44–46/printed13–14,40–42.
\nNone of these readings establishes a general eSafety certification of an operator's entire business, immunity under other laws, or satisfaction of the duties in practice.
\nThe [BOSE amendment](http[local research file] section2 commenced the day after30May registration, hence31May2024. Schedule1 item7 inserts section8A: reasonable steps for end-user safety in design, implementation and maintenance of generative-AI capabilities, and proactively minimising their use for unlawful/harmful material or activity. Examples concern lifecycle risk assessment, training data and harmful prompts. Original PDF3,5–6/printed1,3–4. The expectation is broader in subject matter than the DIS harmful-material classes, but its enforcement form must be distinguished: Online Safety Act section45(4) says a determination under that section does not itself impose a duty enforceable by court proceedings. Part4 separately supplies reporting mechanisms. [Act compilation in force11December2024](http[local research file] sections44–45, scoped reader only.
\nThe current eSafety guidance index dates the BOSE guidance's January2025 revision to the replacement of AAT by ART. A guidance revision date must not be relabelled legislative commencement. The available record does not establish that the respondent meant this update, or knew of it before the email.
\nThe [Deepfake Sexual Material Act](http[local research file] is another genuine2024 AI-adjacent change, but its section2 is explicit:3September2024. New Criminal Code section474.17A concerns using a carriage service to transmit specified sexual material of a person who is or appears18+, with knowledge of lack of consent or recklessness about consent. Altered or wholly technologically created depictions count; transmit includes making available, publication, distribution, advertising and promotion. There are specified exceptions. The creation/alteration aggravated offence under474.17AA(5) also requires the underlying transmission offence. This is not a blanket criminalisation of every act of making an image, and not a prospective eSafety approval system. Schedule1 item7 applies the amendments to material transmitted after commencement irrespective of when created/altered. PDF5–10,13/printed1–6,9.
\nFour official original PDFs were obtained by ordinary public GET and preserved, with acquisition timestamps, finalURLs, bytes and SHA256 in legal-captures.json. Complete text extracts were generated for retrieval, but generation is not full-document reading. DIS read scope: original title, sections1–5, selected section6 definitions, section7, selected section8/9 material, sections10–12 table/overlap rules, selected13/14 clauses,19–20,22–23 thresholds,36–38; substantive pages were inspected through retained extraction and the public reader. Visual checks were PDF1,5,33,34. RES scope: cover and sections1–5 only; visualPDF5. BOSE: all8pages extracted text read; visualPDF3. Deepfake Act: PDF5–10,13 text; visualPDF6. Other rendered pages are not claimed as visual reads.
\nDIS SHA256 afe30304b7094509de84a1845534b346a48c4fc2f2f07e163fd47fd040ce487d; RES75d2ce8834e804815dad8b72102bd61d4a2ec757b3edd336be541bfee0f3b8b9; BOSE39171722f9c258fa07d9e69324f781c0de000b358defc159fb492b2dfe5eae3a; Deepfake Act1ab54fc6a743c991d2db80aee8897206b10000933649364b77d53d6a40a60e1d.
\nThe June eSafety announcement and current guidance-index ordinary GETs timed out. Their public web-reader/indexed-primary passages were available and read; no original-HTML hash is asserted, and no transport retry was pursued. Act45 was a scoped public-reader read, not a retained original. This lane made no service-specific query or target-service visit. No January2025 commencement matching the asserted AI regime was located in this bounded pass; this is not an exhaustive negative over all Australian instruments.
\nCarver's separate service-specific lane has identified official ClothOff correspondence in FOILog141/request25202 that expressly invokes the22December2024 DIS commencement. That is a useful separately sourced counterpart for the clock; this lane has not independently reread that bundle and does not join ClothOff to Drawnudes or promote a request for further information into compliance confirmation.
\nThe next discriminator is the actual eSafety-to-respondent correspondence: dated requirement list, named legal basis/service classification, evidence supplied in reply, and exact wording/scope of any acknowledgement or assessment. Likely custodians are eSafety's relevant industry-standards/compliance team and the operator; there is no outreach in this run. The forest implication is a credible sector-wide compliance mechanism with an unresolved operator-specific assertion, rather than proof of a January AI licensing regime or regulator-to-regulator coordination. Acquisition stops here for root's combined review.
", "path": "esafety-service-wave-2026-09-09/legal-clock.md", "sourceIds": [ "w25-sf-s-dis-standard", "w25-sf-s-res-standard", "w25-sf-s-bose-amendment", "w25-sf-s-deepfake-act" ], "markdown": "9 September 2026. Bounded official-source legal-clock pass, now closed. This note does not determine this operator's compliance or identify an eSafety assurance.\n\nThe closest documented match is the **Online Safety (Designated Internet Services—Class 1A and Class 1B Material) Industry Standard 2024**, which commenced **22 December 2024**, rather than January 2025. Its contemporary announcement expressly discusses some generative-AI and nudification services. That makes it a strong candidate for the instrument being loosely described in the email; it does not establish that this was the operator's intended reference. The separate RES standard has the same clock. The other concrete AI-related changes checked here commenced in May and September 2024.\n\nThe source assertion is the respondent's 25 December 2024 email, section 6, retained Tilak declaration Exhibit 6 PDF21: it says eSafety communicated specific requirements, the service complied, and was confirmed aligned with upcoming AI legislation effective January 2025. It is an interested participant's account, with no instrument name or enclosed Australian assurance on that page. Local source: ../sf-compliance-wave-2026-09-09/tilak-declaration-june13-2025.pdf; this pass read the existing PDF21 text extract, not the rest of the declaration afresh.\n\n## The matching clock and its limits\n\n| Instrument | Making / registration | Commencement | Relevance and boundary |\n| --- | --- | --- | --- |\n| DIS Class1A/1B Standard2024, F2024L00710 | Made19June2024 by Julie Inman Grant; Federal Register registration21June2024 | 22December2024 | Closest substantive match: includes high-impact generative-AI services and model distribution platforms, with duties directed to specified seriously harmful classes. |\n| RES Class1A/1B Standard2024, F2024L00711 | Made19June2024 by Julie Inman Grant; Federal Register registration21June2024 | 22December2024 | Companion instrument for relevant electronic services. Its date is a cross-check, not a service-classification finding. |\n| BOSE Amendment Determination2024, F2024L00590 | Minister Michelle Rowland; registered30May2024 | 31May2024 | Adds express generative-AI expectations. This is a different legal mechanism from the mandatory standard, and an earlier clock. |\n| Criminal Code Amendment (Deepfake Sexual Material) Act2024, C2024A00078 | Assent and registration2September2024 | 3September2024 | Criminal transmission offences covering altered/created sexual depictions; not a January service-approval regime. |\n\nBoth standards' section2 says commencement is six months after the later of the day after registration under the Online Safety Act and the day after registration under the Legislation Act. Original covers/footers establish Federal Register registration; eSafety's published guidance expressly confirms the resulting22December date. The contemporary21June announcement says both were registered that day and would come into force six months after registration, following the disallowance period. The22December date therefore comes from the regulator's explicit account as well as the statutory formula, not an assumption that all registration events must be identical.\n\nSources: DIS original (see source references) RES original (see source references) covers and PDF5/printed1, sections1–5; eSafety21June2024 announcement (see source references) article body, reader lines212–237; eSafety regulatory-guidance index (see source references) Online Safety Codes and Standards and Basic Online Safety Expectations sections (current page, last-updated26May2026).\n\n## What the DIS standard actually requires\n\nSection3 identifies section145 of the Online Safety Act2021 as authority. Section5 applies wherever a designated service is supplied from, only insofar as it serves end-users in Australia, and provides a predominant-purpose exception where another applicable code or standard better aligns with the service. The high-impact generative-AI definition in section6 turns on machine-learning generation capacity and synthetic high-impact material; it excludes a service with controls rendering that risk immaterial. For that category, high-impact material uses X18+/RC film or game and Category2-restricted/RC publication classifications. This is a category trigger, not a finding that every generated adult image breaches the standard. Section4's protective object and the actual duties concern Class1A/1B material. Those classes include child sexual exploitation/pro-terror/extreme crime-and-violence material and specified other crime/violence/drug material. (Original PDF5–7 and12/printed1–3,8.)\n\nThe June announcement specifically says the standards may cover apps generating pornography or nudifying images without effective controls preventing their application to children. That sentence supplies the concrete service-type connection to the operator's description. It still does not classify or clear any named service.\n\nFor a covered high-impact generative-AI DIS, section22(3) requires preventive systems for child-sexual-exploitation/pro-terror outputs; regular model testing/review and prompt mitigation adjustments; differentiation of AI outputs; warnings for Australian users specifically seeking child-abuse images; report/support information for relevant generation terms; and automated detection/action on child-abuse material in training data, prompts and outputs. The notes acknowledge limited model visibility/control and require reliance on available alternatives in those circumstances. PDF33–34/printed29–30, visually checked.\n\nThe instrument also requires contractual provisions governing Class1A/1B use and giving suspension, restriction, termination or material-removal rights for child-exploitation/pro-terror misuse, plus appropriate enforcement in the circumstances specified by section13. Section19 requires management, supervision, internal reporting and sufficient skilled personnel. Section20 requires appropriate known-CSAM detection/removal, with express technical/practicability and encryption/systemic-weakness exceptions and alternative-action duties; it does not displace section22. These are scoped readings of sections13(2),13(5),19 and20, not a claim to have audited every applicable clause. Locators: PDF24–25/printed20–21; PDF30–31/printed26–27.\n\nTwo timing qualifications matter. Section7's ordinary first risk-assessment deadline can be six months after commencement, but section7(6)(f) expressly exempts high-impact generative-AI DIS from subsections(1) and(4), with a material-change qualification in the following page's note. Do not mechanically assign that initial assessment deadline to every AI service. Section36(3) prevents a compliance-report request under that particular section before the first anniversary of commencement, and permits no more than one in12months; subsection(5) gives two months to respond. Those limits do not mean the standard's substantive duties wait a year or that all other reporting powers are unavailable. Section38 requires compliance-action records to be kept for at least two years after the relevant calendar year's end. PDF17–18,44–46/printed13–14,40–42.\n\nNone of these readings establishes a general eSafety certification of an operator's entire business, immunity under other laws, or satisfaction of the duties in practice.\n\n## The two concrete rival clocks\n\nThe BOSE amendment (see source references) section2 commenced the day after30May registration, hence31May2024. Schedule1 item7 inserts section8A: reasonable steps for end-user safety in design, implementation and maintenance of generative-AI capabilities, and proactively minimising their use for unlawful/harmful material or activity. Examples concern lifecycle risk assessment, training data and harmful prompts. Original PDF3,5–6/printed1,3–4. The expectation is broader in subject matter than the DIS harmful-material classes, but its enforcement form must be distinguished: Online Safety Act section45(4) says a determination under that section does not itself impose a duty enforceable by court proceedings. Part4 separately supplies reporting mechanisms. Act compilation in force11December2024 (see source references) sections44–45, scoped reader only.\n\nThe current eSafety guidance index dates the BOSE guidance's January2025 revision to the replacement of AAT by ART. A guidance revision date must not be relabelled legislative commencement. The available record does not establish that the respondent meant this update, or knew of it before the email.\n\nThe Deepfake Sexual Material Act (see source references) is another genuine2024 AI-adjacent change, but its section2 is explicit:3September2024. New Criminal Code section474.17A concerns using a carriage service to transmit specified sexual material of a person who is or appears18+, with knowledge of lack of consent or recklessness about consent. Altered or wholly technologically created depictions count; transmit includes making available, publication, distribution, advertising and promotion. There are specified exceptions. The creation/alteration aggravated offence under474.17AA(5) also requires the underlying transmission offence. This is not a blanket criminalisation of every act of making an image, and not a prospective eSafety approval system. Schedule1 item7 applies the amendments to material transmitted after commencement irrespective of when created/altered. PDF5–10,13/printed1–6,9.\n\n## Provenance, gap and stopping condition\n\nFour official original PDFs were obtained by ordinary public GET and preserved, with acquisition timestamps, finalURLs, bytes and SHA256 in legal-captures.json. Complete text extracts were generated for retrieval, but generation is not full-document reading. DIS read scope: original title, sections1–5, selected section6 definitions, section7, selected section8/9 material, sections10–12 table/overlap rules, selected13/14 clauses,19–20,22–23 thresholds,36–38; substantive pages were inspected through retained extraction and the public reader. Visual checks were PDF1,5,33,34. RES scope: cover and sections1–5 only; visualPDF5. BOSE: all8pages extracted text read; visualPDF3. Deepfake Act: PDF5–10,13 text; visualPDF6. Other rendered pages are not claimed as visual reads.\n\nDIS SHA256 afe30304b7094509de84a1845534b346a48c4fc2f2f07e163fd47fd040ce487d; RES75d2ce8834e804815dad8b72102bd61d4a2ec757b3edd336be541bfee0f3b8b9; BOSE39171722f9c258fa07d9e69324f781c0de000b358defc159fb492b2dfe5eae3a; Deepfake Act1ab54fc6a743c991d2db80aee8897206b10000933649364b77d53d6a40a60e1d.\n\nThe June eSafety announcement and current guidance-index ordinary GETs timed out. Their public web-reader/indexed-primary passages were available and read; no original-HTML hash is asserted, and no transport retry was pursued. Act45 was a scoped public-reader read, not a retained original. This lane made no service-specific query or target-service visit. No January2025 commencement matching the asserted AI regime was located in this bounded pass; this is not an exhaustive negative over all Australian instruments.\n\nCarver's separate service-specific lane has identified official ClothOff correspondence in FOILog141/request25202 that expressly invokes the22December2024 DIS commencement. That is a useful separately sourced counterpart for the clock; this lane has not independently reread that bundle and does not join ClothOff to Drawnudes or promote a request for further information into compliance confirmation.\n\nThe next discriminator is the actual eSafety-to-respondent correspondence: dated requirement list, named legal basis/service classification, evidence supplied in reply, and exact wording/scope of any acknowledgement or assessment. Likely custodians are eSafety's relevant industry-standards/compliance team and the operator; there is no outreach in this run. The forest implication is a credible sector-wide compliance mechanism with an unresolved operator-specific assertion, rather than proof of a January AI licensing regime or regulator-to-regulator coordination. Acquisition stops here for root's combined review.\n", "file": "research/esafety-service-wave-2026-09-09-legal-clock.md" }, { "id": "esafety-service-wave-2026-09-09-service-intervention", "title": "eSafety and drawnudes.io: service-specific intervention check", "description": "Research observed 9 September 2026. This is a bounded public-record result, not a determination of the service's compliance. No service website was visited, tested or contacted.", "html": "Research observed 9 September 2026. This is a bounded public-record result, not a determination of the service's compliance. No service website was visited, tested or contacted.
\nThe inspected official material does not verify that eSafety imposed a service-specific requirement on drawnudes.io before 25 December 2024, or that it approved or confirmed that service's compliance. It does establish contemporaneous outreach to a different named provider, ClothOff. The exact missing record is the eSafety-to-operator correspondence and any subsequent assessment or confirmation naming drawnudes.io (or a legally identified operator demonstrably responsible for it).
\nThe already retained Tilak declaration, Exhibit 6, PDF pages 19–22, reproduces a 25 December 2024 email from a correspondent signing as Ali. Section 6, PDF 21, claims eSafety recently contacted the service, supplied specific requirements, and confirmed alignment with forthcoming January 2025 AI legislation after the operator complied. The 4 January 2025 follow-up, PDF 19, repeats that an Australian regulator provided a list of requirements which the operator fulfilled. The city attorney's 3 January reply, also PDF 19, says the office understands the correspondent operates drawnudes.io and asks whether the described policies apply to other sites. Those are attribution and scope facts, not agency verification. The exhibit contains no attached Australian notice or confirmation.
\nSource: Karun Tilak declaration, filed court material independently republished at http[local research file] . Local original: ../sf-compliance-wave-2026-09-09/tilak-declaration-june13-2025.pdf . SHA-256 bbd7cfa2059c35fd203ce7202b1bb6072d7bbfe6412ea061ffdf7aa58d2183a5. This lane read retained text for PDF 19–22; it did not redownload the document or review all 61 pages. Parent owns the broader declaration analysis.
\nThe [eSafety disclosure log](http[local research file] Log 141 / request 25202, lists a release on 5 December 2025 concerning correspondence with two nudify providers. It says three documents were released fully and three partly, with redactions under FOI sections 22, 47E(d), 47F. That description does not identify the second provider.
\nThe linked [23-page document set](http[local research file] supplies a concrete comparator. Document 2, PDF 3–7, records eSafety's 3 October 2024 outreach to ClothOff, recommending consideration of whether the DIS Standard applies and outlining potential policies, reporting and output-prevention duties. Its 11 November reply says ClothOff is likely covered, with commencement on 22 December 2024. On 5 December, ClothOff claims changes; eSafety points to guidance and asks about those changes. Document 3, PDF 8–13, contains an 8 January 2025 follow-up asking for details and explaining the initial enforcement discretion. It is not a compliance certificate. Document 5, PDF 18–19, is a 23 June 2025 general circular, expressly disavowing any necessary service-specific compliance concern from receipt alone. None of the inspected readable substantive messages names drawnudes.io. This correspondence supports an outreach mechanism, not a transfer of ClothOff's identity or regulatory status to Drawnudes.
\nReading scope: PDF 1–13 substantive correspondence, PDF 14 delivery-failure explanation, PDF 18–20 substantive circular and delivery-failure explanation; remaining technical message headers were not audited. Whole web-reader text was searched for “draw” with no match. Redacted or absent recipient fields were not reconstructed. No claim is made that this is the full agency file or that every recipient of generic circulars has been identified.
\n[eSafety's 8 September 2025 announcement](http[local research file] main text beginning with the date through the media-contact heading, says it formally warned an unnamed UK company operating two services for failing to prevent child sexual exploitation material. It expressly withholds the company name to avoid promotion. This is an agency account of later enforcement, not evidence of a December 2024 requirement or a named Drawnudes action.
\n[The 27 November 2025 follow-up](http[local research file] main text through the media-contact heading, reports withdrawal of Australian access to three services after September enforcement. Its accompanying Hugging Face account concerns a different intermediary. Neither account identifies Drawnudes or independently establishes the operator's earlier claimed approval.
\n[The 20 May 2026 announcement](http[local research file] main text through the media-contact heading, describes a direction to comply within 14 days to an unnamed Argentina-based provider under March 2026 age-restricted-material codes. It again explicitly withholds the name. Its later report that the earlier three services relaunched under a new owner is not an identification bridge. No unnamed provider is assigned to drawnudes.io by this packet.
\nTen search queries are retained in service-query-log.json. Exact service/name variants restricted to esafety.gov.au and aph.gov.au yielded no service-specific official result. Broader nudify/enforcement queries led to the official media-release lists and disclosure log. The lists were discovery surfaces, not an exhaustive reviewed enforcement register. No parliamentary full-corpus review, every-year annual-report audit, or all-notices inspection was performed. An unindexed, unpublished, redacted, or differently named record could exist.
\nOrdinary Python requests GETs to the Log 141 PDF and disclosure page each timed out after 35 seconds. The already successful official web reader was the usable substitute; no original bytes or original SHA-256 were acquired for these records. Neither ordinary route was retried. A bounded visual check requested PDF 3, 8, 18 through web screenshot. PDF 3 and 18 returned “Failed to fetch ... Cache miss”; PDF 8 returned a reference without a visible image. No successful visual inspection is claimed and there was no recovery attempt. Findings rest on the readable official text. service-reader-scope.md is a derived scope/locator record, not a facsimile or original capture; its hash is identified separately in service-captures.json.
\nThe defensible relation is Ali, as reproduced in a filed exhibit, claims prior eSafety contact and compliance confirmation concerning the described service. A second, independently supported relation is eSafety sent ClothOff regulatory guidance and requests for information about claimed changes in 2024–25. They remain separate records. No edge saying eSafety approved Drawnudes, certified its safeguards, endorsed the business, imposed a particular remedy on it, or confirmed compliance is warranted.
\nThe strongest rival explanation is that the operator received ordinary preparatory guidance similar to the released ClothOff correspondence and characterized it as confirmation. Another possibility is a distinct unpublished exchange with a stronger conclusion. The record here discriminates neither. The actual dated outgoing notice/guidance, the operator's response, and eSafety's subsequent assessment are the next useful records; their likely custodians are eSafety and the operator/counsel. This is an identified record gap, not authorization for outreach or a new acquisition branch.
\nForest implication: regulatory contact, legal obligation, claimed implementation, agency assessment and actual approval are different stages. This pass verifies the mechanism for another provider while leaving the service-specific claim unresolved. Hubble's separate legal-instrument packet tests the January 2025 date; it should not be used to backfill an unverified service-level decision.
", "path": "esafety-service-wave-2026-09-09/service-intervention.md", "sourceIds": [ "w25-sf-s-foi-log", "w25-sf-s-foi141", "w25-sf-s-esafety-september", "w25-sf-s-esafety-november", "w25-sf-s-esafety-may" ], "markdown": "Research observed 9 September 2026. This is a bounded public-record result, not a determination of the service's compliance. No service website was visited, tested or contacted.\n\nThe inspected official material does not verify that eSafety imposed a service-specific requirement on drawnudes.io before 25 December 2024, or that it approved or confirmed that service's compliance. It does establish contemporaneous outreach to a different named provider, ClothOff. The exact missing record is the eSafety-to-operator correspondence and any subsequent assessment or confirmation naming drawnudes.io (or a legally identified operator demonstrably responsible for it).\n\n## What the operator actually claimed\n\nThe already retained Tilak declaration, Exhibit 6, PDF pages 19–22, reproduces a 25 December 2024 email from a correspondent signing as Ali. Section 6, PDF 21, claims eSafety recently contacted the service, supplied specific requirements, and confirmed alignment with forthcoming January 2025 AI legislation after the operator complied. The 4 January 2025 follow-up, PDF 19, repeats that an Australian regulator provided a list of requirements which the operator fulfilled. The city attorney's 3 January reply, also PDF 19, says the office understands the correspondent operates drawnudes.io and asks whether the described policies apply to other sites. Those are attribution and scope facts, not agency verification. The exhibit contains no attached Australian notice or confirmation.\n\nSource: Karun Tilak declaration, filed court material independently republished at http[local research file] . Local original: ../sf-compliance-wave-2026-09-09/tilak-declaration-june13-2025.pdf . SHA-256 bbd7cfa2059c35fd203ce7202b1bb6072d7bbfe6412ea061ffdf7aa58d2183a5. This lane read retained text for PDF 19–22; it did not redownload the document or review all 61 pages. Parent owns the broader declaration analysis.\n\n## The closest official correspondence record\n\nThe eSafety disclosure log (see source references) Log 141 / request 25202, lists a release on 5 December 2025 concerning correspondence with two nudify providers. It says three documents were released fully and three partly, with redactions under FOI sections 22, 47E(d), 47F. That description does not identify the second provider.\n\nThe linked 23-page document set (see source references) supplies a concrete comparator. Document 2, PDF 3–7, records eSafety's 3 October 2024 outreach to ClothOff, recommending consideration of whether the DIS Standard applies and outlining potential policies, reporting and output-prevention duties. Its 11 November reply says ClothOff is likely covered, with commencement on 22 December 2024. On 5 December, ClothOff claims changes; eSafety points to guidance and asks about those changes. Document 3, PDF 8–13, contains an 8 January 2025 follow-up asking for details and explaining the initial enforcement discretion. It is not a compliance certificate. Document 5, PDF 18–19, is a 23 June 2025 general circular, expressly disavowing any necessary service-specific compliance concern from receipt alone. None of the inspected readable substantive messages names drawnudes.io. This correspondence supports an outreach mechanism, not a transfer of ClothOff's identity or regulatory status to Drawnudes.\n\nReading scope: PDF 1–13 substantive correspondence, PDF 14 delivery-failure explanation, PDF 18–20 substantive circular and delivery-failure explanation; remaining technical message headers were not audited. Whole web-reader text was searched for “draw” with no match. Redacted or absent recipient fields were not reconstructed. No claim is made that this is the full agency file or that every recipient of generic circulars has been identified.\n\n## Later enforcement cannot silently supply the missing identity or date\n\neSafety's 8 September 2025 announcement (see source references) main text beginning with the date through the media-contact heading, says it formally warned an unnamed UK company operating two services for failing to prevent child sexual exploitation material. It expressly withholds the company name to avoid promotion. This is an agency account of later enforcement, not evidence of a December 2024 requirement or a named Drawnudes action.\n\nThe 27 November 2025 follow-up (see source references) main text through the media-contact heading, reports withdrawal of Australian access to three services after September enforcement. Its accompanying Hugging Face account concerns a different intermediary. Neither account identifies Drawnudes or independently establishes the operator's earlier claimed approval.\n\nThe 20 May 2026 announcement (see source references) main text through the media-contact heading, describes a direction to comply within 14 days to an unnamed Argentina-based provider under March 2026 age-restricted-material codes. It again explicitly withholds the name. Its later report that the earlier three services relaunched under a new owner is not an identification bridge. No unnamed provider is assigned to drawnudes.io by this packet.\n\n## Search and custody limits\n\nTen search queries are retained in service-query-log.json. Exact service/name variants restricted to esafety.gov.au and aph.gov.au yielded no service-specific official result. Broader nudify/enforcement queries led to the official media-release lists and disclosure log. The lists were discovery surfaces, not an exhaustive reviewed enforcement register. No parliamentary full-corpus review, every-year annual-report audit, or all-notices inspection was performed. An unindexed, unpublished, redacted, or differently named record could exist.\n\nOrdinary Python requests GETs to the Log 141 PDF and disclosure page each timed out after 35 seconds. The already successful official web reader was the usable substitute; no original bytes or original SHA-256 were acquired for these records. Neither ordinary route was retried. A bounded visual check requested PDF 3, 8, 18 through web screenshot. PDF 3 and 18 returned “Failed to fetch ... Cache miss”; PDF 8 returned a reference without a visible image. No successful visual inspection is claimed and there was no recovery attempt. Findings rest on the readable official text. service-reader-scope.md is a derived scope/locator record, not a facsimile or original capture; its hash is identified separately in service-captures.json.\n\n## Decision boundary and candidate relations\n\nThe defensible relation is **Ali, as reproduced in a filed exhibit, claims prior eSafety contact and compliance confirmation concerning the described service**. A second, independently supported relation is **eSafety sent ClothOff regulatory guidance and requests for information about claimed changes in 2024–25**. They remain separate records. No edge saying eSafety approved Drawnudes, certified its safeguards, endorsed the business, imposed a particular remedy on it, or confirmed compliance is warranted.\n\nThe strongest rival explanation is that the operator received ordinary preparatory guidance similar to the released ClothOff correspondence and characterized it as confirmation. Another possibility is a distinct unpublished exchange with a stronger conclusion. The record here discriminates neither. The actual dated outgoing notice/guidance, the operator's response, and eSafety's subsequent assessment are the next useful records; their likely custodians are eSafety and the operator/counsel. This is an identified record gap, not authorization for outreach or a new acquisition branch.\n\nForest implication: regulatory contact, legal obligation, claimed implementation, agency assessment and actual approval are different stages. This pass verifies the mechanism for another provider while leaving the service-specific claim unresolved. Hubble's separate legal-instrument packet tests the January 2025 date; it should not be used to backfill an unverified service-level decision.\n", "file": "research/esafety-service-wave-2026-09-09-service-intervention.md" }, { "id": "evaluation-wave-2026-09-09-evaluation-and-acceptance", "title": "CDDP evaluation, acceptance and continuation of funding", "description": "9 September 2026. Root's bounded acquisition after the supplier-platform forest. No final evaluation output was located in the scoped public search. Requirements, announced funding, unpublished deliverables, disclosed copies and observed performance remain separate.", "html": "9 September 2026. Root's bounded acquisition after the supplier-platform forest. No final evaluation output was located in the scoped public search. Requirements, announced funding, unpublished deliverables, disclosed copies and observed performance remain separate.
\nPrimary retained original: ../supplier-platform-wave-2026-09-09/cddp-foi-bundle.pdf, SHA256 71bd6a522a765d53f9aa79b239134226be64891ea3c67711abc9d788fb87748f. Source: http[local research file] . This run read PDF pages 18,34–35,62–68,88–100 and visually inspected pages 18,63,93,95,97,100,193–194. Local PNGs are derivatives of that original. Page18 is almost fully redacted; no hidden text was reconstructed. The prior run read the contract/IP pages separately.
\nOfficial release: http[local research file] . This run read the complete exemption account, not only its document inventory. The public page contains an unreconciled date discrepancy: publication 28 June 2024, request 25 August 2023, response field 25 September 2024. Original bundle filename contains FOI2023 while release path/reference says FOI2024. Preserve those labels without silently fixing the response year.
\nThe October 2021 business case, PDF62–68 (printed61–67), envisages process and impact evaluation, consultation with an independent evaluator and HMT, and sharing supporting evidence with Treasury teams. Process evaluation would examine cross-government coordination and working practices, with KPIs developed after an initial systems audit. Impact evaluation would assess whether improved data products enhanced government understanding and operational response. Specific criteria were still to be agreed. Publicly legible criteria concern funded delivery, government data sharing/analysis, stakeholder benefit and lessons. Detailed portions of both evaluation strands are visibly redacted; absence of a readable accuracy, rights or false-positive metric is not proof that no such metric was used.
\nThe expert-evaluator specification, contract 102384, PDF88–92, asks for independent end-to-end monitoring, value-for-money assessment, assessment of digital practice and the effectiveness of correctly identifying misinformation/disinformation trends. This correctness requirement matters: the available specification is not limited to efficiency. The contract also requires reporting that senior civil servants across Whitehall can use, with successes, failures and long-term recommendations. A final report is an expressly commissioned deliverable. Experimental or quasi-experimental evaluation is conditional; interest in possible A/B testing is not evidence an experiment was conducted, let alone a population manipulation experiment.
\nThe provisional schedule contains a feasibility report to HMT in February 2022, a phase-two decision in March 2022, interim evaluation in August 2022 and final evaluation gateway in February 2023 before phase end in March 2023. Earlier rows carry apparent year inconsistencies (November–February 2021 and December/January 2021); these remain as printed. PDF92 expressly makes subsequent-phase funding conditional on assessment of the discovery feasibility review and approval by HMT officials. This locates an intended funding gate. It does not identify the actual official, decision minute, evidence considered or whether the gate was exercised as planned.
\nThe internal DCMS team administers contracts, supplies governance secretariat and leads cross-government engagement. The separately contracted digital project-management team leads discovery/delivery, while the evaluator examines it. These are specified roles, not names of the actual committee members. PDF62 and68 direct evaluation evidence and financial/performance information toward the Shared Outcomes Fund team, HMT spending team and other government departments.
\nThe release inventory identifies document6 as the DCMS/Oliver Wyman evaluation-and-monitoring contract. PDF95 is order102384, but the supplier name, address and registration number are visibly redacted; the brand identification comes from the department's inventory, not a recovered supplier field. Exact legal-entity registration remains unresolved. No related corporate ownership claim is made.
\nPDF93's pricing table is visible in the rendered page despite failing text extraction: Phase I GBP108,400 excluding VAT; Phase II GBP547,600; total fixed tender price GBP656,000 excluding VAT. The same GBP656,000 appears in PDF97 as the estimated first-year charges used for liability calculation. Payment is in arrears after relevant deliverables satisfy DCMS. This is a disclosed contractual/tender price, not proof of payment. Tables B and C below contain blacked-out fields despite the inventory calling the pricing document disclosed; do not say all prices or rate cards were disclosed.
\nThe order is dated 14 January 2022 (PDF96), while PDF97 visibly prints a 20 December 2022 start, three-month initial period and 31 March 2023 expiry, alongside a phase-one extension reservation. Those terms differ from the specification's December2021-to-March2023 project schedule. A clerical mistake, later amendment or different phase might explain the difference; none is established. PDF100 signature/name/role/date fields are visibly empty. The released copy does not independently establish execution, and it does not prove no executed contract exists.
\nPDF193–194 visibly redact the table of programme outputs/success measures and detailed milestone descriptions/acceptance criteria. The dates 15 July 2022 and 30 September 2022 remain visible. This confirms that the missing substantive cells are redactions, not merely an OCR failure. Conversely PDF93 demonstrates why failed text extraction alone must not be called redaction. The contract variation's reported MVP completion is stronger than a plan, but still not an independent assessment of accuracy, fairness, effectiveness or downstream platform action.
\nThe department's release account applies personal-data, commercial, policy-development, security and public-affairs exemptions to specified parts. It says section36 was applied across every document except the pricing schedule and defends withholding operational capabilities/vulnerabilities. Discovery Report and Product Operating Model are withheld in full under the described security/public-affairs basis after removal would make the remainder illegible. These are the department's reasons; no ICO or tribunal review of this particular release was acquired. The earlier ICO Crisp-report ruling is a different record and does not adjudicate these CDDP redactions. Neither-confirm-nor-deny wording concerning additional security-related records proves neither their existence nor absence.
\nA separate Treasury Shared Outcomes Fund Round2 summary publicly lists CDDP funding of GBP8.4 million and participants DCMS, FCDO, Home Office, Cabinet Office, Ministry of Defence, GCHQ and NCSC. This is announced pilot funding and participation, not a ledger of money received by Faculty or Oliver Wyman, nor proof every participant supplied intelligence data or controlled the platform. Do not add its GBP8.4m to supplier prices, which may be components or differently scoped amounts.
\nRetained original shared-outcomes-round2.pdf: 230,526 bytes, SHA256 c91b2c59cfcafc70bb10f4aab856d4d15c92166957f587599d47a94a671598bc. Read PDF9/printed6, CDDP entry. Source: http[local research file] . Public reader failed once; ordinary direct download succeeded. Do not confuse adjacent Global Supply Chain Intelligence Pilot and its GBP4.7m entry with CDDP.
\nScoped exact-title searches combined CDDP, final report, evaluation, Oliver Wyman and contract102384 across government/parliament/supplier sources. They returned the original commissioning material but no actual final evaluation, agreed KPI set, Treasury gate decision, or completed acceptance certificate. This is an acquisition gap, not evidence that evaluation never occurred or was adverse. No new request was sent.
\nMost revealing likely records are the initial audit/KPI agreement; discovery feasibility report sent to HMT and funding decision; interim/final evaluation report and its supporting evidence; milestone acceptance record; and the log translating analyst findings into subsequent action. Commissioned documents have firmer existence support than guessed internal communications; delivered versions remain unconfirmed.
\nA separate official source surfaced: the Online Safety Data Initiative (OSDI) appointed PA Consulting to evaluate a Faculty/PUBLIC delivery consortium. OSDI is not CDDP, its evaluator is not Oliver Wyman, and its results cannot stand in for CDDP's. It offers a useful next comparison: how the same public funding model and overlapping suppliers handle classification, independent assessment and publication in a less opaque project. Source lead: http[local research file] . Only its indexed official opening was read in this run; further acquisition follows a forest review.
", "path": "evaluation-wave-2026-09-09/evaluation-and-acceptance.md", "sourceIds": [ "w16-supp-s-foi-release", "w16-supp-s-foi-bundle", "w33-con-s-round2" ], "markdown": "9 September 2026. Root's bounded acquisition after the supplier-platform forest. No final evaluation output was located in the scoped public search. Requirements, announced funding, unpublished deliverables, disclosed copies and observed performance remain separate.\n\n## Read scope and custody\n\nPrimary retained original: ../supplier-platform-wave-2026-09-09/cddp-foi-bundle.pdf, SHA256 71bd6a522a765d53f9aa79b239134226be64891ea3c67711abc9d788fb87748f. Source: http[local research file] . This run read PDF pages 18,34–35,62–68,88–100 and visually inspected pages 18,63,93,95,97,100,193–194. Local PNGs are derivatives of that original. Page18 is almost fully redacted; no hidden text was reconstructed. The prior run read the contract/IP pages separately.\n\nOfficial release: http[local research file] . This run read the complete exemption account, not only its document inventory. The public page contains an unreconciled date discrepancy: publication 28 June 2024, request 25 August 2023, response field 25 September 2024. Original bundle filename contains FOI2023 while release path/reference says FOI2024. Preserve those labels without silently fixing the response year.\n\n## What the government wanted evaluated\n\nThe October 2021 business case, PDF62–68 (printed61–67), envisages process and impact evaluation, consultation with an independent evaluator and HMT, and sharing supporting evidence with Treasury teams. Process evaluation would examine cross-government coordination and working practices, with KPIs developed after an initial systems audit. Impact evaluation would assess whether improved data products enhanced government understanding and operational response. Specific criteria were still to be agreed. Publicly legible criteria concern funded delivery, government data sharing/analysis, stakeholder benefit and lessons. Detailed portions of both evaluation strands are visibly redacted; absence of a readable accuracy, rights or false-positive metric is not proof that no such metric was used.\n\nThe expert-evaluator specification, contract 102384, PDF88–92, asks for independent end-to-end monitoring, value-for-money assessment, assessment of digital practice and the effectiveness of correctly identifying misinformation/disinformation trends. This correctness requirement matters: the available specification is not limited to efficiency. The contract also requires reporting that senior civil servants across Whitehall can use, with successes, failures and long-term recommendations. A final report is an expressly commissioned deliverable. Experimental or quasi-experimental evaluation is conditional; interest in possible A/B testing is not evidence an experiment was conducted, let alone a population manipulation experiment.\n\nThe provisional schedule contains a feasibility report to HMT in February 2022, a phase-two decision in March 2022, interim evaluation in August 2022 and final evaluation gateway in February 2023 before phase end in March 2023. Earlier rows carry apparent year inconsistencies (November–February 2021 and December/January 2021); these remain as printed. PDF92 expressly makes subsequent-phase funding conditional on assessment of the discovery feasibility review and approval by HMT officials. This locates an intended funding gate. It does not identify the actual official, decision minute, evidence considered or whether the gate was exercised as planned.\n\nThe internal DCMS team administers contracts, supplies governance secretariat and leads cross-government engagement. The separately contracted digital project-management team leads discovery/delivery, while the evaluator examines it. These are specified roles, not names of the actual committee members. PDF62 and68 direct evaluation evidence and financial/performance information toward the Shared Outcomes Fund team, HMT spending team and other government departments.\n\n## Evaluator identity, price and copy limitations\n\nThe release inventory identifies document6 as the DCMS/Oliver Wyman evaluation-and-monitoring contract. PDF95 is order102384, but the supplier name, address and registration number are visibly redacted; the brand identification comes from the department's inventory, not a recovered supplier field. Exact legal-entity registration remains unresolved. No related corporate ownership claim is made.\n\nPDF93's pricing table is visible in the rendered page despite failing text extraction: Phase I GBP108,400 excluding VAT; Phase II GBP547,600; total fixed tender price GBP656,000 excluding VAT. The same GBP656,000 appears in PDF97 as the estimated first-year charges used for liability calculation. Payment is in arrears after relevant deliverables satisfy DCMS. This is a disclosed contractual/tender price, not proof of payment. Tables B and C below contain blacked-out fields despite the inventory calling the pricing document disclosed; do not say all prices or rate cards were disclosed.\n\nThe order is dated 14 January 2022 (PDF96), while PDF97 visibly prints a 20 December 2022 start, three-month initial period and 31 March 2023 expiry, alongside a phase-one extension reservation. Those terms differ from the specification's December2021-to-March2023 project schedule. A clerical mistake, later amendment or different phase might explain the difference; none is established. PDF100 signature/name/role/date fields are visibly empty. The released copy does not independently establish execution, and it does not prove no executed contract exists.\n\n## What remains hidden about acceptance\n\nPDF193–194 visibly redact the table of programme outputs/success measures and detailed milestone descriptions/acceptance criteria. The dates 15 July 2022 and 30 September 2022 remain visible. This confirms that the missing substantive cells are redactions, not merely an OCR failure. Conversely PDF93 demonstrates why failed text extraction alone must not be called redaction. The contract variation's reported MVP completion is stronger than a plan, but still not an independent assessment of accuracy, fairness, effectiveness or downstream platform action.\n\nThe department's release account applies personal-data, commercial, policy-development, security and public-affairs exemptions to specified parts. It says section36 was applied across every document except the pricing schedule and defends withholding operational capabilities/vulnerabilities. Discovery Report and Product Operating Model are withheld in full under the described security/public-affairs basis after removal would make the remainder illegible. These are the department's reasons; no ICO or tribunal review of this particular release was acquired. The earlier ICO Crisp-report ruling is a different record and does not adjudicate these CDDP redactions. Neither-confirm-nor-deny wording concerning additional security-related records proves neither their existence nor absence.\n\n## Published programme scale and named participants\n\nA separate Treasury Shared Outcomes Fund Round2 summary publicly lists CDDP funding of GBP8.4 million and participants DCMS, FCDO, Home Office, Cabinet Office, Ministry of Defence, GCHQ and NCSC. This is announced pilot funding and participation, not a ledger of money received by Faculty or Oliver Wyman, nor proof every participant supplied intelligence data or controlled the platform. Do not add its GBP8.4m to supplier prices, which may be components or differently scoped amounts.\n\nRetained original shared-outcomes-round2.pdf: 230,526 bytes, SHA256 c91b2c59cfcafc70bb10f4aab856d4d15c92166957f587599d47a94a671598bc. Read PDF9/printed6, CDDP entry. Source: http[local research file] . Public reader failed once; ordinary direct download succeeded. Do not confuse adjacent Global Supply Chain Intelligence Pilot and its GBP4.7m entry with CDDP.\n\n## Search result and highest-value next records\n\nScoped exact-title searches combined CDDP, final report, evaluation, Oliver Wyman and contract102384 across government/parliament/supplier sources. They returned the original commissioning material but no actual final evaluation, agreed KPI set, Treasury gate decision, or completed acceptance certificate. This is an acquisition gap, not evidence that evaluation never occurred or was adverse. No new request was sent.\n\nMost revealing likely records are the initial audit/KPI agreement; discovery feasibility report sent to HMT and funding decision; interim/final evaluation report and its supporting evidence; milestone acceptance record; and the log translating analyst findings into subsequent action. Commissioned documents have firmer existence support than guessed internal communications; delivered versions remain unconfirmed.\n\nA separate official source surfaced: the Online Safety Data Initiative (OSDI) appointed PA Consulting to evaluate a Faculty/PUBLIC delivery consortium. OSDI is not CDDP, its evaluator is not Oliver Wyman, and its results cannot stand in for CDDP's. It offers a useful next comparison: how the same public funding model and overlapping suppliers handle classification, independent assessment and publication in a less opaque project. Source lead: http[local research file] . Only its indexed official opening was read in this run; further acquisition follows a forest review.\n", "file": "research/evaluation-wave-2026-09-09-evaluation-and-acceptance.md" }, { "id": "filmon-classification-wave-2026-09-09-current-product", "title": "DoubleVerify: customer choice, vendor defaults and review", "description": "9 September 2026. Root acquisition is closed. Four original company HTML pages were and their full substantive text read. A current SEC-hosted FY2025 filing was read in specific passages through the primary web reader; no original filing was downloaded. These are company descriptions, not a customer configuration, independent accuracy audit or observed resto", "html": "9 September 2026. Root acquisition is closed. Four original company HTML pages were retained and their full substantive text read. A current SEC-hosted FY2025 filing was read in specific passages through the primary web reader; no original filing was downloaded. These are company descriptions, not a customer configuration, independent accuracy audit or observed restoration. The historical FilmOn case is a separate lane. Current brand/product statements and consolidated issuer disclosures do not by themselves prove the exact contracting entity in the older lawsuit.
\nThe undated current [brand-suitability product page](http[local research file] describes a proprietary classifier using AI and human review across several media types. It supplies more than100 categories which brands can choose to avoid. Its ordinary commercial example includes avoiding otherwise legitimate video-game or motor content to fit a particular campaign. The page distinguishes pre-bid evaluation from post-purchase measurement. Neither classification nor measurement alone demonstrates a completed advertising exclusion. The advertised accreditation and accuracy remain vendor claims here; no accreditor audit or actual campaign was inspected. Original-derived lines59–111, particularly78–98; primary reader77–184.
\nThe [13 February2025 announcement](http[local research file] announces 'Highly Illicit: Do Not Monetize' and 'P2P Sharing and Streaming' avoidance categories and says both will be enabled by default. This establishes an announced default policy; the activation date and customer uptake are unverified. It reports automatic provision of the former to advertisers and distribution to over100 platform partners. A supplied category or default does not prove every recipient enabled it or every listed site's advertising stopped. The latter covers infrastructure that the text acknowledges has legitimate uses but can be misused.
\nDV says it used three years of public NCMEC data, covering over300 domains/providers, cross-checked prior classifications and assessed remaining sites itself. It says it excluded large, well-moderated social/search services. This is DV's account of input selection and transformation, not an independently verified NCMEC judgment or endorsement of DV's resulting list. The underlying NCMEC table and its meaning were not acquired in this run. Named FBI outreach is an offered-assistance account, not acceptance or instructions. Original-derived72–138, decisive96–120; primary reader92–139. The captured original supplies the publication date.
\nA [13 June2024 company statement](http[local research file] describes a publisher classification-review process, access to reports/logs, and separation of assessment from sales/client teams. It claims paying-publisher status does not soften ratings and gives an unnamed customer example. These are safeguards asserted by the company, not independently observed separations or a worked correction. It supplies no decision deadline, independent appeal adjudicator, mandatory advertiser reconsideration or restored-income outcome in the read body. Its automation/objectivity language does not establish absence of model or policy bias. Original-derived72–100; primary reader94–115.
\nThe [25 February2025 URL-reporting announcement](http[local research file] says DV advertisers using Brand Safety and Suitability measurement can request URL-level detail. Direct integration into its interface was 'coming soon' and external integrations were planned later that year. Those are stages in a dated announcement; retrieval in2026 does not turn the plans into completed deployment. The described entitlement is for advertiser customers and does not itself give every affected publisher access to every customer's report or preferences. No request was made. Original-derived72–97; primary reader95–118. Actual dates on the captured originals resolve the web-reader's omitted publication dates; nearby recent-post dates are not used.
\nThe [FY2025 Form10-K](http[local research file] describes customer configuration distributed across the service, proprietary classification and customer feedback shaping product development. It reports90percent of2025 revenue from advertisers and10percent from supply-side customers. Advertiser fees are primarily based on measured transaction volume, commonly per thousand impressions; some percentage-of-spend arrangements are capped. Supply-side arrangements include minimum guarantees and tiered pricing. This supports a two-sided commercial model, not a fee per adverse classification or a proved financial cause of a specific exclusion.
\nExact retained source scope is product-10k.reader.txt: reader403–434 (customer feedback, product teams and configuration/classification) and1152–1225 (MD&A overview, fee basis and revenue sources). The relevant prose was read; no full filing, audit opinion, revenue-table facsimile, individual contract or accounting record was inspected. The fiscal year is2025, not the capture year. No stockholder command, grant finance or external ownership branch was acquired.
\nThe current offering contains several powers that should occupy separate places on the map: selecting source inputs, assigning a classification, choosing a default, changing customer settings and implementing a placement decision. A review can change one without necessarily changing the others. The acquired product description supports customer choice; the dated announcement supplies a vendor-chosen default policy, with implementation not independently observed. Neither cancels the other.
\nThe most revealing next input record is the exact public NCMEC dataset DV says it used, including whether its rows count provider reports, notices received or assessed illegal domains. Those meanings are not interchangeable. The missing implementation records are the category policy/version, a recipient's actual configuration and a dated review/correction followed by recipient action. The historical court lane may illuminate the contested classification process but cannot prove these2025/2026 settings were used then.
\nNo customer login, publisher submission, operational classification query, harmful-content endpoint, payment, outreach or public release occurred. Acquisition closes pending independent review and the combined forest. All original files and derivative scopes are listed in product-captures.json; successful captures are not claims that every link or surrounding page was read.
", "path": "filmon-classification-wave-2026-09-09/current-product.md", "sourceIds": [ "w30-class-s-dv-product", "w30-class-s-dv-review", "w30-class-s-dv-url", "w30-class-s-dv-defaults", "w30-class-s-dv-10k" ], "markdown": "9 September 2026. Root acquisition is closed. Four original company HTML pages were retained and their full substantive text read. A current SEC-hosted FY2025 filing was read in specific passages through the primary web reader; no original filing was downloaded. These are company descriptions, not a customer configuration, independent accuracy audit or observed restoration. The historical FilmOn case is a separate lane. Current brand/product statements and consolidated issuer disclosures do not by themselves prove the exact contracting entity in the older lawsuit.\n\n## Category assignment and customer selection\n\nThe undated current brand-suitability product page (see source references) describes a proprietary classifier using AI and human review across several media types. It supplies more than100 categories which brands can choose to avoid. Its ordinary commercial example includes avoiding otherwise legitimate video-game or motor content to fit a particular campaign. The page distinguishes pre-bid evaluation from post-purchase measurement. Neither classification nor measurement alone demonstrates a completed advertising exclusion. The advertised accreditation and accuracy remain vendor claims here; no accreditor audit or actual campaign was inspected. Original-derived lines59–111, particularly78–98; primary reader77–184.\n\n## An announced default complicates the choice model\n\nThe 13 February2025 announcement (see source references) announces 'Highly Illicit: Do Not Monetize' and 'P2P Sharing and Streaming' avoidance categories and says both will be enabled by default. This establishes an announced default policy; the activation date and customer uptake are unverified. It reports automatic provision of the former to advertisers and distribution to over100 platform partners. A supplied category or default does not prove every recipient enabled it or every listed site's advertising stopped. The latter covers infrastructure that the text acknowledges has legitimate uses but can be misused.\n\nDV says it used three years of public NCMEC data, covering over300 domains/providers, cross-checked prior classifications and assessed remaining sites itself. It says it excluded large, well-moderated social/search services. This is DV's account of input selection and transformation, not an independently verified NCMEC judgment or endorsement of DV's resulting list. The underlying NCMEC table and its meaning were not acquired in this run. Named FBI outreach is an offered-assistance account, not acceptance or instructions. Original-derived72–138, decisive96–120; primary reader92–139. The captured original supplies the publication date.\n\n## Review and evidence access\n\nA 13 June2024 company statement (see source references) describes a publisher classification-review process, access to reports/logs, and separation of assessment from sales/client teams. It claims paying-publisher status does not soften ratings and gives an unnamed customer example. These are safeguards asserted by the company, not independently observed separations or a worked correction. It supplies no decision deadline, independent appeal adjudicator, mandatory advertiser reconsideration or restored-income outcome in the read body. Its automation/objectivity language does not establish absence of model or policy bias. Original-derived72–100; primary reader94–115.\n\nThe 25 February2025 URL-reporting announcement (see source references) says DV advertisers using Brand Safety and Suitability measurement can request URL-level detail. Direct integration into its interface was 'coming soon' and external integrations were planned later that year. Those are stages in a dated announcement; retrieval in2026 does not turn the plans into completed deployment. The described entitlement is for advertiser customers and does not itself give every affected publisher access to every customer's report or preferences. No request was made. Original-derived72–97; primary reader95–118. Actual dates on the captured originals resolve the web-reader's omitted publication dates; nearby recent-post dates are not used.\n\n## The revenue model and configuration authority\n\nThe FY2025 Form10-K (see source references) describes customer configuration distributed across the service, proprietary classification and customer feedback shaping product development. It reports90percent of2025 revenue from advertisers and10percent from supply-side customers. Advertiser fees are primarily based on measured transaction volume, commonly per thousand impressions; some percentage-of-spend arrangements are capped. Supply-side arrangements include minimum guarantees and tiered pricing. This supports a two-sided commercial model, not a fee per adverse classification or a proved financial cause of a specific exclusion.\n\nExact retained source scope is product-10k.reader.txt: reader403–434 (customer feedback, product teams and configuration/classification) and1152–1225 (MD&A overview, fee basis and revenue sources). The relevant prose was read; no full filing, audit opinion, revenue-table facsimile, individual contract or accounting record was inspected. The fiscal year is2025, not the capture year. No stockholder command, grant finance or external ownership branch was acquired.\n\n## Structural result and next discriminator\n\nThe current offering contains several powers that should occupy separate places on the map: selecting source inputs, assigning a classification, choosing a default, changing customer settings and implementing a placement decision. A review can change one without necessarily changing the others. The acquired product description supports customer choice; the dated announcement supplies a vendor-chosen default policy, with implementation not independently observed. Neither cancels the other.\n\nThe most revealing next input record is the exact public NCMEC dataset DV says it used, including whether its rows count provider reports, notices received or assessed illegal domains. Those meanings are not interchangeable. The missing implementation records are the category policy/version, a recipient's actual configuration and a dated review/correction followed by recipient action. The historical court lane may illuminate the contested classification process but cannot prove these2025/2026 settings were used then.\n\nNo customer login, publisher submission, operational classification query, harmful-content endpoint, payment, outreach or public release occurred. Acquisition closes pending independent review and the combined forest. All original files and derivative scopes are listed in product-captures.json; successful captures are not claims that every link or surrounding page was read.\n", "file": "research/filmon-classification-wave-2026-09-09-current-product.md" }, { "id": "filmon-classification-wave-2026-09-09-judicial-case", "title": "FilmOn classification litigation: a paid label, an attempted correction, and a procedural limit", "description": "Bounded acquisition closed, 9 September 2026. This packet concerns the historical FilmOn/DoubleVerify dispute, not current DoubleVerify product operation.", "html": "Bounded acquisition closed, 9 September 2026. This packet concerns the historical FilmOn/DoubleVerify dispute, not current DoubleVerify product operation.
\nThe [official Supreme Court opinion](http[local research file] FilmOn.com Inc. v. DoubleVerify Inc. (2019) 7 Cal.5th133, S244157, was obtained by an ordinary public GET (redirect to www4.courts.ca.gov). The original is judicial-opinion.pdf: 28 PDF pages, SHA256 e6a005c6cbea8dd2a92b5481f5d93658ce940081c0e586b3f9582e84e46ccadb. The entire opinion's substantive text, printed pp1-25/PDF2-26, and identifying/counsel pages were read; PDF5,24,26 were also visually checked. The parser contains occasional repeated phrases from the PDF text layer, so those decisive pages were checked against the visible original.
\nThis is the published California Supreme Court decision of 6 May2019, authored by Justice Cuellar with six concurring justices. It reversed the Second Appellate District, Division Three decision B264074; the trial case was Los Angeles Superior Court BC561987. Unlike the later eGumball opinion, this is a reported Supreme Court opinion, not an unpublished appellate disposition.
\nThe official case page led to the [filed media amicus brief](http[local research file] Its original, judicial-amicus.pdf, has58 scanned pages; SHA c49b1f15d44431457381f9a9eec9774db5e9b3cd90def7da9a5d2d5666e91227. Actual visual scope: PDF1-3 (covers/contents), PDF11-18 (printed10-17, interest and argument summary), PDF55-56 (printed54-55, conclusion/signature/certificate). The whole brief was not read. Additional rendered pages are not claimed as inspected.
\nThe [appellant's opening brief](http[local research file] judicial-appellant.pdf, is44 scanned pages, SHA e9b4c0a8ec9332b151d75e383f1f9a69b592328781a507106743f9c479a747dc. Actual visual scope: PDF1 and11-17 (printed6-12). This narrow read tests the specifically discovered correction attempts and a settlement footnote; it is not a whole-brief or appendix read. The original correspondence and underlying appendix exhibits were not acquired.
\nThe opinion describes DoubleVerify gathering information and selling confidential reports to businesses considering advertising on websites. Its report glossary defined the categories. The Adult Content label encompassed mature material inappropriate for children, including explicit language, sounds and themes; it was not limited to pornography or unlawful material. The copyright streaming/file-sharing label encompassed sites presently or historically associated with distributing or accessing copyrighted material without suitable controls, licensing or permission (printed pp3-4/PDF4-5). The category definition is a provider's classification rule, not an adjudication that every labelled site was currently infringing.
\nThe Supreme Court described the report's three columns: domains/subdomains, total impressions, and thematic categories defined by DoubleVerify (printed p22/PDF23). Paying recipients agreed to confidentiality and used the reports for their business purposes (printed p23/PDF24). These pages establish the described classification/distribution mechanism; they do not identify a universal downstream blocking rule.
\nFilmOn alleged the labels were false and caused existing or prospective advertising partners to refuse business. The opinion states those as allegations (printed p4/PDF5), not a damages finding or independently established causal estimate. This packet has no final calculation of lost revenue, advertiser instruction, restored contract or actual reclassification receipt.
\nThe opening brief bears a filed stamp of 16 January2018. FilmOn's account says it learned of reports through a former client in late2013 and sent a cease-and-desist letter on12 December2013 demanding tag removal and correction. It says DoubleVerify subsequently described Adult Content as akin to an R or TV-MA rating, but did not clarify to advertisers that FilmOn's content was nonpornographic or limit the label to particular channels. It further alleges that after an investigation DoubleVerify maintained the classifications despite letters, calls and meetings (printed pp10-12/PDF15-17).
\nThis is source-specific evidence that FilmOn formally asserted a disputed explanation/correction process, supported in its brief by appendix citations. It does not independently prove the contents of those communications, that the classification was false under its actual glossary, or that the alleged refusal persisted after2018. The distinction matters: a debate over whether a broad category communicated pornography, and whether it should apply to an entire network or individual channels, is more specific than a generic allegation of blacklisting.
\nThe same brief's printed p12/PDF17 footnote2 says AOL was dismissed pursuant to a settlement on9 April2015. That is FilmOn counsel's report about another named defendant at an earlier stage. No settlement agreement, dismissal order, terms or payment was acquired. It is not evidence of a DoubleVerify settlement, a correction or restored advertising.
\nThe trial court granted DoubleVerify's anti-SLAPP motion and the Court of Appeal affirmed. The Supreme Court reversed the latter judgment because DoubleVerify's confidential reports did not qualify under Code of Civil Procedure425.16(e)(4), the public-interest catchall (printed pp2-6,17-25).
\nThe court's inquiry identifies the public issue implicated by the speech, then asks about its functional relationship to public discussion. Audience, speaker and purpose inform that second inquiry. The mere presence of a broad public-interest subject, such as copyright or adult content, was insufficient here. The reports were exchanged privately for the recipients' business purposes and were not intended to participate in public discussion. This is a statutory procedural ruling, not a judicial validation of FilmOn's allegation that the labels were false.
\nThe court expressly refused to exclude all commercially oriented or confidential speech categorically. No single contextual factor was dispositive. It also rejected FilmOn's proposed commercial/noncommercial sorting, even though FilmOn won the ultimate reversal (printed pp15-16/PDF16-17; p23/PDF24). Both parties agreed the distinct section425.17(c) exemption did not apply because the reports concerned FilmOn, not DoubleVerify or a competitor (printed p14 fn4/PDF15). Thus winning the result does not mean every argument was adopted.
\nThe opinion grants judicial notice of copyright-case orders and legislative history (printed p18 fn6/PDF19); that does not convert its narrow anti-SLAPP disposition into an adjudication of the accuracy of all challenged labels. The actual final paragraph simply reverses (printed p25/PDF26). It orders no correction, reclassification, ad purchase, damages payment or service restoration.
\nThe filed amicus was submitted in support of DoubleVerify by Motion Picture Association of America, Inc.; The Hearst Corporation; Tegna Inc.; California News Publishers Association; and First Amendment Coalition. It was filed6 June2018 and dated29 May2018. Davis Wright Tremaine LLP counsel Kelli L. Sager, Thomas R. Burke and Rochelle L. Wilcox are named; Wilcox signed it (PDF1-2,55-56). These are documented amicus/representation roles, not evidence of ownership, category authorship, funding of DoubleVerify or control of its decisions.
\nIn the introductory interest statement, the amici describe their media/creative work and interest in protecting it from piracy. They support keeping rampant-infringement sites and services out of the legitimate advertising market and describe companies such as DoubleVerify as helping that effort (printed pp10-11/PDF11-12). Their summary says brands use assurance-provider information to instruct advertising networks where ads should or should not appear (printed p14/PDF15). That is the amici's stated commercial mechanism and rationale, not independent proof that all listed amici bought this particular report.
\nThey argue that reports are informational products, analogous to news and films, rather than advertising for a product, and fear a broad commercial-context exception could endanger other speakers (printed pp15-17/PDF16-18). Their express request was to reject FilmOn's arguments and affirm protection of DoubleVerify's speech under anti-SLAPP (printed p54/PDF55). The Supreme Court did not grant that requested disposition. Its refusal to adopt a categorical commercial-speech exclusion overlaps part of their concern, but this packet does not infer that their brief caused that reasoning.
\nFour focused discovery queries sought the original and same-case remand/settlement/dismissal. They located the opinion, filed briefing and unrelated settlements/cases citing FilmOn. No later FilmOn-versus-DoubleVerify final merits judgment, correction, reclassification or advertising restoration record was located. This is a bounded search result, not proof those records or events do not exist. The only specifically named settlement found was the earlier AOL footnote, preserved at its attributed scope.
\nMeaningful relations supported now are DoubleVerify-to-confidential-report category authorship/distribution as described by the court; FilmOn-to-DoubleVerify asserted correction demand; named amici-to-Supreme Court filed request for procedural protection; and Supreme Court-to-prior judgment reversal. A potential later remedy requires its own order, agreement or correction/recipient notice. The 2025 eGumball court's reliance on FilmOn is a legal-precedent relationship, not a shared operational database. Acquisition is closed for root review; no new case, current product, ownership, payment or content-site branch was opened.
", "path": "filmon-classification-wave-2026-09-09/judicial-case.md", "sourceIds": [ "w30-class-s-filmon-opinion", "w30-class-s-filmon-amicus", "w30-class-s-filmon-appellant" ], "markdown": "Bounded acquisition closed, 9 September 2026. This packet concerns the historical FilmOn/DoubleVerify dispute, not current DoubleVerify product operation.\n\n## Originals and actual reading\n\nThe official Supreme Court opinion (see source references) FilmOn.com Inc. v. DoubleVerify Inc. (2019) 7 Cal.5th133, S244157, was obtained by an ordinary public GET (redirect to www4.courts.ca.gov). The original is judicial-opinion.pdf: 28 PDF pages, SHA256 e6a005c6cbea8dd2a92b5481f5d93658ce940081c0e586b3f9582e84e46ccadb. The entire opinion's substantive text, printed pp1-25/PDF2-26, and identifying/counsel pages were read; PDF5,24,26 were also visually checked. The parser contains occasional repeated phrases from the PDF text layer, so those decisive pages were checked against the visible original.\n\nThis is the published California Supreme Court decision of 6 May2019, authored by Justice Cuellar with six concurring justices. It reversed the Second Appellate District, Division Three decision B264074; the trial case was Los Angeles Superior Court BC561987. Unlike the later eGumball opinion, this is a reported Supreme Court opinion, not an unpublished appellate disposition.\n\nThe official case page led to the filed media amicus brief (see source references) Its original, judicial-amicus.pdf, has58 scanned pages; SHA c49b1f15d44431457381f9a9eec9774db5e9b3cd90def7da9a5d2d5666e91227. Actual visual scope: PDF1-3 (covers/contents), PDF11-18 (printed10-17, interest and argument summary), PDF55-56 (printed54-55, conclusion/signature/certificate). The whole brief was not read. Additional rendered pages are not claimed as inspected.\n\nThe appellant's opening brief (see source references) judicial-appellant.pdf, is44 scanned pages, SHA e9b4c0a8ec9332b151d75e383f1f9a69b592328781a507106743f9c479a747dc. Actual visual scope: PDF1 and11-17 (printed6-12). This narrow read tests the specifically discovered correction attempts and a settlement footnote; it is not a whole-brief or appendix read. The original correspondence and underlying appendix exhibits were not acquired.\n\n## What the historical labels meant and who used them\n\nThe opinion describes DoubleVerify gathering information and selling confidential reports to businesses considering advertising on websites. Its report glossary defined the categories. The Adult Content label encompassed mature material inappropriate for children, including explicit language, sounds and themes; it was not limited to pornography or unlawful material. The copyright streaming/file-sharing label encompassed sites presently or historically associated with distributing or accessing copyrighted material without suitable controls, licensing or permission (printed pp3-4/PDF4-5). The category definition is a provider's classification rule, not an adjudication that every labelled site was currently infringing.\n\nThe Supreme Court described the report's three columns: domains/subdomains, total impressions, and thematic categories defined by DoubleVerify (printed p22/PDF23). Paying recipients agreed to confidentiality and used the reports for their business purposes (printed p23/PDF24). These pages establish the described classification/distribution mechanism; they do not identify a universal downstream blocking rule.\n\nFilmOn alleged the labels were false and caused existing or prospective advertising partners to refuse business. The opinion states those as allegations (printed p4/PDF5), not a damages finding or independently established causal estimate. This packet has no final calculation of lost revenue, advertiser instruction, restored contract or actual reclassification receipt.\n\n## An actual alleged correction attempt, with its limits\n\nThe opening brief bears a filed stamp of 16 January2018. FilmOn's account says it learned of reports through a former client in late2013 and sent a cease-and-desist letter on12 December2013 demanding tag removal and correction. It says DoubleVerify subsequently described Adult Content as akin to an R or TV-MA rating, but did not clarify to advertisers that FilmOn's content was nonpornographic or limit the label to particular channels. It further alleges that after an investigation DoubleVerify maintained the classifications despite letters, calls and meetings (printed pp10-12/PDF15-17).\n\nThis is source-specific evidence that FilmOn formally asserted a disputed explanation/correction process, supported in its brief by appendix citations. It does not independently prove the contents of those communications, that the classification was false under its actual glossary, or that the alleged refusal persisted after2018. The distinction matters: a debate over whether a broad category communicated pornography, and whether it should apply to an entire network or individual channels, is more specific than a generic allegation of blacklisting.\n\nThe same brief's printed p12/PDF17 footnote2 says AOL was dismissed pursuant to a settlement on9 April2015. That is FilmOn counsel's report about another named defendant at an earlier stage. No settlement agreement, dismissal order, terms or payment was acquired. It is not evidence of a DoubleVerify settlement, a correction or restored advertising.\n\n## What the Supreme Court decided\n\nThe trial court granted DoubleVerify's anti-SLAPP motion and the Court of Appeal affirmed. The Supreme Court reversed the latter judgment because DoubleVerify's confidential reports did not qualify under Code of Civil Procedure425.16(e)(4), the public-interest catchall (printed pp2-6,17-25).\n\nThe court's inquiry identifies the public issue implicated by the speech, then asks about its functional relationship to public discussion. Audience, speaker and purpose inform that second inquiry. The mere presence of a broad public-interest subject, such as copyright or adult content, was insufficient here. The reports were exchanged privately for the recipients' business purposes and were not intended to participate in public discussion. This is a statutory procedural ruling, not a judicial validation of FilmOn's allegation that the labels were false.\n\nThe court expressly refused to exclude all commercially oriented or confidential speech categorically. No single contextual factor was dispositive. It also rejected FilmOn's proposed commercial/noncommercial sorting, even though FilmOn won the ultimate reversal (printed pp15-16/PDF16-17; p23/PDF24). Both parties agreed the distinct section425.17(c) exemption did not apply because the reports concerned FilmOn, not DoubleVerify or a competitor (printed p14 fn4/PDF15). Thus winning the result does not mean every argument was adopted.\n\nThe opinion grants judicial notice of copyright-case orders and legislative history (printed p18 fn6/PDF19); that does not convert its narrow anti-SLAPP disposition into an adjudication of the accuracy of all challenged labels. The actual final paragraph simply reverses (printed p25/PDF26). It orders no correction, reclassification, ad purchase, damages payment or service restoration.\n\n## The media coalition's requested protection\n\nThe filed amicus was submitted in support of DoubleVerify by Motion Picture Association of America, Inc.; The Hearst Corporation; Tegna Inc.; California News Publishers Association; and First Amendment Coalition. It was filed6 June2018 and dated29 May2018. Davis Wright Tremaine LLP counsel Kelli L. Sager, Thomas R. Burke and Rochelle L. Wilcox are named; Wilcox signed it (PDF1-2,55-56). These are documented amicus/representation roles, not evidence of ownership, category authorship, funding of DoubleVerify or control of its decisions.\n\nIn the introductory interest statement, the amici describe their media/creative work and interest in protecting it from piracy. They support keeping rampant-infringement sites and services out of the legitimate advertising market and describe companies such as DoubleVerify as helping that effort (printed pp10-11/PDF11-12). Their summary says brands use assurance-provider information to instruct advertising networks where ads should or should not appear (printed p14/PDF15). That is the amici's stated commercial mechanism and rationale, not independent proof that all listed amici bought this particular report.\n\nThey argue that reports are informational products, analogous to news and films, rather than advertising for a product, and fear a broad commercial-context exception could endanger other speakers (printed pp15-17/PDF16-18). Their express request was to reject FilmOn's arguments and affirm protection of DoubleVerify's speech under anti-SLAPP (printed p54/PDF55). The Supreme Court did not grant that requested disposition. Its refusal to adopt a categorical commercial-speech exclusion overlaps part of their concern, but this packet does not infer that their brief caused that reasoning.\n\n## Outcome frontier and restrained mapping\n\nFour focused discovery queries sought the original and same-case remand/settlement/dismissal. They located the opinion, filed briefing and unrelated settlements/cases citing FilmOn. No later FilmOn-versus-DoubleVerify final merits judgment, correction, reclassification or advertising restoration record was located. This is a bounded search result, not proof those records or events do not exist. The only specifically named settlement found was the earlier AOL footnote, preserved at its attributed scope.\n\nMeaningful relations supported now are DoubleVerify-to-confidential-report category authorship/distribution as described by the court; FilmOn-to-DoubleVerify asserted correction demand; named amici-to-Supreme Court filed request for procedural protection; and Supreme Court-to-prior judgment reversal. A potential later remedy requires its own order, agreement or correction/recipient notice. The 2025 eGumball court's reliance on FilmOn is a legal-precedent relationship, not a shared operational database. Acquisition is closed for root review; no new case, current product, ownership, payment or content-site branch was opened.\n", "file": "research/filmon-classification-wave-2026-09-09-judicial-case.md" }, { "id": "funding-wave-2026-09-09-reset-funding", "title": "Reset.Tech Australia: three documented funding instruments", "description": "Observed 9 September 2026. This bounded pass follows the latest forest review and the user's funding steering. It establishes funding for regulatory advocacy, a specific commission producing policy proposals, and a separately scoped privacy-research award. It does not establish donor authorship or approval of Australia's September 2026 exposure draft.", "html": "Observed 9 September 2026. This bounded pass follows the latest forest review and the user's funding steering. It establishes funding for regulatory advocacy, a specific commission producing policy proposals, and a separately scoped privacy-research award. It does not establish donor authorship or approval of Australia's September 2026 exposure draft.
\nThe most consequential bridge is McKinnon's commission of two identified reports, one of which expressly recommends considering a duty of care and the other of which transmits EU and UK transparency models. Mannifera supplies a quantified award for advocacy during the Online Safety Act review. The Internet Society Foundation award has a different, documented children's-privacy research purpose.
\n| Relationship | \nAmount and time | \nEvidentiary state | \n
|---|---|---|
| Mannifera → Reset Australia | \n$40,000; labelled a 2024 Signature Grant | \nGrantor names recipient and purpose. Currency code is not printed; Australian context, not converted to USD. No bank receipt or exact award/payment date acquired. | \n
| Internet Society Foundation → Reset Tech Australia | \nUS$219,000; 8 January 2024–7 July 2025 | \nProject award record. A separate 2024 return names Reset Tech Australia Limited for US$87,600 monetary support; these figures must not be added as independent awards. | \n
| McKinnon → two Reset reports | \nCommissioned in 2023 and 2024; fee undisclosed | \nFunder identifies both commissions. Recipient's first report also describes the early-2023 approach. No signed contract or payment amount acquired. | \n
Reset's 2024 transparency report, printed/PDF page 37, lists Reset.Tech (US), Internet Society Foundation, Susan McKinnon Foundation, Boundless Foundation, “Manniferra,” an unnamed private donor and NSW Government. The NSW footnote limits the acknowledgment to support attending the social-media summit. No amounts or rights appear there. Page 6 describes duty-of-care advocacy; page 38 identifies future work on children's privacy and the duty. These separate passages do not allocate any donor's money to each activity. The report says annual accounts were being prepared; that historical statement does not establish their present availability. [Reset 2024 transparency report](http[local research file] inspected PDF pp2, 6, 37–38 and relevant report-title references.
\nThe Mannifera identification is supported by the donor explicitly naming Reset, rather than merely correcting a similar spelling. Internet Society Foundation's return gives the legal recipient name. The other recipient-only names remain at their disclosed scope: no exact Boundless legal entity, unnamed donor identity, Reset US award amount or NSW reimbursement amount is resolved by this pass. The closed ACNC route was not reopened.
\nUnder its own Reset Australia heading, Mannifera records a $40,000 2024 Signature Grant from its dis/misinformation round. Its purpose is to bring Reset's research and advocacy on transparency mechanisms into policy development and the social-media committee inquiry. The passage specifically situates the work around the misinformation bill and Online Safety Act review. That supports an advocacy-funding edge; it does not identify an earmark for particular duty-of-care clauses. The preceding Australian Progress section also mentions Reset under a different heading; that ambiguous passage is not counted as a second Reset award. [Mannifera partner record](http[local research file] Reset heading and adjacent section, rendered lines 213–235.
\nMannifera's FY24 account says a national dis/misinformation call opened in February 2024. Its “so far” recipient list does not include Reset. Consequently, the two pages do not fix Reset's award date within FY2023/24 or contradict the later listing's calendar-2024 label. [FY24 account](http[local research file] body paragraphs on the call and listed recipients, lines 5–15.
\nThe undated membership guide describes a formal partnership involving Philanthropy Australia, Australian Communities Foundation (ACF) and Reichstein Foundation, and a funds path through an ACF sub-fund. Members participate in collective funding decisions at annual Musters; a member Taskforce and Executive Officer oversee activities. Signature grants support core operations advancing advocacy projects. These are published programme arrangements, not the signed Reset award terms or evidence that every member approved that award. No individual donor veto, Reset board-appointment right, publication clearance or parliamentary drafting power is specified in the inspected passages. [Membership guide](http[local research file] physical PDF pp6–8, especially grant types, governance/legal structure and member participation. The URL's 2024/04 path is not treated as a verified publication date.
\nThe administering organisation independently says Mannifera members give together through ACF. This corroborates the pooled route, without identifying the trustee/payor on Reset's particular contract. [ACF Annual Report 2023/24](http[local research file] physical PDF p12, printed pp20–21. The report's text search did not locate Reset; that is a search limitation, not proof ACF made no payment.
\nThe Foundation's phase-two project record names Reset Tech Australia and US$219,000 for 8 January 2024–7 July 2025. It examines young people's privacy expectations in educational technology and targeting, and how policymakers and regulators can implement those expectations. Its international scope includes Australia. The record establishes a funded research purpose, not a commission of Australia's duty-of-care legislation; anticipated methods and results are not evidence of completed outputs. [Phase-two grant record](http[local research file] project dates, amount and research summary, rendered lines 87–141.
\nThe 2024 Form 990 identifies Internet Society Foundation, EIN 82-3285688. Schedule A names Reset Tech Australia Limited with US$87,600 monetary support. Schedule F has an unnamed same-amount research row; matching does not establish its recipient. Its supplement specifies accrual accounting. No recipient cash receipt or instalment reconciliation is established. Reported procedures require proposal approval under delegated authority and a board-approved budget, signed agreements, monitoring and completion reports. These are general institutional procedures, not acquired Reset-specific terms. [2024 Form 990](http[local research file] PDF p1; pp23–24 Schedule A table; p39 Schedule F row; p42 monitoring/accounting supplement. Inspected text supports these statements; image rendering failed, so no visual table verification is claimed.
\nMcKinnon says it commissioned Reset in 2023 and 2024 to produce Functioning or Failing? and Achieving Digital Platform Public Transparency in Australia. This identifies particular research commissions rather than general organisational support. It discloses no fee, contractual clearance, intellectual-property allocation or termination provision. [McKinnon account](http[local research file] Reset section, rendered lines 61–66.
\nThe first report is dated May 2024. Its introduction says Susan McKinnon Foundation approached Reset early in 2023 to design a project testing the misinformation code; Reset describes designing the monitoring and experiments using methods from its global network. Its recommendations include considering a duty of care for misinformation/disinformation, EU-style risk assessments and researcher access. This establishes an initiated research commission and an output containing policy proposals. It does not establish that the commissioner required those conclusions or that government adopted them. [Functioning or Failing?](http[local research file] PDF pp1–5, especially recommendations p3 and introduction p5. Research validity and the body experiments were not independently audited here.
\nThe second original is dated June 2024. Its summary explicitly examines the EU Digital Services Act and UK Online Safety Act, proposing Australian risk assessments, prescribed annual transparency reports, independent audits, data portals and researcher access. The proposed application spans online safety and misinformation frameworks. This is an identified commissioned model-transmission artifact, not just ideological similarity. [Achieving Digital Platform Public Transparency in Australia](http[local research file] PDF pp1–3. A public GET succeeded and those pages were extracted from the original 31-page PDF in memory. The funding caption's logo did not appear in text extraction; the commissioning attribution rests on McKinnon's named account, not an unread logo.
\nMcKinnon's current privacy policy identifies The Susan McKinnon Charitable Foundation, ABN 12 653 756 597, together with subsidiaries/affiliates, and also uses the company's “Limited” legal name. The policy is dated September 2025. It identifies a legal entity within the present McKinnon group; it does not determine which entity signed the earlier commissions. [Privacy policy](http[local research file] opening application/definition paragraphs, rendered lines 52–65. The precise historic contracting party remains open.
\nSafe financial verbs are: Mannifera records a 2024 grant for transparency advocacy; Internet Society Foundation lists a phase-two research award and reports annual monetary support; McKinnon says it commissioned two identified reports; Reset describes the commissioning approach and publishes recommendations. None should be replaced by “paid for the bill,” “dictated the duty” or “approved the final clauses.” The strongest supported mechanism is financing research and advocacy that formulates and circulates regulatory models. This pass does not isolate that funding's causal effect on government choices.
\nThe narrow next records are the Mannifera/ACF grant letter and disbursement record identifying Reset's legal counterparty and restrictions; the Foundation's executed phase-two agreement, amendments and instalment reconciliation; and McKinnon's 2023/24 scopes of work, executed contracts and publication-clearance provisions. Likely custodians are the named grantor/administrator and Reset. Their absence here leaves payment timing and substantive approval rights unresolved, not presumed absent. No request or outreach has been made.
\nAcquisition scope: public grantor/recipient pages and scoped PDF readings only, observed 9 September 2026. Web image rendering returned cache misses for the Foundation return and Mannifera guide. The second McKinnon report exceeded the web reader's size limit; one ordinary public GET supplied its opening pages. No originals are retained in this one-file lane and no original-file hashes are claimed. No ACNC, account, paid-access, Butterfield or wider donor branch was opened. This is the financial-instrument breakpoint for the next whole-case reassessment.
", "path": "funding-wave-2026-09-09/reset-funding.md", "sourceIds": [ "w11-fund-s-reset-disclosure", "w11-fund-s-mannifera-partners", "w11-fund-s-mannifera-guide", "w11-fund-s-acf-report", "w11-fund-s-isf-award", "w11-fund-s-isf990", "w11-fund-s-mckinnon-commissions", "w11-fund-s-functioning-report", "w11-fund-s-transparency-report", "w11-fund-s-mckinnon-identity" ], "markdown": "Observed 9 September 2026. This bounded pass follows the latest forest review and the user's funding steering. It establishes funding for regulatory advocacy, a specific commission producing policy proposals, and a separately scoped privacy-research award. It does not establish donor authorship or approval of Australia's September 2026 exposure draft.\n\nThe most consequential bridge is McKinnon's commission of two identified reports, one of which expressly recommends considering a duty of care and the other of which transmits EU and UK transparency models. Mannifera supplies a quantified award for advocacy during the Online Safety Act review. The Internet Society Foundation award has a different, documented children's-privacy research purpose.\n\n| Relationship | Amount and time | Evidentiary state |\n| --- | --- | --- |\n| Mannifera → Reset Australia | $40,000; labelled a 2024 Signature Grant | Grantor names recipient and purpose. Currency code is not printed; Australian context, not converted to USD. No bank receipt or exact award/payment date acquired. |\n| Internet Society Foundation → Reset Tech Australia | US$219,000; 8 January 2024–7 July 2025 | Project award record. A separate 2024 return names Reset Tech Australia Limited for US$87,600 monetary support; these figures must not be added as independent awards. |\n| McKinnon → two Reset reports | Commissioned in 2023 and 2024; fee undisclosed | Funder identifies both commissions. Recipient's first report also describes the early-2023 approach. No signed contract or payment amount acquired. |\n\n## Recipient disclosure and entity boundaries\n\nReset's 2024 transparency report, printed/PDF page 37, lists Reset.Tech (US), Internet Society Foundation, Susan McKinnon Foundation, Boundless Foundation, “Manniferra,” an unnamed private donor and NSW Government. The NSW footnote limits the acknowledgment to support attending the social-media summit. No amounts or rights appear there. Page 6 describes duty-of-care advocacy; page 38 identifies future work on children's privacy and the duty. These separate passages do not allocate any donor's money to each activity. The report says annual accounts were being prepared; that historical statement does not establish their present availability. Reset 2024 transparency report (see source references) inspected PDF pp2, 6, 37–38 and relevant report-title references.\n\nThe Mannifera identification is supported by the donor explicitly naming Reset, rather than merely correcting a similar spelling. Internet Society Foundation's return gives the legal recipient name. The other recipient-only names remain at their disclosed scope: no exact Boundless legal entity, unnamed donor identity, Reset US award amount or NSW reimbursement amount is resolved by this pass. The closed ACNC route was not reopened.\n\n## Mannifera: a quantified advocacy grant and an identifiable pooled-funding route\n\nUnder its own **Reset Australia** heading, Mannifera records a $40,000 2024 Signature Grant from its dis/misinformation round. Its purpose is to bring Reset's research and advocacy on transparency mechanisms into policy development and the social-media committee inquiry. The passage specifically situates the work around the misinformation bill and Online Safety Act review. That supports an advocacy-funding edge; it does not identify an earmark for particular duty-of-care clauses. The preceding Australian Progress section also mentions Reset under a different heading; that ambiguous passage is not counted as a second Reset award. Mannifera partner record (see source references) Reset heading and adjacent section, rendered lines 213–235.\n\nMannifera's FY24 account says a national dis/misinformation call opened in February 2024. Its “so far” recipient list does not include Reset. Consequently, the two pages do not fix Reset's award date within FY2023/24 or contradict the later listing's calendar-2024 label. FY24 account (see source references) body paragraphs on the call and listed recipients, lines 5–15.\n\nThe undated membership guide describes a formal partnership involving Philanthropy Australia, Australian Communities Foundation (ACF) and Reichstein Foundation, and a funds path through an ACF sub-fund. Members participate in collective funding decisions at annual Musters; a member Taskforce and Executive Officer oversee activities. Signature grants support core operations advancing advocacy projects. These are published programme arrangements, not the signed Reset award terms or evidence that every member approved that award. No individual donor veto, Reset board-appointment right, publication clearance or parliamentary drafting power is specified in the inspected passages. Membership guide (see source references) physical PDF pp6–8, especially grant types, governance/legal structure and member participation. The URL's 2024/04 path is not treated as a verified publication date.\n\nThe administering organisation independently says Mannifera members give together through ACF. This corroborates the pooled route, without identifying the trustee/payor on Reset's particular contract. ACF Annual Report 2023/24 (see source references) physical PDF p12, printed pp20–21. The report's text search did not locate Reset; that is a search limitation, not proof ACF made no payment.\n\n## Internet Society Foundation: restricted research purpose, separate annual support figure\n\nThe Foundation's phase-two project record names Reset Tech Australia and US$219,000 for 8 January 2024–7 July 2025. It examines young people's privacy expectations in educational technology and targeting, and how policymakers and regulators can implement those expectations. Its international scope includes Australia. The record establishes a funded research purpose, not a commission of Australia's duty-of-care legislation; anticipated methods and results are not evidence of completed outputs. Phase-two grant record (see source references) project dates, amount and research summary, rendered lines 87–141.\n\nThe 2024 Form 990 identifies **Internet Society Foundation, EIN 82-3285688**. Schedule A names **Reset Tech Australia Limited** with **US$87,600 monetary support**. Schedule F has an unnamed same-amount research row; matching does not establish its recipient. Its supplement specifies accrual accounting. No recipient cash receipt or instalment reconciliation is established. Reported procedures require proposal approval under delegated authority and a board-approved budget, signed agreements, monitoring and completion reports. These are general institutional procedures, not acquired Reset-specific terms. 2024 Form 990 (see source references) PDF p1; pp23–24 Schedule A table; p39 Schedule F row; p42 monitoring/accounting supplement. Inspected text supports these statements; image rendering failed, so no visual table verification is claimed.\n\n## McKinnon: a funded route into concrete regulatory proposals\n\nMcKinnon says it commissioned Reset in 2023 and 2024 to produce **Functioning or Failing?** and **Achieving Digital Platform Public Transparency in Australia**. This identifies particular research commissions rather than general organisational support. It discloses no fee, contractual clearance, intellectual-property allocation or termination provision. McKinnon account (see source references) Reset section, rendered lines 61–66.\n\nThe first report is dated May 2024. Its introduction says Susan McKinnon Foundation approached Reset early in 2023 to design a project testing the misinformation code; Reset describes designing the monitoring and experiments using methods from its global network. Its recommendations include considering a duty of care for misinformation/disinformation, EU-style risk assessments and researcher access. This establishes an initiated research commission and an output containing policy proposals. It does not establish that the commissioner required those conclusions or that government adopted them. Functioning or Failing? (see source references) PDF pp1–5, especially recommendations p3 and introduction p5. Research validity and the body experiments were not independently audited here.\n\nThe second original is dated June 2024. Its summary explicitly examines the EU Digital Services Act and UK Online Safety Act, proposing Australian risk assessments, prescribed annual transparency reports, independent audits, data portals and researcher access. The proposed application spans online safety and misinformation frameworks. This is an identified commissioned model-transmission artifact, not just ideological similarity. Achieving Digital Platform Public Transparency in Australia (see source references) PDF pp1–3. A public GET succeeded and those pages were extracted from the original 31-page PDF in memory. The funding caption's logo did not appear in text extraction; the commissioning attribution rests on McKinnon's named account, not an unread logo.\n\nMcKinnon's current privacy policy identifies **The Susan McKinnon Charitable Foundation, ABN 12 653 756 597**, together with subsidiaries/affiliates, and also uses the company's “Limited” legal name. The policy is dated September 2025. It identifies a legal entity within the present McKinnon group; it does not determine which entity signed the earlier commissions. Privacy policy (see source references) opening application/definition paragraphs, rendered lines 52–65. The precise historic contracting party remains open.\n\n## Integration limit and next obtainable records\n\nSafe financial verbs are: Mannifera **records a 2024 grant for transparency advocacy**; Internet Society Foundation **lists a phase-two research award** and **reports annual monetary support**; McKinnon **says it commissioned two identified reports**; Reset **describes the commissioning approach and publishes recommendations**. None should be replaced by “paid for the bill,” “dictated the duty” or “approved the final clauses.” The strongest supported mechanism is financing research and advocacy that formulates and circulates regulatory models. This pass does not isolate that funding's causal effect on government choices.\n\nThe narrow next records are the Mannifera/ACF grant letter and disbursement record identifying Reset's legal counterparty and restrictions; the Foundation's executed phase-two agreement, amendments and instalment reconciliation; and McKinnon's 2023/24 scopes of work, executed contracts and publication-clearance provisions. Likely custodians are the named grantor/administrator and Reset. Their absence here leaves payment timing and substantive approval rights unresolved, not presumed absent. No request or outreach has been made.\n\nAcquisition scope: public grantor/recipient pages and scoped PDF readings only, observed 9 September 2026. Web image rendering returned cache misses for the Foundation return and Mannifera guide. The second McKinnon report exceeded the web reader's size limit; one ordinary public GET supplied its opening pages. No originals are retained in this one-file lane and no original-file hashes are claimed. No ACNC, account, paid-access, Butterfield or wider donor branch was opened. This is the financial-instrument breakpoint for the next whole-case reassessment.\n", "file": "research/funding-wave-2026-09-09-reset-funding.md" }, { "id": "funding-wave-2026-09-09-uk-model-funding", "title": "UK model funding: money, work and remaining decision rights", "description": "Observed 9 September 2026. Bounded acquisition after the duty-care forest review. Primary records below establish financed policy work and institutional succession. They do not establish donor approval of specific statutory clauses. No outreach or public release.", "html": "Observed 9 September 2026. Bounded acquisition after the duty-care forest review. Primary records below establish financed policy work and institutional succession. They do not establish donor approval of specific statutory clauses. No outreach or public release.
\nThe recipient is The Carnegie United Kingdom Trust, Scottish charity SC012799 / Irish charity 20142957, incorporated by Royal Charter. Do not merge it with Carnegie Corporation of New York or other similarly named Carnegie foundations.
\nThe 2021 accounts, notes 4 and 19 (printed26 and40; PDF26 and40), report GBP122,500 Minderoo grant income and GBP123,330 Luminate partnership income. Note19 expressly assigns both to Harm Regulation and Social Media. Printed8/PDF8 describes that project as promoting a duty-of-care regulatory system with Woods and Walsh, including a reworked bill, proposed amendments, parliamentary evidence and advice to the UN Special Rapporteur. These are reported restricted project income, not new independent donor cash confirmations or personal payments to the authors.
\n2021 balances and spending: Luminate opening108,357 + income123,330 - expenditure107,756 = closing123,931; Minderoo opening0 + income122,500 - expenditure32,468 = closing90,032. The 2022 comparative within the 2023 accounts reports no new income to these funds: Luminate resources expended and transfers122,500, closing1,431; Minderoo resources expended and transfers54,572, closing35,460. These are movements in the same funds, not additional grants to sum indiscriminately.
\nThe 2023 accounts note4 reports GBP168,844 Luminate partnership income for2023. Note16 records the online-safety fund spending/transfer line as170,275 for Luminate,35,460 for Minderoo and3,600 for an unnamed donation, closing at zero. Narrative printed8 calls the financial support Luminate/Reset; this does not resolve the contracting legal entity or mean two separate grants. Note16 states the funding supported completion of work on the Online Safety Act.
\nImportant source anomaly: note16's current-year opening heading actually prints 1 Jan2022. Local visual inspection of PDF28 confirms this is in the document, not a parser error. Its opening balances match the preceding2022 closing balances; the page says year to31Dec2023, its closing column31Dec2023, and note4 independently labels168,844 as2023. Preserve the anomaly; infer the opening heading likely retained the previous year. Do not silently correct a quoted table.
\nCarnegie's2023 report separately identifies its endowment and its own governing board as resources and decision-makers. Perrin was a trustee and Audit and Risk Committee member in that report. The trustees set strategy/budgets, with execution delegated to management. A donor's restricted purpose constrains permissible expenditure; the published accounts do not reveal a donor veto over conclusions, ministerial appointments, draft wording or publication.
\n2021 source: http[local research file] . Web PDF accessible,46pages; scoped project/finance sections read. 2023 source: http[local research file] .31pages; finance/governance/project sections read; original retained and note16 visually inspected. 2023 report signed5June2024. Local original SHA256 e3ff7704f72954aa1df542aad2aa30c8395806068a4fcd9f2edbd3877fa57c7f,788306bytes. Do not infer a2026 grant from website upload/index dates.
\nCarnegie's22November2021 announcement credits Frontier Technology, a Minderoo Foundation initiative, for financial support of its advice to the UN Special Rapporteur on Minority Issues. Woods/Perrin carried the systems/risk approach into that advice. This is attributable financial support for a named international output. It is not evidence the UN adopted all its recommendations, or that Minderoo dictated them. Amount and agreement remain absent; do not count this as a new additional grant outside the accounts' project funding.
\nhttp[local research file] . Primary announcement body read; no UN adoption audit in this pass.
\nCarnegie's26October2023 transition statement says its programme ends and Woods/Walsh continue through the new Online Safety Act Network, supported by former partner Reset. The2023 report independently says Carnegie supported development of the new network. This establishes deliberate succession of expertise and convening work after enactment, not ownership of the network or transfer of every employee/asset.
\nhttp[local research file] . Primary body read. Indexed older homepage at http[local research file] repeats Reset support, but a fresh open redirects to http[local research file] and the fetched body no longer mentions Reset. Thus rely on the dated transition statement for historical support, not a claim about the exact current funding arrangement. The renamed network's2026 legal operator, grant recipient, current funders, budget and agreement rights remain consequential gaps. The current about-page route returns an internal error; stop that route.
\nLuminate's current funding-process page describes strategy alignment, internally sourced opportunities (no unsolicited applications), due diligence, approval and legal funding agreements setting structure and reporting. It also offers unrestricted funding and ongoing programmatic contact. This makes grant proposal, internal approval, signed agreement and reporting schedule specific plausible evidence targets. Generic present-day procedure cannot establish the terms of a2021/2023 Carnegie award.
\nhttp[local research file] . Primary body available. Luminate Foundation Inc's2023 990-PF was accessible: http[local research file] . Web-text searches for Carnegie and Reset found no match. This is a limited search result, not proof of no grant; the correct contracting entity and grant year remain unresolved. Do not merge Luminate Foundation Inc automatically with every Luminate/Reset-branded award.
\nThe2021 accounts separately identify GBP90,000 Scottish Government funding for National Digital Ethics; this is a different restricted project, not evidence the government funded the Harm Regulation project. Note5 reports GBP76,437 as the value of Google advertising credits used, not cash, and describes their general charitable promotion use. The2020 comparative names Oak Foundation under Affordable Credit, not online safety. These are real relations with different purposes; attaching them all to a single online-regulation campaign would manufacture the causal conclusion.
\nCarnegie's reserve/governance disclosures support a mixed-resource organization, not a foundation merely passing through outside instructions. Strategic patronage can still shape which work receives sustained capacity. The unanswered question is its extent at identifiable decisions.
\nObtain the signed Carnegie Luminate/Reset and Minderoo grant agreements and schedules, including contracting entities, grant period, restricted deliverables, amendment/publication approval, reporting and clawback terms. Compare Carnegie trustee/programme approval records to those schedules. Obtain the2023 network transition agreement and its legal/payroll/fiscal-host arrangement. These should distinguish topic selection and capacity support from approval of outputs. Existence of some agreement/reporting documents is strongly implied by accounts and published funding procedure, but their availability is unverified.
\nAcquisition stopped at this coherent funding-to-output breakpoint. Root must integrate Reset's parallel return and explicitly reassess the whole case before selecting another acquisition run.
\nRoot checked the primary Mannifera partner page while verifying Reset's grant. Its Human Rights Law Centre section explicitly records a2024 $60,000 dis/misinformation Signature Grant (currency code unspecified) and says supported work aligns legal/tech experts around reform positions, then trains civil-society advocates and makes the public case. It identifies the misinformation bill, Online Safety Act reforms and privacy reforms as the policy opportunities. This is a published purpose of funded coalition-building; it does not identify all participants, a particular shared clause, an executed payment or a direction to Reset. The previously listed2023/2022 grants must not be folded into the2024 purpose without further evidence. No additional acquisition branch was opened. Source: http[local research file] , Human Rights Law Centre heading, rendered lines9-33, observed9September2026.
", "path": "funding-wave-2026-09-09/uk-model-funding.md", "sourceIds": [ "w11-fund-s-carnegie2021", "w11-fund-s-carnegie2023", "w11-fund-s-un-advice", "w11-fund-s-succession", "w11-fund-s-luminate-process" ], "markdown": "Observed 9 September 2026. Bounded acquisition after the duty-care forest review. Primary records below establish financed policy work and institutional succession. They do not establish donor approval of specific statutory clauses. No outreach or public release.\n\n## Carnegie UK: recipient identity and restricted purpose\n\nThe recipient is The Carnegie United Kingdom Trust, Scottish charity SC012799 / Irish charity 20142957, incorporated by Royal Charter. Do not merge it with Carnegie Corporation of New York or other similarly named Carnegie foundations.\n\nThe 2021 accounts, notes 4 and 19 (printed26 and40; PDF26 and40), report GBP122,500 Minderoo grant income and GBP123,330 Luminate partnership income. Note19 expressly assigns both to Harm Regulation and Social Media. Printed8/PDF8 describes that project as promoting a duty-of-care regulatory system with Woods and Walsh, including a reworked bill, proposed amendments, parliamentary evidence and advice to the UN Special Rapporteur. These are reported restricted project income, not new independent donor cash confirmations or personal payments to the authors.\n\n2021 balances and spending: Luminate opening108,357 + income123,330 - expenditure107,756 = closing123,931; Minderoo opening0 + income122,500 - expenditure32,468 = closing90,032. The 2022 comparative within the 2023 accounts reports no new income to these funds: Luminate resources expended and transfers122,500, closing1,431; Minderoo resources expended and transfers54,572, closing35,460. These are movements in the same funds, not additional grants to sum indiscriminately.\n\nThe 2023 accounts note4 reports GBP168,844 Luminate partnership income for2023. Note16 records the online-safety fund spending/transfer line as170,275 for Luminate,35,460 for Minderoo and3,600 for an unnamed donation, closing at zero. Narrative printed8 calls the financial support Luminate/Reset; this does not resolve the contracting legal entity or mean two separate grants. Note16 states the funding supported completion of work on the Online Safety Act.\n\nImportant source anomaly: note16's current-year opening heading actually prints 1 Jan2022. Local visual inspection of PDF28 confirms this is in the document, not a parser error. Its opening balances match the preceding2022 closing balances; the page says year to31Dec2023, its closing column31Dec2023, and note4 independently labels168,844 as2023. Preserve the anomaly; infer the opening heading likely retained the previous year. Do not silently correct a quoted table.\n\nCarnegie's2023 report separately identifies its endowment and its own governing board as resources and decision-makers. Perrin was a trustee and Audit and Risk Committee member in that report. The trustees set strategy/budgets, with execution delegated to management. A donor's restricted purpose constrains permissible expenditure; the published accounts do not reveal a donor veto over conclusions, ministerial appointments, draft wording or publication.\n\n2021 source: http[local research file] . Web PDF accessible,46pages; scoped project/finance sections read. 2023 source: http[local research file] .31pages; finance/governance/project sections read; original retained and note16 visually inspected. 2023 report signed5June2024. Local original SHA256 e3ff7704f72954aa1df542aad2aa30c8395806068a4fcd9f2edbd3877fa57c7f,788306bytes. Do not infer a2026 grant from website upload/index dates.\n\n## An identifiable financed international output\n\nCarnegie's22November2021 announcement credits Frontier Technology, a Minderoo Foundation initiative, for financial support of its advice to the UN Special Rapporteur on Minority Issues. Woods/Perrin carried the systems/risk approach into that advice. This is attributable financial support for a named international output. It is not evidence the UN adopted all its recommendations, or that Minderoo dictated them. Amount and agreement remain absent; do not count this as a new additional grant outside the accounts' project funding.\n\nhttp[local research file] . Primary announcement body read; no UN adoption audit in this pass.\n\n## Succession and Reset support\n\nCarnegie's26October2023 transition statement says its programme ends and Woods/Walsh continue through the new Online Safety Act Network, supported by former partner Reset. The2023 report independently says Carnegie supported development of the new network. This establishes deliberate succession of expertise and convening work after enactment, not ownership of the network or transfer of every employee/asset.\n\nhttp[local research file] . Primary body read. Indexed older homepage at http[local research file] repeats Reset support, but a fresh open redirects to http[local research file] and the fetched body no longer mentions Reset. Thus rely on the dated transition statement for historical support, not a claim about the exact current funding arrangement. The renamed network's2026 legal operator, grant recipient, current funders, budget and agreement rights remain consequential gaps. The current about-page route returns an internal error; stop that route.\n\n## Funder process: a map to records, not the missing agreement\n\nLuminate's current funding-process page describes strategy alignment, internally sourced opportunities (no unsolicited applications), due diligence, approval and legal funding agreements setting structure and reporting. It also offers unrestricted funding and ongoing programmatic contact. This makes grant proposal, internal approval, signed agreement and reporting schedule specific plausible evidence targets. Generic present-day procedure cannot establish the terms of a2021/2023 Carnegie award.\n\nhttp[local research file] . Primary body available. Luminate Foundation Inc's2023 990-PF was accessible: http[local research file] . Web-text searches for Carnegie and Reset found no match. This is a limited search result, not proof of no grant; the correct contracting entity and grant year remain unresolved. Do not merge Luminate Foundation Inc automatically with every Luminate/Reset-branded award.\n\n## Counterexamples that must survive the map\n\nThe2021 accounts separately identify GBP90,000 Scottish Government funding for National Digital Ethics; this is a different restricted project, not evidence the government funded the Harm Regulation project. Note5 reports GBP76,437 as the value of Google advertising credits used, not cash, and describes their general charitable promotion use. The2020 comparative names Oak Foundation under Affordable Credit, not online safety. These are real relations with different purposes; attaching them all to a single online-regulation campaign would manufacture the causal conclusion.\n\nCarnegie's reserve/governance disclosures support a mixed-resource organization, not a foundation merely passing through outside instructions. Strategic patronage can still shape which work receives sustained capacity. The unanswered question is its extent at identifiable decisions.\n\n## High-information records now specified\n\nObtain the signed Carnegie Luminate/Reset and Minderoo grant agreements and schedules, including contracting entities, grant period, restricted deliverables, amendment/publication approval, reporting and clawback terms. Compare Carnegie trustee/programme approval records to those schedules. Obtain the2023 network transition agreement and its legal/payroll/fiscal-host arrangement. These should distinguish topic selection and capacity support from approval of outputs. Existence of some agreement/reporting documents is strongly implied by accounts and published funding procedure, but their availability is unverified.\n\nAcquisition stopped at this coherent funding-to-output breakpoint. Root must integrate Reset's parallel return and explicitly reassess the whole case before selecting another acquisition run.\n\n## Incidental same-page coordination evidence, scoped for next review\n\nRoot checked the primary Mannifera partner page while verifying Reset's grant. Its Human Rights Law Centre section explicitly records a2024 $60,000 dis/misinformation Signature Grant (currency code unspecified) and says supported work aligns legal/tech experts around reform positions, then trains civil-society advocates and makes the public case. It identifies the misinformation bill, Online Safety Act reforms and privacy reforms as the policy opportunities. This is a published purpose of funded coalition-building; it does not identify all participants, a particular shared clause, an executed payment or a direction to Reset. The previously listed2023/2022 grants must not be folded into the2024 purpose without further evidence. No additional acquisition branch was opened. Source: http[local research file] , Human Rights Law Centre heading, rendered lines9-33, observed9September2026.\n", "file": "research/funding-wave-2026-09-09-uk-model-funding.md" }, { "id": "gdi-classification-recourse-wave-2026-09-09-recourse-findings", "title": "GDI classification recourse: an appeal is described, but downstream correction remains unproved", "description": "Bounded acquisition closed on 9 September 2026. Three primary records were read: GDI's April 2024 parliamentary evidence, its privacy/research policy updated April 2026, and IAS's dated announcement that it was discontinuing GDI input. Two original HTML files and one official-reader derivative were retained. The exact procedures differ in date and scope; non", "html": "Bounded acquisition closed on 9 September 2026. Three primary records were read: GDI's April 2024 parliamentary evidence, its privacy/research policy updated April 2026, and IAS's dated announcement that it was discontinuing GDI input. Two original HTML files and one official-reader derivative were retained. The exact procedures differ in date and scope; none supplies an executed distribution agreement or proves an individual appeal changed advertising decisions.
\nGDI CEO Clare Melford's written evidence FON0071, dated 29 April 2024, expressly says publishers can appeal their ratings through a linked review form on GDI's contact page. Under GDI's Review Process, it describes a two-to-three-week review and regular site reassessment, with ratings able to change when content changes. This is an explicit publisher-facing procedure described by the organisation to Parliament, stronger evidence than an invitation for academics to critique methodology. It is not a tribunal finding, a tested service-level guarantee, or proof of a successful appeal.
\nThe preceding GDI's Service section describes risk data produced using AI and human review. It says advertising technology companies use that data in purchasing, and attributes ultimate purchasing and data-use control to licensees. GDI also says it charges for the data. These are GDI's statements about its commercial mechanism, not independently read licence terms. They support separating classification, licensing and the recipient's advertising decision; they do not establish that GDI could compel a customer to undo a decision after a rating changed. The evidence does not publish an evidentiary test for appeal, reviewer independence, reasons requirement, interim suspension, deletion deadline, update delivery protocol or recipient acknowledgement.
\nSource: [FON0071 official parliamentary publication](http[local research file] headings GDI's Service and GDI's Review Process; retained reader lines 41–61, date at line 74. Full substantive reader text was read. The submission refers to a screenshot and links a review form; neither the screenshot nor the target form was acquired or exercised. Do not infer that the same form or review duration remains current in September 2026. The statement is not a detailed, versioned DEL methodology or an identification of the criteria governing the 2024 Oracle segment.
\nGDI's Privacy Policy, last updated 20 April 2026, identifies Disinformation Index Ltd as controller. Section 15 expressly accepts questions, concerns or complaints about personal-data processing or any research outputs, and says it will consider representations and, where appropriate, review, clarify or update findings. Section 19.3 acknowledges that some projects may have no advance notice to platforms or stakeholders; it describes a post-publication enquiry route, prompt review/response and corrective action where appropriate. This gives a public route to raise a research-output concern, although it does not promise the requested substantive result.
\nThe boundaries matter. Section 1 excludes personal-data processing for services under a separate agreement from this notice's coverage. Sections 14–16 describe data-subject rights and supervisory complaints, subject to applicable law and limitations; those do not automatically turn a corporate publisher's disagreement with a rating into a successful data-protection claim. Section 19.1 promises deletion of unnecessary or inaccurate personal data without undue delay, which is not an explicit promise to delete a domain rating or notify every advertising recipient. Section 19.3 says dedicated project notices describe retention and sharing but are not routinely published. No such project notice was obtained, and no contact was made.
\nThis policy does not supply a numerical research-complaint deadline equivalent to the historical two-to-three-week account, an independent appeal tier, or a rule requiring downstream customers to replace cached classifications or resume advertising. It must not be retroactively applied as the DEL's 2024 operating rules. Its statement in section 18 about personal data not being subject to automated decision-making with legal or significant effects is GDI's current self-description, not an audited conclusion about historical commercial outcomes.
\nSource: [GDI Privacy Policy](http[local research file] sections 1, 14–16, 18, 19.1–19.4 and final update date. Entire substantive text sections 1–20 was read through the official reader. Two embedded collection/retention tables were not visually read; no table contents are inferred.
\nIAS's IAS Enhancements to Context Control Avoidance announces a decision to discontinue the limited GDI input in that product and says its past use of GDI would not affect that solution going forward. The page describes customers selecting exclusions by categories and risk levels, but it does not publish the exact former GDI fields, a customer migration log, deletion protocol or an individual corrected classification. This is a directly named distributor's reported product change, not an observed technical removal or evidence that a particular publisher won an appeal. Its stated benefit claims, including neutrality and independence, are provider assertions rather than independent judgments.
\nThe headline and dateline say 18 December 2025. Retained HTML explicitly separates datePublished 2025-12-18T18:17:32Z from dateModified 2026-07-06T16:30:34Z, and the posted-date elements display both dates. The September 2026 captured wording cannot be assumed identical to the original December wording; no version comparison was acquired. Do not assign an exact implementation date from those page dates alone.
Source: [IAS announcement](http[local research file] full substantive announcement, especially the paragraph beginning its continuous-improvement explanation. Full body was reader-read; original HTML date elements and JSON-LD were locally inspected. This is distinct from Oracle's announced 30 September 2024 service end in the prior accepted packet, and from current GDI research products.
\nFON0071's GDI and Government Funding section says GDI had received no UK government funding since 2022; describes FCDO Russia-unit Countering Disinformation and Media Development work in 2018–2022 to develop large-scale detection and country studies; and says the supporting grant expired in March 2023. The receipt-period statement and later grant-expiry date are not a self-evident contradiction. It also attributes then-current resources to private foundations, commercial data licensing and a small EU Horizon Europe grant. This is a historical recipient disclosure, not donor-side confirmation, acquired conditions, proof of payment dates or government approval of a particular rating. No funding genealogy was pursued.
\nThe evidence now defeats the broad claim that no publisher appeal was publicly described. It also leaves a distinct question unanswered: what obliges a data recipient to stop using or replace a successfully challenged signal? GDI may revise a rating while an advertiser independently declines the inventory; the acquired documents do not reconcile those steps or prove any such instance occurred.
\nThe most useful expected instruments are held by GDI and its relevant licensee: the dated publisher-review procedure behind the 2024 form; a service-specific notice or licence annex defining corrections, updates and deletion; and an actual recipient instruction/acknowledgement following a resolved rating. IAS's versioned discontinuation/migration specification would test its broader no-continuing-effect statement. These are missing records, not proof that no rules or corrections exist.
\nThe 2019 GDI methodology PDF reader returned an unsafe-open non-retryable error and was closed without retry. It was not used as methodological evidence. The previously closed 2022 policy PDF and Jounce routes were not reopened even where search results resurfaced them. No named-site score audit, customer system, operational list, form submission, account, or harmful-content endpoint was accessed. The packet closes at three read primary records; acquisition and custody details are in recourse-captures.json.
Acquisition closed on 9 September 2026 at three substantive donor-side records: written answer 149420, the CDMD 2022–2023 programme summary, and one September 2022 spending entry. The programme original is retained; the parliamentary answer and spending entry have scoped official-reader derivatives. The findings do not identify a signed GDI grant agreement, its grant number, or clauses controlling publication, intellectual property or commercial reuse.
\nIn written answer 149420, answered 24 February 2023, FCDO minister Leo Docherty says the Counter Disinformation and Media Development Programme had funded GDI since 2019 for activities outside the US. He reports GBP 1,999,026 between 2019 and 2022, with a further GBP 600,797 in the then-current financial year. These are donor-reported amounts provided, not an independently reconciled payment ledger or merely a future award announcement. The answer does not identify individual disbursements or the precise period boundaries of the first amount. Neither amount should be added to the separate spending entry below.
\nAsked about discussions, the answer says there “will have been frequent contact” between the programme team and GDI, and regular performance reports and reviews. This describes the department's expected programme relationship; it supplies neither an actual meeting count nor the reports, instructions or review decisions. It does not establish a right to approve a particular rating, selection of a particular publisher, or customer purchasing decisions. Its explanation places GDI's technological work in the context of countering Russian disinformation.
\nSource: [official answer 149420](http[local research file] question and entire answer; reader lines 24–37, date line 32. Original GET returned 403 and was closed. fcdo-answer149420.reader.txt retains the working official-reader text; the retained error HTML is not substantive source content.
The outside-US wording is a consequential statement of funded-activity scope. Without the agreement, it is not a demonstrated contractual prohibition on every later use of knowledge, code or data in US markets. Conversely, commercial availability elsewhere does not show that FCDO commissioned US exclusions.
\nThe official Counter Disinformation and Media Development Programme Summary 2022 to 2023 gives identifier CSSF-05-000006, a seven-year programme running April 2016–March 2023, with FCDO and Cabinet Office partners. It identifies the Baltic States, Central Europe and Eastern Neighbourhood as the programme's operating region. Its FY 2022–2023 allocation is GBP 12 million ODA and GBP 21.1 million non-ODA. These are programme allocations; they are not GDI's income, an additional GDI award or demonstrated expenditure.
\nIts component table covers independent media/civil society, vulnerable audiences, partner capabilities, strategic communications with monitoring/evaluation/research, and programme delivery. Most implementers and contract dates are recorded generically as Various; the partner-capability component lists HMG and 1 April 2022–31 March 2023. GDI is not named in the summary. The objectives include analysing and exposing information operations, strengthening independent media and audience resilience, and international coordination. Expected results are aims, not evidence they occurred.
\nThis summary establishes the parent programme's identity and strategic envelope. It cannot allocate GDI to a component, establish its country list, or supply a GDI-specific deliverable or approval condition. Its duration also cannot substitute for the individual grant term even though its March 2023 endpoint is consistent with the recipient's reported grant expiry.
\nSource: [original ODT](http[local research file] entire document text, particularly identity/duration/budget fields, component table and support/results headings. Retained fcdo-cdmd2022-23.odt, SHA-256 c384de3682c5328fdb40642321b3733fcfcae23b6c0f7aee183e2785b9d5bc94. Read from its XML text; no rendered-page or visual claim. Metadata says three pages and 4 November 2024 creation/modification. The [official attachment index](http[local research file] was published 30 January 2025. This is a retrospectively published summary, not proof that these exact bytes were the governing instrument in 2022. The index's general sensitivity rationale for withholding some programme detail is not proof that a specific GDI clause was withheld.
FCDO's September 2022 spending-over-GBP-25,000 publication contains this selected row:
\n| Field | \nRecorded value | \n
|---|---|
| Department / entity | \nFCDO / FCDO | \n
| GL.Date | \n05/09/2022 | \n
| Unique.Identifier | \n300,000,130,671,861 | \n
| Invoice.Line.Number | \n1 | \n
| Supplier.Name | \nGLOBAL DISINFORMATION INDEX | \n
| Invoice.GBP.Line.AMT | \nGBP 130,506.00 | \n
| Category | \nProject Management | \n
This is a donor ledger/invoice-line record with a general-ledger date, not a recipient bank statement or proof of the exact settlement date. The row supplies no grant/programme identifier, deliverable or country. Its generic category must not be converted into an invented project-management contract. The identifier is a transaction-record identifier, not the grant number. The selected row is not a full annual reconciliation or evidence of the last payment.
\nSource: [official CSV preview](http[local research file] header and sole matched GDI row, reader lines 58–62 and 741. Preview says updated 28 November 2024, distinct from the September 2022 ledger period. fcdo-spend-sept2022.reader.txt retains these scoped passages. The CSV download reader failed; no original CSV was retained or another transport attempted.
The prior accepted FON0071 recipient disclosure says no UK funding since 2022, while its grant expired in March 2023. A September 2022 ledger date lies in both calendar 2022 and UK financial year 2022–2023. This demonstrates why the labels can coexist; it does not reconcile the two parliamentary amounts to receipts, allocate this row to either amount, or prove an exact last-funding date. The donor answer and the recipient account also describe programme participation using different start-period formulations; no unsupported reconciliation is imposed.
\nRead together with the earlier accepted product and recipient evidence, public funding of GDI's work and a commercial licensing operation are both documented at their respective source strengths. The missing causal instrument is what assets the commissioned work created and what reuse rights attached. The acquired records do not establish that donor-funded code or classifications were the same assets later licensed to Oracle or another buyer. Neither a tightly confined non-reusable commission nor unrestricted commercial reuse is proved.
\nThe next useful record would be FCDO's executed GDI award/variation and deliverable schedule under CDMD, with the financial schedule, acceptance/review rights and IP/data/publication provisions; corresponding performance returns would distinguish reserved rights from exercised instructions. FCDO and GDI are the expected holders. A supplier-licence annex would then be needed to join any retained assets to commercial use. No request or contact has been made.
\nA finite official-domain search covered parliamentary answers, GOV.UK programme disclosures, DevTracker and exact GDI/grant-agreement/IP terms. Answer 171377 was surfaced for country scope, but its ordinary reader failed and direct GET returned 403; the route was closed and the indexed answer was not promoted to body-read evidence. Other spending months, programme versions, third-party discussions and government-country studies remain discovery only. Previously closed GDI/Jounce routes were not revisited. Full transport, read scopes and hashes are in fcdo-captures.json; no case, graph or publication was changed.
Closed bounded acquisition, 9 September 2026. Three primary originals retained; one deliverable read selectively. This note makes no finding about the current operation of an advertising exclusion list.
\nThe identified project is AI4TRUST, grant 101070190, running 1 January 2023–28 February 2026, with an EC signature dated 6 December 2022. CORDIS, last updated 17 June 2026, labels it closed. The registered German participant is GDI GLOBAL DISINFORMATION INDEX GUGHAFTUNGSBESCHRANKT. Its published net EU contribution and total cost are each EUR 385,275. The whole-project contribution/cost is EUR 5,950,682.50 and must not become GDI income. Fondazione Bruno Kessler coordinates the Horizon research action. [CORDIS, project header, programme and German GDI participant entry](http[local research file]
\nGDI appears in the participant section, separate from the section describing partners that do not sign the grant. This supports a funded participant characterization rather than an associated-partner or UK-guarantee claim. The financial field is the registry's published EU contribution; no dated transfer ledger was acquired. Neither ownership of the German entity by UK Disinformation Index Ltd nor an IP transfer between them is established. The existing UKRI identity in the case is not joined to this project.
\nThis is a confirmed GDI Horizon project active during the April 2024 FON0071 disclosure. It is the strongest identified match, not an explicit identification of that disclosure's unnamed grant or proof that it was GDI's only Horizon project. The earlier statement and this record are distinct sources with different specificity.
\nD2.1, submitted 1 November 2023, records coordinator and quality-manager review (PDF 2–3). Its selected definitions distinguish project emphasis on intent from GDI's risk-based alternative (11–12). GDI's described NLP method uses analyst-coded examples and semantic matching to flag websites for human review (34–35). Existing GDI keyword lists contributed to the project's search seeds (55). Planned collection is purpose-limited, and consortium interface access depends on data-sharing agreements (65). The internal GDI policies mentioned at 69 are not reproduced. [D2.1 original](http[local research file]
\nThe document's internal acceptance/submission labels do not prove Commission acceptance. Query refinement is not publisher recourse. The NLP licensing guidance crossing pages 68–69 is introduced for NCSR-D/FBK; it must not be assigned to GDI as an executed licence. This is a design record, not a runtime audit.
\nIn its 20 April 2026 retrospective, GDI reports developing the decision-support system within the Monitoring Dashboard: combined detection signals, trained using consortium fact-checker annotations, surface risk patterns for human review. It also requests faster Article 40 vetting, mandatory platform API compatibility with approved third-party detection systems, and researcher audit access. [GDI retrospective, “What AI4TRUST delivered” and “Recommendations”](http[local research file]
\nThat is a meaningful reported research contribution → request for institutional access. The platform refusals described in the same article remain GDI's account: the applications and responses were not acquired. The article is not a legal instrument, award of access, or proof that a platform implemented the requested compatibility.
\nThe strongest actual knowledge-flow evidence runs from GDI's existing resources into funded methodology. The retrospective supplies a later, attributed output claim. No selected record identifies that output with DEL, the former Oracle segment, IAS's former input, or the prospective OHMD. Similar risk terminology does not establish shared code, training data, licences, recipients or commercial deployment. Nothing here establishes that FCDO commissioned these inputs or directed subsequent exclusions.
\nFor later integration, useful relations are the EU contribution recorded against the German participant; that participant's research role; GDI's pre-existing resource contribution to D2.1; the documented internal report review; the attributed Monitoring Dashboard contribution; and GDI's request for new access obligations. Keep the April 2024 grant match qualified, financial allocation separate from cash, and requested authority separate from an awarded right. No graph edits were made.
\nThe strongest rival explanation is ordinary consortium research using participants' prior expertise, followed by advocacy about research access. The present records support that account at least as well as a theory of publicly funded commercial exclusion infrastructure. The latter requires an additional documented transfer or use.
\nThe highest-value missing record is the AI4TRUST consortium agreement's GDI background/results and exploitation or access-rights schedules, or a specific licence tying its decision-support output to a named commercial product. Likely holders are the project coordinator and relevant GDI legal participant. D2.1 also points to data-management/platform specifications, but this run did not acquire another deliverable or treat generic grant rules as exercised powers. A grant accession/payment record would settle finer legal and money-state questions; a corporate record would settle the UK/German relationship. The named annotation codebook and final decision-support thresholds/validation are further precise records needed to test classification choices. The codebook and internal policies are expected records, not evidence of bias. These are reopening conditions, not claims that the documents do not exist.
\nOriginals: horizon-cordis.html, horizon-gdi2026.html, horizon-d21.pdf. horizon-cordis.full-dom.txt preserves header/date fields omitted by the narrower main-text extraction. The 90-page PDF was not read in full: text scope is PDF 1–5, 9–12, 34–35, 53–55, 60–65, 68–69. Images actually viewed: 2, 3, 11, 35, 55, 65. PDF 2 visually emphasizes Public, UCAM reviewer, Coordinator accepted and Ready for submission; its boxes are not filled checkmarks. No Commission approval is inferred. No keyword annex, live dataset or operational interface was used.
The PDF reader returned an Internal Error; one ordinary GET then retained the original successfully. CORDIS and GDI originals also returned 200. No further acquisition route was opened. Exact files, hashes, dates and source scopes are in horizon-captures.json; transport evidence remains in horizon-transport.json.
Closed acquisition: 9 September 2026. Three primary originals were retained. This packet concerns classification, distribution and recourse; Oak funding is handled separately in oak-gdi-support.md. It establishes an advertised GDI-to-Oracle product route in a 2024 guide, followed by an official Oracle service-end policy. It does not establish current delivery through Oracle, an advertiser's actual configuration, or a measured loss of publisher revenue.
Oracle's 2024 Contextual Intelligence Segment Guide describes segments that advertisers can use to target suitable contexts or avoid unsuitable ones. Its introduction says Oracle maintains its own contextual segments through an editorial team and describes the disinformation solution as powered by its GDI partnership. The decisive supplier row on PDF/printed page 129 identifies Disinformation protection from GDI, machine name gi_gdi_pfi. Oracle attributes the high-risk news-website ratings to GDI's combination of machine learning and expert human review. This supports GDI supplying a classification input and Oracle packaging it as an advertising segment; it does not show a particular advertiser selecting it, a bid being rejected, or a publisher's income changing.
The general branded-context introduction on PDF 128 describes third-party data used for targeting or avoiding matching content. The selected reading does not establish the executed licence, data fields, update cadence, downstream customer list or payment terms. Oracle's separate Floor-tier table on PDF 3 says that tier is blocked by default without a toggle. Its comparison to URL-based gv_pfi construction is not an explicit default rule for the different GDI segment gi_gdi_pfi. The guide therefore cannot support universal or non-overridable activation of the GDI segment. A PDF 10 text-layer row is ambiguous and was not used to assign the GDI identifier or a default.
Source: [Oracle guide](http[local research file] PDF 1–3, PDF 128 branded-context introduction, PDF 129 entire GDI table. Page 3 and page 129 were visually checked from retained original bytes. The document bears copyright 2024; metadata records creation 30 May and modification 3 June 2024, which are not independently established publication or deployment dates.
\nOracle's Advertising EOL Frequently Asked Questions, 12 August 2024, Version 3.0, explicitly includes Contextual Intelligence and supporting integrations in the services ending 30 September 2024. It says Oracle would take no new advertising agreements or renewals, end data-provider relationships, and terminate access even when an order extended beyond that date. This is a published provider policy, not evidence from a runtime test. It is sufficient to prevent presenting the historical guide as a current Oracle delivery channel.
\nThe same FAQ makes the residual boundary concrete. Rights to use previously supplied data depend on the product service description. Reporting and payment obligations may continue while data remains in use, and Oracle says payments for post-service data usage follow executed provider contracts. These are general provisions, not proof that GDI had a particular continuing licence, was paid after the end date, or that every stored classification was deleted. The FAQ itself says it is informational and does not form contractual obligations. The executed GDI/Oracle agreement and relevant service description remain unacquired.
\nSource: [Oracle EOL FAQ](http[local research file] PDF 1 date/version; PDF 2 product and end-date table; PDF 3 termination, prior-data and privacy questions; PDF 4 continuing reporting/provider-payment questions; PDF 5 legal qualification. All five pages were text-read; PDF 2–4 were visually checked.
\nThe [current About page](http[local research file] observed 9 September 2026, frames GDI's work as mapping revenues and producing evidence for policymakers and industry. Its definition extends beyond individual false statements: harmful, false or misleading adversarial narratives may target at-risk groups, current scientific consensus, democracy, civic institutions and governing bodies, including through contextual omissions. This is the current page's classificatory scope; the selected sources do not establish that the same criteria governed the historical DEL or Oracle segment. Its statement that it accepts no funding from media companies or technology platforms is also a current self-description. Historic licensing alone neither resolves nor establishes a contradiction without the relevant fee/payment definitions and dates. It describes the Online Harms Monetisation Dashboard in prospective language: it will estimate revenues across the open web and social platforms. This is a stated product direction, not proof of a deployed dashboard. Its approximately USD 200 million annual redirection figure is GDI's estimate attributed to its and partner organisations' work, not an independently tested outcome here. The page neither confirms current DEL distribution through another supplier nor establishes that all such distribution ended.
\nIts invitations to peers and academics to critique methodology are not an affected publisher's appeal instrument. The three selected records do not disclose a publisher notification entitlement, reasons/evidence procedure, correction or removal deadline, independent appeal, or propagation of corrected ratings to customers. Those instruments are unacquired, not proved nonexistent. Oracle's privacy opt-out/DSAR continuity during wind-down concerns a different process and cannot fill this gap.
\n| Direction | \nSupported relationship | \nStage / boundary | \n
|---|---|---|
| GDI → Oracle Contextual Intelligence | \nSupplies high-risk website ratings packaged in gi_gdi_pfi | \nOracle's 2024 product description; no executed licence or customer activation inspected | \n
| Oracle segment → advertisers | \nOffers a classification input for contextual targeting/avoidance | \nProduct capability; actual selection, default and restriction effects unproved for this segment | \n
| Oracle → advertising customers and integrations | \nAnnounces end of Contextual Intelligence service/access on 30 September 2024 | \nDated official provider policy; retained-data rights depend on unacquired instruments | \n
| Oracle → data providers | \nDescribes continuing payment calculation under executed contracts for post-EOL usage | \nGeneral policy only; no GDI-specific amount, payment or continuing use established | \n
| GDI → OHMD | \nDescribes a forthcoming revenue-estimation product | \nCurrent stated plan, not verified operational deployment | \n
The strongest rival to a present-tense exclusion-network account is documented channel retirement combined with an unresolved current distribution route. Conversely, the retirement notice cannot establish that historical signals had no continuing use. A high-value next record would be the historical GDI–Oracle licence/service description or a current distribution agreement specifying publisher review, update/deletion propagation and recipient discretion. No further acquisition is undertaken in this packet.
\nOriginals: mechanism-gdi-about.html (SHA-256 7e26f1638af8a92ea94193ca8d4011d6735dacf6f56716d7afd63a31fdbc1e1f); mechanism-oracle-guide2024.pdf (47b531a6967b09c9d9fd714cf209a7cbb49263686c1db2405399a71ffe6b86da); mechanism-oracle-eol2024.pdf (179f16108f5f783a79092c4b4835e4100557cd3a04fb0c96fe92bf5ee0ca5f2a). Full extraction of the 129-page guide was retained, but only the scoped passages above were substantively read. One large local display was truncated; it is not counted as a full read. Exact transport, derivative/image hashes and reading scopes are in mechanism-captures.json.
The indexed GDI 14 April 2022 Jounce announcement redirected to the current blog on ordinary web-reader opening; the historical article was not acquired or used to create a partnership edge. An indexed 2022 GDI policy PDF returned InternalError in the web reader; the parent replaced that planned third record with Oracle's consequential end-of-service policy, and that PDF route was closed without another attempt. No secondary controversy accounts, list entries, harmful-content endpoints or operational customer systems were read. The root's Oak article is a separate source and was not duplicated.
", "path": "gdi-resource-to-advertising-wave-2026-09-09/gdi-mechanism.md", "sourceIds": [ "w32-gdi-s-gdiabout", "w32-gdi-s-oracle", "w32-gdi-s-oracleend" ], "markdown": "Closed acquisition: 9 September 2026. Three primary originals were retained. This packet concerns classification, distribution and recourse; Oak funding is handled separately in `oak-gdi-support.md`. It establishes an advertised GDI-to-Oracle product route in a 2024 guide, followed by an official Oracle service-end policy. It does not establish current delivery through Oracle, an advertiser's actual configuration, or a measured loss of publisher revenue.\n\n## Exact product handoff\n\nOracle's 2024 *Contextual Intelligence Segment Guide* describes segments that advertisers can use to target suitable contexts or avoid unsuitable ones. Its introduction says Oracle maintains its own contextual segments through an editorial team and describes the disinformation solution as powered by its GDI partnership. The decisive supplier row on PDF/printed page 129 identifies **Disinformation protection from GDI**, machine name **`gi_gdi_pfi`**. Oracle attributes the high-risk news-website ratings to GDI's combination of machine learning and expert human review. This supports GDI supplying a classification input and Oracle packaging it as an advertising segment; it does not show a particular advertiser selecting it, a bid being rejected, or a publisher's income changing.\n\nThe general branded-context introduction on PDF 128 describes third-party data used for targeting or avoiding matching content. The selected reading does not establish the executed licence, data fields, update cadence, downstream customer list or payment terms. Oracle's separate Floor-tier table on PDF 3 says that tier is blocked by default without a toggle. Its comparison to URL-based `gv_pfi` construction is **not an explicit default rule for the different GDI segment `gi_gdi_pfi`**. The guide therefore cannot support universal or non-overridable activation of the GDI segment. A PDF 10 text-layer row is ambiguous and was not used to assign the GDI identifier or a default.\n\nSource: Oracle guide (see source references) PDF 1–3, PDF 128 branded-context introduction, PDF 129 entire GDI table. Page 3 and page 129 were visually checked from retained original bytes. The document bears copyright 2024; metadata records creation 30 May and modification 3 June 2024, which are not independently established publication or deployment dates.\n\n## Oracle's subsequent service-end policy changes the time claim\n\nOracle's *Advertising EOL Frequently Asked Questions*, **12 August 2024, Version 3.0**, explicitly includes Contextual Intelligence and supporting integrations in the services ending **30 September 2024**. It says Oracle would take no new advertising agreements or renewals, end data-provider relationships, and terminate access even when an order extended beyond that date. This is a published provider policy, not evidence from a runtime test. It is sufficient to prevent presenting the historical guide as a current Oracle delivery channel.\n\nThe same FAQ makes the residual boundary concrete. Rights to use previously supplied data depend on the product service description. Reporting and payment obligations may continue while data remains in use, and Oracle says payments for post-service data usage follow executed provider contracts. These are general provisions, not proof that GDI had a particular continuing licence, was paid after the end date, or that every stored classification was deleted. The FAQ itself says it is informational and does not form contractual obligations. The executed GDI/Oracle agreement and relevant service description remain unacquired.\n\nSource: Oracle EOL FAQ (see source references) PDF 1 date/version; PDF 2 product and end-date table; PDF 3 termination, prior-data and privacy questions; PDF 4 continuing reporting/provider-payment questions; PDF 5 legal qualification. All five pages were text-read; PDF 2–4 were visually checked.\n\n## What the current GDI page does—and does not—establish\n\nThe current About page (see source references) observed 9 September 2026, frames GDI's work as mapping revenues and producing evidence for policymakers and industry. Its definition extends beyond individual false statements: harmful, false or misleading adversarial narratives may target at-risk groups, current scientific consensus, democracy, civic institutions and governing bodies, including through contextual omissions. This is the current page's classificatory scope; the selected sources do not establish that the same criteria governed the historical DEL or Oracle segment. Its statement that it accepts no funding from media companies or technology platforms is also a current self-description. Historic licensing alone neither resolves nor establishes a contradiction without the relevant fee/payment definitions and dates. It describes the Online Harms Monetisation Dashboard in prospective language: it **will** estimate revenues across the open web and social platforms. This is a stated product direction, not proof of a deployed dashboard. Its approximately USD 200 million annual redirection figure is GDI's estimate attributed to its and partner organisations' work, not an independently tested outcome here. The page neither confirms current DEL distribution through another supplier nor establishes that all such distribution ended.\n\nIts invitations to peers and academics to critique methodology are not an affected publisher's appeal instrument. The three selected records do not disclose a publisher notification entitlement, reasons/evidence procedure, correction or removal deadline, independent appeal, or propagation of corrected ratings to customers. Those instruments are **unacquired**, not proved nonexistent. Oracle's privacy opt-out/DSAR continuity during wind-down concerns a different process and cannot fill this gap.\n\n## Candidate directional relationships and limits\n\n| Direction | Supported relationship | Stage / boundary |\n| --- | --- | --- |\n| GDI → Oracle Contextual Intelligence | Supplies high-risk website ratings packaged in `gi_gdi_pfi` | Oracle's 2024 product description; no executed licence or customer activation inspected |\n| Oracle segment → advertisers | Offers a classification input for contextual targeting/avoidance | Product capability; actual selection, default and restriction effects unproved for this segment |\n| Oracle → advertising customers and integrations | Announces end of Contextual Intelligence service/access on 30 September 2024 | Dated official provider policy; retained-data rights depend on unacquired instruments |\n| Oracle → data providers | Describes continuing payment calculation under executed contracts for post-EOL usage | General policy only; no GDI-specific amount, payment or continuing use established |\n| GDI → OHMD | Describes a forthcoming revenue-estimation product | Current stated plan, not verified operational deployment |\n\nThe strongest rival to a present-tense exclusion-network account is documented channel retirement combined with an unresolved current distribution route. Conversely, the retirement notice cannot establish that historical signals had no continuing use. A high-value next record would be the historical GDI–Oracle licence/service description or a current distribution agreement specifying publisher review, update/deletion propagation and recipient discretion. No further acquisition is undertaken in this packet.\n\n## Custody and bounded failures\n\nOriginals: `mechanism-gdi-about.html` (SHA-256 `7e26f1638af8a92ea94193ca8d4011d6735dacf6f56716d7afd63a31fdbc1e1f`); `mechanism-oracle-guide2024.pdf` (`47b531a6967b09c9d9fd714cf209a7cbb49263686c1db2405399a71ffe6b86da`); `mechanism-oracle-eol2024.pdf` (`179f16108f5f783a79092c4b4835e4100557cd3a04fb0c96fe92bf5ee0ca5f2a`). Full extraction of the 129-page guide was retained, but only the scoped passages above were substantively read. One large local display was truncated; it is not counted as a full read. Exact transport, derivative/image hashes and reading scopes are in `mechanism-captures.json`.\n\nThe indexed GDI 14 April 2022 Jounce announcement redirected to the current blog on ordinary web-reader opening; the historical article was not acquired or used to create a partnership edge. An indexed 2022 GDI policy PDF returned InternalError in the web reader; the parent replaced that planned third record with Oracle's consequential end-of-service policy, and that PDF route was closed without another attempt. No secondary controversy accounts, list entries, harmful-content endpoints or operational customer systems were read. The root's Oak article is a separate source and was not duplicated.\n", "file": "research/gdi-resource-to-advertising-wave-2026-09-09-gdi-mechanism.md" }, { "id": "gdi-resource-to-advertising-wave-2026-09-09-oak-gdi-support", "title": "Oak support for GDI: specified capacity and a commercial intervention", "description": "9 September 2026. Root donor acquisition closed. The product/counterpart lane is separate. This note establishes donor-reported awards and purpose, not an executed licence, traced payment or campaign instruction.", "html": "9 September 2026. Root donor acquisition closed. The product/counterpart lane is separate. This note establishes donor-reported awards and purpose, not an executed licence, traced payment or campaign instruction.
\nOak's 2023 Annual Report, PDF/printed26, lists Global Disinformation Index — USD903614 (3 years), described as core support to expand technological solutions addressing online disinformation. The report's PDF6 frames its lists as 2023 grant-making. It supplies no exact approval/start/end dates, instalments, contracting legal entity or GDI grant conditions. The three-year amount is not annual expenditure or cash received. PDF4 identifies Oak as a group of philanthropic organisations in multiple countries; the .dk hosting location does not identify a Danish legal grantor. [Existing donor-report citation](http[local research file]
\nThe same original page26 lists three already relevant rights organisations: National Council for Civil Liberties (Liberty), USD783133/3years core support for public/policymaker human-rights work; Open Rights Group, USD437349/3years to support refugee/migrant-sector opposition to data collection/sharing and technologies driving immigration controls; and Privacy International, USD1138554/3years for privacy protections and safeguards against surveillance/data exploitation. These earlier awards strengthen the common-funder/divergent-intervention evidence. None is a verified earmark for the later TCN complaint or Statewatch/ORG letter. Do not calculate the litigation budget from these amounts or claim the June2026 DFF grant was the only possible prior source of resources.
\nOak's [3 July2024 programme story](http[local research file] repeats the report's GDI account: it explicitly supports reducing the commercial incentives of disinformation and describes the Dynamic Exclusion List being licensed to agencies, brand-safety firms and other businesses. It portrays the product as an input to advertisers' choices. Its descriptions of neutrality/independence are Oak's descriptions, not independently established ratings of GDI. The article attributes an estimated USD200million annual revenue reduction and an80percent fall to co-founder Clare Melford, who expressly disclaims sole credit. No underlying measurement, counterfactual or transactions were acquired. The surrounding coverage/list figures are historical, including a June2023 assessment date; none is a verified September2026 operating statistic.
\nThis establishes a deliberate donor-supported capacity aimed at the revenue channel. It does not show Oak selecting websites, approving classifier rules, ordering a particular advertiser to stop spending, or commanding a government. The counterpart/product records are needed to show how the classifications become selectable restrictions, and whether a historical delivery route still operates. A common donor can support efforts to constrain commercial misinformation and efforts to limit surveillance without holding a single uniform position on every technology or legal power. That observation does not establish universal independence from grant conditions.
\nThe .org report URL found in search returned404 in one ordinary GET; its error body is retained and the route closed. Existing case source src-4a4bb675ee5f supplied the actual .dk URL, acquired with one successful GET: oak-annual2023-dk.pdf, SHA256 bc576201d168bc53091be98e1b227a714bbff5bdf0cd5239f2bbd205867a51ac. Root read PDF1,4,6,24–26 as extracted text and visually checked the full grant page26. Other pages were extracted but not silently classified as read. This refreshes an existing source with new reading scope; prior pp4–5/p40 context must be preserved in later integration.
\nThe July2024 article original is oak-information2024.html, SHAbbc34e8233ca5c69946da666075b1ccd8f6399132c354c1c44044a54aaec5f4c. Root read its substantive official-reader body. The article and annual report repeat the same narrative and attributed GDI estimates; they are not two independent outcome confirmations. Their references to IPIE and Foxglove are context, not an opened research branch or independently verified litigation/UNESCO findings.
\nA finite exact-term search of retained Oak2025 report text found no GDI match; this does not prove funding or the organisation ceased. A whole-case label/source/lead search found no exact GDI record before this run. Initial substring matching for IPIE falsely matched recipient; a corrected word-boundary search removed those irrelevant results. No conclusions were drawn from them. Original transports, error custody, page render and hashes are in donor-captures.json. No account, outreach, graph, visible UI or packaging changes.
", "path": "gdi-resource-to-advertising-wave-2026-09-09/oak-gdi-support.md", "sourceIds": [ "src-4a4bb675ee5f", "w32-gdi-s-oakstory" ], "markdown": "9 September 2026. Root donor acquisition closed. The product/counterpart lane is separate. This note establishes donor-reported awards and purpose, not an executed licence, traced payment or campaign instruction.\n\nOak's 2023 Annual Report, PDF/printed26, lists **Global Disinformation Index — USD903614 (3 years)**, described as core support to expand technological solutions addressing online disinformation. The report's PDF6 frames its lists as 2023 grant-making. It supplies no exact approval/start/end dates, instalments, contracting legal entity or GDI grant conditions. The three-year amount is not annual expenditure or cash received. PDF4 identifies Oak as a group of philanthropic organisations in multiple countries; the .dk hosting location does not identify a Danish legal grantor. Existing donor-report citation (see source references)\n\nThe same original page26 lists three already relevant rights organisations: National Council for Civil Liberties (Liberty), USD783133/3years core support for public/policymaker human-rights work; Open Rights Group, USD437349/3years to support refugee/migrant-sector opposition to data collection/sharing and technologies driving immigration controls; and Privacy International, USD1138554/3years for privacy protections and safeguards against surveillance/data exploitation. These earlier awards strengthen the common-funder/divergent-intervention evidence. None is a verified earmark for the later TCN complaint or Statewatch/ORG letter. Do not calculate the litigation budget from these amounts or claim the June2026 DFF grant was the only possible prior source of resources.\n\nOak's 3 July2024 programme story (see source references) repeats the report's GDI account: it explicitly supports reducing the commercial incentives of disinformation and describes the Dynamic Exclusion List being licensed to agencies, brand-safety firms and other businesses. It portrays the product as an input to advertisers' choices. Its descriptions of neutrality/independence are Oak's descriptions, not independently established ratings of GDI. The article attributes an estimated USD200million annual revenue reduction and an80percent fall to co-founder Clare Melford, who expressly disclaims sole credit. No underlying measurement, counterfactual or transactions were acquired. The surrounding coverage/list figures are historical, including a June2023 assessment date; none is a verified September2026 operating statistic.\n\nThis establishes a deliberate donor-supported capacity aimed at the revenue channel. It does not show Oak selecting websites, approving classifier rules, ordering a particular advertiser to stop spending, or commanding a government. The counterpart/product records are needed to show how the classifications become selectable restrictions, and whether a historical delivery route still operates. A common donor can support efforts to constrain commercial misinformation and efforts to limit surveillance without holding a single uniform position on every technology or legal power. That observation does not establish universal independence from grant conditions.\n\n## Custody and bounded reading\n\nThe .org report URL found in search returned404 in one ordinary GET; its error body is retained and the route closed. Existing case source src-4a4bb675ee5f supplied the actual .dk URL, acquired with one successful GET: oak-annual2023-dk.pdf, SHA256 bc576201d168bc53091be98e1b227a714bbff5bdf0cd5239f2bbd205867a51ac. Root read PDF1,4,6,24–26 as extracted text and visually checked the full grant page26. Other pages were extracted but not silently classified as read. This refreshes an existing source with new reading scope; prior pp4–5/p40 context must be preserved in later integration.\n\nThe July2024 article original is oak-information2024.html, SHAbbc34e8233ca5c69946da666075b1ccd8f6399132c354c1c44044a54aaec5f4c. Root read its substantive official-reader body. The article and annual report repeat the same narrative and attributed GDI estimates; they are not two independent outcome confirmations. Their references to IPIE and Foxglove are context, not an opened research branch or independently verified litigation/UNESCO findings.\n\nA finite exact-term search of retained Oak2025 report text found no GDI match; this does not prove funding or the organisation ceased. A whole-case label/source/lead search found no exact GDI record before this run. Initial substring matching for IPIE falsely matched recipient; a corrected word-boundary search removed those irrelevant results. No conclusions were drawn from them. Original transports, error custody, page render and hashes are in donor-captures.json. No account, outreach, graph, visible UI or packaging changes.\n", "file": "research/gdi-resource-to-advertising-wave-2026-09-09-oak-gdi-support.md" }, { "id": "governance-wave-2026-09-09-mannifera-selection", "title": "Mannifera selection, fiscal authority and the HRLC reform output", "description": "Observed 9 September 2026. Bounded continuation selected by the latest forest review. This adds a published selection mechanism, fiscal-trustee authority, one member's disclosed financial participation, and an identifiable Mannifera-funded HRLC policy output. It does not establish how any individual voted on the HRLC grant or that its promised common expert ", "html": "Observed 9 September 2026. Bounded continuation selected by the latest forest review. This adds a published selection mechanism, fiscal-trustee authority, one member's disclosed financial participation, and an identifiable Mannifera-funded HRLC policy output. It does not establish how any individual voted on the HRLC grant or that its promised common expert position was formally agreed.
\nThe distinction that matters is between three decisions: choosing a project for collective support, authorising a charitable grant from the hosted pool, and endorsing particular policy recommendations. The documents allocate the first two functions; the report and subsequent advocacy show policy production, while collective endorsement remains unresolved.
\nMannifera's Membership Guide places membership contributions into an Australian Communities Foundation (ACF) sub-fund and pooled fund. It describes a formal enabling partnership with Philanthropy Australia, ACF and Reichstein Foundation, and a member Taskforce plus Executive Officer. Members commit to at least $65,000 annually for three years. This is a published membership model, not a receipt showing every member paid that minimum in 2024. [Membership Guide](http[local research file] physical PDF pp7–8, printed governance/joining spreads; relevant passages read in the preceding funding pass and reused here. No publication date is inferred from its URL.
\nThe Granting Guide provides the actual staged process. The Executive Officer handles enquiries and assesses proposals; up to six collective members recommend a shortlist; the Taskforce endorses that shortlist; voting members decide by consensus at the annual Muster. Interests are recorded and assessed for conflicts or management. Criteria include strategic fit, a credible theory of change, political opportunity, an appropriate team/coalition and charity eligibility. It explicitly supports legal reform, inside-track advocacy and coalition building. However, the guide contains a 2023 timetable, despite its 2024 URL path. It establishes published historical procedure, not proof that every step was followed for the special 2024 round. It does not specify an automatic recusal rule or document anyone's particular vote. [Granting Guide](http[local research file] PDF pp6–8, particularly printed pp10–11, stages 1–5 and 2023 timeline; pp8–9 assessment criteria.
\nThe current approach page calls the governing body a rotating member Board alongside the Executive Officer and says members have equal opportunity to influence strategy and grants. That terminology differs from the older Taskforce guide; it is not, by itself, proof of a newly incorporated legal entity or a precise governance-change date. Nor does equal opportunity alone establish equal votes. [Current governance description](http[local research file] Governance paragraph, rendered lines 81–83.
\nACF's current FAQ identifies Australian Communities Foundation Limited as trustee. Named funds are management accounts rather than separate legal entities; donors request grants, but the trustee must approve them. The FAQ describes a practice of actioning requests unless a legal obstacle exists, while explicitly declining to guarantee donor directions. This shows practical donor influence alongside retained legal responsibility. [ACF FAQ](http[local research file] questions on named funds, trustee identity and donor choice, lines 78–95, 118–119, 153–155.
\nThe public ACF establishment template identifies the trustee's ABN 20 077 830 347. Section 12 says gifts join the trust assets; the trustee need not follow preferences and may grant without asking first. Advisory committees can only recommend grants to ACF's board. These are substantial retained powers in a public template, not an acquired Mannifera-signed agreement. The exact underlying ACF trust and any negotiated Mannifera arrangements remain unidentified. [Sub-fund establishment form](http[local research file] PDF p1, p3 section 5(iii), p8 section 12. The dated URL is not treated as a verified execution or publication date.
\nOne concrete member-money example is The Myer Foundation. Its FY24 report lists a three-year $165,000 Mannifera membership award, approved FY2022 and labelled membership 2021–2024, with a dash under commitment from FY24. Separate FY24 entries are $5,000 general operating support in Directors' Grants and $9,100 general operating support through its Family Grants Program. These are recorded values/commitments, not demonstrated cash payments; the old membership award is not a new $165,000 FY24 commitment. The report names Jemima Myer as the Foundation's Mannifera representative and a Mannifera board member. No HRLC earmark or individual vote follows. Currency is printed as dollars in Australian context, without a currency code. [The Myer Foundation FY24 report](http[local research file] PDF p4 Directors' Grants, p8 Family Grants, p9 representation biography. Only these scoped entries were examined; a separate official annual-reports landing-page route failed and was not retried.
\nMannifera's Executive Officer describes a dedicated open round following the 2023 Voice referendum, selecting seven organisations and providing $370,000 with an unnamed outside co-funder. It names $60,000 awarded to HRLC to align legal/technical experts and equip advocates with a shared reform position. Reset's $40,000 was for transparency advocacy, with HRLC and Digital Rights Watch identified as partners. This establishes a deliberately complementary programme rather than a connection inferred solely from shared donors. The unnamed co-funder's amount, payment date and influence are not disclosed; no share is assigned to Myer or any other member. [Executive Officer's round account](http[local research file] lines 13–16 and 27–32. The body supplies context but no visible precise publication date; award amounts are not bank receipts.
\nThe standing partner record labels HRLC's award a 2024 Signature Grant, connects it to misinformation, Online Safety Act and privacy reform opportunities, and distinguishes older grants. That supports a 2024 award-purpose relation, not the assertion that a particular unpublished expert agreement was completed. [HRLC partner entry](http[local research file] Human Rights Law Centre—Disinformation section.
\nRights-First: Principles for Digital Platform Regulation is authored by David Mejia-Canales and dated 4 October 2024. It expressly acknowledges a Mannifera grant. Named assistance includes Jackie Turner and the Australian Conservation Foundation, which is distinct from the fiscal host Australian Communities Foundation. Its main proposals combine a legal platform duty with human-rights constraints, limited removal powers, user data control and court oversight. Crucially, it attributes views to the authors and says they need not reflect HRLC, affiliates or supporters. It is a funded policy output, not an endorsement register. The acknowledgment does not identify the grant number, amount or budget line; assigning the whole $60,000 award to this report would go beyond it. [Rights-First PDF](http[local research file] physical PDF pp2–3, printed pp2–5, acknowledgment/author statement and executive summary. No substantive audit of all legal reasoning was undertaken.
\nThe disclaimer does not erase HRLC's subsequent institutional advocacy. Its public launch page, labelled 7 October 2024, calls for an independent rule-setting regulator, legislated duty, rights assessment and substantial penalties. It thus provides a recipient-level policy position. It still does not turn other consultees, the donor or every coalition partner into signatories. [HRLC launch](http[local research file] dated heading and recommendations, read through the indexed primary page.
\nAn HRLC update labelled 8 October 2024 says Mejia-Canales represented it at Parliament House and met with the Albanese Government. Its URL contains a different 2025 date string, so the displayed date and URL must not be silently conflated. The account establishes HRLC's report of engagement, not an independently acquired meeting minute or government acceptance of each recommendation. [HRLC parliamentary-engagement update](http[local research file] lines 40–49.
\nHRLC's 2024–25 annual report says the publication was followed by parliamentary lobbying and claims its intervention helped secure the November 2024 duty commitment. This is participant attribution, not independent causal proof. Its stated preference for a duty over blanket youth social-media bans also limits any claim that this advocacy endorsed every age restriction. The donor section lists Mannifera among donations/grant payments received during 1 July 2024–30 June 2025; the visual amount-band mapping was not verified here, so no exact receipt is inferred. [HRLC Annual Report 2024–25](http[local research file] p38 and p66. The amount, period and restriction of the 2024 grant still need reconciliation.
\nMannifera's 2025 impact account names HRLC, Australian Democracy Network and Reset as misinformation-bill advocates, reports collective advocacy, and describes interest in a digital duty. It also admits the misinformation bill did not proceed. Its claims of securing amendments and increasing parliamentary support are retrospective funder/partner assessments; no amendment number or independent counterfactual is supplied in the passage. The report further describes networking, communications and convening support. None identifies a signed common platform spanning every funded group. [Mannifera Impact Report 2025](http[local research file] PDF p5/printed pp6–7 and PDF p8/printed p12. The strategy-coordination passage on the following physical page concerns another funding area and is not transferred into this digital-reform claim.
\nThe supported chain is: a member contributes or commits to the pool; Mannifera's published process selects aligned projects; ACF retains grant authority under its general arrangements; an HRLC publication credits Mannifera and articulates a regulatory position; HRLC reports taking that work to Parliament. The actual 2024 selection file, grant contract and coalition endorsement record are separate missing links. No single member's control of the project, common donor instruction over report text, or government adoption of all recommendations has been established.
\nThe narrow next records are the special-round assessment/shortlist and conflict-management record; the Mannifera–ACF executed arrangement and HRLC approval/grant letter; and HRLC's expert-convening agenda, participant list, agreed statement/version history and grant-completion report. Those would distinguish a jointly endorsed reform position from a named author's funded framework and later compatible advocacy. Likely custodians are Mannifera/ACF for selection and grants, and HRLC for the convening and output. No request has been sent.
\nThis pass stops at that decision breakpoint for the root forest review. Public pages and scoped reader extracts were used; no original PDFs or hashes are retained in this single-file lane. Search results about other causes, member biographies and unrelated coalitions were not treated as this project's evidence. Butterfield, ACNC, account access and outreach remain outside this run.
", "path": "governance-wave-2026-09-09/mannifera-selection.md", "sourceIds": [ "w12-gov-s-mann-granting", "w12-gov-s-acf-faq", "w12-gov-s-acf-template", "w12-gov-s-myer2024", "w12-gov-s-mann-round", "w12-gov-s-rights-first", "w12-gov-s-hrlc-launch", "w12-gov-s-hrlc-parliament" ], "markdown": "Observed 9 September 2026. Bounded continuation selected by the latest forest review. This adds a published selection mechanism, fiscal-trustee authority, one member's disclosed financial participation, and an identifiable Mannifera-funded HRLC policy output. It does not establish how any individual voted on the HRLC grant or that its promised common expert position was formally agreed.\n\nThe distinction that matters is between three decisions: choosing a project for collective support, authorising a charitable grant from the hosted pool, and endorsing particular policy recommendations. The documents allocate the first two functions; the report and subsequent advocacy show policy production, while collective endorsement remains unresolved.\n\n## Selection and money custody\n\nMannifera's Membership Guide places membership contributions into an Australian Communities Foundation (ACF) sub-fund and pooled fund. It describes a formal enabling partnership with Philanthropy Australia, ACF and Reichstein Foundation, and a member Taskforce plus Executive Officer. Members commit to at least $65,000 annually for three years. This is a published membership model, not a receipt showing every member paid that minimum in 2024. Membership Guide (see source references) physical PDF pp7–8, printed governance/joining spreads; relevant passages read in the preceding funding pass and reused here. No publication date is inferred from its URL.\n\nThe Granting Guide provides the actual staged process. The Executive Officer handles enquiries and assesses proposals; up to six collective members recommend a shortlist; the Taskforce endorses that shortlist; voting members decide by consensus at the annual Muster. Interests are recorded and assessed for conflicts or management. Criteria include strategic fit, a credible theory of change, political opportunity, an appropriate team/coalition and charity eligibility. It explicitly supports legal reform, inside-track advocacy and coalition building. However, the guide contains a **2023 timetable**, despite its 2024 URL path. It establishes published historical procedure, not proof that every step was followed for the special 2024 round. It does not specify an automatic recusal rule or document anyone's particular vote. Granting Guide (see source references) PDF pp6–8, particularly printed pp10–11, stages 1–5 and 2023 timeline; pp8–9 assessment criteria.\n\nThe current approach page calls the governing body a rotating member **Board** alongside the Executive Officer and says members have equal opportunity to influence strategy and grants. That terminology differs from the older Taskforce guide; it is not, by itself, proof of a newly incorporated legal entity or a precise governance-change date. Nor does equal opportunity alone establish equal votes. Current governance description (see source references) Governance paragraph, rendered lines 81–83.\n\nACF's current FAQ identifies **Australian Communities Foundation Limited** as trustee. Named funds are management accounts rather than separate legal entities; donors request grants, but the trustee must approve them. The FAQ describes a practice of actioning requests unless a legal obstacle exists, while explicitly declining to guarantee donor directions. This shows practical donor influence alongside retained legal responsibility. ACF FAQ (see source references) questions on named funds, trustee identity and donor choice, lines 78–95, 118–119, 153–155.\n\nThe public ACF establishment template identifies the trustee's **ABN 20 077 830 347**. Section 12 says gifts join the trust assets; the trustee need not follow preferences and may grant without asking first. Advisory committees can only recommend grants to ACF's board. These are substantial retained powers in a public template, not an acquired Mannifera-signed agreement. The exact underlying ACF trust and any negotiated Mannifera arrangements remain unidentified. Sub-fund establishment form (see source references) PDF p1, p3 section 5(iii), p8 section 12. The dated URL is not treated as a verified execution or publication date.\n\nOne concrete member-money example is **The Myer Foundation**. Its FY24 report lists a three-year **$165,000** Mannifera membership award, approved FY2022 and labelled membership 2021–2024, with a dash under commitment from FY24. Separate FY24 entries are **$5,000** general operating support in Directors' Grants and **$9,100** general operating support through its Family Grants Program. These are recorded values/commitments, not demonstrated cash payments; the old membership award is not a new $165,000 FY24 commitment. The report names Jemima Myer as the Foundation's Mannifera representative and a Mannifera board member. No HRLC earmark or individual vote follows. Currency is printed as dollars in Australian context, without a currency code. The Myer Foundation FY24 report (see source references) PDF p4 Directors' Grants, p8 Family Grants, p9 representation biography. Only these scoped entries were examined; a separate official annual-reports landing-page route failed and was not retried.\n\n## The 2024 round and HRLC's intended role\n\nMannifera's Executive Officer describes a dedicated open round following the 2023 Voice referendum, selecting seven organisations and providing **$370,000** with an unnamed outside co-funder. It names **$60,000 awarded to HRLC** to align legal/technical experts and equip advocates with a shared reform position. Reset's **$40,000** was for transparency advocacy, with HRLC and Digital Rights Watch identified as partners. This establishes a deliberately complementary programme rather than a connection inferred solely from shared donors. The unnamed co-funder's amount, payment date and influence are not disclosed; no share is assigned to Myer or any other member. Executive Officer's round account (see source references) lines 13–16 and 27–32. The body supplies context but no visible precise publication date; award amounts are not bank receipts.\n\nThe standing partner record labels HRLC's award a **2024 Signature Grant**, connects it to misinformation, Online Safety Act and privacy reform opportunities, and distinguishes older grants. That supports a 2024 award-purpose relation, not the assertion that a particular unpublished expert agreement was completed. HRLC partner entry (see source references) Human Rights Law Centre—Disinformation section.\n\n## What is actually identifiable as an output\n\n**Rights-First: Principles for Digital Platform Regulation** is authored by David Mejia-Canales and dated **4 October 2024**. It expressly acknowledges a Mannifera grant. Named assistance includes Jackie Turner and the **Australian Conservation Foundation**, which is distinct from the fiscal host **Australian Communities Foundation**. Its main proposals combine a legal platform duty with human-rights constraints, limited removal powers, user data control and court oversight. Crucially, it attributes views to the authors and says they need not reflect HRLC, affiliates or supporters. It is a funded policy output, not an endorsement register. The acknowledgment does not identify the grant number, amount or budget line; assigning the whole $60,000 award to this report would go beyond it. Rights-First PDF (see source references) physical PDF pp2–3, printed pp2–5, acknowledgment/author statement and executive summary. No substantive audit of all legal reasoning was undertaken.\n\nThe disclaimer does not erase HRLC's subsequent institutional advocacy. Its public launch page, labelled **7 October 2024**, calls for an independent rule-setting regulator, legislated duty, rights assessment and substantial penalties. It thus provides a recipient-level policy position. It still does not turn other consultees, the donor or every coalition partner into signatories. HRLC launch (see source references) dated heading and recommendations, read through the indexed primary page.\n\nAn HRLC update labelled **8 October 2024** says Mejia-Canales represented it at Parliament House and met with the Albanese Government. Its URL contains a different 2025 date string, so the displayed date and URL must not be silently conflated. The account establishes HRLC's report of engagement, not an independently acquired meeting minute or government acceptance of each recommendation. HRLC parliamentary-engagement update (see source references) lines 40–49.\n\nHRLC's 2024–25 annual report says the publication was followed by parliamentary lobbying and claims its intervention helped secure the November 2024 duty commitment. This is participant attribution, not independent causal proof. Its stated preference for a duty over blanket youth social-media bans also limits any claim that this advocacy endorsed every age restriction. The donor section lists Mannifera among donations/grant payments received during 1 July 2024–30 June 2025; the visual amount-band mapping was not verified here, so no exact receipt is inferred. HRLC Annual Report 2024–25 (see source references) p38 and p66. The amount, period and restriction of the 2024 grant still need reconciliation.\n\nMannifera's 2025 impact account names HRLC, Australian Democracy Network and Reset as misinformation-bill advocates, reports collective advocacy, and describes interest in a digital duty. It also admits the misinformation bill did not proceed. Its claims of securing amendments and increasing parliamentary support are retrospective funder/partner assessments; no amendment number or independent counterfactual is supplied in the passage. The report further describes networking, communications and convening support. None identifies a signed common platform spanning every funded group. Mannifera Impact Report 2025 (see source references) PDF p5/printed pp6–7 and PDF p8/printed p12. The strategy-coordination passage on the following physical page concerns another funding area and is not transferred into this digital-reform claim.\n\n## Record-to-decision conclusion\n\nThe supported chain is: a member contributes or commits to the pool; Mannifera's published process selects aligned projects; ACF retains grant authority under its general arrangements; an HRLC publication credits Mannifera and articulates a regulatory position; HRLC reports taking that work to Parliament. The actual 2024 selection file, grant contract and coalition endorsement record are separate missing links. No single member's control of the project, common donor instruction over report text, or government adoption of all recommendations has been established.\n\nThe narrow next records are the special-round assessment/shortlist and conflict-management record; the Mannifera–ACF executed arrangement and HRLC approval/grant letter; and HRLC's expert-convening agenda, participant list, agreed statement/version history and grant-completion report. Those would distinguish a jointly endorsed reform position from a named author's funded framework and later compatible advocacy. Likely custodians are Mannifera/ACF for selection and grants, and HRLC for the convening and output. No request has been sent.\n\nThis pass stops at that decision breakpoint for the root forest review. Public pages and scoped reader extracts were used; no original PDFs or hashes are retained in this single-file lane. Search results about other causes, member biographies and unrelated coalitions were not treated as this project's evidence. Butterfield, ACNC, account access and outreach remain outside this run.\n", "file": "research/governance-wave-2026-09-09-mannifera-selection.md" }, { "id": "governance-wave-2026-09-09-mckinnon-governance", "title": "McKinnon's commissioning governance", "description": "Observed 9 September 2026. Bounded public-source follow-through from funding-wave-2026-09-09/reset-funding.md and the latest forest review. No canonical case changes, outreach or ACNC access.", "html": "Observed 9 September 2026. Bounded public-source follow-through from funding-wave-2026-09-09/reset-funding.md and the latest forest review. No canonical case changes, outreach or ACNC access.
The new finding is an identifiable internal research-governance mechanism, alongside an important entity transition. The historical Research Centre reports Research Committee oversight of research spending and publication/dissemination activities. McKinnon separately identifies its Reset commissions. The public records read here do not join those two facts into a Reset-specific committee decision, contracting entity or editorial approval right.
\nThe name McKinnon covers several entities. Its current funding page lists the Foundation trust, Research Centre and Charitable Foundation separately. The page says Grant Rule and Sophie Oh fund the organisation and describes related investment and administration arrangements. That is a group disclosure, not the execution page of either Reset contract. [Funding and Financials](http[local research file] “Our Funding” and “Our entities”; source gov-mck-s-funding, read.
| Entity and identifier | \nWhat the primary record establishes | \nConsequence for the commissions | \n
|---|---|---|
| The Trustee for Susan McKinnon Foundation, ABN 28 739 100 711 | \nABR identifies a discretionary investment trust, active from 21 April 2015, with private ancillary fund status. [ABR record](http[local research file] entity and DGR fields; gov-mck-s-abr-trust, read. | \nThis identifies the trust, not its corporate trustee's exact legal name or the contract signatory. | \n
| Susan McKinnon Research Centre Ltd, ABN 40 656 129 127 / ACN 656 129 127 | \nABR history carries this name from 16 December 2021 to 14 April 2026. The same identifier now belongs to OH-RULE FAMILY PHILANTHROPY LTD. [History](http[local research file] entity-name rows, and [current record](http[local research file] gov-mck-s-abr-research, read. | \nPreserve historical name and current alias on the same entity. Renaming does not establish contract assignment or a completed wind-up. | \n
| Susan McKinnon Charitable Foundation Ltd, ABN 12 653 756 597 | \nABR records an Australian public company active from 17 September 2021. [ABR record](http[local research file] entity fields; gov-mck-s-abr-charitable, read. | \nIts existence during 2023/24 does not show it commissioned these reports. The September 2025 privacy policy is not a historical contract. | \n
McKinnon's consolidated 2025 program report says the Charitable Foundation became the group's primary charitable vehicle during 2025, following specific listing. Its financial year ends on 31 December. This is a dated operating transition, not simply a new public brand. The cover is dated 18 February 2026, while the website link says published 17 February 2026; retain both labels rather than silently choosing one. [Program Expenditure Report 2025](http[local research file] PDF pp1–2; gov-mck-s-program2025, read. The wider donation tables were inspected but are not used to assign a Reset payment.
The trust's FY2024 directors' report names Grant Rule and Michael Schoenfeld as the trustee's directors throughout the reporting year and to the report date, unless otherwise stated. Its stated activity is grants aligned with policy and governance change. Note 12 identifies the Research Centre as delivering programs towards the trust's goals and providing grant-program marketing/strategic advisory services at cost. It reports a $19,530,300 donation to the Research Centre in FY2024. This is intra-group capacity, not a Reset fee or a new amount to add to downstream spending. [Trust FY2024 financial report](http[local research file] PDF p3/printed1 and PDF pp16–17/printed14–15; gov-mck-s-trust2024, read in relevant scope.
The Research Centre's FY2024 report names Colin Galbraith, Grant Rule and Michael Schoenfeld as directors. It expressly places application of research funds and activities connected with publishing and disseminating research under Research Committee oversight. It reports one member at 30 June 2024, without naming that member. Its Note 11 reports the same FY2024 intra-group donation. The accounts identify an accountable institutional route; they do not disclose committee appointments, voting thresholds, delegated limits or a Reset decision. [Research Centre FY2024 report](http[local research file] PDF pp3–4/printed1–2, PDF p17/printed15 and PDF p18/printed16; gov-mck-s-research2024, read in relevant scope.
The FY2023 report already names the same directors, Research Committee oversight and one member. Its adjacent history statements are internally awkward: printed1 describes funding in June 2023 and commencement in July 2023; printed2 describes the first operational year after funding in FY2022. Do not use either sentence to fix a confident commissioning chronology. [Research Centre FY2023 report](http[local research file] PDF pp3–4/printed1–2; gov-mck-s-research2023, read in that scope.
Today's general website board is a separate, later observation. It lists John Thwaites as interim executive chair and describes Acting CEO Leigh O'Neill as leading strategy, operations and programs; Tom Mooney leads the democracy work. These roles are not evidence of who approved the early-2023 Reset approach. About McKinnon.
\nMcKinnon's current account gives the commission's broad purpose: evaluate Australia's regulatory system and recommend improvements. It names the May 2024 Functioning or Failing? and June 2024 Achieving Digital Platform Public Transparency in Australia, commissioned in 2023 and 2024. It links readers to both outputs. That is a public report-use route, but not a contractual publication-clearance provision. [Commissioning account](http[local research file] Reset section; gov-mck-s-commissions, read. Reuse the existing canonical source/commission edges if already present.
The May report names Rys Farthing and Alice Dawkins under authors/contributors. Its introduction says the Foundation approached Reset early in 2023 to design tests of the misinformation code; Reset attributes the research design to itself and methods from its global network. Its recommendations include considering a duty of care, risk assessments and researcher access. The report thus attributes question initiation to the funder and methodological work to the recipient. Neither attribution resolves who cleared publication or mandated conclusions. [Functioning or Failing?](http[local research file] PDF pp2–3 and5–6; gov-mck-s-functioning2024, read in relevant scope. The experiments' validity was not independently audited.
The June report's first three pages confirm its date and EU/UK comparison and recommend risk assessments, prescribed transparency reports, independent audits, data portals and researcher access. Those opening pages do not disclose an executed scope or approval arrangement. [Achieving Digital Platform Public Transparency in Australia](http[local research file] PDF pp1–3; gov-mck-s-transparency2024, read in that scope. PDF pp30–31 were also inspected for end matter. The funder logo is not legible in the extracted text; commissioning attribution rests on McKinnon's named account, not a visually unread logo.
A separate, more revealing process record is the April 2024 briefing Regulating for Transparency. It describes a March 2024 roundtable of 20 academic/civil-society experts, held under the Chatham House Rule, and translates discussion into regulatory recommendations. It acknowledges McKinnon support, assigns errors and omissions to Reset and explicitly separates attendance from endorsement. Its acknowledgements include HRLC contributors and a named independent consultant, among others; these are bounded contributions, not evidence all participants approved the final text. [Original briefing](http[local research file] PDF pp1,4,14–15/printed same; gov-mck-s-roundtable2024, read in relevant scope. This supports an actual convening-to-output path. Do not automatically label it one of the two commissions or assert a separate payment.
| Candidate relation | \nType / evidence | \nBoundary | \n
|---|---|---|
| Foundation trust → Research Centre: reports FY2024 intra-group donation | \nFinance; gov-mck-s-trust2024, gov-mck-s-research2024 | \n$19,530,300, dollar label; Australian context, currency code not printed in cited passages. One transaction reported by both sides, not two awards or Reset money. | \n
| Research Committee → Research Centre research spending and dissemination activities: oversees | \nGovernance; gov-mck-s-research2023, gov-mck-s-research2024 | \nHistorical general mandate; no Reset-specific decision or final editorial veto established. | \n
| Rule/Schoenfeld → trustee board; Galbraith/Rule/Schoenfeld → Research Centre board: named directors | \nGovernance; respective FY2024 reports | \nHistorical entity-specific roles; no disclosed member-appointment or dismissal right. | \n
| Research Centre → OH-RULE FAMILY PHILANTHROPY LTD: later registered name | \nIdentity, dated 14 April 2026; gov-mck-s-abr-research | \nSame ABN/ACN; not an inferred ownership transfer. History extraction renders the first character as “0”; current record renders “O”. | \n
| McKinnon → Reset May/June reports: commissions named evaluations and recommendations | \nCommission; gov-mck-s-commissions, supported by recipient May introduction | \nGroup identity and purpose established; exact contracting company, fee and approval rights open. | \n
| McKinnon → April briefing: acknowledged support; March roundtable → briefing: informs | \nSupport / convening-to-output; gov-mck-s-roundtable2024 | \nSupport amount/form unknown. Participant attendance is expressly not endorsement. | \n
The strongest supported explanation is financed agenda-setting with identifiable research production and publication responsibilities. A compatible rival is that the funder selected an already aligned specialist, leaving methods and conclusions with Reset. The records do not distinguish that from exercised donor editorial direction. Research Committee oversight is a useful new institutional fact, not the missing evidence of a committee approving the particular reports.
\nThe next discriminating records are the 2023/24 executed Reset engagement letters and scopes, identifying legal counterparties, fee, milestones, publication rights and approval/termination terms; Research Committee terms of reference, membership/appointment resolutions and relevant minutes; and the single-member register and constitutions in force when those decisions were made. Custodians are McKinnon's historical entities/current records holder and Reset. Appointment resolutions would distinguish founders' public sponsorship of appointments from their actual legal power. Contract allocation records would determine whether the Research Centre route applies to these commissions. None was obtained, and no contact was made.
\nAll URLs above are primary issuer/recipient/ABR records. Access state is read only for the stated passages; no whole-report content audit is claimed. Source ancestry: existing funding note supplied discovery; McKinnon's page links the commissioned recipient outputs; financial statements are published on McKinnon's Storyblok asset host; ABR independently supplies registered identity/history. These are not ACNC mirrors accessed through a workaround. Search results on ACNC/storage and unrelated biographies were excluded without opening.
Original PDF bytes were retrieved into memory, extracted with pypdf and hashed; they are not retained on disk because this lane owns only this note. Reproducible identifiers:
\n| Original | \nByte count | \nSHA-256 | \n
|---|---|---|
| Trust FY2024 | \n1,503,445 | \n58777019dee161291be34314936208c79e9d1085239cac5a2cb2768475bd5cfc | \n
| Research Centre FY2024 | \n1,538,073 | \n3fa0f6819f77c4170111318109d34d28a38328cf7c4725bd3d4e7d7fec2663da | \n
| Research Centre FY2023 | \n777,035 | \n45cce3230034c975633aea56db54d6de32656779b8e4ae355d814c56c9cd44e7 | \n
| June2024 transparency report | \n11,375,061 | \naaf1eb0c85f044f03e7db25d64a1647e67de5c22534f17fd4f4550df22082fcb | \n
Trust FY2024 additional inspected passages: PDF4,10,16–18 (p18 text extraction blank). Research Centre FY2024: PDF3–4,11,17–18. No visual verification of signatures or logos is claimed. FY2023 web reader returned Internal Error; one ordinary GET succeeded. An independently published Monash2022 annual report was discovered as a possible strategy-development counterpart, but web reading exceeded the size limit and one ordinary GET returned HTTP403; it remains inaccessible and is not a findings source. Woolworths' Holly Kramer biography returned Internal Error and was not used. No further retries on these branches.
\nA bounded search of exact entity names with constitution, Research Committee, governing council, commissioning and agreement terms did not locate the missing charter or Reset contract. This is a precise public-record frontier, not a claim those records do not exist. Acquisition stops here for the parent's forest assessment.
", "path": "governance-wave-2026-09-09/mckinnon-governance.md", "sourceIds": [ "w12-gov-s-mckinnon-entities", "w12-gov-s-mckinnon-trust-abr", "w12-gov-s-research-abr-history", "w12-gov-s-research-abr-current", "w12-gov-s-charitable-abr", "w12-gov-s-mckinnon-trust2024", "w12-gov-s-research2024", "w12-gov-s-reset-roundtable" ], "markdown": "Observed 9 September 2026. Bounded public-source follow-through from `funding-wave-2026-09-09/reset-funding.md` and the latest forest review. No canonical case changes, outreach or ACNC access.\n\nThe new finding is an identifiable internal research-governance mechanism, alongside an important entity transition. The historical Research Centre reports Research Committee oversight of research spending and publication/dissemination activities. McKinnon separately identifies its Reset commissions. The public records read here do **not** join those two facts into a Reset-specific committee decision, contracting entity or editorial approval right.\n\n## Legal entities and the historical decision layer\n\nThe name McKinnon covers several entities. Its current funding page lists the Foundation trust, Research Centre and Charitable Foundation separately. The page says Grant Rule and Sophie Oh fund the organisation and describes related investment and administration arrangements. That is a group disclosure, not the execution page of either Reset contract. Funding and Financials (see source references) “Our Funding” and “Our entities”; source `gov-mck-s-funding`, read.\n\n| Entity and identifier | What the primary record establishes | Consequence for the commissions |\n| --- | --- | --- |\n| The Trustee for Susan McKinnon Foundation, ABN 28 739 100 711 | ABR identifies a discretionary investment trust, active from 21 April 2015, with private ancillary fund status. ABR record (see source references) entity and DGR fields; `gov-mck-s-abr-trust`, read. | This identifies the trust, not its corporate trustee's exact legal name or the contract signatory. |\n| Susan McKinnon Research Centre Ltd, ABN 40 656 129 127 / ACN 656 129 127 | ABR history carries this name from 16 December 2021 to 14 April 2026. The same identifier now belongs to OH-RULE FAMILY PHILANTHROPY LTD. History (see source references) entity-name rows, and current record (see source references) `gov-mck-s-abr-research`, read. | Preserve historical name and current alias on the same entity. Renaming does not establish contract assignment or a completed wind-up. |\n| Susan McKinnon Charitable Foundation Ltd, ABN 12 653 756 597 | ABR records an Australian public company active from 17 September 2021. ABR record (see source references) entity fields; `gov-mck-s-abr-charitable`, read. | Its existence during 2023/24 does not show it commissioned these reports. The September 2025 privacy policy is not a historical contract. |\n\nMcKinnon's consolidated 2025 program report says the Charitable Foundation became the group's primary charitable vehicle during 2025, following specific listing. Its financial year ends on 31 December. This is a dated operating transition, not simply a new public brand. The cover is dated **18 February 2026**, while the website link says published **17 February 2026**; retain both labels rather than silently choosing one. Program Expenditure Report 2025 (see source references) PDF pp1–2; `gov-mck-s-program2025`, read. The wider donation tables were inspected but are not used to assign a Reset payment.\n\nThe trust's FY2024 directors' report names **Grant Rule and Michael Schoenfeld** as the trustee's directors throughout the reporting year and to the report date, unless otherwise stated. Its stated activity is grants aligned with policy and governance change. Note 12 identifies the Research Centre as delivering programs towards the trust's goals and providing grant-program marketing/strategic advisory services at cost. It reports a **$19,530,300 donation to the Research Centre in FY2024**. This is intra-group capacity, not a Reset fee or a new amount to add to downstream spending. Trust FY2024 financial report (see source references) PDF p3/printed1 and PDF pp16–17/printed14–15; `gov-mck-s-trust2024`, read in relevant scope.\n\nThe Research Centre's FY2024 report names **Colin Galbraith, Grant Rule and Michael Schoenfeld** as directors. It expressly places application of research funds and activities connected with publishing and disseminating research under **Research Committee oversight**. It reports one member at 30 June 2024, without naming that member. Its Note 11 reports the same FY2024 intra-group donation. The accounts identify an accountable institutional route; they do not disclose committee appointments, voting thresholds, delegated limits or a Reset decision. Research Centre FY2024 report (see source references) PDF pp3–4/printed1–2, PDF p17/printed15 and PDF p18/printed16; `gov-mck-s-research2024`, read in relevant scope.\n\nThe FY2023 report already names the same directors, Research Committee oversight and one member. Its adjacent history statements are internally awkward: printed1 describes funding in June 2023 and commencement in July 2023; printed2 describes the first operational year after funding in FY2022. Do not use either sentence to fix a confident commissioning chronology. Research Centre FY2023 report (see source references) PDF pp3–4/printed1–2; `gov-mck-s-research2023`, read in that scope.\n\nToday's general website board is a separate, later observation. It lists John Thwaites as interim executive chair and describes Acting CEO Leigh O'Neill as leading strategy, operations and programs; Tom Mooney leads the democracy work. These roles are not evidence of who approved the early-2023 Reset approach. About McKinnon.\n\n## From commission to published advocacy\n\nMcKinnon's current account gives the commission's broad purpose: evaluate Australia's regulatory system and recommend improvements. It names the May 2024 *Functioning or Failing?* and June 2024 *Achieving Digital Platform Public Transparency in Australia*, commissioned in 2023 and 2024. It links readers to both outputs. That is a public report-use route, but not a contractual publication-clearance provision. Commissioning account (see source references) Reset section; `gov-mck-s-commissions`, read. Reuse the existing canonical source/commission edges if already present.\n\nThe May report names **Rys Farthing and Alice Dawkins** under authors/contributors. Its introduction says the Foundation approached Reset early in 2023 to design tests of the misinformation code; Reset attributes the research design to itself and methods from its global network. Its recommendations include considering a duty of care, risk assessments and researcher access. The report thus attributes question initiation to the funder and methodological work to the recipient. Neither attribution resolves who cleared publication or mandated conclusions. Functioning or Failing? (see source references) PDF pp2–3 and5–6; `gov-mck-s-functioning2024`, read in relevant scope. The experiments' validity was not independently audited.\n\nThe June report's first three pages confirm its date and EU/UK comparison and recommend risk assessments, prescribed transparency reports, independent audits, data portals and researcher access. Those opening pages do not disclose an executed scope or approval arrangement. Achieving Digital Platform Public Transparency in Australia (see source references) PDF pp1–3; `gov-mck-s-transparency2024`, read in that scope. PDF pp30–31 were also inspected for end matter. The funder logo is not legible in the extracted text; commissioning attribution rests on McKinnon's named account, not a visually unread logo.\n\nA separate, more revealing process record is the **April 2024** briefing *Regulating for Transparency*. It describes a **March 2024 roundtable of 20 academic/civil-society experts**, held under the Chatham House Rule, and translates discussion into regulatory recommendations. It acknowledges McKinnon support, assigns errors and omissions to Reset and explicitly separates attendance from endorsement. Its acknowledgements include HRLC contributors and a named independent consultant, among others; these are bounded contributions, not evidence all participants approved the final text. Original briefing (see source references) PDF pp1,4,14–15/printed same; `gov-mck-s-roundtable2024`, read in relevant scope. This supports an actual convening-to-output path. Do not automatically label it one of the two commissions or assert a separate payment.\n\n## Candidate directional relations and limits\n\n| Candidate relation | Type / evidence | Boundary |\n| --- | --- | --- |\n| Foundation trust → Research Centre: reports FY2024 intra-group donation | Finance; `gov-mck-s-trust2024`, `gov-mck-s-research2024` | $19,530,300, dollar label; Australian context, currency code not printed in cited passages. One transaction reported by both sides, not two awards or Reset money. |\n| Research Committee → Research Centre research spending and dissemination activities: oversees | Governance; `gov-mck-s-research2023`, `gov-mck-s-research2024` | Historical general mandate; no Reset-specific decision or final editorial veto established. |\n| Rule/Schoenfeld → trustee board; Galbraith/Rule/Schoenfeld → Research Centre board: named directors | Governance; respective FY2024 reports | Historical entity-specific roles; no disclosed member-appointment or dismissal right. |\n| Research Centre → OH-RULE FAMILY PHILANTHROPY LTD: later registered name | Identity, dated 14 April 2026; `gov-mck-s-abr-research` | Same ABN/ACN; not an inferred ownership transfer. History extraction renders the first character as “0”; current record renders “O”. |\n| McKinnon → Reset May/June reports: commissions named evaluations and recommendations | Commission; `gov-mck-s-commissions`, supported by recipient May introduction | Group identity and purpose established; exact contracting company, fee and approval rights open. |\n| McKinnon → April briefing: acknowledged support; March roundtable → briefing: informs | Support / convening-to-output; `gov-mck-s-roundtable2024` | Support amount/form unknown. Participant attendance is expressly not endorsement. |\n\nThe strongest supported explanation is financed agenda-setting with identifiable research production and publication responsibilities. A compatible rival is that the funder selected an already aligned specialist, leaving methods and conclusions with Reset. The records do not distinguish that from exercised donor editorial direction. Research Committee oversight is a useful new institutional fact, not the missing evidence of a committee approving the particular reports.\n\nThe next discriminating records are the **2023/24 executed Reset engagement letters and scopes**, identifying legal counterparties, fee, milestones, publication rights and approval/termination terms; **Research Committee terms of reference, membership/appointment resolutions and relevant minutes**; and the **single-member register and constitutions** in force when those decisions were made. Custodians are McKinnon's historical entities/current records holder and Reset. Appointment resolutions would distinguish founders' public sponsorship of appointments from their actual legal power. Contract allocation records would determine whether the Research Centre route applies to these commissions. None was obtained, and no contact was made.\n\n## Reading and custody\n\nAll URLs above are primary issuer/recipient/ABR records. Access state is `read` only for the stated passages; no whole-report content audit is claimed. Source ancestry: existing funding note supplied discovery; McKinnon's page links the commissioned recipient outputs; financial statements are published on McKinnon's Storyblok asset host; ABR independently supplies registered identity/history. These are not ACNC mirrors accessed through a workaround. Search results on ACNC/storage and unrelated biographies were excluded without opening.\n\nOriginal PDF bytes were retrieved into memory, extracted with pypdf and hashed; they are **not retained on disk** because this lane owns only this note. Reproducible identifiers:\n\n| Original | Byte count | SHA-256 |\n| --- | --- | --- |\n| Trust FY2024 | 1,503,445 | `58777019dee161291be34314936208c79e9d1085239cac5a2cb2768475bd5cfc` |\n| Research Centre FY2024 | 1,538,073 | `3fa0f6819f77c4170111318109d34d28a38328cf7c4725bd3d4e7d7fec2663da` |\n| Research Centre FY2023 | 777,035 | `45cce3230034c975633aea56db54d6de32656779b8e4ae355d814c56c9cd44e7` |\n| June2024 transparency report | 11,375,061 | `aaf1eb0c85f044f03e7db25d64a1647e67de5c22534f17fd4f4550df22082fcb` |\n\nTrust FY2024 additional inspected passages: PDF4,10,16–18 (p18 text extraction blank). Research Centre FY2024: PDF3–4,11,17–18. No visual verification of signatures or logos is claimed. FY2023 web reader returned Internal Error; one ordinary GET succeeded. An independently published Monash2022 annual report was discovered as a possible strategy-development counterpart, but web reading exceeded the size limit and one ordinary GET returned HTTP403; it remains inaccessible and is not a findings source. Woolworths' Holly Kramer biography returned Internal Error and was not used. No further retries on these branches.\n\nA bounded search of exact entity names with constitution, Research Committee, governing council, commissioning and agreement terms did not locate the missing charter or Reset contract. This is a precise public-record frontier, not a claim those records do not exist. Acquisition stops here for the parent's forest assessment.\n", "file": "research/governance-wave-2026-09-09-mckinnon-governance.md" }, { "id": "governance-wave-2026-09-09-uk-governance-articles", "title": "Reset Tech UK: constitutional decision rights", "description": "Observed9September2026. Bounded primary-filing review of company14982650, distinct from the informal Online Safety Network and from other Reset legal entities. The May2025 constitution reveals a director-member structure and an explicit cross-entity conflict provision. It does not reconcile the current PSC categories or establish donor control of policy outp", "html": "Observed9September2026. Bounded primary-filing review of company14982650, distinct from the informal Online Safety Network and from other Reset legal entities. The May2025 constitution reveals a director-member structure and an explicit cross-entity conflict provision. It does not reconcile the current PSC categories or establish donor control of policy outputs.
\nCompanies House filing history records articles and adoption resolution filed21May2025. The resolution is headed7May2025 and replaces the previous articles immediately; it bears Dale Benjamin Scott's agreement as the voting member. Poppy Wood's PSC notification, filed19May2025, gives7May2025 as the date she became registrable. The resolution is evidence of this adoption decision, not a permanent personal veto. Its signature date is printed5/7/2025 alongside the explicit7May heading; do not reinterpret it as a later July action. A separate source anomaly is that its header prints company number4982650 while the official docket and attached articles identify14982650. Retain this discrepancy rather than silently correcting a quotation or inventing another company.
\n[Official filing history](http[local research file] [Filed articles](http[local research file] [Filed resolution](http[local research file] [Wood PSC01](http[local research file]
\nAll20 articles pages, both resolution pages and both PSC01 pages were visually inspected as scans. The articles contain17 numbered substantive pages following a cover and two index pages; below, PDF page equals printed page plus3. Relevant role/date/control passages were read on current official officers and PSC pages; unrelated personal-data fields are excluded from this note. These official web reads were observed9September but their reader cache reports an earlier crawl; they are not a freshly certified member register.
\nDirectors and members are the same constitutional constituency. Article22, PDF14/printed11, makes directors the only members: appointment automatically admits them and their names must enter the member register; membership ends with directorship and is nontransferable. The filed constitution names no special membership class or named-person veto. Article24.2 gives each member one vote on a written resolution. Ordinary written decisions require a simple majority; special written resolutions at least75% of eligible voting rights. Removing a director or auditor through the statutory member-resolution route cannot use a written resolution (24.4, PDF15). Do not silently generalize the written-vote clause into a fully audited account of every statutory voting mechanism.
\nThe board manages and can retract delegations. Articles9-12, PDF9/printed6, give directors general management authority. They can delegate functions and day-to-day work, permit further delegation, and change or terminate those arrangements. Committees cannot knowingly incur expenditure or liability without authorization or a board-approved budget (10.4). Board decisions normally use a majority at a quorate meeting; an eligible chair has a casting vote on a tie. Decisions without a meeting require a majority sharing the view, not unanimity (13,15.5,16; PDF10-11). Board quorum cannot be below two and otherwise defaults to two or one-third, whichever is greater (15.2).
\nAppointments and removals are collective powers in the text. Article20.2, PDF13/printed10, lets directors appoint eligible directors. Article21 includes specified cessation/removal grounds;21.6 permits a board removal resolution at a meeting attended by at least half the directors after14clear days' notice and an opportunity for representations. The constitution also recognizes statutory cessation/removal routes. It does not name Scott, Wood, a donor or an outside organization as sole appointer. Article19 says the company should seek always to have at least three directors; the public officers page lists two. That mismatch is a useful record question, not proof the two-person board's decisions are invalid. The articles expressly allow limited appointments when the number falls below quorum.
\nSome matters are entrenched. Article8, PDF8/printed5, requires unanimity of eligible members to amend purposes, asset lock, winding-up provisions or the entrenchment article itself. Other special resolutions have the stated75% threshold. This is a real member-level reserved power. No donor-specific publication-clearance, budget-veto or board-appointment provision was identified in the complete articles; separate contracts could carry rights that this document does not disclose.
\nSchedule1, definition1.11, PDF19-20/printed16-17, expressly includes Reset Tech, Reset Tech Action, Reset Tech Australia Limited, Reset Tech GmbH and Reset Tech Action GmbH within its defined Group Company category, with further companies addable by members. This is an explicit constitutional list. It is not by itself proof of a parent/subsidiary relationship, consolidated ownership, each entity's exact registry identity, or an instruction from the US entity to Australia.
\nArticle17, PDF11-12/printed8-9, requires declaration of direct/indirect transaction interests and conflicting duties. Potentially conflicted directors may ordinarily participate, count in the quorum and vote unless a majority of the other participating directors decides otherwise or the non-exempt personal-benefit rule applies. Article17.4.2's exceptions include specified remuneration/expense decisions and decisions involving a defined Group Company with which a director is connected. Thus a group connection does not automatically require recusal under these articles. If17.5 applies, participation is limited to informing debate, the director is excluded from the relevant quorum and withdraws without voting. Whether any particular director disclosed, voted, was excluded or improperly benefited requires the actual decision record; statutory duties/enforceability were not independently audited.
\nDefinition1.8, PDF19, supplies a connection test involving specified equity/voting interests or statutory connected-person relationships. Do not substitute a shared conference, job title or common purpose for that definition. Article17.6 also permits specified withholding of confidential information where disclosure would breach another duty, after compliance with the conflict provisions. That is a permitted confidentiality boundary, not proof that information was withheld in any concrete case.
\nThe company is limited by guarantee without share capital. Article4 caps the member's specified winding-up guarantee atGBP1; that is not a share price or the limit of all possible liabilities. Articles1-2 expressly authorize public-benefit technology/data-policy work, legislative advocacy and efforts to influence opinion and public institutions. That supplies a declared organizational purpose, not proof of a particular policy win.
\nArticles3,6-8, PDF6-8, restrict private profit distribution and constrain assets/winding-up, with stated exceptions. Reasonable remuneration/services/rent/interest and director expenses are permitted; do not describe the organization as unable to pay insiders. Net winding-up assets cannot be distributed among members. Its non-share structure and these limits make an equity-ownership interpretation of the PSC percentages especially inappropriate.
\nThe [official officers page](http[local research file] lists Scott and Wood as the two active directors, both appointed5July2023. The [PSC page](http[local research file] lists Wood at more than25% up to50% of voting rights, notified7May2025; Scott at75% or more plus appointment/removal rights, notified5July2023. Wood's original PSC01 repeats her voting band. These are reported control categories, not additive shareholding percentages. Companies House itself warns that it does not check the accuracy of filed information.
\nIf the two listed directors are the complete relevant membership, the new article22 membership rule plus one-member-one-written-vote rule gives an apparent equal written-vote structure. That conditional calculation is not a finding that the PSC register is definitively wrong or that all current rights are50/50. The constitution does not disclose how Scott's broader registered control category is currently supported. A membership-rights transition associated with the May2025 replacement, stale/incomplete reporting, attribution through another arrangement or another unacquired rights instrument are possible explanations. The present records do not choose between them. The original pre-May2025 articles, current statutory member register, PSC basis/changes and any joint-voting or appointment arrangement would discriminate; they were not acquired in this bounded pass.
\nUse company14982650 → board for constitutional management/delegation; board → director/member role for appointment/admission; eligible members → entrenched provisions for unanimous amendment consent. Record UK company → each listed Group Company as defined constitutional grouping for conflict rules, not equity ownership or parent control. Scott → May2025 adoption is an observed resolution agreement. Keep PSC assertions as attributed register records pending reconciliation, separate from the constitution's general rules. None establishes operational control of OSN merely because a network uses Reset services, or a donor veto over a funded article.
\nThe highest-information next record is the company's current member register and the instrument explaining Scott's registered appointment/voting rights after the new articles; the next operational record is a specific board delegation, conflict declaration/minute, or spending/publication approval connecting these powers to funded network work. Article29 requires minutes of appointments/resolutions/proceedings, and30 requires the member register and accounts. Their existence as required records is distinct from public availability. Root owns the next forest review; no outreach or subsequent acquisition branch was opened.
\nThe web reader's articles click returned an expired signed-S3 redirect; resolution and PSC01 clicks returned cache misses. One ordinary GET of each observed stable public filing endpoint succeeded, without reusing signed tokens. PDFs are scanned: pypdf returned no text and startxref repair warnings for articles/resolution. The bundled visual PDF renderer displayed all pages. An initial unavailable fitz import was replaced with already available pypdf/pypdfium2; no installation occurred. No repeated failed web route or access bypass was used.
\nRetained originals under governance-wave-2026-09-09/originals:
\n| File | \nBytes | \nSHA256 | \n
|---|---|---|
| uk-articles-2025.pdf | \n685147 | \n240587be6f798183ddcf3f651cd6dcdd3d208cbb71410761c7a89171942147da | \n
| uk-articles-resolution-2025.pdf | \n44909 | \n7b3ba5028ece428b39984dffe269c708b947fbb7da1296f2fd49eedbdb110278 | \n
| uk-articles-psc-notification-2025.pdf | \n20825 | \n97c6aa6109c98696dd39bac685763ff827abb393312b716973571f961ccc87f8 | \n
The same uk-articles prefixes identify derived page PNGs; they are visual-reading aids, not independent sources. No canonical case or other worker's file was changed. This is the constitutional decision-rights breakpoint.
", "path": "governance-wave-2026-09-09/uk-governance-articles.md", "sourceIds": [ "w12-gov-s-uk-articles", "w12-gov-s-uk-resolution", "w12-gov-s-uk-psc", "w12-gov-s-uk-directors" ], "markdown": "Observed9September2026. Bounded primary-filing review of company14982650, distinct from the informal Online Safety Network and from other Reset legal entities. The May2025 constitution reveals a director-member structure and an explicit cross-entity conflict provision. It does not reconcile the current PSC categories or establish donor control of policy outputs.\n\n## Instruments and reading scope\n\nCompanies House filing history records articles and adoption resolution filed21May2025. The resolution is headed7May2025 and replaces the previous articles immediately; it bears Dale Benjamin Scott's agreement as the voting member. Poppy Wood's PSC notification, filed19May2025, gives7May2025 as the date she became registrable. The resolution is evidence of this adoption decision, not a permanent personal veto. Its signature date is printed5/7/2025 alongside the explicit7May heading; do not reinterpret it as a later July action. A separate source anomaly is that its header prints company number4982650 while the official docket and attached articles identify14982650. Retain this discrepancy rather than silently correcting a quotation or inventing another company.\n\nOfficial filing history (see source references) Filed articles (see source references) Filed resolution (see source references) Wood PSC01 (see source references)\n\nAll20 articles pages, both resolution pages and both PSC01 pages were visually inspected as scans. The articles contain17 numbered substantive pages following a cover and two index pages; below, PDF page equals printed page plus3. Relevant role/date/control passages were read on current official officers and PSC pages; unrelated personal-data fields are excluded from this note. These official web reads were observed9September but their reader cache reports an earlier crawl; they are not a freshly certified member register.\n\n## Who can decide\n\n**Directors and members are the same constitutional constituency.** Article22, PDF14/printed11, makes directors the only members: appointment automatically admits them and their names must enter the member register; membership ends with directorship and is nontransferable. The filed constitution names no special membership class or named-person veto. Article24.2 gives each member one vote on a **written** resolution. Ordinary written decisions require a simple majority; special written resolutions at least75% of eligible voting rights. Removing a director or auditor through the statutory member-resolution route cannot use a written resolution (24.4, PDF15). Do not silently generalize the written-vote clause into a fully audited account of every statutory voting mechanism.\n\n**The board manages and can retract delegations.** Articles9-12, PDF9/printed6, give directors general management authority. They can delegate functions and day-to-day work, permit further delegation, and change or terminate those arrangements. Committees cannot knowingly incur expenditure or liability without authorization or a board-approved budget (10.4). Board decisions normally use a majority at a quorate meeting; an eligible chair has a casting vote on a tie. Decisions without a meeting require a majority sharing the view, not unanimity (13,15.5,16; PDF10-11). Board quorum cannot be below two and otherwise defaults to two or one-third, whichever is greater (15.2).\n\n**Appointments and removals are collective powers in the text.** Article20.2, PDF13/printed10, lets directors appoint eligible directors. Article21 includes specified cessation/removal grounds;21.6 permits a board removal resolution at a meeting attended by at least half the directors after14clear days' notice and an opportunity for representations. The constitution also recognizes statutory cessation/removal routes. It does not name Scott, Wood, a donor or an outside organization as sole appointer. Article19 says the company should seek always to have at least three directors; the public officers page lists two. That mismatch is a useful record question, not proof the two-person board's decisions are invalid. The articles expressly allow limited appointments when the number falls below quorum.\n\n**Some matters are entrenched.** Article8, PDF8/printed5, requires unanimity of eligible members to amend purposes, asset lock, winding-up provisions or the entrenchment article itself. Other special resolutions have the stated75% threshold. This is a real member-level reserved power. No donor-specific publication-clearance, budget-veto or board-appointment provision was identified in the complete articles; separate contracts could carry rights that this document does not disclose.\n\n## Cross-entity conflicts: an actual governance bridge\n\nSchedule1, definition1.11, PDF19-20/printed16-17, expressly includes **Reset Tech, Reset Tech Action, Reset Tech Australia Limited, Reset Tech GmbH and Reset Tech Action GmbH** within its defined Group Company category, with further companies addable by members. This is an explicit constitutional list. It is not by itself proof of a parent/subsidiary relationship, consolidated ownership, each entity's exact registry identity, or an instruction from the US entity to Australia.\n\nArticle17, PDF11-12/printed8-9, requires declaration of direct/indirect transaction interests and conflicting duties. Potentially conflicted directors may ordinarily participate, count in the quorum and vote unless a majority of the other participating directors decides otherwise or the non-exempt personal-benefit rule applies. Article17.4.2's exceptions include specified remuneration/expense decisions and decisions involving a defined Group Company with which a director is connected. Thus a group connection does **not automatically require recusal under these articles**. If17.5 applies, participation is limited to informing debate, the director is excluded from the relevant quorum and withdraws without voting. Whether any particular director disclosed, voted, was excluded or improperly benefited requires the actual decision record; statutory duties/enforceability were not independently audited.\n\nDefinition1.8, PDF19, supplies a connection test involving specified equity/voting interests or statutory connected-person relationships. Do not substitute a shared conference, job title or common purpose for that definition. Article17.6 also permits specified withholding of confidential information where disclosure would breach another duty, after compliance with the conflict provisions. That is a permitted confidentiality boundary, not proof that information was withheld in any concrete case.\n\n## Purpose and asset limits\n\nThe company is limited by guarantee without share capital. Article4 caps the member's specified winding-up guarantee atGBP1; that is not a share price or the limit of all possible liabilities. Articles1-2 expressly authorize public-benefit technology/data-policy work, legislative advocacy and efforts to influence opinion and public institutions. That supplies a declared organizational purpose, not proof of a particular policy win.\n\nArticles3,6-8, PDF6-8, restrict private profit distribution and constrain assets/winding-up, with stated exceptions. Reasonable remuneration/services/rent/interest and director expenses are permitted; do not describe the organization as unable to pay insiders. Net winding-up assets cannot be distributed among members. Its non-share structure and these limits make an equity-ownership interpretation of the PSC percentages especially inappropriate.\n\n## PSC reconciliation remains open\n\nThe official officers page (see source references) lists Scott and Wood as the two active directors, both appointed5July2023. The PSC page (see source references) lists Wood at more than25% up to50% of voting rights, notified7May2025; Scott at75% or more plus appointment/removal rights, notified5July2023. Wood's original PSC01 repeats her voting band. These are reported control categories, not additive shareholding percentages. Companies House itself warns that it does not check the accuracy of filed information.\n\nIf the two listed directors are the complete relevant membership, the new article22 membership rule plus one-member-one-written-vote rule gives an apparent equal written-vote structure. **That conditional calculation is not a finding that the PSC register is definitively wrong or that all current rights are50/50.** The constitution does not disclose how Scott's broader registered control category is currently supported. A membership-rights transition associated with the May2025 replacement, stale/incomplete reporting, attribution through another arrangement or another unacquired rights instrument are possible explanations. The present records do not choose between them. The original pre-May2025 articles, current statutory member register, PSC basis/changes and any joint-voting or appointment arrangement would discriminate; they were not acquired in this bounded pass.\n\n## Suggested typed relations and next record\n\nUse company14982650 → board for constitutional management/delegation; board → director/member role for appointment/admission; eligible members → entrenched provisions for unanimous amendment consent. Record UK company → each listed Group Company as **defined constitutional grouping for conflict rules**, not equity ownership or parent control. Scott → May2025 adoption is an observed resolution agreement. Keep PSC assertions as attributed register records pending reconciliation, separate from the constitution's general rules. None establishes operational control of OSN merely because a network uses Reset services, or a donor veto over a funded article.\n\nThe highest-information next record is the company's current member register and the instrument explaining Scott's registered appointment/voting rights after the new articles; the next operational record is a specific board delegation, conflict declaration/minute, or spending/publication approval connecting these powers to funded network work. Article29 requires minutes of appointments/resolutions/proceedings, and30 requires the member register and accounts. Their existence as required records is distinct from public availability. Root owns the next forest review; no outreach or subsequent acquisition branch was opened.\n\n## Custody and route closure\n\nThe web reader's articles click returned an expired signed-S3 redirect; resolution and PSC01 clicks returned cache misses. One ordinary GET of each observed stable public filing endpoint succeeded, without reusing signed tokens. PDFs are scanned: pypdf returned no text and startxref repair warnings for articles/resolution. The bundled visual PDF renderer displayed all pages. An initial unavailable fitz import was replaced with already available pypdf/pypdfium2; no installation occurred. No repeated failed web route or access bypass was used.\n\nRetained originals under governance-wave-2026-09-09/originals:\n\n| File | Bytes | SHA256 |\n| --- | --- | --- |\n| uk-articles-2025.pdf | 685147 | 240587be6f798183ddcf3f651cd6dcdd3d208cbb71410761c7a89171942147da |\n| uk-articles-resolution-2025.pdf | 44909 | 7b3ba5028ece428b39984dffe269c708b947fbb7da1296f2fd49eedbdb110278 |\n| uk-articles-psc-notification-2025.pdf | 20825 | 97c6aa6109c98696dd39bac685763ff827abb393312b716973571f961ccc87f8 |\n\nThe same uk-articles prefixes identify derived page PNGs; they are visual-reading aids, not independent sources. No canonical case or other worker's file was changed. This is the constitutional decision-rights breakpoint.\n", "file": "research/governance-wave-2026-09-09-uk-governance-articles.md" }, { "id": "governance-wave-2026-09-09-uk-network-operator", "title": "Online Safety Network operator and Reset UK company records", "description": "Observed 9 September 2026. Public institutional records and scoped visual inspection of the original 2024 company accounts. This packet establishes operational custody and several corporate disclosure boundaries. It does not establish donor instructions, a network-specific budget or exercised control over a particular policy recommendation.", "html": "Observed 9 September 2026. Public institutional records and scoped visual inspection of the original 2024 company accounts. This packet establishes operational custody and several corporate disclosure boundaries. It does not establish donor instructions, a network-specific budget or exercised control over a particular policy recommendation.
\nThe [Online Safety Network privacy notice](http[local research file] says the network has no formal membership or association. Reset Tech operates its website, newsletter and shared email inbox. This is an explicit operational relationship, stronger than inferred affiliation. The notice covers network activity and identifies a federation of Reset legal entities, including Reset Tech UK Limited. It describes them as a single organization for the notice's purposes, with shared staff and systems. That wording is not a legal merger, a corporate ownership tree or proof that the UK company alone contracts for every network expense.
\nThe notice describes records of professional contacts, expertise, opinions, correspondence, meeting notes, workshops and roundtables. Individual organizations' conversations outside the shared inbox are governed by their own notices. Mailchimp handles the newsletter; listed website providers include Matomo, Heroku and Cloudflare. Conditional provisions permit sharing with co-funders or transferring the network/site to another organization. They do not establish that a particular transfer or disclosure occurred. Last updated is labelled October 2025 despite the currently displayed network name; do not date every sentence to a verified 2026 revision.
\nAn important counterconstraint is explicit advice that readers concerned about government surveillance may use Tor or VPNs. The notice also states a 48-hour technical-data retention rule with testing/security exceptions. These are stated policies, not independently verified practice, but they are inconsistent with attributing an undifferentiated anti-anonymity doctrine to every aspect of this organization.
\nThe [general Reset privacy notice](http[local research file] separately describes partner onboarding, sanctions checks via Plaid Monitor, human review, grant-progress records and information systems. It says partner funding records are retained for ten years after funding ends. These provisions identify types of records likely to exist and possible custodians; no particular grant file has been obtained. Sanctions screening is not evidence of ideological vetting. These general partner terms must not be substituted for an OSN contract.
\nThe [Companies House overview](http[local research file] identifies RESET TECH UK LIMITED, company 14982650, incorporated 5 July 2023, as a company limited by guarantee without share capital. The fresh [filing history](http[local research file] includes 2026 filings absent from an older indexed view. Accounts to 31 December 2025 are due 30 September 2026 and were not yet overdue at observation.
\nThe [PSC register](http[local research file] lists Poppy Wood with more than 25% but no more than 50% of voting rights, and Dale Benjamin Scott with at least 75% and a right to appoint/remove directors. Report these as separate register statements, not additive equity percentages. The company has no share capital. Membership, timing or arrangements need reconciliation against the constitution and member register; the apparent discrepancy is not proof of misconduct. The companion uk-governance-articles.md reads the actual constitutional provisions.
\nThe official [2024 accounts PDF](http[local research file] filed 24 September 2025, reports GBP 1,571,393 turnover; GBP 70,631 operating expenses; GBP 1,451,987 administration expenses; and GBP 39,507 after-tax profit. These are whole-company figures, not OSN costs or identified donor receipts. The source's pretax/tax/after-tax printed numbers differ arithmetically by one pound; figures are preserved rather than corrected. No material inference is made from the difference.
\nNote 4 breaks administration expenses into salaries/wages/payroll GBP 76,881; benefits/insurance GBP 122,254; professional/consulting services GBP 402,105; subscriptions/software GBP 47,504; travel GBP 17,557; and other expenses GBP 785,686. The statement says five employees during 2024, without establishing five full-time equivalents for the whole year. It would be misleading to calculate an average annual salary by division. Expense names alone do not resolve outsourced staff, group recharge or grant pass-through amounts.
\nYear-end cash is GBP 40,997 in HSBC UK accounts. This identifies a banking relationship, not bank direction over advocacy. Note 7 puts GBP 1,156 receivable from Reset Tech Action gGmbH and GBP 10,743 directors' loan under debtors. The latter is owed to the company, not a director's funding contribution; individual identity and terms are not disclosed there. The named intercompany receivable is a concrete accounting relation, but does not identify all group transactions or establish parent control.
\nNote 9 records members' contingent GBP 1 winding-up guarantee. This is not invested capital or a measure of voting power. Note 11 says the company avails itself of FRS 102 section 33.1A exemption for disclosure of transactions with group undertakings. Note 1 claims the small-company cash-flow statement exemption. Report what the accounts invoke; the applicability of each exemption has not been independently audited. These provisions explain specific disclosure limits, not unlawful concealment. The filing does not provide a complete donor-to-project cash trail.
\nThe accounts name Scott and Wood as directors and are approved 17 September 2025. Root visually inspected physical PDF pages 3, 4, 7, 8, 11, 12, 13 and 14; printed pages are one lower. These cover the directors, directors' narrative, signatures/responsibilities, income statement, accounting policies, expenses, debtors and related-party notes. No external audit opinion is claimed. Scanned pages have no usable pypdf text; the local text file contains page markers only and must not be treated as OCR.
\nThe supported path is informal collaboration hosted through Reset's shared operating infrastructure, with identifiable UK company directors, constitutional decision rules and some group accounting relations. It is not yet a demonstrated single legal entity controlling all federated activity. The missing records with high value are the OSN operating/transition agreement, division of employer and contracting responsibilities, related-party service arrangements, member register and basis for PSC notices, plus any network budget or delegated editorial authority. Public procedures make these plausible targets but do not prove a particular document or clause exists.
\nThe network notice was read both through indexed primary text and an ordinary successful GET, then locally extracted with scripts/navigation removed. The official accounts link initially produced a stale expired signed redirect in the web reader. One ordinary GET of the observed stable public Companies House endpoint succeeded; no credentials, access restrictions or signed-token reuse were involved. The original is retained. No new OCR dependency was installed. Rendered PNGs are for document inspection, not artwork.
", "path": "governance-wave-2026-09-09/uk-network-operator.md", "sourceIds": [ "w12-gov-s-osn-notice", "w12-gov-s-uk-overview", "w12-gov-s-uk-accounts" ], "markdown": "Observed 9 September 2026. Public institutional records and scoped visual inspection of the original 2024 company accounts. This packet establishes operational custody and several corporate disclosure boundaries. It does not establish donor instructions, a network-specific budget or exercised control over a particular policy recommendation.\n\n## Who maintains the network\n\nThe Online Safety Network privacy notice (see source references) says the network has no formal membership or association. Reset Tech operates its website, newsletter and shared email inbox. This is an explicit operational relationship, stronger than inferred affiliation. The notice covers network activity and identifies a federation of Reset legal entities, including Reset Tech UK Limited. It describes them as a single organization for the notice's purposes, with shared staff and systems. That wording is not a legal merger, a corporate ownership tree or proof that the UK company alone contracts for every network expense.\n\nThe notice describes records of professional contacts, expertise, opinions, correspondence, meeting notes, workshops and roundtables. Individual organizations' conversations outside the shared inbox are governed by their own notices. Mailchimp handles the newsletter; listed website providers include Matomo, Heroku and Cloudflare. Conditional provisions permit sharing with co-funders or transferring the network/site to another organization. They do not establish that a particular transfer or disclosure occurred. Last updated is labelled October 2025 despite the currently displayed network name; do not date every sentence to a verified 2026 revision.\n\nAn important counterconstraint is explicit advice that readers concerned about government surveillance may use Tor or VPNs. The notice also states a 48-hour technical-data retention rule with testing/security exceptions. These are stated policies, not independently verified practice, but they are inconsistent with attributing an undifferentiated anti-anonymity doctrine to every aspect of this organization.\n\nThe general Reset privacy notice (see source references) separately describes partner onboarding, sanctions checks via Plaid Monitor, human review, grant-progress records and information systems. It says partner funding records are retained for ten years after funding ends. These provisions identify types of records likely to exist and possible custodians; no particular grant file has been obtained. Sanctions screening is not evidence of ideological vetting. These general partner terms must not be substituted for an OSN contract.\n\n## The identifiable UK company\n\nThe Companies House overview (see source references) identifies RESET TECH UK LIMITED, company 14982650, incorporated 5 July 2023, as a company limited by guarantee without share capital. The fresh filing history (see source references) includes 2026 filings absent from an older indexed view. Accounts to 31 December 2025 are due 30 September 2026 and were not yet overdue at observation.\n\nThe PSC register (see source references) lists Poppy Wood with more than 25% but no more than 50% of voting rights, and Dale Benjamin Scott with at least 75% and a right to appoint/remove directors. Report these as separate register statements, not additive equity percentages. The company has no share capital. Membership, timing or arrangements need reconciliation against the constitution and member register; the apparent discrepancy is not proof of misconduct. The companion uk-governance-articles.md reads the actual constitutional provisions.\n\n## What the 2024 accounts expose and leave unresolved\n\nThe official 2024 accounts PDF (see source references) filed 24 September 2025, reports GBP 1,571,393 turnover; GBP 70,631 operating expenses; GBP 1,451,987 administration expenses; and GBP 39,507 after-tax profit. These are whole-company figures, not OSN costs or identified donor receipts. The source's pretax/tax/after-tax printed numbers differ arithmetically by one pound; figures are preserved rather than corrected. No material inference is made from the difference.\n\nNote 4 breaks administration expenses into salaries/wages/payroll GBP 76,881; benefits/insurance GBP 122,254; professional/consulting services GBP 402,105; subscriptions/software GBP 47,504; travel GBP 17,557; and other expenses GBP 785,686. The statement says five employees during 2024, without establishing five full-time equivalents for the whole year. It would be misleading to calculate an average annual salary by division. Expense names alone do not resolve outsourced staff, group recharge or grant pass-through amounts.\n\nYear-end cash is GBP 40,997 in HSBC UK accounts. This identifies a banking relationship, not bank direction over advocacy. Note 7 puts GBP 1,156 receivable from Reset Tech Action gGmbH and GBP 10,743 directors' loan under debtors. The latter is owed to the company, not a director's funding contribution; individual identity and terms are not disclosed there. The named intercompany receivable is a concrete accounting relation, but does not identify all group transactions or establish parent control.\n\nNote 9 records members' contingent GBP 1 winding-up guarantee. This is not invested capital or a measure of voting power. Note 11 says the company avails itself of FRS 102 section 33.1A exemption for disclosure of transactions with group undertakings. Note 1 claims the small-company cash-flow statement exemption. Report what the accounts invoke; the applicability of each exemption has not been independently audited. These provisions explain specific disclosure limits, not unlawful concealment. The filing does not provide a complete donor-to-project cash trail.\n\nThe accounts name Scott and Wood as directors and are approved 17 September 2025. Root visually inspected physical PDF pages 3, 4, 7, 8, 11, 12, 13 and 14; printed pages are one lower. These cover the directors, directors' narrative, signatures/responsibilities, income statement, accounting policies, expenses, debtors and related-party notes. No external audit opinion is claimed. Scanned pages have no usable pypdf text; the local text file contains page markers only and must not be treated as OCR.\n\n## Evidential consequence\n\nThe supported path is informal collaboration hosted through Reset's shared operating infrastructure, with identifiable UK company directors, constitutional decision rules and some group accounting relations. It is not yet a demonstrated single legal entity controlling all federated activity. The missing records with high value are the OSN operating/transition agreement, division of employer and contracting responsibilities, related-party service arrangements, member register and basis for PSC notices, plus any network budget or delegated editorial authority. Public procedures make these plausible targets but do not prove a particular document or clause exists.\n\n## Acquisition custody\n\nThe network notice was read both through indexed primary text and an ordinary successful GET, then locally extracted with scripts/navigation removed. The official accounts link initially produced a stale expired signed redirect in the web reader. One ordinary GET of the observed stable public Companies House endpoint succeeded; no credentials, access restrictions or signed-token reuse were involved. The original is retained. No new OCR dependency was installed. Rendered PNGs are for document inspection, not artwork.\n", "file": "research/governance-wave-2026-09-09-uk-network-operator.md" }, { "id": "implementation-instructions-wave-2026-09-09-uk-commissioning", "title": "Who instructs NSOIT's monitoring suppliers?", "description": "Bounded implementation investigation, 9 September 2026. The anchor remains DSIT's published 5 February 2025 payment, GBP 56,736, transaction 654609, to Crisp Thinking (UK) Ltd. No record inspected here joins that payment to a particular report, monitoring specification or older contract. See the preceding commercial-participant.md and its original CSV; the p", "html": "Bounded implementation investigation, 9 September 2026. The anchor remains DSIT's published 5 February 2025 payment, GBP 56,736, transaction 654609, to Crisp Thinking (UK) Ltd. No record inspected here joins that payment to a particular report, monitoring specification or older contract. See the preceding commercial-participant.md and its retained original CSV; the payment is not counted again.
\nThe newly visible mechanism is ministerial scope-setting, departmental commissioning and assessment, supplier analysis, then platform review under private terms. An acquired draft policy specifies documented instructions and agreed search methods for contractors. Contemporary parliamentary testimony describes both referrals and warnings about lawful content. The exact escalation approvals and classification thresholds remain partly redacted, and an executed Crisp SOW remains unestablished.
\nA public FOI-associated copy of the National Security and Online Information Team Data Protection and Compliance Policy supplies detailed instructions. It is hosted by Big Brother Watch at a path identifying FOI2024-00576 and May 2024. Every page is marked DRAFT, OFFICIAL-SENSITIVE. No approval date, signed adoption, final revision or incorporation into the February 2025 engagement was established. Describe the following as provisions of this acquired draft, rather than proof every analyst or supplier followed them. [Publicly hosted policy copy](http[local research file]
\n| Who / decision | \nDraft provision | \nExact PDF locator and limit | \n
|---|---|---|
| NSOIT or central government: commission work | \nNSOIT may commission within its remit or be tasked through central command structures, with Cabinet Office National Security Secretariat given as an example. Sustained monitoring requires appropriate ministerial agreement. | \n§3.2, PDF 4. This is a route for tasking, not an actual task order. | \n
| NSOIT Analysis team: assess a task | \nEvery collection activity must have a legitimate policy objective, be proportionate, meet data-protection requirements and avoid directed surveillance. | \n§3.3, PDF 4; §§3.7–3.9, PDF 5. | \n
| DSIT: instruct and onboard suppliers | \nDelivery partners must follow NSOIT instructions. DSIT takes controller responsibility where it processes personal data; third parties are to be processors under Article 28, operating on documented instructions. Legal, Commercial and Operational Data teams participate in onboarding. | \n§§3.4–3.6, PDF 4–5. This does not identify which Crisp entity signed any agreement. | \n
| NSOIT and partner: agree methods | \nCase-specific methodology must cover data sources, search terms, task subject/purpose and minimisation. Contractual assurance and periodic review are specified. | \n§3.6(b)–(d), PDF 5. The actual search dictionary, classification specification and completed assurance record were not acquired. | \n
| Analysts: collection limits | \nPublicly accessible sources only; no name-based searching or building a picture of private individuals/groups by following activity. These instructions are to be communicated to partners. | \n§3.1, PDF 4; Schedule 1, PDF 9. The draft permits access to public sites requiring login, while prohibiting fake personas. Schedule 2 Part B §§1.4–1.6, PDF 15. | \n
| NSOIT: decide on referral | \nTeams identify potentially harmful material within ministerially agreed remit. NSOIT reviews against agreed thresholds. Authorized content may be shared for platforms to act at their discretion, with information minimised. | \n§§4.1–4.2, PDF 7. The threshold and authorization detail is partly blacked out: do not invent a named approving officer or automatic approval rule. | \n
| Politicians, news sources and journalists: an exclusion | \nThe draft excludes their content from platform referrals; its journalist definition depends on a news-publishing entity, editorial control and other institutional criteria. | \n§§1.6–1.7, PDF 3; §4.2(b), PDF 7; Schedule 1, PDF 9. This is not a blanket exemption for every citizen publishing commentary. | \n
| Head of NSOIT and DPO: exceptional retention | \nRaw samples normally up to three months; output retention/review and justified extensions involve the Head and Data Protection Officer. Thematic leads manage expiry of escalation links. | \n§3.13, PDF 6; §4.4(A), PDF 7; Schedule 2 §2.5, PDF 13. It does not show retention complied with in an actual case. | \n
| Legal advisers: surveillance boundary | \nThe draft says DSIT has no directed-surveillance RIPA authorization; activity risking that threshold must not be authorized and uncertainty goes to legal advisers. | \nSchedule 2 Part B §§1.5–1.6, PDF 15. This is the draft's legal/operational account, not an independent determination about each collection. | \n
The data-collection checklist, Schedule 3, PDF 16–19, is almost entirely redacted. PDF 7 also visibly blacks out portions of escalation and review procedure. These are real redactions, confirmed visually at PDF 7 and 16, not merely failed text extraction. The draft says a DPIA has been completed and is regularly reviewed (Schedule 2 Part A §2.2(d), PDF 11); the actual DPIA and its conclusions were not acquired. No attempt was made to recover redacted text.
\nIn HL3811, answered 30 April 2024, the government says DSIT has platform trusted-flagger status and may refer material within ministerial steers/remit that appears to violate platform terms; platforms decide the outcome. This establishes the department's publicly described referral channel, without proving a DSA statutory designation. [Answer](http[local research file]
\nHL7610, answered 3 June 2025, identifies foreign-state threats, election risks and AI/deepfakes among the team's subjects, with ministers agreeing high-risk work. 9335, answered 22 June 2026, restates the public-safety/national-security remit and continuing ministerial review. Neither answer publishes a commissioned topic list, the exact search terms or referral threshold scores. [2025 answer](http[local research file] [2026 answer](http[local research file]
\nThe current official privacy notice names DSIT as controller. It describes aggregate analysis of public information, incidental personal/special-category data, minimisation, no private-page review and no automated decision-making/profiling. It describes platform referrals, public-interest processing bases, retention limits and DPO/ICO complaint routes. It now names The Global Strategy Network Limited as an appointed analysis contractor. This is not an exclusive-provider statement, proof Crisp was replaced, or a specification for the February 2025 payment. The page displays publication on16 April 2024, but a revision date for this contractor paragraph was not established. Do not backdate the current wording. [Notice, controller, purpose, sharing, retention and rights sections](http[local research file]
\nThe distinction between public trend analysis and processing individual posts must stay visible: the policy's stated restrictions do not mean personal data is never processed. Equally, the existence of handles, links or political opinions in collected public content does not alone prove an individualized surveillance programme.
\nICO decision IC-374825-D9K4, 11 November 2025, confirms a request for Crisp reports within the week beginning 24 June 2024. DSIT said all but one had been deleted under retention policy; on 24 October 2025 it partially disclosed the remaining report and released a June referral count of ten (§§8–14, PDF 2–3). The ICO inspected the unredacted report (§18). DSIT described keyword tracking, behavioural analytics, platform strategies and threat-identification criteria (§36, PDF 11–14). Those are attributed submissions, not methods personally inspected here. The Commissioner upheld the remaining national-security redactions after considering public interest (§§39–43, PDF 15), with tribunal appeal available (§44, PDF 16). This is independent review of disclosure, not certification of NSOIT's operating legality. The underlying report, its exact date and any invoice linkage remain unread. Paragraph 36 calls it the week of the general election; that wording must not silently replace the request's week beginning 24 June 2024. [Original decision](http[local research file]
\nThe 14 April 2026 parliamentary hearing, Q25–30, PDF 8–11, supplies an unusually explicit operating account. Minister Kanishka Narayan described a public-safety/national-security threshold, declined the requested activity/removal figures, and offered future briefings subject to security constraints. At Q29, Talitha Rowland described Southport-era monitoring of lawful but harmful narratives, early warnings even before content appeared on a platform, and requests for review against each platform's own terms. She distinguished that work from law enforcement's illegal-content response. At Q30 she said they lacked specific foreign-interference indicators at the time; do not recast this example as proven foreign-state content. This supports a government-to-platform influence channel affecting lawful material, while platforms retain the described action decision. It does not demonstrate a removal order, Crisp's role in that example, or the resulting removal rate. [Hearing, Q25–30](http[local research file]
\nThe Science, Innovation and Technology Committee's July 2025 report, §§60–62, recommended statutory footing and Intelligence and Security Committee oversight. Its October government response, received 18 September and published 17 October 2025, instead defended ministerial oversight, existing parliamentary questioning, FOI and subject-access routes; it did not commit to the requested statutory/ISC reform. A general commitment to accountability is therefore not acceptance of that institutional safeguard. [Committee §62](http[local research file] [response, Appendix 1, Disinformation Campaigns, reply to §62](http[local research file]
\nThis is not an absence-of-all-review finding: MPs questioned ministers, the ICO examined withheld material, and stated data-rights routes exist. The missing layer is a published independent operational audit and the actual task/approval records needed to test compliance. Nor does the government's statement that platforms decide action settle the separate question of how official flagging, privileged access and regulatory relationships influence those decisions.
\n| Candidate ID | \nTyped relation | \nEvidence stage | \n
|---|---|---|
| commission-e-ministers | \nDSIT ministers → NSOIT: set/review remit | \nPublic ministerial account, 2025 and 2026. | \n
| commission-e-tasking | \nNSOIT / Cabinet Office NSS → collection activity: proposed tasking route | \nAcquired disclosed draft §3.2; no actual order. | \n
| commission-e-partner-instructions | \nDSIT/NSOIT → delivery partners: documented instruction/method agreement | \nDraft §§3.5–3.6; applicability to paid Crisp engagement unestablished. | \n
| commission-e-crisp-report | \nCrisp → NSOIT: produced report | \nIndependent disclosure adjudication; report itself unread. | \n
| commission-e-platform-referral | \nNSOIT → platforms: flags or warns for terms review | \nDepartmental accounts, 2024 and 2026; voluntary action described. | \n
| commission-e-ico-review | \nICO → DSIT disclosure decision: upheld remaining redactions | \n11 November 2025; not an operational audit. | \n
| commission-e-oversight-request | \nCommittee → government: requests statutory/ISC oversight | \nJuly 2025 recommendation; response offers existing channels without commitment to requested reform. | \n
| commission-e-current-contractor | \nNSOIT → Global Strategy Network Ltd: reports appointment for analysis | \nCurrent notice; appointment date and exact contract unestablished. | \n
The strongest rival to a private supplier choosing what citizens may say is an instructed contractor performing bounded public-source analysis, with government retaining tasking/referral responsibility and platforms retaining action. The strongest concern supported by these records is more specific: a government-selected monitoring and privileged referral process can reach lawful material, while precise criteria and authorization steps are not fully visible and a proposed stronger oversight structure was not adopted in the inspected response.
\nThe next discriminating record is the executed SOW/data-processing schedule/tasking instruction and completed assessment matching transaction 654609, followed by the escalation log and platform response for a specified case. A published final compliance policy, its adoption/version history, the DPIA and an independent assurance report would test whether the draft's constraints actually governed delivery. Likely holders are DSIT Commercial, NSOIT, the Operational Data/DPO team and the actual supplier. These are precise public-record gaps, not a claim that the records do not exist. No outreach was undertaken. Acquisition stops here for forest review.
\nThe original payment remains in the prior wave. This wave retained the following bytes; .txt files are extraction derivatives, not originals. The capture manifest is commission-captures.json.
\n| Source/file | \nCapture and SHA-256 | \nReading scope / state | \n
|---|---|---|
| commission-ico-decision.pdf | \n2026-09-09T09:01:46Z; 172637 bytes;1e33d2fcaa65f09cb4d045b04f85689e65db3779604fdcaa2e175477acd3ae56 | \nFull 16-page original text read, scoped findings above. | \n
| commission-privacy.html | \n2026-09-09T09:01:46Z; 96332 bytes;5891f9aa42d6c159d0c1a8b34ad286bbe536a53d93c94dcada1f3c46d3f5690c | \nComplete substantive current notice read through official reader; original HTML retained. | \n
| commission-compliance-redacted.pdf | \n2026-09-09T09:03:33Z; 264507 bytes;af10011b6bab2a29cc2c59225b9100aa37a9a9df3e14176dea5abcfbe8b26a79 | \nFull 19-page available text read; PDF 7 and 16 visually inspected. Public FOI-associated custodian copy; marked draft; no underlying DSIT file comparison. | \n
| 2026 oral hearing | \nExact parliamentary URL above; reader turn985view2 / turn987view1–4 | \nQ25–30 plus institutional context Q7 read. Ordinary GET 403; no original retained and no retry. | \n
| 2025 committee report / response | \nExact URLs above; report §§60–62; response metadata and relevant reply | \nPrimary reader text read; response ordinary GET 403. No original retained, no retry. | \n
| Parliamentary answers | \nHL3811, HL7610, 9335 exact URLs above | \nComplete question and answer text read through official reader. | \n
The BBW PDF reader initially returned a one-line 'One moment, please...' page. One ordinary GET substitute succeeded with an actual PDF; no challenge solving or access-control evasion occurred. The prior Contracts Finder 403, ACNC restriction and other closed routes were not retried. Searches also surfaced unrelated MOD 'New Style of IT' records; these were excluded because the acronym denotes a different programme. Additional spending months were not opened or totalled. No WEF taxonomy specification, supplier-controlled censorship right or February 2025 invoice match was found in the inspected material.
", "path": "implementation-instructions-wave-2026-09-09/uk-commissioning.md", "sourceIds": [ "w15-impl-s-ico-disclosure", "w15-impl-s-nsoit-draft-policy", "w15-impl-s-nsoit-privacy", "w15-impl-s-nsoit-oral", "w15-impl-s-trusted-flagger", "w15-impl-s-minister-scope", "w15-impl-s-remit-review", "w15-impl-s-oversight-request", "w15-impl-s-oversight-response" ], "markdown": "Bounded implementation investigation, 9 September 2026. The anchor remains DSIT's published **5 February 2025 payment, GBP 56,736, transaction 654609**, to Crisp Thinking (UK) Ltd. No record inspected here joins that payment to a particular report, monitoring specification or older contract. See the preceding commercial-participant.md and its retained original CSV; the payment is not counted again.\n\nThe newly visible mechanism is ministerial scope-setting, departmental commissioning and assessment, supplier analysis, then platform review under private terms. An acquired draft policy specifies documented instructions and agreed search methods for contractors. Contemporary parliamentary testimony describes both referrals and warnings about lawful content. The exact escalation approvals and classification thresholds remain partly redacted, and an executed Crisp SOW remains unestablished.\n\n## The instruction instrument, with its status intact\n\nA public FOI-associated copy of the **National Security and Online Information Team Data Protection and Compliance Policy** supplies detailed instructions. It is hosted by Big Brother Watch at a path identifying FOI2024-00576 and May 2024. Every page is marked **DRAFT, OFFICIAL-SENSITIVE**. No approval date, signed adoption, final revision or incorporation into the February 2025 engagement was established. Describe the following as provisions of this acquired draft, rather than proof every analyst or supplier followed them. Publicly hosted policy copy (see source references)\n\n| Who / decision | Draft provision | Exact PDF locator and limit |\n| --- | --- | --- |\n| NSOIT or central government: commission work | NSOIT may commission within its remit or be tasked through central command structures, with Cabinet Office National Security Secretariat given as an example. Sustained monitoring requires appropriate ministerial agreement. | §3.2, PDF 4. This is a route for tasking, not an actual task order. |\n| NSOIT Analysis team: assess a task | Every collection activity must have a legitimate policy objective, be proportionate, meet data-protection requirements and avoid directed surveillance. | §3.3, PDF 4; §§3.7–3.9, PDF 5. |\n| DSIT: instruct and onboard suppliers | Delivery partners must follow NSOIT instructions. DSIT takes controller responsibility where it processes personal data; third parties are to be processors under Article 28, operating on documented instructions. Legal, Commercial and Operational Data teams participate in onboarding. | §§3.4–3.6, PDF 4–5. This does not identify which Crisp entity signed any agreement. |\n| NSOIT and partner: agree methods | Case-specific methodology must cover data sources, search terms, task subject/purpose and minimisation. Contractual assurance and periodic review are specified. | §3.6(b)–(d), PDF 5. The actual search dictionary, classification specification and completed assurance record were not acquired. |\n| Analysts: collection limits | Publicly accessible sources only; no name-based searching or building a picture of private individuals/groups by following activity. These instructions are to be communicated to partners. | §3.1, PDF 4; Schedule 1, PDF 9. The draft permits access to public sites requiring login, while prohibiting fake personas. Schedule 2 Part B §§1.4–1.6, PDF 15. |\n| NSOIT: decide on referral | Teams identify potentially harmful material within ministerially agreed remit. NSOIT reviews against agreed thresholds. Authorized content may be shared for platforms to act at their discretion, with information minimised. | §§4.1–4.2, PDF 7. The threshold and authorization detail is partly blacked out: do not invent a named approving officer or automatic approval rule. |\n| Politicians, news sources and journalists: an exclusion | The draft excludes their content from platform referrals; its journalist definition depends on a news-publishing entity, editorial control and other institutional criteria. | §§1.6–1.7, PDF 3; §4.2(b), PDF 7; Schedule 1, PDF 9. This is not a blanket exemption for every citizen publishing commentary. |\n| Head of NSOIT and DPO: exceptional retention | Raw samples normally up to three months; output retention/review and justified extensions involve the Head and Data Protection Officer. Thematic leads manage expiry of escalation links. | §3.13, PDF 6; §4.4(A), PDF 7; Schedule 2 §2.5, PDF 13. It does not show retention complied with in an actual case. |\n| Legal advisers: surveillance boundary | The draft says DSIT has no directed-surveillance RIPA authorization; activity risking that threshold must not be authorized and uncertainty goes to legal advisers. | Schedule 2 Part B §§1.5–1.6, PDF 15. This is the draft's legal/operational account, not an independent determination about each collection. |\n\nThe data-collection checklist, Schedule 3, PDF 16–19, is almost entirely redacted. PDF 7 also visibly blacks out portions of escalation and review procedure. These are real redactions, confirmed visually at PDF 7 and 16, not merely failed text extraction. The draft says a DPIA has been completed and is regularly reviewed (Schedule 2 Part A §2.2(d), PDF 11); the actual DPIA and its conclusions were not acquired. No attempt was made to recover redacted text.\n\n## What current public accounts corroborate\n\nIn **HL3811, answered 30 April 2024**, the government says DSIT has platform trusted-flagger status and may refer material within ministerial steers/remit that appears to violate platform terms; platforms decide the outcome. This establishes the department's publicly described referral channel, without proving a DSA statutory designation. Answer (see source references)\n\n**HL7610, answered 3 June 2025**, identifies foreign-state threats, election risks and AI/deepfakes among the team's subjects, with ministers agreeing high-risk work. **9335, answered 22 June 2026**, restates the public-safety/national-security remit and continuing ministerial review. Neither answer publishes a commissioned topic list, the exact search terms or referral threshold scores. 2025 answer (see source references) 2026 answer (see source references)\n\nThe current official privacy notice names DSIT as controller. It describes aggregate analysis of public information, incidental personal/special-category data, minimisation, no private-page review and no automated decision-making/profiling. It describes platform referrals, public-interest processing bases, retention limits and DPO/ICO complaint routes. It now names **The Global Strategy Network Limited** as an appointed analysis contractor. This is not an exclusive-provider statement, proof Crisp was replaced, or a specification for the February 2025 payment. The page displays publication on16 April 2024, but a revision date for this contractor paragraph was not established. Do not backdate the current wording. Notice, controller, purpose, sharing, retention and rights sections (see source references)\n\nThe distinction between public trend analysis and processing individual posts must stay visible: the policy's stated restrictions do not mean personal data is never processed. Equally, the existence of handles, links or political opinions in collected public content does not alone prove an individualized surveillance programme.\n\n## Actual supplier output and disclosure review\n\n**ICO decision IC-374825-D9K4, 11 November 2025**, confirms a request for Crisp reports within the week beginning 24 June 2024. DSIT said all but one had been deleted under retention policy; on 24 October 2025 it partially disclosed the remaining report and released a June referral count of ten (§§8–14, PDF 2–3). The ICO inspected the unredacted report (§18). DSIT described keyword tracking, behavioural analytics, platform strategies and threat-identification criteria (§36, PDF 11–14). Those are attributed submissions, not methods personally inspected here. The Commissioner upheld the remaining national-security redactions after considering public interest (§§39–43, PDF 15), with tribunal appeal available (§44, PDF 16). This is independent review of disclosure, not certification of NSOIT's operating legality. The underlying report, its exact date and any invoice linkage remain unread. Paragraph 36 calls it the week of the general election; that wording must not silently replace the request's week beginning 24 June 2024. Original decision (see source references)\n\n## Referral, lawful speech and the limits of oversight\n\nThe **14 April 2026 parliamentary hearing**, Q25–30, PDF 8–11, supplies an unusually explicit operating account. Minister Kanishka Narayan described a public-safety/national-security threshold, declined the requested activity/removal figures, and offered future briefings subject to security constraints. At Q29, Talitha Rowland described Southport-era monitoring of lawful but harmful narratives, early warnings even before content appeared on a platform, and requests for review against each platform's own terms. She distinguished that work from law enforcement's illegal-content response. At Q30 she said they lacked specific foreign-interference indicators at the time; do not recast this example as proven foreign-state content. This supports a government-to-platform influence channel affecting lawful material, while platforms retain the described action decision. It does not demonstrate a removal order, Crisp's role in that example, or the resulting removal rate. Hearing, Q25–30 (see source references)\n\nThe Science, Innovation and Technology Committee's **July 2025 report**, §§60–62, recommended statutory footing and Intelligence and Security Committee oversight. Its October government response, received 18 September and published 17 October 2025, instead defended ministerial oversight, existing parliamentary questioning, FOI and subject-access routes; it did not commit to the requested statutory/ISC reform. A general commitment to accountability is therefore not acceptance of that institutional safeguard. Committee §62 (see source references) response, Appendix 1, Disinformation Campaigns, reply to §62 (see source references)\n\nThis is not an absence-of-all-review finding: MPs questioned ministers, the ICO examined withheld material, and stated data-rights routes exist. The missing layer is a published independent operational audit and the actual task/approval records needed to test compliance. Nor does the government's statement that platforms decide action settle the separate question of how official flagging, privileged access and regulatory relationships influence those decisions.\n\n## Candidate relations and the discriminating gap\n\n| Candidate ID | Typed relation | Evidence stage |\n| --- | --- | --- |\n| commission-e-ministers | DSIT ministers → NSOIT: set/review remit | Public ministerial account, 2025 and 2026. |\n| commission-e-tasking | NSOIT / Cabinet Office NSS → collection activity: proposed tasking route | Acquired disclosed draft §3.2; no actual order. |\n| commission-e-partner-instructions | DSIT/NSOIT → delivery partners: documented instruction/method agreement | Draft §§3.5–3.6; applicability to paid Crisp engagement unestablished. |\n| commission-e-crisp-report | Crisp → NSOIT: produced report | Independent disclosure adjudication; report itself unread. |\n| commission-e-platform-referral | NSOIT → platforms: flags or warns for terms review | Departmental accounts, 2024 and 2026; voluntary action described. |\n| commission-e-ico-review | ICO → DSIT disclosure decision: upheld remaining redactions | 11 November 2025; not an operational audit. |\n| commission-e-oversight-request | Committee → government: requests statutory/ISC oversight | July 2025 recommendation; response offers existing channels without commitment to requested reform. |\n| commission-e-current-contractor | NSOIT → Global Strategy Network Ltd: reports appointment for analysis | Current notice; appointment date and exact contract unestablished. |\n\nThe strongest rival to a private supplier choosing what citizens may say is an instructed contractor performing bounded public-source analysis, with government retaining tasking/referral responsibility and platforms retaining action. The strongest concern supported by these records is more specific: a government-selected monitoring and privileged referral process can reach lawful material, while precise criteria and authorization steps are not fully visible and a proposed stronger oversight structure was not adopted in the inspected response.\n\nThe next discriminating record is the executed SOW/data-processing schedule/tasking instruction and completed assessment matching transaction 654609, followed by the escalation log and platform response for a specified case. A published final compliance policy, its adoption/version history, the DPIA and an independent assurance report would test whether the draft's constraints actually governed delivery. Likely holders are DSIT Commercial, NSOIT, the Operational Data/DPO team and the actual supplier. These are precise public-record gaps, not a claim that the records do not exist. No outreach was undertaken. Acquisition stops here for forest review.\n\n## Source custody and access limits\n\nThe original payment remains in the prior wave. This wave retained the following bytes; .txt files are extraction derivatives, not originals. The capture manifest is commission-captures.json.\n\n| Source/file | Capture and SHA-256 | Reading scope / state |\n| --- | --- | --- |\n| commission-ico-decision.pdf | 2026-09-09T09:01:46Z; 172637 bytes;1e33d2fcaa65f09cb4d045b04f85689e65db3779604fdcaa2e175477acd3ae56 | Full 16-page original text read, scoped findings above. |\n| commission-privacy.html | 2026-09-09T09:01:46Z; 96332 bytes;5891f9aa42d6c159d0c1a8b34ad286bbe536a53d93c94dcada1f3c46d3f5690c | Complete substantive current notice read through official reader; original HTML retained. |\n| commission-compliance-redacted.pdf | 2026-09-09T09:03:33Z; 264507 bytes;af10011b6bab2a29cc2c59225b9100aa37a9a9df3e14176dea5abcfbe8b26a79 | Full 19-page available text read; PDF 7 and 16 visually inspected. Public FOI-associated custodian copy; marked draft; no underlying DSIT file comparison. |\n| 2026 oral hearing | Exact parliamentary URL above; reader turn985view2 / turn987view1–4 | Q25–30 plus institutional context Q7 read. Ordinary GET 403; no original retained and no retry. |\n| 2025 committee report / response | Exact URLs above; report §§60–62; response metadata and relevant reply | Primary reader text read; response ordinary GET 403. No original retained, no retry. |\n| Parliamentary answers | HL3811, HL7610, 9335 exact URLs above | Complete question and answer text read through official reader. |\n\nThe BBW PDF reader initially returned a one-line 'One moment, please...' page. One ordinary GET substitute succeeded with an actual PDF; no challenge solving or access-control evasion occurred. The prior Contracts Finder 403, ACNC restriction and other closed routes were not retried. Searches also surfaced unrelated MOD 'New Style of IT' records; these were excluded because the acronym denotes a different programme. Additional spending months were not opened or totalled. No WEF taxonomy specification, supplier-controlled censorship right or February 2025 invoice match was found in the inspected material.\n", "file": "research/implementation-instructions-wave-2026-09-09-uk-commissioning.md" }, { "id": "implementation-wave-2026-09-09-commercial-participant", "title": "Crisp, Resolver and the commercial implementation layer", "description": "Bounded public-record investigation, 9 September 2026, following the taxonomy forest review. This packet separates legal control, brand integration, advertised services and an actual departmental spending entry. It does not establish that WEF participation produced a contract or that any regulator requires a Resolver product.", "html": "Bounded public-record investigation, 9 September 2026, following the taxonomy forest review. This packet separates legal control, brand integration, advertised services and an actual departmental spending entry. It does not establish that WEF participation produced a contract or that any regulator requires a Resolver product.
\nThe decisive addition is a primary DSIT record of GBP 56,736 paid to Crisp Thinking (UK) Ltd on 5 February 2025, transaction 654609, within Security and Online Harms. This supplies a concrete government-to-supplier financial relation. It does not disclose the invoice, deliverables, classification rules or contract that generated that payment. Crisp's contribution to the 2023 typology, its later Resolver branding, and the 2025 payment must remain separate dated relations. [Official February dataset](http[local research file]
\n| Date / entity | \nWhat the primary record establishes | \nLimit | \n
|---|---|---|
| Crisp Thinking Group Limited, 05636614 | \nThe current Crisp privacy policy names this company, trading as Crisp. Companies House identifies the same numbered company. | \nA current privacy-policy controller is not automatically the signatory of every historic sales contract. [Policy](http[local research file] [register](http[local research file] | \n
| Crisp Thinking (UK) Limited, 06442193 | \nCompanies House records Crisp Thinking Group Ltd, 05636614, as the active corporate person with significant control, notified 6 April 2016, with 75% or more of shares. | \nThe displayed record does not establish an exact 100% holding or separately display voting-control rights. [UK company PSC record](http[local research file] | \n
| Crisp Thinking (US) Inc. | \nThe current Crisp policy identifies it as another group company, alongside Group Limited and UK Limited. | \nUS registration identifier, jurisdiction and equity chain were not independently established. | \n
| 30 March 2022: Kroll / Resolver | \nKroll announces that it acquired Resolver; the subsequent Crisp announcement expressly calls the earlier target Resolver Inc. | \nThese are acquisition announcements, not share-purchase instruments. Price, exact buyer entity and acquired percentage are not disclosed in the inspected text. [Resolver acquisition announcement](http[local research file] | \n
| 25 May 2022: Crisp Group control | \nThe register records Kroll International (UK) Limited, 09897805, notified on this date, with 75% or more of shares and voting rights. | \nRegistered notification date is distinct from the later public announcement. [Group PSC record](http[local research file] | \n
| 31 May 2022: Kroll / Crisp announcement | \nKroll announces acquisition of Crisp and an intended combination of its real-time risk intelligence with the previously acquired Resolver software. Hildreth and Andrew Burke were to join Kroll's digital-services leadership. | \nThis identifies an announced integration mechanism, not a transaction value or a mandate to public bodies. [Crisp acquisition announcement](http[local research file] | \n
| 29 December 2022: Crisp Group control changes | \nKroll International (UK) Limited is recorded as ceasing control; Kroll Advisory Ltd., 05568550, is the active recorded controller, notified that day, with 75% or more of shares and voting rights. | \nThis is the directly evidenced intermediate control chain. It does not establish ultimate ownership, exact percentages, consideration or the legal mechanism of the transfer. | \n
| 19/20 March 2024: brand alignment | \nResolver's page is posted 19 March, while its body dateline is Toronto, 20 March. It says Crisp will operate under the Resolver banner and describes unification of brands, including Platform Trust & Safety and Brand Equity Protection. | \nPreserve both displayed dates. This is not evidence of a new 2024 share acquisition. [Brand announcement](http[local research file] | \n
| Resolver Inc.: currently served contract form | \nResolver's terms identify Resolver Inc. and applicable affiliates; the customer is identified on an Order Form. The served version is 3.0, dated 20 August 2021. | \nAn unsigned standard form does not resolve the party to a particular Crisp or government engagement. Its Toronto notice address alone does not establish incorporation jurisdiction. [Terms, preamble and §§1, 5, 17](http[local research file] | \n
Companies House expressly cautions that it does not check the accuracy of filed information. The control findings above are what its public record reports, rather than a beneficial-ownership audit. Do not replace the exact entities with an undifferentiated Kroll parent or treat the 2026 footer as evidence that Resolver Inc. authored or contracted for the 2023 workstream.
\nCrisp's current policy describes collecting and indexing publicly available material for harm detection, channel moderation and crisis detection. It distinguishes acting as a processor on customer-owned channels from acting as a controller for its crisis-detection search/index. It describes algorithms and human experts, customer-specified automatic removal, escalation of more consequential cases, and ways to challenge decisions. These are the supplier's published descriptions, not an independent audit of deployment, data access or lawfulness. The policy supplies a concrete route from classification to operational action, but does not identify a WEF definition as binding on its customers. No policy revision date was established; its current text is not backdated to 2023. [Crisp policy, Our Services and automated-decision paragraphs](http[local research file]
\nThe Resolver form explains how a specific purchase would become accountable: an Order Form and, where used, a Statement of Work specify services, responsibilities, deliverables and payment terms; written agreement governs changes. The missing executed order, SOW and processing instructions would therefore discriminate between general capability and the actual system a customer bought. [Resolver terms, §§1 and 5](http[local research file]
\nA separate 18 March 2026 announcement documents a collaboration between Resolver Trust & Safety and Illuminate Tech. Illuminate supplies OSCAR, a tool for translating online-safety requirements into compliance workflows; Resolver supplies harm assessment, intelligence, assurance and remediation expertise. The named market includes platforms facing UK OSA and EU DSA enforcement. The statement says joint industry engagement had begun. It describes Illuminate as founded by former Ofcom experts, but supplies neither an official endorsement nor a named purchaser, compulsory tool, contract value or WEF-related procurement decision. This is direct evidence of a commercial compliance offering and a supplier-reported collaboration. It is not an award. [Announcement, opening through joint-engagement paragraph](http[local research file]
\nThe original February 2025 CSV contains one matching row in this file: data row 94, transaction 654609, payment date 05/02/2025, supplier Crisp Thinking (Uk) Ltd, amount 56736, supplier type Vendor. Expense type is Other Professional Services; expense area is DSIT's Digital and Technology Group / Security and Online Harms. The description assigns it to the Security Team NSOIT & Tech Programme. The supplier name matches the separately registered UK company, but this dataset supplies no company number. Record that cross-source identity matching rather than pretending the CSV itself contains 06442193.
\nThe amount is GBP 56,736, published departmental payment, dated 5 February 2025. It is neither total contract value nor verified net revenue; VAT treatment, invoice period and exact purchased outputs are unestablished. No join to a particular 2023 award, WEF workstream, Australian duty or OSCAR deployment is supported. The collection was first published 29 January 2026 and last updated 30 July 2026; those publication dates are distinct from payment date. The update history says July added later months and does not establish a revision to this specific row. [Original CSV](http[local research file] [publisher and update history](http[local research file]
\nA separate secondary listing discovers an award called Mis/Disinformation Monitoring and Analysis, reference 103528, supplier abbreviated Crisp Thinking, award publication 27 September 2023, delivery 24 August 2023–31 March 2024, listed value GBP 350,960. It describes DCMS handling the notice for DSIT. These remain secondary-reported lead fields, not a primary-verified award or money paid. The actual notice is [Contracts Finder 47f6e657-8370-4133-98df-ee2f2194144e](http[local research file] its web reader failed and the one ordinary GET substitute returned HTTP 403. The route is closed. [Discovery listing](http[local research file]
\nA DCMS FOI response, dated 3 October 2025 and published 7 October, refused a broad request naming Crisp and many other suppliers under the cost limit. The request's list is not evidence of agreements held with every named company, and the refusal does not prove no agreement exists. It is excluded from positive procurement findings. [FOI2025/08916](http[local research file]
\nThe previous packet establishes Hildreth's Crisp affiliation and workstream co-chair role through the participant's 24 August 2023 account. This packet establishes a control chain, a later brand change, advertised compliance services and an independently published departmental payment. Those relations can coexist without a causal procurement edge. [Dated typology account, opening](http[local research file] see the prior taxonomy-producers.md for its full reading scope.
\n| Candidate ID | \nDirection / type | \nStage and limit | \n
|---|---|---|
| commercial-e-group-uk | \nCrisp Thinking Group Limited → Crisp Thinking (UK) Limited: recorded share control, ≥75% | \nRegistry report, notified 6 April 2016; exact percentage unknown. | \n
| commercial-e-kroll-international | \nKroll International (UK) Limited → Crisp Thinking Group Limited: recorded shares/votes ≥75% | \nNotified 25 May 2022; ceased 29 December 2022. | \n
| commercial-e-kroll-advisory | \nKroll Advisory Ltd. → Crisp Thinking Group Limited: recorded shares/votes ≥75% | \nNotified 29 December 2022; active in inspected register. | \n
| commercial-e-resolver-acquisition | \nKroll → Resolver: announced acquisition | \n30 March 2022; buyer legal entity and consideration not resolved. | \n
| commercial-e-crisp-acquisition | \nKroll → Crisp: announced acquisition and integration | \n31 May 2022; use the separate registry edges for exact entities and rights. | \n
| commercial-e-brand | \nCrisp → Resolver: announced brand alignment | \nPosted 19 March / datelined 20 March 2024; not a new share-purchase claim. | \n
| commercial-e-dsit-payment | \nDSIT → Crisp Thinking (UK) Ltd: published supplier payment | \nAmount object: currency GBP; value 56736; kind departmental payment; date 2025-02-05; VAT treatment unknown; transaction 654609. Identity matched by supplier name to register. | \n
| commercial-e-illuminate | \nResolver Trust & Safety ↔ Illuminate Tech: announced compliance-service collaboration | \n18 March 2026 participant statement; OSCAR creator is Illuminate; no procurement or regulatory approval edge. | \n
| commercial-e-service | \nCrisp → customers: describes harm-monitoring/moderation services | \nCurrent published policy and customer instructions; actual customer configuration not observed. | \n
The strongest rival to a workstream-to-revenue causal account is an existing commercial provider contributing expertise to a voluntary taxonomy while independently selling services and later entering a government payment relationship. Its commercial interest is real and relevant. The acquired entities and payment make that interest more concrete, but a roster, chronology and payment cannot determine why a buyer selected it or which taxonomy governed the work.
\nThe decisive missing records are an executed DSIT contract/SOW and invoice matching transaction 654609, supplier-selection or evaluation records, and any deliverable or contractual classification specification that expressly adopts the WEF typology. Likely holders are DSIT's finance/commercial and Security and Online Harms teams and the relevant Crisp contracting company. A public contract, invoice disclosure, evaluation report or published operating specification would reopen that test. No outreach is authorized or performed. The secondary 2023 notice is a locator for a future genuinely accessible primary record, not permission to retry a closed route.
\nAccess date: 9 September 2026. Only the DSIT HTML and CSV original bytes were retained in this lane. Other listed sources were read through the primary web reader, not downloaded; no original hash is claimed for them. The prior WEF original-PDF and ACNC restrictions remain closed.
\n| Source alias | \nExact locator / reading scope | \nState | \n
|---|---|---|
| commercial-s-crisp-policy | \nhttp[local research file] — identity paragraphs and complete Services section; reader turn951view1 lines 10–14, 104–151 | \nRead current primary policy; undated revision. | \n
| commercial-s-psc-group | \nhttp[local research file] — Kroll corporate entries, reader turn960view0 lines 27–81 | \nRead register; original filed PSC forms not read. | \n
| commercial-s-psc-uk | \nhttp[local research file] — current corporate entry, turn960view1 lines 26–51 | \nRead register; no exact 100% inference. | \n
| commercial-s-acquire-resolver | \nKroll Resolver acquisition URL above; dated body, turn963view2 lines 6–18 | \nRead company announcement. | \n
| commercial-s-acquire-crisp | \nKroll Crisp acquisition URL above; dated substantive body, turn954view0 | \nRead company announcement. | \n
| commercial-s-brand | \nResolver brand announcement URL above; turn963view3 lines 5–19 | \nRead company announcement; two displayed dates preserved. | \n
| commercial-s-terms | \nResolver terms URL above; preamble, §§1, 5, 15, 17; turn953view1/turn956view3 | \nRead standard form, version 3.0 dated 20 August 2021; no executed agreement. | \n
| commercial-s-illuminate | \nResolver partnership URL above; turn963view1 lines 5–33 | \nRead complete substantive announcement; no customer list audited. | \n
| commercial-s-dsit | \nOfficial CSV/preview URLs above; header and complete matching row; preview turn961view0 line 62 and turn962view2 line 157; CSV data row 94 | \nRead original CSV selected row; full bytes retained, other rows not substantively audited. | \n
| commercial-s-dsit-publisher | \nDSIT 2025 spending collection above; turn963view0 lines 53–64, 146–187 | \nRead title, publication/update metadata and explanation. | \n
| commercial-s-award-lead | \nContracts Finder notice above, discovered through Bidstats listing | \nPrimary inaccessible; web reader Internal Error, ordinary requests GET HTTPError 403. No original retained and no retry. | \n
| commercial-s-foi | \nFOI2025/08916 above, request and response | \nRead primary release; excluded as evidence of any specific contract held. | \n
Original captures at 2026-09-09T08:56:13Z (start of capture operation):
\n| Local file | \nBytes | \nSHA-256 | \n
|---|---|---|
| commercial-dsit-feb2025-preview.html | \n186361 | \nf6d1e7955190b7099d6968efc0dd41a272e2c51eb15a259fe464b173e2033d53 | \n
| commercial-dsit-feb2025.csv | \n43426 | \nee37e522067dcb3304a3e77857f5cb37f5c7d960b7898128a4b245d795f83391 | \n
The reader's Download CSV click returned Internal Error; the one ordinary publisher-HTML retrieval succeeded, exposed the actual download link, and the linked original CSV succeeded. This was an independent DSIT spending publication, not a substitute route around Contracts Finder's 403. No later month, historical funder roster or unrelated client hunt was opened after this substantive breakpoint. Return to whole-case forest assessment before another acquisition run.
", "path": "implementation-wave-2026-09-09/commercial-participant.md", "sourceIds": [ "w15-impl-s-crisp-policy", "w15-impl-s-crisp-group-psc", "w15-impl-s-crisp-uk-psc", "w15-impl-s-crisp-acquisition", "w15-impl-s-resolver-brand", "w15-impl-s-resolver-terms", "w15-impl-s-illuminate", "w15-impl-s-dsit-payment", "w15-impl-s-dsit-publisher", "w15-impl-s-award-unavailable", "w15-impl-s-award-discovery" ], "markdown": "Bounded public-record investigation, 9 September 2026, following the taxonomy forest review. This packet separates legal control, brand integration, advertised services and an actual departmental spending entry. It does not establish that WEF participation produced a contract or that any regulator requires a Resolver product.\n\nThe decisive addition is a primary DSIT record of **GBP 56,736 paid to Crisp Thinking (UK) Ltd on 5 February 2025**, transaction 654609, within Security and Online Harms. This supplies a concrete government-to-supplier financial relation. It does not disclose the invoice, deliverables, classification rules or contract that generated that payment. Crisp's contribution to the 2023 typology, its later Resolver branding, and the 2025 payment must remain separate dated relations. Official February dataset (see source references)\n\n## Legal entities, rights and dates\n\n| Date / entity | What the primary record establishes | Limit |\n| --- | --- | --- |\n| Crisp Thinking Group Limited, 05636614 | The current Crisp privacy policy names this company, trading as Crisp. Companies House identifies the same numbered company. | A current privacy-policy controller is not automatically the signatory of every historic sales contract. Policy (see source references) register (see source references) |\n| Crisp Thinking (UK) Limited, 06442193 | Companies House records Crisp Thinking Group Ltd, 05636614, as the active corporate person with significant control, notified 6 April 2016, with **75% or more of shares**. | The displayed record does not establish an exact 100% holding or separately display voting-control rights. UK company PSC record (see source references) |\n| Crisp Thinking (US) Inc. | The current Crisp policy identifies it as another group company, alongside Group Limited and UK Limited. | US registration identifier, jurisdiction and equity chain were not independently established. |\n| 30 March 2022: Kroll / Resolver | Kroll announces that it acquired Resolver; the subsequent Crisp announcement expressly calls the earlier target Resolver Inc. | These are acquisition announcements, not share-purchase instruments. Price, exact buyer entity and acquired percentage are not disclosed in the inspected text. Resolver acquisition announcement (see source references) |\n| 25 May 2022: Crisp Group control | The register records **Kroll International (UK) Limited, 09897805**, notified on this date, with **75% or more of shares and voting rights**. | Registered notification date is distinct from the later public announcement. Group PSC record (see source references) |\n| 31 May 2022: Kroll / Crisp announcement | Kroll announces acquisition of Crisp and an intended combination of its real-time risk intelligence with the previously acquired Resolver software. Hildreth and Andrew Burke were to join Kroll's digital-services leadership. | This identifies an announced integration mechanism, not a transaction value or a mandate to public bodies. Crisp acquisition announcement (see source references) |\n| 29 December 2022: Crisp Group control changes | Kroll International (UK) Limited is recorded as ceasing control; **Kroll Advisory Ltd., 05568550**, is the active recorded controller, notified that day, with **75% or more of shares and voting rights**. | This is the directly evidenced intermediate control chain. It does not establish ultimate ownership, exact percentages, consideration or the legal mechanism of the transfer. |\n| 19/20 March 2024: brand alignment | Resolver's page is posted 19 March, while its body dateline is Toronto, 20 March. It says Crisp will operate under the Resolver banner and describes unification of brands, including Platform Trust & Safety and Brand Equity Protection. | Preserve both displayed dates. This is not evidence of a new 2024 share acquisition. Brand announcement (see source references) |\n| Resolver Inc.: currently served contract form | Resolver's terms identify Resolver Inc. and applicable affiliates; the customer is identified on an Order Form. The served version is 3.0, dated 20 August 2021. | An unsigned standard form does not resolve the party to a particular Crisp or government engagement. Its Toronto notice address alone does not establish incorporation jurisdiction. Terms, preamble and §§1, 5, 17 (see source references) |\n\nCompanies House expressly cautions that it does not check the accuracy of filed information. The control findings above are what its public record reports, rather than a beneficial-ownership audit. Do not replace the exact entities with an undifferentiated Kroll parent or treat the 2026 footer as evidence that Resolver Inc. authored or contracted for the 2023 workstream.\n\n## What is sold, and who sets operational instructions\n\nCrisp's current policy describes collecting and indexing publicly available material for harm detection, channel moderation and crisis detection. It distinguishes acting as a processor on customer-owned channels from acting as a controller for its crisis-detection search/index. It describes algorithms and human experts, customer-specified automatic removal, escalation of more consequential cases, and ways to challenge decisions. These are the supplier's published descriptions, not an independent audit of deployment, data access or lawfulness. The policy supplies a concrete route from classification to operational action, but does not identify a WEF definition as binding on its customers. No policy revision date was established; its current text is not backdated to 2023. Crisp policy, Our Services and automated-decision paragraphs (see source references)\n\nThe Resolver form explains how a specific purchase would become accountable: an Order Form and, where used, a Statement of Work specify services, responsibilities, deliverables and payment terms; written agreement governs changes. The missing executed order, SOW and processing instructions would therefore discriminate between general capability and the actual system a customer bought. Resolver terms, §§1 and 5 (see source references)\n\nA separate **18 March 2026** announcement documents a collaboration between Resolver Trust & Safety and **Illuminate Tech**. Illuminate supplies **OSCAR**, a tool for translating online-safety requirements into compliance workflows; Resolver supplies harm assessment, intelligence, assurance and remediation expertise. The named market includes platforms facing UK OSA and EU DSA enforcement. The statement says joint industry engagement had begun. It describes Illuminate as founded by former Ofcom experts, but supplies neither an official endorsement nor a named purchaser, compulsory tool, contract value or WEF-related procurement decision. This is direct evidence of a commercial compliance offering and a supplier-reported collaboration. It is not an award. Announcement, opening through joint-engagement paragraph (see source references)\n\n## A departmental payment, and a separate unverified award lead\n\nThe original February 2025 CSV contains one matching row in this file: **data row 94**, transaction **654609**, payment date **05/02/2025**, supplier **Crisp Thinking (Uk) Ltd**, amount **56736**, supplier type **Vendor**. Expense type is Other Professional Services; expense area is DSIT's Digital and Technology Group / Security and Online Harms. The description assigns it to the Security Team NSOIT & Tech Programme. The supplier name matches the separately registered UK company, but this dataset supplies no company number. Record that cross-source identity matching rather than pretending the CSV itself contains 06442193.\n\nThe amount is **GBP 56,736, published departmental payment, dated 5 February 2025**. It is neither total contract value nor verified net revenue; VAT treatment, invoice period and exact purchased outputs are unestablished. No join to a particular 2023 award, WEF workstream, Australian duty or OSCAR deployment is supported. The collection was first published 29 January 2026 and last updated 30 July 2026; those publication dates are distinct from payment date. The update history says July added later months and does not establish a revision to this specific row. Original CSV (see source references) publisher and update history (see source references)\n\nA separate secondary listing discovers an award called **Mis/Disinformation Monitoring and Analysis**, reference 103528, supplier abbreviated Crisp Thinking, award publication 27 September 2023, delivery 24 August 2023–31 March 2024, listed value GBP 350,960. It describes DCMS handling the notice for DSIT. These remain **secondary-reported lead fields**, not a primary-verified award or money paid. The actual notice is Contracts Finder 47f6e657-8370-4133-98df-ee2f2194144e (see source references) its web reader failed and the one ordinary GET substitute returned HTTP 403. The route is closed. Discovery listing (see source references)\n\nA DCMS FOI response, dated 3 October 2025 and published 7 October, refused a broad request naming Crisp and many other suppliers under the cost limit. The request's list is not evidence of agreements held with every named company, and the refusal does not prove no agreement exists. It is excluded from positive procurement findings. FOI2025/08916 (see source references)\n\n## The join test and candidate relations\n\nThe previous packet establishes Hildreth's Crisp affiliation and workstream co-chair role through the participant's 24 August 2023 account. This packet establishes a control chain, a later brand change, advertised compliance services and an independently published departmental payment. Those relations can coexist without a causal procurement edge. Dated typology account, opening (see source references) see the prior taxonomy-producers.md for its full reading scope.\n\n| Candidate ID | Direction / type | Stage and limit |\n| --- | --- | --- |\n| commercial-e-group-uk | Crisp Thinking Group Limited → Crisp Thinking (UK) Limited: recorded share control, ≥75% | Registry report, notified 6 April 2016; exact percentage unknown. |\n| commercial-e-kroll-international | Kroll International (UK) Limited → Crisp Thinking Group Limited: recorded shares/votes ≥75% | Notified 25 May 2022; ceased 29 December 2022. |\n| commercial-e-kroll-advisory | Kroll Advisory Ltd. → Crisp Thinking Group Limited: recorded shares/votes ≥75% | Notified 29 December 2022; active in inspected register. |\n| commercial-e-resolver-acquisition | Kroll → Resolver: announced acquisition | 30 March 2022; buyer legal entity and consideration not resolved. |\n| commercial-e-crisp-acquisition | Kroll → Crisp: announced acquisition and integration | 31 May 2022; use the separate registry edges for exact entities and rights. |\n| commercial-e-brand | Crisp → Resolver: announced brand alignment | Posted 19 March / datelined 20 March 2024; not a new share-purchase claim. |\n| commercial-e-dsit-payment | DSIT → Crisp Thinking (UK) Ltd: published supplier payment | Amount object: currency GBP; value 56736; kind departmental payment; date 2025-02-05; VAT treatment unknown; transaction 654609. Identity matched by supplier name to register. |\n| commercial-e-illuminate | Resolver Trust & Safety ↔ Illuminate Tech: announced compliance-service collaboration | 18 March 2026 participant statement; OSCAR creator is Illuminate; no procurement or regulatory approval edge. |\n| commercial-e-service | Crisp → customers: describes harm-monitoring/moderation services | Current published policy and customer instructions; actual customer configuration not observed. |\n\nThe strongest rival to a workstream-to-revenue causal account is an existing commercial provider contributing expertise to a voluntary taxonomy while independently selling services and later entering a government payment relationship. Its commercial interest is real and relevant. The acquired entities and payment make that interest more concrete, but a roster, chronology and payment cannot determine why a buyer selected it or which taxonomy governed the work.\n\nThe decisive missing records are an executed DSIT contract/SOW and invoice matching transaction 654609, supplier-selection or evaluation records, and any deliverable or contractual classification specification that expressly adopts the WEF typology. Likely holders are DSIT's finance/commercial and Security and Online Harms teams and the relevant Crisp contracting company. A public contract, invoice disclosure, evaluation report or published operating specification would reopen that test. No outreach is authorized or performed. The secondary 2023 notice is a locator for a future genuinely accessible primary record, not permission to retry a closed route.\n\n## Source scope and custody\n\nAccess date: 9 September 2026. Only the DSIT HTML and CSV original bytes were retained in this lane. Other listed sources were read through the primary web reader, not downloaded; no original hash is claimed for them. The prior WEF original-PDF and ACNC restrictions remain closed.\n\n| Source alias | Exact locator / reading scope | State |\n| --- | --- | --- |\n| commercial-s-crisp-policy | http[local research file] — identity paragraphs and complete Services section; reader turn951view1 lines 10–14, 104–151 | Read current primary policy; undated revision. |\n| commercial-s-psc-group | http[local research file] — Kroll corporate entries, reader turn960view0 lines 27–81 | Read register; original filed PSC forms not read. |\n| commercial-s-psc-uk | http[local research file] — current corporate entry, turn960view1 lines 26–51 | Read register; no exact 100% inference. |\n| commercial-s-acquire-resolver | Kroll Resolver acquisition URL above; dated body, turn963view2 lines 6–18 | Read company announcement. |\n| commercial-s-acquire-crisp | Kroll Crisp acquisition URL above; dated substantive body, turn954view0 | Read company announcement. |\n| commercial-s-brand | Resolver brand announcement URL above; turn963view3 lines 5–19 | Read company announcement; two displayed dates preserved. |\n| commercial-s-terms | Resolver terms URL above; preamble, §§1, 5, 15, 17; turn953view1/turn956view3 | Read standard form, version 3.0 dated 20 August 2021; no executed agreement. |\n| commercial-s-illuminate | Resolver partnership URL above; turn963view1 lines 5–33 | Read complete substantive announcement; no customer list audited. |\n| commercial-s-dsit | Official CSV/preview URLs above; header and complete matching row; preview turn961view0 line 62 and turn962view2 line 157; CSV data row 94 | Read original CSV selected row; full bytes retained, other rows not substantively audited. |\n| commercial-s-dsit-publisher | DSIT 2025 spending collection above; turn963view0 lines 53–64, 146–187 | Read title, publication/update metadata and explanation. |\n| commercial-s-award-lead | Contracts Finder notice above, discovered through Bidstats listing | Primary inaccessible; web reader Internal Error, ordinary requests GET HTTPError 403. No original retained and no retry. |\n| commercial-s-foi | FOI2025/08916 above, request and response | Read primary release; excluded as evidence of any specific contract held. |\n\nOriginal captures at 2026-09-09T08:56:13Z (start of capture operation):\n\n| Local file | Bytes | SHA-256 |\n| --- | --- | --- |\n| commercial-dsit-feb2025-preview.html | 186361 | f6d1e7955190b7099d6968efc0dd41a272e2c51eb15a259fe464b173e2033d53 |\n| commercial-dsit-feb2025.csv | 43426 | ee37e522067dcb3304a3e77857f5cb37f5c7d960b7898128a4b245d795f83391 |\n\nThe reader's Download CSV click returned Internal Error; the one ordinary publisher-HTML retrieval succeeded, exposed the actual download link, and the linked original CSV succeeded. This was an independent DSIT spending publication, not a substitute route around Contracts Finder's 403. No later month, historical funder roster or unrelated client hunt was opened after this substantive breakpoint. Return to whole-case forest assessment before another acquisition run.\n", "file": "research/implementation-wave-2026-09-09-commercial-participant.md" }, { "id": "kybc-policy-wave-2026-09-09-disposition-rules", "title": "UK KYBC: government consideration, narrower advertising proposals, and a 2026 industry pathway", "description": "9 September 2026. Bounded acquisition CLOSED. No canonical case, reader or ZIP changed. This is a source-scoped policy comparison, not a legal opinion or an exhaustive current UK statute search.", "html": "9 September 2026. Bounded acquisition CLOSED. No canonical case, reader or ZIP changed. This is a source-scoped policy comparison, not a legal opinion or an exhaustive current UK statute search.
\nThe acquired government record establishes that KYBC reached the policy agenda and cross-departmental discussion. It does not establish adoption of the IP groups' broader proposal requiring underlying commercial infrastructure providers to verify all business customers and stop serving those they cannot verify. A related but narrower advertiser-verification proposal appears in the 2023 Online Advertising Programme response. Separately, the March 2026 Fraud Strategy describes a current government/IAB UK partnership with an early-2027 report-back and conditional legislation, alongside a distinct statutory fraudulent-advertising regime whose commencement was then expected in 2027. Do not merge these into one enacted KYBC instrument or attribute the advertising pathway to Alliance without a transmission record.
\nSource: House of Commons DCMS Committee, HC719, Economics of music streaming: Government and Competition and Markets Authority Responses, response received 15 September and published 22 September 2021. [Official document](http[local research file] Read title/introduction and recommendation23 with its complete government response, PDF10-11 / printed8-9, readerlines364-415.
\nRecommendation23 asks government to address recording-industry concerns about enforcement of existing KYBC obligations. The appended GOVERNMENT response says KYBC for IP enforcement warrants further attention and the upcoming infringement strategy will consider it; IPO plans work with other departments on possible improvements. This is an identifiable government response to a committee recommendation, rather than an inference from shared vocabulary. It does not name a new verifier, required identification method, service-suspension trigger, sanction or appeal. Its surrounding discussion of licensing, EU Article17 and private civil action must not be treated as adoption of the broader KYBC proposal.
\nCustody: ordinary original GET returned403 and is closed. Error bytes retained separately; their hash does not authenticate the report. The successful official webreader is the credible substitute, retained as disposition-music2021.reader.txt. The cover/date was separately read through that reader, not recovered as original PDF bytes. See manifest.
\nReuse the accepted IPO2022 strategy Policy section and IPO2022/23 Counter infringement section; no new download. The strategy describes industry work on KYBC and a cross-departmental forum. The annual report says IPO hosted several cross-Whitehall forum meetings looking at KYBC rules. Neither names an agreed text or resulting duty. Retained source: ../pipcu-priorities-wave-2026-09-09/priority-ipo2023.html, SHA256 b7f1c94f0df2269776c32aeff112d2f7af7a3ee9da51c1f2046164b40486ab16; the accepted root-strategy-comparison.md holds the exact relevant extract. Earlier strategy original ../pipcu-resources-wave-2026-09-09/referral-strategy2022.html, SHA256 9dcb6d69ad37942b0b4005543be05e9f4561eb854659d149e0f95510067e0654. These are reused previously scoped reads, not new whole-document reads in this run.
\nThe June2023 Creative Industries Sector Vision, CP863, makes the next intention explicit: following previous roundtables, government will engage tech and creative sectors to explore potential KYBC requirements and consider CMA-highlighted issues, including proactive steps against infringing content. Relevant original: [Sector Vision PDF](http[local research file] PDF21 / printed18. The immediately preceding item names DSIT, IPO and DCMS on the IP framework, but the KYBC item says government; it does not allocate final rulemaking powers among these bodies. No roundtable attendee list, agreed minutes, regulatory text, commencement date or customer remedy appears in this item. Read cover and complete relevant page only, not the full report. Original disposition-vision2023.pdf SHA256 0be393e3229e41fbc2fda3fa452c809330f86e1d9c5f8ff82240d2b655450bd6.
\n[DCMS government response to Online Advertising Programme consultation](http[local research file] updated25July2023. Read original executive summary and nextsteps, Part1 Our response, and complete Part4 including respondent views and government response. Original disposition-oap2023.html SHA256 be9abb21a66d4e85f6835d3fcc84fc766d9e9fad2a10d0abd1116e714d083721.
\nThe government envisages statutory duties on platforms, intermediaries and publishers (PIPs) in paid-for advertising, including reciprocal arrangements, aimed at illegal advertising and separately protecting under18s from adverts for products/services illegal to sell to them. It excludes owned media such as a brand's own website and free-to-post classified ads from this programme. Advertisers themselves are essentially carved out of the proposed new regimes, yet would have to align with PIP measures, with advertiser identity verification given as an example used only when necessary/appropriate and with regulators deciding extent. Further consultation would settle scope/functions, powers and proportionate responsibilities, considering control over ads, size and reach. Legislation is promised subject to parliamentary time; this response is not the legislation.
\nPart4 records disagreements over pre-vetting, business burdens and sensitive commercial data, and the government says it will work with regulators to incentivise good practice rather than punish every breach. Those are actual policy constraints and competing stakeholder positions, not an unconditional demand to admit no unverified customer. The same source attributes strengthened financial-advertiser checks on most large platforms to FCA engagement; that is a government-reported narrower antecedent, not proof those firms implemented this IP industry's proposal. No specific universal identifier, identity-document requirement, duty to stop hosting/payment services, or customer appeal is established by the inspected OAP proposal.
\nThis is a mechanism comparison with KYBC, not a proven direct descendant of the IPO forum or Alliance/MPA text. The programme's separate child-protection aim is not evidence that business verification is consumer age-ID.
\n[26February2026 Online Advertising Taskforce minutes](http[local research file] entire2pages read. Meeting date is explicit; do not use it as a separately verified publication date. Original disposition-taskforce2026.pdf SHA256 3be65d983a4ff7d0c911580d628c6952f6f837f87f46b4aabe4c30370e792ee6.
\nSection4 records that Sinead Coogan-Jobes of IAB UK presented the new Ad Fraud and Standards Working Group's draft terms of reference and yearly plan, which taskforce members agreed. The group had been set up following the November2025 meeting. Its forthcoming mapping exercise would identify standards/mechanisms to detect, disrupt and prevent malicious advertising entering the supply chain. This is actual reported collective approval of a work plan; the precise terms themselves were not acquired, and these minutes do not grant the group powers to compel verification or disconnection.
\nSection2 separately reports pilot participants sharing fraud-related signals, with initially modest volume. That is a reported pilot action, not evidence of identity verification or a demonstrated enforcement outcome. Section5 separately records DSIT/ISBA updates about legitimate advertising funding legal-but-harmful content and evidence gathering with platforms, civil society, industry and brand-safety/verification services. This is an explicit adjacent policy discussion, not an adopted KYBC criterion or a takedown order. The age-assurance pilot in section3 has a different subject and is not used to support business-verification claims.
\n[Home Office Fraud Strategy2026-2029](http[local research file] March2026, CP1523. Read cover/imprint and originalPDF26-28, printed22-24, section4.4. Original disposition-fraud2026.pdf SHA256 71e581136a25391c1722208bae114213965f2736fb8a2b920e34aee8732d0864.
\nParagraph55 says wider programmatic advertising largely falls outside the then-current Online Safety Act scope. It reports Home Office and DCMS launching a partnership with IAB UK under the taskforce, reporting to ministers in early2027. If industry partnership and market incentives are insufficient, the government says it will legislate within this Parliament. This supports a current-at-publication policy pathway and a conditional legislative threat, not a commenced blanket KYBC rule. It does not identify the IPO forum as author or describe the broader infrastructure proposal as accepted or rejected wholesale.
\nKeep paragraph54 separate. It describes enacted Online Safety Act fraudulent-advertising duties for services designated Category1 or Category2A, with proportionate systems/processes to prevent paid-for fraudulent adverts. As of this March strategy, Ofcom consultation was aimed at summer2026 and commencement expected2027. It describes fines and, in the most serious non-compliance cases, court-ordered business-disruption measures potentially involving payment, advertising or internet service providers. The threshold is service non-compliance with Act duties, not merely an unverified commercial customer's identity. Paragraph56 separately dates illegal user-generated-content duties to March2025. This packet has not independently audited the underlying statutory provisions, commencement instruments or post-March2026 Ofcom timetable: it cannot certify September2026 commencement or transplant that separate authority into KYBC. The exact current operational advertising-code or court-remedy question remains a later instrument-level task if selected.
\nThe proved sequence is government acknowledgment -> official meetings -> further exploration. The acquired material has no final accepted cross-infrastructure KYBC text. A narrower, related advertising proposal supplies an example of identity verification with regulatory discretion; a later official partnership supplies actual plan approval and report-back accountability. These may represent delay, alternative targeting or parallel policy origins. This finite record does not distinguish those explanations through private negotiation or prove that no other UK verification duties exist.
\nCandidate typed relations, not case additions: UKGovernment -> committee recommendation23 [policy response/consideration, September2021]; IPO -> cross-Whitehall forum [reported convening inFY2022/23, reuse]; SectorVision -> tech/creative sectors [planned engagement on potentialKYBC, June2023]; DCMS OAP proposal -> PIPs [proposed systems obligations and conditional advertiser-verification example, July2023]; OATmembers -> AdFraudandStandardsWG workplan [reported collective approval,26February2026]; HomeOffice/DCMS -> IABUK partnership [reported launch, March2026]; partnership -> ministers [planned report-back, early2027]; government -> further advertising legislation [conditional intention]. A comparison edge between the IP proposal and advertising instruments should say related mechanism, not authorship or implementation. Do not infer a paid grant from the word partnership.
\nThe highest-value missing record is the IPO/DCMS cross-Whitehall KYBC options or disposition paper, with the actual draft text/meeting decision showing whether infrastructure-wide identity checks and stop-service consequences were accepted, narrowed, rejected or parked, and why. Its public institutional holders are IPO/DSIT/DCMS as evidenced by the strategy and vision; this is a record target, not a claim a particular unpublished document exists. The February2026 working-group terms/accepted plan and eventual ministerial report would settle the newer advertising authority path; taskforce minutes alone do not provide enforcement rights or customer remedies.
\nManifest disposition-captures.json identifies four successful official original captures, one403 error response, and the official-reader substitute for that failed original. Scope-limited extracts carry separate hashes/ancestry. Full-file extraction is not full-file reading. Originalvision webreader failed once with InternalError; ordinary publicGET succeeded. Music originalGET403 is closed; only the already-successful officialreader was used as the substitute. No closed LegInfo, LinkedIn, PIPCU, UN or other routes were retried.
\nFinite discovery comprised exact KYBC/business-customer searches limited to gov.uk, Parliament and Hansard, then targeted OnlineAdvertisingProgramme/verification and2025/2026 disposition searches. Several noisy search returns were cut off by the tool; only relevant visible/recovered items were assessed. Irrelevant trademark-journal results and general AML/gambling material were excluded. Search-only leads for Ofcom telecom checks, the OnlineFraudCharter,2023/24 taskforce progress,2025minutes and private industry commentary were not promoted into facts or independently acquired here. No EU rules, company-registration duties or consumer-age scheme was substituted for the requested policy trace. The exact selected originalURLs above, and scoped notes rather than search snippets, support the findings. This is a bounded negative frontier, not proof of universal legislative absence.
\nAcquisition closed at the original2026 strategy and2-page minutes. No further branch opened.
", "path": "kybc-policy-wave-2026-09-09/disposition-rules.md", "sourceIds": [ "w29-enf-s-music-response2021", "w29-enf-s-sector-vision2023", "w29-enf-s-oap-response2023", "w29-enf-s-fraud-strategy2026", "w29-enf-s-oat-minutes2026" ], "markdown": "9 September 2026. Bounded acquisition CLOSED. No canonical case, reader or ZIP changed. This is a source-scoped policy comparison, not a legal opinion or an exhaustive current UK statute search.\n\nThe acquired government record establishes that KYBC reached the policy agenda and cross-departmental discussion. It does not establish adoption of the IP groups' broader proposal requiring underlying commercial infrastructure providers to verify all business customers and stop serving those they cannot verify. A related but narrower advertiser-verification proposal appears in the 2023 Online Advertising Programme response. Separately, the March 2026 Fraud Strategy describes a current government/IAB UK partnership with an early-2027 report-back and conditional legislation, alongside a distinct statutory fraudulent-advertising regime whose commencement was then expected in 2027. Do not merge these into one enacted KYBC instrument or attribute the advertising pathway to Alliance without a transmission record.\n\n## 1. Explicit government acknowledgment: September 2021\n\nSource: House of Commons DCMS Committee, HC719, Economics of music streaming: Government and Competition and Markets Authority Responses, response received 15 September and published 22 September 2021. Official document (see source references) Read title/introduction and recommendation23 with its complete government response, PDF10-11 / printed8-9, readerlines364-415.\n\nRecommendation23 asks government to address recording-industry concerns about enforcement of existing KYBC obligations. The appended GOVERNMENT response says KYBC for IP enforcement warrants further attention and the upcoming infringement strategy will consider it; IPO plans work with other departments on possible improvements. This is an identifiable government response to a committee recommendation, rather than an inference from shared vocabulary. It does not name a new verifier, required identification method, service-suspension trigger, sanction or appeal. Its surrounding discussion of licensing, EU Article17 and private civil action must not be treated as adoption of the broader KYBC proposal.\n\nCustody: ordinary original GET returned403 and is closed. Error bytes retained separately; their hash does not authenticate the report. The successful official webreader is the credible substitute, retained as disposition-music2021.reader.txt. The cover/date was separately read through that reader, not recovered as original PDF bytes. See manifest.\n\n## 2. Policy intention becomes reported meetings, then an explicit further commitment\n\nReuse the accepted IPO2022 strategy Policy section and IPO2022/23 Counter infringement section; no new download. The strategy describes industry work on KYBC and a cross-departmental forum. The annual report says IPO hosted several cross-Whitehall forum meetings looking at KYBC rules. Neither names an agreed text or resulting duty. Retained source: ../pipcu-priorities-wave-2026-09-09/priority-ipo2023.html, SHA256 b7f1c94f0df2269776c32aeff112d2f7af7a3ee9da51c1f2046164b40486ab16; the accepted root-strategy-comparison.md holds the exact relevant extract. Earlier strategy original ../pipcu-resources-wave-2026-09-09/referral-strategy2022.html, SHA256 9dcb6d69ad37942b0b4005543be05e9f4561eb854659d149e0f95510067e0654. These are reused previously scoped reads, not new whole-document reads in this run.\n\nThe June2023 Creative Industries Sector Vision, CP863, makes the next intention explicit: following previous roundtables, government will engage tech and creative sectors to explore potential KYBC requirements and consider CMA-highlighted issues, including proactive steps against infringing content. Relevant original: Sector Vision PDF (see source references) PDF21 / printed18. The immediately preceding item names DSIT, IPO and DCMS on the IP framework, but the KYBC item says government; it does not allocate final rulemaking powers among these bodies. No roundtable attendee list, agreed minutes, regulatory text, commencement date or customer remedy appears in this item. Read cover and complete relevant page only, not the full report. Original disposition-vision2023.pdf SHA256 0be393e3229e41fbc2fda3fa452c809330f86e1d9c5f8ff82240d2b655450bd6.\n\n## 3. A narrower advertiser-verification proposal: July 2023\n\nDCMS government response to Online Advertising Programme consultation (see source references) updated25July2023. Read original executive summary and nextsteps, Part1 Our response, and complete Part4 including respondent views and government response. Original disposition-oap2023.html SHA256 be9abb21a66d4e85f6835d3fcc84fc766d9e9fad2a10d0abd1116e714d083721.\n\nThe government envisages statutory duties on platforms, intermediaries and publishers (PIPs) in paid-for advertising, including reciprocal arrangements, aimed at illegal advertising and separately protecting under18s from adverts for products/services illegal to sell to them. It excludes owned media such as a brand's own website and free-to-post classified ads from this programme. Advertisers themselves are essentially carved out of the proposed new regimes, yet would have to align with PIP measures, with advertiser identity verification given as an example used only when necessary/appropriate and with regulators deciding extent. Further consultation would settle scope/functions, powers and proportionate responsibilities, considering control over ads, size and reach. Legislation is promised subject to parliamentary time; this response is not the legislation.\n\nPart4 records disagreements over pre-vetting, business burdens and sensitive commercial data, and the government says it will work with regulators to incentivise good practice rather than punish every breach. Those are actual policy constraints and competing stakeholder positions, not an unconditional demand to admit no unverified customer. The same source attributes strengthened financial-advertiser checks on most large platforms to FCA engagement; that is a government-reported narrower antecedent, not proof those firms implemented this IP industry's proposal. No specific universal identifier, identity-document requirement, duty to stop hosting/payment services, or customer appeal is established by the inspected OAP proposal.\n\nThis is a mechanism comparison with KYBC, not a proven direct descendant of the IPO forum or Alliance/MPA text. The programme's separate child-protection aim is not evidence that business verification is consumer age-ID.\n\n## 4. An implemented deliberative mechanism, not enacted KYBC: February/March 2026\n\n26February2026 Online Advertising Taskforce minutes (see source references) entire2pages read. Meeting date is explicit; do not use it as a separately verified publication date. Original disposition-taskforce2026.pdf SHA256 3be65d983a4ff7d0c911580d628c6952f6f837f87f46b4aabe4c30370e792ee6.\n\nSection4 records that Sinead Coogan-Jobes of IAB UK presented the new Ad Fraud and Standards Working Group's draft terms of reference and yearly plan, which taskforce members agreed. The group had been set up following the November2025 meeting. Its forthcoming mapping exercise would identify standards/mechanisms to detect, disrupt and prevent malicious advertising entering the supply chain. This is actual reported collective approval of a work plan; the precise terms themselves were not acquired, and these minutes do not grant the group powers to compel verification or disconnection.\n\nSection2 separately reports pilot participants sharing fraud-related signals, with initially modest volume. That is a reported pilot action, not evidence of identity verification or a demonstrated enforcement outcome. Section5 separately records DSIT/ISBA updates about legitimate advertising funding legal-but-harmful content and evidence gathering with platforms, civil society, industry and brand-safety/verification services. This is an explicit adjacent policy discussion, not an adopted KYBC criterion or a takedown order. The age-assurance pilot in section3 has a different subject and is not used to support business-verification claims.\n\nHome Office Fraud Strategy2026-2029 (see source references) March2026, CP1523. Read cover/imprint and originalPDF26-28, printed22-24, section4.4. Original disposition-fraud2026.pdf SHA256 71e581136a25391c1722208bae114213965f2736fb8a2b920e34aee8732d0864.\n\nParagraph55 says wider programmatic advertising largely falls outside the then-current Online Safety Act scope. It reports Home Office and DCMS launching a partnership with IAB UK under the taskforce, reporting to ministers in early2027. If industry partnership and market incentives are insufficient, the government says it will legislate within this Parliament. This supports a current-at-publication policy pathway and a conditional legislative threat, not a commenced blanket KYBC rule. It does not identify the IPO forum as author or describe the broader infrastructure proposal as accepted or rejected wholesale.\n\nKeep paragraph54 separate. It describes enacted Online Safety Act fraudulent-advertising duties for services designated Category1 or Category2A, with proportionate systems/processes to prevent paid-for fraudulent adverts. As of this March strategy, Ofcom consultation was aimed at summer2026 and commencement expected2027. It describes fines and, in the most serious non-compliance cases, court-ordered business-disruption measures potentially involving payment, advertising or internet service providers. The threshold is service non-compliance with Act duties, not merely an unverified commercial customer's identity. Paragraph56 separately dates illegal user-generated-content duties to March2025. This packet has not independently audited the underlying statutory provisions, commencement instruments or post-March2026 Ofcom timetable: it cannot certify September2026 commencement or transplant that separate authority into KYBC. The exact current operational advertising-code or court-remedy question remains a later instrument-level task if selected.\n\n## Consequential comparison and candidate relations\n\nThe proved sequence is government acknowledgment -> official meetings -> further exploration. The acquired material has no final accepted cross-infrastructure KYBC text. A narrower, related advertising proposal supplies an example of identity verification with regulatory discretion; a later official partnership supplies actual plan approval and report-back accountability. These may represent delay, alternative targeting or parallel policy origins. This finite record does not distinguish those explanations through private negotiation or prove that no other UK verification duties exist.\n\nCandidate typed relations, not case additions: UKGovernment -> committee recommendation23 [policy response/consideration, September2021]; IPO -> cross-Whitehall forum [reported convening inFY2022/23, reuse]; SectorVision -> tech/creative sectors [planned engagement on potentialKYBC, June2023]; DCMS OAP proposal -> PIPs [proposed systems obligations and conditional advertiser-verification example, July2023]; OATmembers -> AdFraudandStandardsWG workplan [reported collective approval,26February2026]; HomeOffice/DCMS -> IABUK partnership [reported launch, March2026]; partnership -> ministers [planned report-back, early2027]; government -> further advertising legislation [conditional intention]. A comparison edge between the IP proposal and advertising instruments should say related mechanism, not authorship or implementation. Do not infer a paid grant from the word partnership.\n\nThe highest-value missing record is the IPO/DCMS cross-Whitehall KYBC options or disposition paper, with the actual draft text/meeting decision showing whether infrastructure-wide identity checks and stop-service consequences were accepted, narrowed, rejected or parked, and why. Its public institutional holders are IPO/DSIT/DCMS as evidenced by the strategy and vision; this is a record target, not a claim a particular unpublished document exists. The February2026 working-group terms/accepted plan and eventual ministerial report would settle the newer advertising authority path; taskforce minutes alone do not provide enforcement rights or customer remedies.\n\n## Reading and acquisition frontier\n\nManifest disposition-captures.json identifies four successful official original captures, one403 error response, and the official-reader substitute for that failed original. Scope-limited extracts carry separate hashes/ancestry. Full-file extraction is not full-file reading. Originalvision webreader failed once with InternalError; ordinary publicGET succeeded. Music originalGET403 is closed; only the already-successful officialreader was used as the substitute. No closed LegInfo, LinkedIn, PIPCU, UN or other routes were retried.\n\nFinite discovery comprised exact KYBC/business-customer searches limited to gov.uk, Parliament and Hansard, then targeted OnlineAdvertisingProgramme/verification and2025/2026 disposition searches. Several noisy search returns were cut off by the tool; only relevant visible/recovered items were assessed. Irrelevant trademark-journal results and general AML/gambling material were excluded. Search-only leads for Ofcom telecom checks, the OnlineFraudCharter,2023/24 taskforce progress,2025minutes and private industry commentary were not promoted into facts or independently acquired here. No EU rules, company-registration duties or consumer-age scheme was substituted for the requested policy trace. The exact selected originalURLs above, and scoped notes rather than search snippets, support the findings. This is a bounded negative frontier, not proof of universal legislative absence.\n\nAcquisition closed at the original2026 strategy and2-page minutes. No further branch opened.\n", "file": "research/kybc-policy-wave-2026-09-09-disposition-rules.md" }, { "id": "kybc-policy-wave-2026-09-09-origins-transmission", "title": "KYBC: dated proposals and their UK transmission routes", "description": "Closed 9 September 2026. Four submissions now establish a public UK advocacy trail before the previously inspected IPO forum account. The earliest internally dated Alliance document in this pass is August 2020. This is not a claim that the Alliance invented KYBC or was its first proponent.", "html": "Closed 9 September 2026. Four submissions now establish a public UK advocacy trail before the previously inspected IPO forum account. The earliest internally dated Alliance document acquired in this pass is August 2020. This is not a claim that the Alliance invented KYBC or was its first proponent.
\n| Document and date boundary | \nProposer → recorded recipient/venue | \nDecisive request and locator | \n
|---|---|---|
| [Alliance response to Digital Markets Taskforce](http[local research file] internally August 2020; CMA publication 8 December 2020 | \nAlliance for Intellectual Property → CMA Digital Markets Taskforce call for information | \nPDF 3, response to questions 5–8, seeks verification duties across critical business infrastructure, including providers without Strategic Market Status; it proposes hosting/payment-access consequences for illicit businesses ignoring self-disclosure. PDF 4 proposes a code-of-conduct/enforcement route. | \n
| [BPI response to the same call](http[local research file] labelled Redacted version November 2020 at PDF 18; CMA publication 8 December 2020 | \nBPI → CMA Digital Markets Taskforce | \nPDF 17, §4.2.2, proposes regular verification and service termination for withheld or incorrect information. PDF 6, §3, identifies intermediary categories; PDF 16 places the remedies under proposed codes and a duty of care. | \n
| [Alliance OSB0016](http[local research file] internally September 2021 | \nAlliance → Joint Committee on the Draft Online Safety Bill | \nPDF 2–3, The problem & potential solution, seeks KYBC duties for all business-infrastructure intermediaries, beyond the Bill's scope; an appropriate legislative vehicle and hosting/payment-access consequences. PDF 3 Summary asks Parliamentarians to press government. | \n
| [MPA FDF0068](http[local research file] internally 10 May 2022 at PDF 21; contains a briefing labelled September 2021 | \nMPA → Lords Fraud Act 2006 and Digital Fraud Committee; transmits a briefing from rightsholders on the Creative Industries Council IP subgroup | \nMain §§19–28, PDF 5–7, proposes identity verification and asks the Committee to encourage named departments into the IPO forum. Annex 2, PDF 9–13, supplies the rightsholder proposal, coverage and sanctions. | \n
The [CMA's original case index](http[local research file] Responses to call for information, explicitly links both 2020 responses with the 8 December publication date. Its timetable distinguishes the 1 July call launch and 31 July deadline from publication. The deadline is not the proven date of either submission. Alliance's August document date and BPI's November redacted-version date remain separate from the unacquired initial submission/receipt records. Publication by the recipient establishes entry into a policy process; it does not establish that the recipient accepted the requested rules.
\nAlliance's August 2020 response, PDF 1–2, explains why it sought a regulatory route: it regarded voluntary industry roundtables as slow and Online Harms proposals as excluding IP infringement. Its response to questions 5–8 expands beyond firms with Strategic Market Status; it seeks verification coupled with denial of critical infrastructure. The September 2021 submission repeats that demand in a different parliamentary venue and cites the Digital Markets Taskforce's December 2020 recommendation, G22(13b), to strengthen powers against unlawful activity causing economic detriment. This is an explicit citation and request for scrutiny, not proof that the Taskforce specifically adopted Alliance's KYBC design. Its assessment of the then-current e-commerce law remains attributed advocacy, not this packet's legal conclusion.
\nBPI's §4.2.2 specifies verified name, physical address, email and telephone; for legal entities, current director/responsible-person information, with bank-detail checks suggested where appropriate. Records would be checked regularly. Its proposed consequence attaches to withheld or incorrect information, without making a separately proved infringement a stated prerequisite in that clause. Section 3, PDF 6, includes CDNs, hosting, payments, domain registrars/registries, advertising networks and app stores among intermediaries. This is a broad business-customer proposal, not a requirement for government ID or age checks on every consumer. BPI's proposed information access for legitimate rightsholder enforcement also deserves distinction from publication of every customer's details.
\nThe CIC briefing's Q3–5, PDF 12–13 of the MPA submission, extends across interactive platforms, content delivery, hosting/CDNs, registrars/registries, DNS, proxies, email, financial and advertising services. Privacy-protection companies would themselves have duties rather than count as the end customer. Failure or inability to verify the commercial customer would require service cessation; direct obligations and financial penalties would coexist with existing liability privileges. This is a briefing from listed rightsholders on the IP subgroup, not demonstrated approval by the full CIC or government. Its September 2021 dating/tabling is attributed through MPA's later submission. Main §26 asks the Lords to encourage BEIS, DCMS, Home Office, Treasury and Justice into the proposed forum. Commercial customers need not be incorporated companies; sole traders or commercial social-media users are not expressly excluded. The inspected demand sections supply no detailed challenge procedure for an unsuccessful verification.
\nThree distinct relations are now supported: named organizations submitted identifiable proposals; the CMA and parliamentary sites published them; and proponents cited and sought action through government policy processes. The MPA carries an earlier rightsholder briefing into its later fraud evidence. Those are actual document paths, stronger than a shared vocabulary or contemporary roster. They still do not prove whose argument caused a later decision.
\nThe retained IPO 2022 strategy, under the cross-departmental policy discussion, calls KYBC an industry-developed potential solution and says action requires departmental collaboration. The accepted IPO 2022/23 account reports forum meetings. They do not name the Alliance as originator. This pass makes the chronology possible without filling that attribution gap. The undated current Alliance webpage can now be compared with its dated 2020/2021 positions; it cannot establish the date of the webpage itself.
\nThe strongest rival to an Alliance-led causal account is convergent advocacy by several rightsholder groups, drawing on existing identity/enforcement ideas and approaching multiple forums. Root's separate research-input.md examines the explicitly cited RUSI input. No EU legislative survey, corporate genealogy or new agency branch was pursued here. The MPA's annex references to other reports and coalitions are references by the submitter, not independent verification of each underlying publication. Industry case allegations were not adopted as verified findings about the named services.
The remaining discriminating records are recipient minutes or correspondence identifying which paper was considered; the CIC subgroup's tabling/adoption record; and a government decision that accepts, narrows or rejects specific verification, service-cessation and sanction features. Root/Linnaeus own the disposition comparison. This packet does not infer enacted duties from any submission, a pilot-funded priority or a universal consumer-identification rule.
\norigins-captures.json records exact URLs, hashes, dates and read scopes. Two original CMA-published PDFs and the original CMA index HTML are retained. Alliance's four-page response was read in full; BPI was read at the recorded sections, not all nineteen pages. Local renders of Alliance PDF 3 and BPI PDF 17 were visually inspected and agree with the decisive extracted text. BPI's public file identifies itself as a redacted version; no unredacted counterpart was sought.
The two parliamentary PDF preservation GETs returned 403. Their working public readers supplied the substantive text; the error HTML is not the source document. MPA's PDF reader exposes the Annex 1 and summary text absent from the HTML rendition, but a single batched screenshot request for PDF pages 8, 9, 12 and 13 returned Cache miss. No image inspection or original-byte custody is asserted for those documents, and the routes were not retried. origins-reader-scopes.json preserves the exact reading boundaries and locator notes; it is a derivative, not a facsimile.
Acquisition closed at this dated-submission/transmission breakpoint. No case, atlas, ZIP, current-work or forest file was changed.
", "path": "kybc-policy-wave-2026-09-09/origins-transmission.md", "sourceIds": [ "w29-enf-s-alliance-cma2020", "w29-enf-s-bpi-cma2020", "w29-enf-s-cma-index", "w29-enf-s-alliance-osb2021", "w29-enf-s-mpa-fdf0068" ], "markdown": "Closed 9 September 2026. Four submissions now establish a public UK advocacy trail before the previously inspected IPO forum account. The earliest internally dated Alliance document acquired in this pass is August 2020. This is not a claim that the Alliance invented KYBC or was its first proponent.\n\n## What reached which institution\n\n| Document and date boundary | Proposer → recorded recipient/venue | Decisive request and locator |\n| --- | --- | --- |\n| Alliance response to Digital Markets Taskforce (see source references) internally **August 2020**; CMA publication **8 December 2020** | Alliance for Intellectual Property → CMA Digital Markets Taskforce call for information | PDF 3, response to questions 5–8, seeks verification duties across critical business infrastructure, including providers without Strategic Market Status; it proposes hosting/payment-access consequences for illicit businesses ignoring self-disclosure. PDF 4 proposes a code-of-conduct/enforcement route. |\n| BPI response to the same call (see source references) labelled **Redacted version November 2020** at PDF 18; CMA publication **8 December 2020** | BPI → CMA Digital Markets Taskforce | PDF 17, §4.2.2, proposes regular verification and service termination for withheld or incorrect information. PDF 6, §3, identifies intermediary categories; PDF 16 places the remedies under proposed codes and a duty of care. |\n| Alliance OSB0016 (see source references) internally **September 2021** | Alliance → Joint Committee on the Draft Online Safety Bill | PDF 2–3, The problem & potential solution, seeks KYBC duties for all business-infrastructure intermediaries, beyond the Bill's scope; an appropriate legislative vehicle and hosting/payment-access consequences. PDF 3 Summary asks Parliamentarians to press government. |\n| MPA FDF0068 (see source references) internally **10 May 2022** at PDF 21; contains a briefing labelled **September 2021** | MPA → Lords Fraud Act 2006 and Digital Fraud Committee; transmits a briefing from rightsholders on the Creative Industries Council IP subgroup | Main §§19–28, PDF 5–7, proposes identity verification and asks the Committee to encourage named departments into the IPO forum. Annex 2, PDF 9–13, supplies the rightsholder proposal, coverage and sanctions. |\n\nThe CMA's original case index (see source references) Responses to call for information, explicitly links both 2020 responses with the 8 December publication date. Its timetable distinguishes the 1 July call launch and 31 July deadline from publication. The deadline is **not** the proven date of either submission. Alliance's August document date and BPI's November redacted-version date remain separate from the unacquired initial submission/receipt records. Publication by the recipient establishes entry into a policy process; it does not establish that the recipient accepted the requested rules.\n\n## The proposed intervention was more specific than a slogan\n\nAlliance's August 2020 response, PDF 1–2, explains why it sought a regulatory route: it regarded voluntary industry roundtables as slow and Online Harms proposals as excluding IP infringement. Its response to questions 5–8 expands beyond firms with Strategic Market Status; it seeks verification coupled with denial of critical infrastructure. The September 2021 submission repeats that demand in a different parliamentary venue and cites the Digital Markets Taskforce's December 2020 recommendation, G22(13b), to strengthen powers against unlawful activity causing economic detriment. This is an explicit citation and request for scrutiny, not proof that the Taskforce specifically adopted Alliance's KYBC design. Its assessment of the then-current e-commerce law remains attributed advocacy, not this packet's legal conclusion.\n\nBPI's §4.2.2 specifies verified name, physical address, email and telephone; for legal entities, current director/responsible-person information, with bank-detail checks suggested where appropriate. Records would be checked regularly. Its proposed consequence attaches to withheld or incorrect information, without making a separately proved infringement a stated prerequisite in that clause. Section 3, PDF 6, includes CDNs, hosting, payments, domain registrars/registries, advertising networks and app stores among intermediaries. This is a broad business-customer proposal, not a requirement for government ID or age checks on every consumer. BPI's proposed information access for legitimate rightsholder enforcement also deserves distinction from publication of every customer's details.\n\nThe CIC briefing's Q3–5, PDF 12–13 of the MPA submission, extends across interactive platforms, content delivery, hosting/CDNs, registrars/registries, DNS, proxies, email, financial and advertising services. Privacy-protection companies would themselves have duties rather than count as the end customer. Failure or inability to verify the commercial customer would require service cessation; direct obligations and financial penalties would coexist with existing liability privileges. This is a briefing **from listed rightsholders on the IP subgroup**, not demonstrated approval by the full CIC or government. Its September 2021 dating/tabling is attributed through MPA's later submission. Main §26 asks the Lords to encourage BEIS, DCMS, Home Office, Treasury and Justice into the proposed forum. Commercial customers need not be incorporated companies; sole traders or commercial social-media users are not expressly excluded. The inspected demand sections supply no detailed challenge procedure for an unsuccessful verification.\n\n## Transmission evidence and limits\n\nThree distinct relations are now supported: named organizations **submitted** identifiable proposals; the CMA and parliamentary sites **published** them; and proponents **cited and sought action through** government policy processes. The MPA carries an earlier rightsholder briefing into its later fraud evidence. Those are actual document paths, stronger than a shared vocabulary or contemporary roster. They still do not prove whose argument caused a later decision.\n\nThe retained IPO 2022 strategy, under the cross-departmental policy discussion, calls KYBC an industry-developed potential solution and says action requires departmental collaboration. The accepted IPO 2022/23 account reports forum meetings. They do not name the Alliance as originator. This pass makes the chronology possible without filling that attribution gap. The undated current Alliance webpage can now be compared with its dated 2020/2021 positions; it cannot establish the date of the webpage itself.\n\nThe strongest rival to an Alliance-led causal account is convergent advocacy by several rightsholder groups, drawing on existing identity/enforcement ideas and approaching multiple forums. Root's separate `research-input.md` examines the explicitly cited RUSI input. No EU legislative survey, corporate genealogy or new agency branch was pursued here. The MPA's annex references to other reports and coalitions are references by the submitter, not independent verification of each underlying publication. Industry case allegations were not adopted as verified findings about the named services.\n\nThe remaining discriminating records are recipient minutes or correspondence identifying which paper was considered; the CIC subgroup's tabling/adoption record; and a government decision that accepts, narrows or rejects specific verification, service-cessation and sanction features. Root/Linnaeus own the disposition comparison. This packet does not infer enacted duties from any submission, a pilot-funded priority or a universal consumer-identification rule.\n\n## Custody and scope\n\n`origins-captures.json` records exact URLs, hashes, dates and read scopes. Two original CMA-published PDFs and the original CMA index HTML are retained. Alliance's four-page response was read in full; BPI was read at the recorded sections, not all nineteen pages. Local renders of Alliance PDF 3 and BPI PDF 17 were visually inspected and agree with the decisive extracted text. BPI's public file identifies itself as a redacted version; no unredacted counterpart was sought.\n\nThe two parliamentary PDF preservation GETs returned 403. Their working public readers supplied the substantive text; the error HTML is not the source document. MPA's PDF reader exposes the Annex 1 and summary text absent from the HTML rendition, but a single batched screenshot request for PDF pages 8, 9, 12 and 13 returned Cache miss. No image inspection or original-byte custody is asserted for those documents, and the routes were not retried. `origins-reader-scopes.json` preserves the exact reading boundaries and locator notes; it is a derivative, not a facsimile.\n\nAcquisition closed at this dated-submission/transmission breakpoint. No case, atlas, ZIP, current-work or forest file was changed.\n", "file": "research/kybc-policy-wave-2026-09-09-origins-transmission.md" }, { "id": "kybc-policy-wave-2026-09-09-research-input", "title": "Research input: named co-funding, data provision and later KYBC citation", "description": "9 September2026. Bounded root acquisition CLOSED after one original report directly cited by MPA's parliamentary evidence. This follows a specific source/input edge within the KYBC origins run, not a general RUSI or donor investigation.", "html": "9 September2026. Bounded root acquisition CLOSED after one original report directly cited by MPA's parliamentary evidence. This follows a specific source/input edge within the KYBC origins run, not a general RUSI or donor investigation.
\nThe [March2021 RUSI Whitehall Report4-20, Taking the Profit Out of Intellectual Property Crime: Piracy and Organised Crime](http[local research file] by Ardi Janjeva, Alexandria Reid and Anton Moiseienko, names five co-funders in its Acknowledgements, PDF8/printedvii: UK Intellectual Property Office, Alliance for Intellectual Property, Motion Picture Association, Industry Trust and Premier League. It supplies no amounts, allocation, payment dates, executed agreement or funder approval rights. Industry Trust remains the source's label, without a new legal-identity reconciliation. The authors also credit primary information supplied by funders, PIPCU, BBC Studios and other named/unnamed organizations. Co-funding and supplying material are separate resource relations.
\nThe same acknowledgment says findings and recommendations were independently reached and not influenced by funders. Methodology, PDF18–19/printed5–6, describes a targeted literature review, interviews with77experts between February and September2020, and primary-data review. Recruitment used literature, snowball sampling and invited self-nominations; interview categories include government, enforcement, providers, financial compliance, rightsholders/associations, protection vendors and academics. Two one-hour January2021 validation workshops excluded representatives of funders and prompted revisions. Some contributed evidence cannot be reproduced because of sensitivity. This establishes the authors' reported method and independence measures; it is not an independent check of all underlying interviews/data, a representative sample or a reviewed funder contract. The unreproduced material limits replication but does not establish fabrication or improper direction.
\nThe executive summary and Recommendations, PDF82–83/printed69–70, combine several proposed mechanisms. Recommendation2 calls for an IPO/PIPCU public-private intelligence partnership including financial and online-service actors. Recommendation4 calls for business-customer identity recording/verification to the extent possible in the online-harms review, with compliance resources. Recommendation14 urges financial institutions to use named infringement lists when assessing exposure and taking appropriate action, including ending relationships or filing suspicious-activity reports. Recommendation15 urges providers to verify customers as best practice regardless of whether formal KYBC rules are introduced, qualified by what is reasonably possible. These are recommendations, not acquired regulator orders, provider policies or executed payment denials.
\nOther recommendations call for more financial investigation, acquiring-bank scrutiny and education of crypto-asset service providers. Those are particular anti-piracy proposals in2021; they do not establish later implementation, a general control system over spending or a stablecoin-debt scheme. The report distinguishes counterfeiting from audiovisual piracy and this scoped note does not import its piracy findings into every counterfeit-domain suspension. Its industry/consumer harm estimates and alleged case facts have not been independently verified by this root read and are not adopted here.
\nA specific downstream use is observable. [MPA evidence FDF0068, dated10May2022](http[local research file] paragraph18 and footnote12, cites this report's KYBC recommendation as support for action. MPA is also a named co-funder in the report. That establishes a co-funder citing the resulting research in parliamentary advocacy. It does not establish that MPA dictated the finding or that Parliament/government accepted it. Alliance's separately located August2020 submission predates this publication; the report is not established as the first origin of the proposal. The funding, inputs, publication, citation and actual government decision should be drawn as different edges.
\nThe copyright/version page, PDF3, identifies March2021 and records textual changes on10and19March2021. Original retained root-rusi2021.pdf SHA25603d13de395767d5c5db7e5419ea60b6caa795e955468af7953814742dabbdce0,84pages,2164009bytes. Root read original-extracted PDF2,3,8,18,19,80,81,82,83; executive summaryPDF10–12 was read through the official public reader. Exact scope and derivative ancestry are in root-rusi2021-capture.json. An extracted page is not automatically substantively read. No facsimile/graph audit or full84-page review is claimed.
\nRoot's MPA reading covered substantive main paragraphs1–28, Annex2 policy questions1–5 and signatory list, source date and footnotes through the official reader. Image-only Annex1/summary/case material was not visually inspected. Other returned case allegations were seen as source claims but not independently verified or used to establish ownership/criminal liability. Carver's separately scoped originals/reader custody and source chronology govern the origins packet. No operational content, accounts, inquiries or funder contact occurred. Acquisition is closed for combined review/forest.
", "path": "kybc-policy-wave-2026-09-09/research-input.md", "sourceIds": [ "w29-enf-s-rusi2021" ], "markdown": "9 September2026. Bounded root acquisition CLOSED after one original report directly cited by MPA's parliamentary evidence. This follows a specific source/input edge within the KYBC origins run, not a general RUSI or donor investigation.\n\nThe March2021 RUSI Whitehall Report4-20, Taking the Profit Out of Intellectual Property Crime: Piracy and Organised Crime (see source references) by Ardi Janjeva, Alexandria Reid and Anton Moiseienko, names five co-funders in its Acknowledgements, PDF8/printedvii: UK Intellectual Property Office, Alliance for Intellectual Property, Motion Picture Association, Industry Trust and Premier League. It supplies no amounts, allocation, payment dates, executed agreement or funder approval rights. Industry Trust remains the source's label, without a new legal-identity reconciliation. The authors also credit primary information supplied by funders, PIPCU, BBC Studios and other named/unnamed organizations. Co-funding and supplying material are separate resource relations.\n\nThe same acknowledgment says findings and recommendations were independently reached and not influenced by funders. Methodology, PDF18–19/printed5–6, describes a targeted literature review, interviews with77experts between February and September2020, and primary-data review. Recruitment used literature, snowball sampling and invited self-nominations; interview categories include government, enforcement, providers, financial compliance, rightsholders/associations, protection vendors and academics. Two one-hour January2021 validation workshops excluded representatives of funders and prompted revisions. Some contributed evidence cannot be reproduced because of sensitivity. This establishes the authors' reported method and independence measures; it is not an independent check of all underlying interviews/data, a representative sample or a reviewed funder contract. The unreproduced material limits replication but does not establish fabrication or improper direction.\n\nThe executive summary and Recommendations, PDF82–83/printed69–70, combine several proposed mechanisms. Recommendation2 calls for an IPO/PIPCU public-private intelligence partnership including financial and online-service actors. Recommendation4 calls for business-customer identity recording/verification to the extent possible in the online-harms review, with compliance resources. Recommendation14 urges financial institutions to use named infringement lists when assessing exposure and taking appropriate action, including ending relationships or filing suspicious-activity reports. Recommendation15 urges providers to verify customers as best practice regardless of whether formal KYBC rules are introduced, qualified by what is reasonably possible. These are recommendations, not acquired regulator orders, provider policies or executed payment denials.\n\nOther recommendations call for more financial investigation, acquiring-bank scrutiny and education of crypto-asset service providers. Those are particular anti-piracy proposals in2021; they do not establish later implementation, a general control system over spending or a stablecoin-debt scheme. The report distinguishes counterfeiting from audiovisual piracy and this scoped note does not import its piracy findings into every counterfeit-domain suspension. Its industry/consumer harm estimates and alleged case facts have not been independently verified by this root read and are not adopted here.\n\nA specific downstream use is observable. MPA evidence FDF0068, dated10May2022 (see source references) paragraph18 and footnote12, cites this report's KYBC recommendation as support for action. MPA is also a named co-funder in the report. That establishes a co-funder citing the resulting research in parliamentary advocacy. It does not establish that MPA dictated the finding or that Parliament/government accepted it. Alliance's separately located August2020 submission predates this publication; the report is not established as the first origin of the proposal. The funding, inputs, publication, citation and actual government decision should be drawn as different edges.\n\nThe copyright/version page, PDF3, identifies March2021 and records textual changes on10and19March2021. Original retained root-rusi2021.pdf SHA25603d13de395767d5c5db7e5419ea60b6caa795e955468af7953814742dabbdce0,84pages,2164009bytes. Root read original-extracted PDF2,3,8,18,19,80,81,82,83; executive summaryPDF10–12 was read through the official public reader. Exact scope and derivative ancestry are in root-rusi2021-capture.json. An extracted page is not automatically substantively read. No facsimile/graph audit or full84-page review is claimed.\n\nRoot's MPA reading covered substantive main paragraphs1–28, Annex2 policy questions1–5 and signatory list, source date and footnotes through the official reader. Image-only Annex1/summary/case material was not visually inspected. Other returned case allegations were seen as source claims but not independently verified or used to establish ownership/criminal liability. Carver's separately scoped originals/reader custody and source chronology govern the origins packet. No operational content, accounts, inquiries or funder contact occurred. Acquisition is closed for combined review/forest.\n", "file": "research/kybc-policy-wave-2026-09-09-research-input.md" }, { "id": "liberty-tcn-proceedings-wave-2026-09-09-proceedings", "title": "Liberty's TCN challenge predates the identified DFF grant", "description": "Closed on 9 September 2026 after the recipient account and one official judgment, including its publication landing page. The official record confirms proceedings existed in March 2025; the identified DFF grant was reported approved in June 2026. It therefore supports an existing challenge. These records do not establish how the case was initially financed, ", "html": "Closed on 9 September 2026 after the recipient account and one official judgment, including its publication landing page. The official record confirms proceedings existed in March 2025; the identified DFF grant was reported approved in June 2026. It therefore supports an existing challenge. These records do not establish how the case was initially financed, whether DFF supported it earlier, or whether the later grant reimbursed particular work.
\nLiberty's [13 March 2025 announcement](http[local research file] says legal complaints were filed that day by Liberty, Privacy International, Gus Hosein and Ben Wizner. It distinguishes challenges to the reported notice and the TCN process from an urgent request for an open hearing. The organisations instructed Leigh Day. Their account identifies the alleged notice power as section 253 of the Investigatory Powers Act 2016 and argues incompatibility with the statutory purpose, regulations, privacy and expression rights. Those arguments are the claimants' position, not findings of unlawfulness. The reported notice itself was not acquired.
\nThey also sought coordination with Apple's proceedings, attendance, disclosure and public scrutiny. No original pleading was acquired, so the precise pleaded causes of action, full remedies and jurisdictional statutory provisions remain unverified beyond this account and the court's limited description. Do not equate the asserted notice power with the statutory route by which every claim was brought. Source scope: official-reader lines 48–96, retained in proceedings-liberty2025.reader.txt. The ordinary HTML GET returned 403; only its error body was captured, and that route is closed.
[Apple Inc v Secretary of State for the Home Department, [2025] UKIPTrib 1](http[local research file] case IPT/25/68/CH, is dated 7 April 2025, following a 14 March hearing. Paragraph 33 (PDF8) confirms that Liberty, Privacy International and two individuals filed their separate complaint the day before that hearing: 13 March 2025. It identifies potential overlap and invites case-management proposals, not completed joinder (§§33–35).
\nThe tribunal dismissed the government's application to keep the basic case details private (§§32,42–43). It considered Liberty/PI's written arguments but kept the March hearing private (§37); intervention, pleading disclosure and future hearing arrangements remained unresolved (§§38,40). It also lacked power to grant US legislators' requested permission to discuss the alleged notice (§39).
\nThis was a procedural ruling, not a decision that encryption-access powers were unlawful. The public judgment is a summary/extract of a fuller private judgment (§6), and expressly neither confirms nor refutes the media account (§20). Its procedural framework discusses RIPA 2000 section 69 and Tribunal Rule 7 (§§26–30); that is not the missing pleading's complete jurisdictional basis.
\nThe judicial acknowledgment independently strengthens the initiation date; it does not independently prove a grant payment or DFF's causal role. The previously accepted DFF announcement and linked EUR 36,000 case listing remain the sources for the reported June 2026 allocation. They were not reacquired or counted as new corroboration here.
\nThe useful distinction is support for continuing litigation, participation in a procedural contest, and success on that procedural question. It would be wrong to turn these into a claim that DFF initiated the March 2025 action, that Liberty obtained a merits victory over TCNs, or that every hearing became public. Multiple parties made open-justice arguments, so acknowledgment of Liberty/PI's submissions is not proof they alone caused the result. The historic decision is not a current docket or finality check.
\nThe next discriminating record is the actual public Liberty/PI pleading or a later published tribunal disposition, if made available. That would establish the exact statutory claim route, pleaded remedies and any substantive ruling. No private judgment, secret notice, protected-party identification or further Apple chronology was sought.
\nThe [Judiciary publication page](http[local research file] was captured as proceedings-judiciary.html; its actual PDF href led to proceedings-public-judgment2025.pdf, SHA-256 d19a2c0e350f5bb7f279c7d5cb8eb3d0563a46458a214fc4f327a905d2197f3b. Both ordinary GETs succeeded. The PDF reader returned zero text lines and local text extraction was empty. The nine PDF pages were all rendered and visually read, including the caption and paragraphs 1–43. Images are proceedings-judgment-p01.png through p09.png. The .derived.txt file contains only page markers and is not a substantive text source. No OCR transcript is claimed.
proceedings-captures.json records original/error/reader distinctions, hashes, access and precise page scopes. The source record excludes unrelated Liberty cases and indexed commentary. No second litigation instrument was necessary for this bounded result. No graph, forest, current-work, visual presentation, package, account or outreach changes were made. Acquisition is closed for root review.
Single-source acquisition closed on 9 September 2026. The [regulator-hosted response](http[local research file] is dated 17 October 2025 on PDF20. Ofcom's response index displays 11 December 2025, a separate publication/listing date.
\nFull Fact advocated a broader crisis framework and stronger Ofcom supervision, coupled with constraints on discretion. The cited page 17 principally requests expansion/alignment of the crisis definition with section 175. It does not expressly propose automatic activation of a platform protocol on receipt of a ministerial notice, nor supply the wording eventually added by Ofcom. It also is not merely a request to constrain government: the wider response seeks materially broader regulatory coverage and coordination.
\nAt paragraph76, PDF17, Full Fact argues that the proposed definition is too narrowly directed at public safety. It asks to include public health and national security, aligning the definition with section 175 and comparing the EU Digital Services Act's crisis definition. It describes the existing Secretary of State route to Ofcom's media-literacy priorities or public statement notices. This is a scope-alignment argument using an existing power as a benchmark, not a newly drafted executive activation power.
\nAt paragraph77, PDF17, it asks for a clear, transparent, widely understood crisis definition, evidence and proportionality, and severity indicators. It also proposes covering clusters of misleading claims/narratives and slow-onset incidents, as well as discrete sudden events.
\nAt paragraph78, PDF17–18, it says government bodies likely already have incident definitions, criticizes the lack of transparency around that framework, and recommends Ofcom seek insight and align definitions to improve public/private coordination. The government-practice claims are Full Fact's account; the underlying sources were not acquired here.
\nAt paragraph79, PDF18, it recognizes that Ofcom's judgment about whether an incident occurred may differ from a provider's judgment. It proposes shared indicators to make those judgments more understandable. It does not confer authority on Full Fact to declare a binding crisis.
\n| Request within the same submission | \nPrecise location | \n
|---|---|
| Legislative review to address systemic and collective misinformation harms, including legal-but-harmful adult content, with expression safeguards. Full Fact acknowledges wider legal reform lies beyond this consultation. | \nParagraphs12–15, PDF4 | \n
| A centralized Ofcom information-incident framework; advance communication channels with police, government, other providers and civil society; provider notifications to Ofcom and users when protocols start/end. | \nParagraphs29–34, PDF7–8 | \n
| Stronger Ofcom supervision of effectiveness and proportionality; remedial steps where necessary, including protection of expression. Provider discretion should sit within clearer command structures and accountability. | \nParagraphs38–44, PDF9–10 | \n
| Written decision logs, evidence preservation where appropriate, indicators/risk assessments for de-escalation and ending a crisis, and clear responsibility for starting/stopping it. The stated purpose includes avoiding indefinite crises and misuse. | \nParagraphs62–67, PDF14–15 | \n
| Routine submission of post-crisis analyses to Ofcom, ongoing regulatory review, and publicly available summaries of Ofcom's assessment/recommendations. | \nParagraphs71–72, PDF16 | \n
These are proposed process safeguards, not evidence that an affected user's post or account would be restored or that a particular appeal succeeded. Full Fact's proportionality argument also criticizes a focus on minimizing provider costs and asks that the severity of harms be weighed (paragraphs86–90, PDF20). Its argument combines stronger intervention with accountability; either half alone would misrepresent the response.
\nA related institutional boundary appears in paragraph52, PDF12: Full Fact acknowledges that Ofcom cannot currently mandate changes to platforms' partnership policies to require working with fact checkers. Paragraphs53–54, PDF12 separately seek guaranteed, meaningful data access for accredited fact checkers and enforceable consequences for non-cooperation. This is advocacy for a future entitlement, not an existing award, partnership mandate or acquired data-access right. No donor or organizational genealogy was pursued.
\nThe accepted root operational note records Ofcom's statement 3.46–47/footnote29, PDF17, identifying Full Fact's response p17 and an Ofcom/DSIT meeting on 4 August 2025 as inputs to questions about the interaction. Ofcom's resulting amendment says providers should consider an Ofcom public statement notice alongside their other indicators.
\nReading the actual response narrows that attribution: Full Fact supplied a definition/coordination argument involving section 175, while the regulator's eventual notice-consideration wording is not found as an explicit Full Fact instruction in the inspected text. The response's 17 October date is later than the 4 August meeting. It cannot by itself prove Full Fact initiated that meeting's question, attended it, co-drafted the amendment or controlled the wording. No such inference is made.
\nThe strongest rival to a direct ministerial-trigger lobbying account is that Ofcom resolved an interaction surfaced by differently framed stakeholder inputs. The remaining discriminating record is the 4 August2025 Ofcom/DSIT meeting note or drafting record showing what was asked, by whom, and how the final wording was chosen. Root owns that separate public search.
\nThe ordinary GET returned 403 and was closed. No original PDF was acquired. fullfact-reader.txt retains 825 of 829 official-reader lines, covering substantive paragraphs 1–90 across 20 PDF pages; omitted lines 577–580 are explicitly marked. Decisive paragraphs76–79 and surrounding authority/safeguard passages were read in full. The extracted PDF has irregular internal spacing, preserved rather than silently normalized. No images or linked case evidence, DSA text, Full Fact framework, government framework or external report were independently inspected in this lane.
\nfullfact-captures.json records source/date/scope, reader custody and the retained 403 error. The reader derivative is not original-byte custody. Acquisition is complete; no broader branch was opened.
", "path": "ministerial-notice-origins-wave-2026-09-09/fullfact-origin.md", "sourceIds": [ "w31-fa-s-fullfact" ], "markdown": "Single-source acquisition closed on 9 September 2026. The regulator-hosted response (see source references) is dated **17 October 2025** on PDF20. Ofcom's response index displays 11 December 2025, a separate publication/listing date.\n\nFull Fact advocated a broader crisis framework and stronger Ofcom supervision, coupled with constraints on discretion. The cited page 17 principally requests expansion/alignment of the crisis definition with section 175. It does **not expressly propose automatic activation of a platform protocol on receipt of a ministerial notice**, nor supply the wording eventually added by Ofcom. It also is not merely a request to constrain government: the wider response seeks materially broader regulatory coverage and coordination.\n\n## What page 17 actually says\n\nAt **paragraph76, PDF17**, Full Fact argues that the proposed definition is too narrowly directed at public safety. It asks to include public health and national security, aligning the definition with section 175 and comparing the EU Digital Services Act's crisis definition. It describes the existing Secretary of State route to Ofcom's media-literacy priorities or public statement notices. This is a scope-alignment argument using an existing power as a benchmark, not a newly drafted executive activation power.\n\nAt **paragraph77, PDF17**, it asks for a clear, transparent, widely understood crisis definition, evidence and proportionality, and severity indicators. It also proposes covering clusters of misleading claims/narratives and slow-onset incidents, as well as discrete sudden events.\n\nAt **paragraph78, PDF17–18**, it says government bodies likely already have incident definitions, criticizes the lack of transparency around that framework, and recommends Ofcom seek insight and align definitions to improve public/private coordination. The government-practice claims are Full Fact's account; the underlying sources were not acquired here.\n\nAt **paragraph79, PDF18**, it recognizes that Ofcom's judgment about whether an incident occurred may differ from a provider's judgment. It proposes shared indicators to make those judgments more understandable. It does not confer authority on Full Fact to declare a binding crisis.\n\n## Surrounding authority and safeguards\n\n| Request within the same submission | Precise location |\n| --- | --- |\n| Legislative review to address systemic and collective misinformation harms, including legal-but-harmful adult content, with expression safeguards. Full Fact acknowledges wider legal reform lies beyond this consultation. | Paragraphs12–15, PDF4 |\n| A centralized Ofcom information-incident framework; advance communication channels with police, government, other providers and civil society; provider notifications to Ofcom and users when protocols start/end. | Paragraphs29–34, PDF7–8 |\n| Stronger Ofcom supervision of effectiveness and proportionality; remedial steps where necessary, including protection of expression. Provider discretion should sit within clearer command structures and accountability. | Paragraphs38–44, PDF9–10 |\n| Written decision logs, evidence preservation where appropriate, indicators/risk assessments for de-escalation and ending a crisis, and clear responsibility for starting/stopping it. The stated purpose includes avoiding indefinite crises and misuse. | Paragraphs62–67, PDF14–15 |\n| Routine submission of post-crisis analyses to Ofcom, ongoing regulatory review, and publicly available summaries of Ofcom's assessment/recommendations. | Paragraphs71–72, PDF16 |\n\nThese are proposed process safeguards, not evidence that an affected user's post or account would be restored or that a particular appeal succeeded. Full Fact's proportionality argument also criticizes a focus on minimizing provider costs and asks that the severity of harms be weighed (paragraphs86–90, PDF20). Its argument combines stronger intervention with accountability; either half alone would misrepresent the response.\n\nA related institutional boundary appears in **paragraph52, PDF12**: Full Fact acknowledges that Ofcom cannot currently mandate changes to platforms' partnership policies to require working with fact checkers. **Paragraphs53–54, PDF12** separately seek guaranteed, meaningful data access for accredited fact checkers and enforceable consequences for non-cooperation. This is advocacy for a future entitlement, not an existing award, partnership mandate or acquired data-access right. No donor or organizational genealogy was pursued.\n\n## What can be joined to Ofcom's drafting change\n\nThe accepted [root operational note](ofcom-crisis-authority-wave-2026-09-09-crisis-authority.md) records Ofcom's statement 3.46–47/footnote29, PDF17, identifying Full Fact's response p17 and an Ofcom/DSIT meeting on 4 August 2025 as inputs to questions about the interaction. Ofcom's resulting amendment says providers should consider an Ofcom public statement notice alongside their other indicators.\n\nReading the actual response narrows that attribution: Full Fact supplied a definition/coordination argument involving section 175, while the regulator's eventual notice-consideration wording is not found as an explicit Full Fact instruction in the inspected text. The response's 17 October date is later than the 4 August meeting. It cannot by itself prove Full Fact initiated that meeting's question, attended it, co-drafted the amendment or controlled the wording. No such inference is made.\n\nThe strongest rival to a direct ministerial-trigger lobbying account is that Ofcom resolved an interaction surfaced by differently framed stakeholder inputs. The remaining discriminating record is the 4 August2025 Ofcom/DSIT meeting note or drafting record showing what was asked, by whom, and how the final wording was chosen. Root owns that separate public search.\n\n## Custody and reading limits\n\nThe ordinary GET returned 403 and was closed. No original PDF was acquired. fullfact-reader.txt retains 825 of 829 official-reader lines, covering substantive paragraphs 1–90 across 20 PDF pages; omitted lines 577–580 are explicitly marked. Decisive paragraphs76–79 and surrounding authority/safeguard passages were read in full. The extracted PDF has irregular internal spacing, preserved rather than silently normalized. No images or linked case evidence, DSA text, Full Fact framework, government framework or external report were independently inspected in this lane.\n\nfullfact-captures.json records source/date/scope, reader custody and the retained 403 error. The reader derivative is not original-byte custody. Acquisition is complete; no broader branch was opened.\n", "file": "research/ministerial-notice-origins-wave-2026-09-09-fullfact-origin.md" }, { "id": "ministerial-notice-origins-wave-2026-09-09-meeting-origin", "title": "The cited DSIT meeting and confirmed planning records", "description": "Root's bounded public acquisition closed9September2026. Ofcom's June crisis statement3.46-47/footnote29, PDF17, cites an Ofcom/DSIT meeting on4August2025, alongside Full Fact's responsep17, as questions about the interaction of section175 directions and the crisis measures. The statement then identifies its amendment requiring consideration of an issued publ", "html": "Root's bounded public acquisition closed9September2026. Ofcom's June crisis statement3.46-47/footnote29, PDF17, cites an Ofcom/DSIT meeting on4August2025, alongside Full Fact's responsep17, as questions about the interaction of section175 directions and the crisis measures. The statement then identifies its amendment requiring consideration of an issued public statement notice. The dated regulator attribution is established; an agenda, attendee list, minute, submitted briefing or agreed action from that meeting was not obtained.
\nFive bounded exact-date/actor searches, plus the existing Ofcom FOI-response index searched for DSIT/date/crisis, did not locate the meeting document. An unrelated spectrum document's printing date, broad ministerial meeting lists from a different quarter, and unrelated spending records were not treated as meeting evidence. There is no inference that the meeting was ministerial rather than official-level, involved a named minister, was secret, or was held at Full Fact's request. A dated citation establishes a claimed exchange; it does not identify each participant's position or approval rights.
\nThe FOI index did expose a directly relevant existing response. Ofcom reference02210914, dated19June2026, replies to a request received12June about the protocol's drafting timeline and emergency-enforcement criteria. Original URL: http[local research file] . All95reader lines across four pages were read; crisis-planning-foi.reader.txt retains the official-reader derivative. No original PDF bytes acquired and no new request sent.
\nIn its own answer to1(a)-(c), PDF2, Ofcom confirms an initial internal planning document, says work commenced23October2024 following17October parliamentary and22October Secretary of State correspondence, and says final approval was completed between21May and26May2026. These are agency statements about its records, not the underlying planning document, letters or signed approvals. They establish a precise prior planning/approval trail to seek; they do not disclose the4August2025meeting or who approved the text. The requester-provided Belfast narrative onPDF1 is separately attributed and is not Ofcom's confirmation of every background claim.
\nQuestion3 asks for a decision record explaining why the10June2026open letter was selected instead of formal enforcement. Ofcom says it does not hold information within that question's scope, explaining that sending a conducive letter did not exclude other actions and no such either/or decision was made. This is a scoped response, not a claim that no briefing or approval for the letter exists at all. The response directs emergency-enforcement criteria to its published guidance3.8-3.9; that guidance was not reacquired in this lane. Its statement that a separate Grok investigation remained open is historical19June status and is not adopted as current.
\nMost revealing remaining records: the meeting note/action log or briefing for4August2025; the confirmed initial planning document; the21-26May2026approval material and comparison against the consulted draft. Ofcom and DSIT are indicated holders for the meeting; Ofcom expressly confirms holding the planning record. Public acquisition is finite and closed. Any later information request requires separate outreach authority. The Full Fact passage is acquired in the independent companion lane, including its actual17October2025document date; that later date prevents treating the submission itself as proof it initiated the August meeting.
", "path": "ministerial-notice-origins-wave-2026-09-09/meeting-origin.md", "sourceIds": [ "w31-fa-s-planningfoi" ], "markdown": "Root's bounded public acquisition closed9September2026. Ofcom's June crisis statement3.46-47/footnote29, PDF17, cites an Ofcom/DSIT meeting on4August2025, alongside Full Fact's responsep17, as questions about the interaction of section175 directions and the crisis measures. The statement then identifies its amendment requiring consideration of an issued public statement notice. The dated regulator attribution is established; an agenda, attendee list, minute, submitted briefing or agreed action from that meeting was not obtained.\n\nFive bounded exact-date/actor searches, plus the existing Ofcom FOI-response index searched for DSIT/date/crisis, did not locate the meeting document. An unrelated spectrum document's printing date, broad ministerial meeting lists from a different quarter, and unrelated spending records were not treated as meeting evidence. There is no inference that the meeting was ministerial rather than official-level, involved a named minister, was secret, or was held at Full Fact's request. A dated citation establishes a claimed exchange; it does not identify each participant's position or approval rights.\n\nThe FOI index did expose a directly relevant existing response. Ofcom reference02210914, dated19June2026, replies to a request received12June about the protocol's drafting timeline and emergency-enforcement criteria. Original URL: http[local research file] . All95reader lines across four pages were read; crisis-planning-foi.reader.txt retains the official-reader derivative. No original PDF bytes acquired and no new request sent.\n\nIn its own answer to1(a)-(c), PDF2, Ofcom confirms an initial internal planning document, says work commenced23October2024 following17October parliamentary and22October Secretary of State correspondence, and says final approval was completed between21May and26May2026. These are agency statements about its records, not the underlying planning document, letters or signed approvals. They establish a precise prior planning/approval trail to seek; they do not disclose the4August2025meeting or who approved the text. The requester-provided Belfast narrative onPDF1 is separately attributed and is not Ofcom's confirmation of every background claim.\n\nQuestion3 asks for a decision record explaining why the10June2026open letter was selected instead of formal enforcement. Ofcom says it does not hold information within that question's scope, explaining that sending a conducive letter did not exclude other actions and no such either/or decision was made. This is a scoped response, not a claim that no briefing or approval for the letter exists at all. The response directs emergency-enforcement criteria to its published guidance3.8-3.9; that guidance was not reacquired in this lane. Its statement that a separate Grok investigation remained open is historical19June status and is not adopted as current.\n\nMost revealing remaining records: the meeting note/action log or briefing for4August2025; the confirmed initial planning document; the21-26May2026approval material and comparison against the consulted draft. Ofcom and DSIT are indicated holders for the meeting; Ofcom expressly confirms holding the planning record. Public acquisition is finite and closed. Any later information request requires separate outreach authority. The Full Fact passage is acquired in the independent companion lane, including its actual17October2025document date; that later date prevents treating the submission itself as proof it initiated the August meeting.\n", "file": "research/ministerial-notice-origins-wave-2026-09-09-meeting-origin.md" }, { "id": "naag-payment-letter-wave-2026-09-09-letter-mechanism", "title": "Public officials ask payment networks to use private rules", "description": "9 September 2026. Bounded run: who originated the August 2025 multistate NCII intervention, what action it sought, and whether a response or consequence is established. NCII means nonconsensual intimate imagery. This packet does not investigate or link to the offending services themselves.", "html": "9 September 2026. Bounded run: who originated the August 2025 multistate NCII intervention, what action it sought, and whether a response or consequence is established. NCII means nonconsensual intimate imagery. This packet does not investigate or link to the offending services themselves.
\nThe [22 August 2025 combined letters](http[local research file] contain separate payment and search interventions. Payment addressees are the legal officers of Visa, Mastercard, American Express, PayPal, Google and Apple, addressed as Google Pay and Apple Pay. Block/Cash App is not a payment addressee. The payment body asks for current controls and further action, using existing terms and acceptable-use policies to identify and remove sellers of deepfake NCII tools/content proactively. This is seller-level exclusion, not merely deletion of an individual image.
\nThe payment body does not supply an operative statutory provision, case number, sanction, compulsory response deadline, seller list or named bank instruction. Its First Amendment assurance is the authors' position, not a judicial decision. Prior Pornhub and Civitai cutoffs appear as cited precedents; they cannot be outcomes caused by this later letter. The separate search letter asks for restrictions, warnings and redirects and refers to the Take It Down Act. Neither instrument establishes a recipient's actual subsequent decision. See PDF1-3 and9-11.
\nPennsylvania's 26 August announcement.
\nThe original visibly uses NAAG letterhead and the 2025 officers: John Formella, William Tong, Marty Jackley and Letitia James, with Brian Kane as executive director. The current September 2026 leadership page belongs to a different observation date. Published May 2026 signature procedures describe a later process; without the actual 2025 circulation record they do not establish the historical approval steps for these letters. Actual sponsor instructions, edits, signature confirmations and transmission/receipt records remain unacquired.
\nIn his [19 September 2025 official newsletter](http[local research file] signatory Raul Labrador describes the coalition letters as a choice between safeguards and legal consequences. Root read the complete substantive body, including the links to the two letters. This makes an account of entirely consequence-free persuasion inadequate: at least one participating AG publicly attached threatened legal consequences to the demand. The newsletter still supplies no filed case, operative order, meeting, recipient response or completed company change. It is not evidence that every co-signatory adopted that later characterization. Its statements about search behavior and prevalence are advocacy claims, not independently tested results in this run.
\nThe supported mechanism is coordinated public-office pressure directed at company counsel, seeking decisions under private network rules. This offers a concrete conversion point between government advocacy and commercial exclusion. It does not require proving that an AG owns or controls a payment company. The July Cash App NAAG reimbursement belongs to a different investigation; no allocation or expense record connects that money to this letter campaign.
\nAt the same time, the presence of company discretion does not establish an unconstrained or wholly voluntary choice. Official threats, existing law, reputational exposure and company policy may interact. We need a recipient's reply, counsel assessment, merchant/acquirer instruction, actual restriction or later enforcement record to determine which interaction occurred here. We have no basis to assign these letters responsibility for Steam removals, a general pornography prohibition, or universal identity requirements. The seller-level scope makes collateral reach a concrete question, but actual overbreadth requires a worked case.
\nRoot read all 15 pages of the NAAG combined-letter web-reader representation in the preceding acquisition phase and all returned text of the Pennsylvania indexed copy in this run. Root reread the retained live Pennsylvania payment body and visually inspected PDF1; the content and 15-page structure match the inspected representations at the substantive clauses, without claiming byte identity across hosts. Root also read the entire retained Pennsylvania announcement main body. Vermont's complete substantive announcement was returned in search; direct reader open returned403, so this packet does not pretend that route succeeded.
\nThe indexed Pennsylvania /2025/09/ PDF URL returned404 through ordinary GET despite a working indexed reader. Its retained pa-letters.pdf is an error response, explicitly marked false in captures-root.json. The exact PDF href in the successful announcement led to pa-letters-live.pdf, a successful original: 1,397,552 bytes, SHA256 75db0aba797ff32ebf992254f169571c2385611f171fb87b6f055e6e09c88a97. PDF metadata shows a September2 modification, not a proven publication or drafting date. No metadata-based authorship inference is made. Pennsylvania announcement: 122,926 bytes, SHA256 c1780ae668b126d546a9dece7107b5171c7bf666a91a3cafdf09cfaa0ed9b120. The failed path is closed, not repeatedly retried.
\nIndependent letter review is in independent-letter-review.md; response-search.md records the bounded outcome lane and any secondary follow-up leads. The letter's cited prevalence infographic has not been audited; its percentages are not adopted as this campaign's findings. No outreach, target-service use, paid route, account access or submission occurred.
", "path": "naag-payment-letter-wave-2026-09-09/letter-mechanism.md", "sourceIds": [ "w24-ag-s-pa-authorship", "w24-ag-s-letters" ], "markdown": "9 September 2026. Bounded run: who originated the August 2025 multistate NCII intervention, what action it sought, and whether a response or consequence is established. NCII means nonconsensual intimate imagery. This packet does not investigate or link to the offending services themselves.\n\n## The instrument and the action\n\nThe 22 August 2025 combined letters (see source references) contain separate payment and search interventions. Payment addressees are the legal officers of Visa, Mastercard, American Express, PayPal, Google and Apple, addressed as Google Pay and Apple Pay. Block/Cash App is not a payment addressee. The payment body asks for current controls and further action, using existing terms and acceptable-use policies to identify and remove sellers of deepfake NCII tools/content proactively. This is seller-level exclusion, not merely deletion of an individual image.\n\nThe payment body does not supply an operative statutory provision, case number, sanction, compulsory response deadline, seller list or named bank instruction. Its First Amendment assurance is the authors' position, not a judicial decision. Prior Pornhub and Civitai cutoffs appear as cited precedents; they cannot be outcomes caused by this later letter. The separate search letter asks for restrictions, warnings and redirects and refers to the Take It Down Act. Neither instrument establishes a recipient's actual subsequent decision. See PDF1-3 and9-11.\n\n## Origin and collective endorsement\n\nPennsylvania's 26 August announcement.\n\nThe original visibly uses NAAG letterhead and the 2025 officers: John Formella, William Tong, Marty Jackley and Letitia James, with Brian Kane as executive director. The current September 2026 leadership page belongs to a different observation date. Published May 2026 signature procedures describe a later process; without the actual 2025 circulation record they do not establish the historical approval steps for these letters. Actual sponsor instructions, edits, signature confirmations and transmission/receipt records remain unacquired.\n\n## A later threat changes the interpretation\n\nIn his 19 September 2025 official newsletter (see source references) signatory Raul Labrador describes the coalition letters as a choice between safeguards and legal consequences. Root read the complete substantive body, including the links to the two letters. This makes an account of entirely consequence-free persuasion inadequate: at least one participating AG publicly attached threatened legal consequences to the demand. The newsletter still supplies no filed case, operative order, meeting, recipient response or completed company change. It is not evidence that every co-signatory adopted that later characterization. Its statements about search behavior and prevalence are advocacy claims, not independently tested results in this run.\n\n## What this joins, and what it does not\n\nThe supported mechanism is coordinated public-office pressure directed at company counsel, seeking decisions under private network rules. This offers a concrete conversion point between government advocacy and commercial exclusion. It does not require proving that an AG owns or controls a payment company. The July Cash App NAAG reimbursement belongs to a different investigation; no allocation or expense record connects that money to this letter campaign.\n\nAt the same time, the presence of company discretion does not establish an unconstrained or wholly voluntary choice. Official threats, existing law, reputational exposure and company policy may interact. We need a recipient's reply, counsel assessment, merchant/acquirer instruction, actual restriction or later enforcement record to determine which interaction occurred here. We have no basis to assign these letters responsibility for Steam removals, a general pornography prohibition, or universal identity requirements. The seller-level scope makes collateral reach a concrete question, but actual overbreadth requires a worked case.\n\n## Reading and custody\n\nRoot read all 15 pages of the NAAG combined-letter web-reader representation in the preceding acquisition phase and all returned text of the Pennsylvania indexed copy in this run. Root reread the retained live Pennsylvania payment body and visually inspected PDF1; the content and 15-page structure match the inspected representations at the substantive clauses, without claiming byte identity across hosts. Root also read the entire retained Pennsylvania announcement main body. Vermont's complete substantive announcement was returned in search; direct reader open returned403, so this packet does not pretend that route succeeded.\n\nThe indexed Pennsylvania /2025/09/ PDF URL returned404 through ordinary GET despite a working indexed reader. Its retained pa-letters.pdf is an error response, explicitly marked false in captures-root.json. The exact PDF href in the successful announcement led to pa-letters-live.pdf, a successful original: 1,397,552 bytes, SHA256 75db0aba797ff32ebf992254f169571c2385611f171fb87b6f055e6e09c88a97. PDF metadata shows a September2 modification, not a proven publication or drafting date. No metadata-based authorship inference is made. Pennsylvania announcement: 122,926 bytes, SHA256 c1780ae668b126d546a9dece7107b5171c7bf666a91a3cafdf09cfaa0ed9b120. The failed path is closed, not repeatedly retried.\n\nIndependent letter review is in independent-letter-review.md; response-search.md records the bounded outcome lane and any secondary follow-up leads. The letter's cited prevalence infographic has not been audited; its percentages are not adopted as this campaign's findings. No outreach, target-service use, paid route, account access or submission occurred.\n", "file": "research/naag-payment-letter-wave-2026-09-09-letter-mechanism.md" }, { "id": "naag-payment-letter-wave-2026-09-09-response-search", "title": "Response/outcome search: 22 August 2025 NAAG deepfake-NCII letters", "description": "Observed 9 September 2026. Bounded public-source search closed. This lane did not acquire a company reply, an attributed policy change, meeting record or AG compliance determination tied specifically to the 22 August 2025 letters. That is a scoped result, not evidence that no private reply, meeting or resulting change exists. Parent independently owns acquis", "html": "Observed 9 September 2026. Bounded public-source search closed. This lane did not acquire a company reply, an attributed policy change, meeting record or AG compliance determination tied specifically to the 22 August 2025 letters. That is a scoped result, not evidence that no private reply, meeting or resulting change exists. Parent independently owns acquisition and reading of the original combined letters and Pennsylvania announcement.
\nIdaho Attorney General Raul Labrador's 19 September 2025 newsletter returns to the 47-AG letters, restates demands on payment processors and search providers, and presents implementation or legal consequences as the choice facing recipients. It links separate search/payment letters. This is documented subsequent public pressure by a signatory. It reports no company reply, agreed commitment, meeting, particular enforcement filing or implemented change. Its assertions about continuing company conduct remain the AG's assertions, not an independent audit of platform behavior. The newsletter's separately claimed ICAC results are not an outcome of these letters and are not used here.
\n[Idaho AG newsletter](http[local research file] complete substantive body read, web lines97-116, especially108-115. Publisher date19September2025. Official web-reader substantive read, followed by one successful ordinary GET of the original newsletter. Exact linked-letter hrefs were extracted from that HTML; linked PDF bytes were not independently compared with parent's original.
\nCalifornia's 26 August 2025 announcement describes the letters as seeking information and opening communication with companies. That is an intention to engage, not proof a conversation happened. Read with Idaho's later threat framing rather than treating either summary as the exact operative instrument. [California AG announcement](http[local research file] complete substantive announcement body lines157-184 read; relevant paragraphs166-173. This is contemporaneous framing, not a later outcome. Official web-reader only; no original bytes captured.
\nA defensible new relation is Idaho AG -> the letter campaign: publicly reiterates demands and warns of legal consequences on19September2025. It is a public warning/advocacy act, not an order, filing, adjudication or observed compliance. No company-response edge should be added from this pass.
\nThe San Francisco Chronicle's 17 July 2026 report says Chiu's July16 Apple/Google notices cite the 47-AG warning as prior notice. This is a reported legal-use connection; the original notices were not acquired, so their wording and proposed legal significance remain unverified here. [Reported later citation](http[local research file]
\nThe official former office home redirects from http[local research file] to http[local research file] with an April24,2026 archived-site boundary. Home and news links were inspected; http[local research file] returned only three web-reader lines. Targeted searches of both official domains did not locate the notices. Stop this route here; the gap is the actual July16 notices, attachments and subsequent response correspondence held by the office and recipients, not proof the reported notices do not exist.
\n[TechCrunch's separate July report](http[local research file] reports responses to San Francisco's action, not a reply to NAAG. No original response acquired.
\nMagica's derivative summary was inspected solely to look for links to originals; its citation of the earlier warning leads to the Chronicle, not an original notice. Its link list does not independently corroborate that lineage. Scope: date/title and main body77-124 plus source links125-138. [Derivative discovery page](http[local research file] Do not promote its app-count, financial or legal analyses from this search.
\nThe 29 August 2025 Payment Expert article repeats the letters' requests and prior interventions, and contains no company response in its inspected substantive body. Scope: lines165-221. It is secondary reporting, not an official reply or outcome. Its incidental legal summary was not adopted. [Contemporaneous reporting](http[local research file]
\nCivitai/Pornhub interventions cited as precedents in the letters must stay antecedent examples. This pass did not freshly investigate their events or infer they were caused by the later letters. Steam/adult-game restrictions, separate AI-company letters, xAI actions, vaping and other branches were excluded. Search results about those subjects were discovery noise, not accepted evidence. Company measures or legislative activity that predate22August2025 cannot be counted as responses merely because search returned them; no complete policy-version history was reconstructed.
\nOne focused pass used the following18 query strings, without a recency restriction. Search engine indexing and compound-domain query interpretation limit completeness; these are not exhaustive searches of every company or AG record system.
\n"August 22, 2025" "attorneys general" "payment" deepfake response\n"NAAG" "47" "deepfake" "response"\n"attorneys general" "deepfake" "August" "2025" Mastercard reply\n"attorneys general" "deepfake" "August" "2025" Google response\nsite.sfcityattorney.org Apple Google nudify "2026" "August"\nsite.mastercard.com "attorneys general" "deepfake"\nsite.visa.com "attorneys general" "deepfake"\n"August 22, 2025" "deepfake" "respond"\n"Chiu" "July 17, 2026" "Apple" "cityattorney"\nsite.sfcityattorney.org "nudify"\nsite.sfcityattorney.org "nonconsensual" "Apple"\n"47" "deepfake" "response" "PayPal"\nsite.sf.gov "Chiu" "Apple" "nudify"\nsite.blog.google "attorneys general" "deepfake"\nsite.naag.org "deepfake" "response" "payment"\n"deepfake" "attorneys general" response site:paypal.com OR site:americanexpress.com OR site:apple.com\n"deepfake" "attorneys general" response site:microsoft.com OR site:blogs.bing.com OR site:yahooinc.com\n"August 22" "2025" "NCII" "response"\nCompany-domain searches did not surface a specific original reply to these letters. They did produce pre-existing material and unrelated hosted content; domain appearance alone was not treated as corporate authorship. Exact-date searches also produced unrelated meanings of NCII. No broad archive enumeration, authenticated access, public-record request, contact, spending or case mutation occurred. Web-reader locators are retrieval-specific; this file is a scoped reading/search note, not a preserved set of website originals.
\nFor response or compliance: recipient general-counsel replies explicitly identifying these letters, the lead AGs' correspondence/meeting log and any agreed implementation or escalation document. A dated internal or public policy change must identify this intervention or a matched correspondence chain before causal attribution. Lack of public outcome reporting could reflect private engagement, refusal, incomplete public indexing, no action, or action for other reasons; this search does not select among them. For the narrower reported SF lineage, obtain the actual July16 notices before mapping the 2025 warning into that later legal theory. No subsequent branch is started by this return.
\nAt parent request, one ordinary unauthenticated GET succeeded with HTTP200 at2026-09-09T11:40:24.462431+00:00. Retained response-idaho-newsletter.html,62,317bytes, SHA-2564e8a67b3aafe342a0dfd53fa853aae662a4ddda5bdd9b17b32d24ce6cbf78608. response-idaho-capture.json records scope and exact source hrefs. These substantiate where the newsletter directs readers; the linked PDFs were not newly fetched or treated as independent outcomes.
\nSearch-letter href: http[local research file]
\nPayment-letter href: http[local research file]
\nThe September upload directory does not independently establish a different letter date. Parent has separately retained the Pennsylvania combined original at its current exact href; this lane makes no byte-identity claim about the separate Idaho PDFs. Acquisition remains closed.
", "path": "naag-payment-letter-wave-2026-09-09/response-search.md", "sourceIds": [ "w24-ag-s-idaho" ], "markdown": "Observed 9 September 2026. Bounded public-source search closed. This lane did not acquire a company reply, an attributed policy change, meeting record or AG compliance determination tied specifically to the 22 August 2025 letters. That is a scoped result, not evidence that no private reply, meeting or resulting change exists. Parent independently owns acquisition and reading of the original combined letters and Pennsylvania announcement.\n\n## Strongest primary follow-up\n\nIdaho Attorney General Raul Labrador's 19 September 2025 newsletter returns to the 47-AG letters, restates demands on payment processors and search providers, and presents implementation or legal consequences as the choice facing recipients. It links separate search/payment letters. This is documented subsequent public pressure by a signatory. It reports no company reply, agreed commitment, meeting, particular enforcement filing or implemented change. Its assertions about continuing company conduct remain the AG's assertions, not an independent audit of platform behavior. The newsletter's separately claimed ICAC results are not an outcome of these letters and are not used here.\n\nIdaho AG newsletter (see source references) complete substantive body read, web lines97-116, especially108-115. Publisher date19September2025. Official web-reader substantive read, followed by one successful ordinary GET of the original newsletter. Exact linked-letter hrefs were extracted from that HTML; linked PDF bytes were not independently compared with parent's original.\n\nCalifornia's 26 August 2025 announcement describes the letters as seeking information and opening communication with companies. That is an intention to engage, not proof a conversation happened. Read with Idaho's later threat framing rather than treating either summary as the exact operative instrument. California AG announcement (see source references) complete substantive announcement body lines157-184 read; relevant paragraphs166-173. This is contemporaneous framing, not a later outcome. Official web-reader only; no original bytes captured.\n\nA defensible new relation is Idaho AG -> the letter campaign: publicly reiterates demands and warns of legal consequences on19September2025. It is a public warning/advocacy act, not an order, filing, adjudication or observed compliance. No company-response edge should be added from this pass.\n\n## Promising later lead, original still missing\n\nThe San Francisco Chronicle's 17 July 2026 report says Chiu's July16 Apple/Google notices cite the 47-AG warning as prior notice. This is a reported legal-use connection; the original notices were not acquired, so their wording and proposed legal significance remain unverified here. Reported later citation (see source references)\n\nThe official former office home redirects from http[local research file] to http[local research file] with an April24,2026 archived-site boundary. Home and news links were inspected; http[local research file] returned only three web-reader lines. Targeted searches of both official domains did not locate the notices. Stop this route here; the gap is the actual July16 notices, attachments and subsequent response correspondence held by the office and recipients, not proof the reported notices do not exist.\n\nTechCrunch's separate July report (see source references) reports responses to San Francisco's action, not a reply to NAAG. No original response acquired.\n\nMagica's derivative summary was inspected solely to look for links to originals; its citation of the earlier warning leads to the Chronicle, not an original notice. Its link list does not independently corroborate that lineage. Scope: date/title and main body77-124 plus source links125-138. Derivative discovery page (see source references) Do not promote its app-count, financial or legal analyses from this search.\n\n## False-positive and chronology boundaries\n\nThe 29 August 2025 Payment Expert article repeats the letters' requests and prior interventions, and contains no company response in its inspected substantive body. Scope: lines165-221. It is secondary reporting, not an official reply or outcome. Its incidental legal summary was not adopted. Contemporaneous reporting (see source references)\n\nCivitai/Pornhub interventions cited as precedents in the letters must stay antecedent examples. This pass did not freshly investigate their events or infer they were caused by the later letters. Steam/adult-game restrictions, separate AI-company letters, xAI actions, vaping and other branches were excluded. Search results about those subjects were discovery noise, not accepted evidence. Company measures or legislative activity that predate22August2025 cannot be counted as responses merely because search returned them; no complete policy-version history was reconstructed.\n\n## Exact search scope\n\nOne focused pass used the following18 query strings, without a recency restriction. Search engine indexing and compound-domain query interpretation limit completeness; these are not exhaustive searches of every company or AG record system.\n\n```\n\"August 22, 2025\" \"attorneys general\" \"payment\" deepfake response\n\"NAAG\" \"47\" \"deepfake\" \"response\"\n\"attorneys general\" \"deepfake\" \"August\" \"2025\" Mastercard reply\n\"attorneys general\" \"deepfake\" \"August\" \"2025\" Google response\nsite.sfcityattorney.org Apple Google nudify \"2026\" \"August\"\nsite.mastercard.com \"attorneys general\" \"deepfake\"\nsite.visa.com \"attorneys general\" \"deepfake\"\n\"August 22, 2025\" \"deepfake\" \"respond\"\n\"Chiu\" \"July 17, 2026\" \"Apple\" \"cityattorney\"\nsite.sfcityattorney.org \"nudify\"\nsite.sfcityattorney.org \"nonconsensual\" \"Apple\"\n\"47\" \"deepfake\" \"response\" \"PayPal\"\nsite.sf.gov \"Chiu\" \"Apple\" \"nudify\"\nsite.blog.google \"attorneys general\" \"deepfake\"\nsite.naag.org \"deepfake\" \"response\" \"payment\"\n\"deepfake\" \"attorneys general\" response site:paypal.com OR site:americanexpress.com OR site:apple.com\n\"deepfake\" \"attorneys general\" response site:microsoft.com OR site:blogs.bing.com OR site:yahooinc.com\n\"August 22\" \"2025\" \"NCII\" \"response\"\n```\n\nCompany-domain searches did not surface a specific original reply to these letters. They did produce pre-existing material and unrelated hosted content; domain appearance alone was not treated as corporate authorship. Exact-date searches also produced unrelated meanings of NCII. No broad archive enumeration, authenticated access, public-record request, contact, spending or case mutation occurred. Web-reader locators are retrieval-specific; this file is a scoped reading/search note, not a preserved set of website originals.\n\n## Next record that would change the inference\n\nFor response or compliance: recipient general-counsel replies explicitly identifying these letters, the lead AGs' correspondence/meeting log and any agreed implementation or escalation document. A dated internal or public policy change must identify this intervention or a matched correspondence chain before causal attribution. Lack of public outcome reporting could reflect private engagement, refusal, incomplete public indexing, no action, or action for other reasons; this search does not select among them. For the narrower reported SF lineage, obtain the actual July16 notices before mapping the 2025 warning into that later legal theory. No subsequent branch is started by this return.\n\n## Final Idaho custody capture\n\nAt parent request, one ordinary unauthenticated GET succeeded with HTTP200 at2026-09-09T11:40:24.462431+00:00. Retained response-idaho-newsletter.html,62,317bytes, SHA-2564e8a67b3aafe342a0dfd53fa853aae662a4ddda5bdd9b17b32d24ce6cbf78608. response-idaho-capture.json records scope and exact source hrefs. These substantiate where the newsletter directs readers; the linked PDFs were not newly fetched or treated as independent outcomes.\n\nSearch-letter href: http[local research file]\n\nPayment-letter href: http[local research file]\n\nThe September upload directory does not independently establish a different letter date. Parent has separately retained the Pennsylvania combined original at its current exact href; this lane makes no byte-identity claim about the separate Idaho PDFs. Acquisition remains closed.\n", "file": "research/naag-payment-letter-wave-2026-09-09-response-search.md" }, { "id": "naag-resources-wave-2026-09-09-finance", "title": "NAAG settlement resources: bounded finance return", "description": "Observed 9 September 2026. Acquisition is closed at the fund-instrument and legal-identity boundary. This packet does not establish receipt of the Cash App reimbursement, a fund allocation of it, or motive for selecting an enforcement target.", "html": "Observed 9 September 2026. Acquisition is closed at the fund-instrument and legal-identity boundary. This packet does not establish receipt of the Cash App reimbursement, a fund allocation of it, or motive for selecting an enforcement target.
\nPublished NAAG fund instruments document a real mechanism through which earlier settlements finance later investigation capacity. They authorize segregated funds, grants from principal, paid experts and other costs, conditional replenishment after successful litigation, and NAAG administration fees. They also specify applications, receipts and reporting records that should exist. Their dates and committee structures matter: the 2012 and 2016 rules cannot simply be assigned to the seven-person consolidated fund committee currently described in the separately acquired governance packet without an amendment or delegation record.
\nThe July 2026 Cash App instrument separately provides $264,153.73 to NAAG as investigation-cost reimbursement within the $45 million state allocation (Oregon stipulation, paragraph 48, PDF15-16; retained in the prior July packet). That is an obligation in the acquired instrument, not a payment receipt. No record in this run assigns that amount to either fund below. It could reimburse already-incurred coordination costs without becoming a new investigation grant.
\nOriginal URL: http[local research file]
\nThe entire five-page official web-reader rendition was read, including Attachment 1. Ordinary unauthenticated GET returned 403 once; the retained error response is not the PDF. The accessible text and this scoped reading note are derivative evidence. The URL upload directory is 2020; section I dates Executive Committee approval to 2016.
\nSection A, PDF1, recites a June 2016 partial settlement providing approximately $631 million to participating states and $20 million to NAAG for consumer-protection oversight, training and enforcement, including automobile fraud, and investigation-cost reimbursement. These are the rule's description of settlement provisions, not newly acquired cash receipts.
\nSections B-C, PDF1-2, vest administration in a three-AG special committee: NAAG's president appoints from the Executive Committee, which approves appointments. At most two may share a political party. Two affirmative votes approve grants, including principal; decisions are final without appeal. Money must be in a new segregated account, not commingled; additional deposits must lawfully fit the purpose. NAAG invests under the special committee's approved policy. If the special committee determines the fund purpose no longer exists, the Executive Committee may dissolve it and direct remaining money to NAAG's General Fund. This is a conditional authority, not an observed transfer.
\nSection D, PDF3, requires recipients, to the extent applicable and permitted by law, to promise repayment when funded litigation succeeds, limited by available recovery. Conferences, education, training and travel-related events are exempt. Accounting is due by June 30 following disbursement. These are not uniformly repayable loans. Section E requires a temporary replacement when a special-committee member's office applies; implementation of that recusal was not inspected.
\nSection F, PDF3, authorizes NAAG a fee of 1.0% of total fund value at inception and annually thereafter, covering administration and third-party costs. The current fee amount and any Cash App linkage are unknown. Section H, PDF4, includes the fund within NAAG's annual audit procedures; no resulting audit was acquired.
\nSection B treats fund statements as public records and calls for balances, applications, deposits and withdrawals to be supplied at the annual summer meeting or on a NAAG member's request. Attachment 1, PDF5, requires written applications signed by an AG or, for NAAG itself, its executive director. Applications identify the purpose, action, amount and payee; payments may be directed to vendors, AG offices, NAAG or a designated multistate office. Bills and receipts go to NAAG finance for disbursement. The text does not establish an unrestricted public inspection procedure for confidential applications.
\nOriginal URL: http[local research file]
\nThe entire official seven-page web-reader rendition, including appendices, was read. One ordinary GET returned 403; no original PDF bytes were retained. This is the Financial Services and Consumer Protection Enforcement, Education, and Training Fund.
\nPDF1 describes all five 4 April 2012 consent judgments against Bank of America, JPMorgan Chase, Wells Fargo, Citi and Ally/GMAC as containing an Exhibit B provision directing $15 million to NAAG for the fund. Do not silently describe that as the aggregate fund, or reconstruct $75 million received: the underlying five judgments and aggregate receipts were not acquired. The MOU describes the State Monitoring Committee's development of the rules; that body is distinct from the grant special committee.
\nSections A-B, PDF2, authorize investigation/prosecution expenses, experts, consultants, information exchange and education concerning fraud and illegal or deceptive financial/consumer practices. A five-AG special committee is appointed by the NAAG president and approved by its Executive Committee, with one-year terms and at most three from one party. Grants of $5,000 or more require three votes; smaller grants require two votes from different parties. Principal can be spent, and decisions are final without appeal. Public fund statements and member-request reporting are specified.
\nSections C-D, PDF3, establish segregated purpose-limited money and committee-approved investment policy, conditional litigation-success repayment, June 30 accounting, and the education/travel exceptions. This MOU does not contain the VW instrument's dissolution-to-General-Fund clause; that authority must not be transplanted. Sections E-H, PDF4, prescribe temporary replacement of an applicant committee member, annual audit inclusion, and a 1.0% administration fee on total fund value at NAAG's fiscal-year close. Attachment 2, PDF7, identifies written application and NAAG finance bills/receipts as records of approval and expenditure. Neither present balances nor applied fees were acquired.
\nThe official GAO report GAO-26-107956, published 18 June 2026, reports that NAAG was not required to file Form 990 in the August 2025 IRS exempt-organization data it examined. It groups NAAG with organizations coded as governmental units for this filing analysis. This does not establish a broader legal conclusion that NAAG is a government department, nor transfer NAIC's detailed IRS history or financial transparency to NAAG. Scoped read: report date; section on filing requirements of similar organizations; footnotes 41-42. Original HTML retained: http[local research file]
\nAn IRS-hosted original 2024 return separately identifies National Association of Attorneys General Mission Foundation Inc, EIN 30-0088843. PDF1 states a calendar 2024 period, 501(c)(3) status and education/research/training purpose benefiting NAAG members. Its return signature date is 14 November 2025, not an independently established public-release date. Schedule A, PDF16, identifies a Type I supporting organization, sole supported organization NAAG, EIN 52-1322260, with monetary support of $0 and other support of $5,192,939. These are the Foundation's reported support categories; other support is not a cash payment and neither figure is an association-wide budget. Schedule R, PDF44, also lists NAAG as a related exempt organization. The form's tax categories should not be repurposed as a complete operational-control map.
\nOriginal: http[local research file]
\nForty-eight original pages retained; visual substantive reads were PDF1,13-16,44,47-48 only. PDF47-48 were blank schedule fields; PDF13-15 described training, Supreme Court advocacy and member services. The original is image-only for the attempted text extractor; the empty derived text is not evidence of a read. Other rendered pages were not substantively inspected. No Cash App item was sought by pretending those unread pages were searched successfully.
\nNo NAAG association audited financial statement, current segregated fund balance, annual fund statement, Cash App remittance advice, deposit ledger or cost invoice was acquired. The two official PDF byte routes each stopped after one 403, using the readable official web rendition as the credible substitute. The IRS original succeeded by ordinary public GET after web-parser failures. No private/account or paid route was used. Failure to locate a Form 990 is not suspicious by itself given GAO's filing analysis. Membership dues, settlement reimbursements and restricted grants must remain separate, and a Foundation return cannot substitute for association accounts.
\nThe useful next records are the current consolidated fund committee charter/amendments; latest fund-specific statements and audit notes; and, for Cash App, remittance plus NAAG finance's underlying invoices, coding and allocation. These would distinguish reimbursement of past expenses, retention in general operating resources, and an authorized restricted-fund deposit. A fund application, committee approval and disbursement record would be needed to name a later investigation actually financed. The instruments establish the mechanism, not that any particular later case was purchased or directed by an earlier defendant.
", "path": "naag-resources-wave-2026-09-09/finance.md", "sourceIds": [ "w24-ag-s-vw-rules", "w24-ag-s-fsf-rules", "w24-ag-s-gao", "w24-ag-s-foundation-return" ], "markdown": "Observed 9 September 2026. Acquisition is closed at the fund-instrument and legal-identity boundary. This packet does not establish receipt of the Cash App reimbursement, a fund allocation of it, or motive for selecting an enforcement target.\n\n## Finding\n\nPublished NAAG fund instruments document a real mechanism through which earlier settlements finance later investigation capacity. They authorize segregated funds, grants from principal, paid experts and other costs, conditional replenishment after successful litigation, and NAAG administration fees. They also specify applications, receipts and reporting records that should exist. Their dates and committee structures matter: the 2012 and 2016 rules cannot simply be assigned to the seven-person consolidated fund committee currently described in the separately acquired governance packet without an amendment or delegation record.\n\nThe July 2026 Cash App instrument separately provides $264,153.73 to NAAG as investigation-cost reimbursement within the $45 million state allocation (Oregon stipulation, paragraph 48, PDF15-16; retained in the prior July packet). That is an obligation in the acquired instrument, not a payment receipt. No record in this run assigns that amount to either fund below. It could reimburse already-incurred coordination costs without becoming a new investigation grant.\n\n## VW Settlement Fund: 2016 rules, five pages\n\nOriginal URL: http[local research file]\n\nThe entire five-page official web-reader rendition was read, including Attachment 1. Ordinary unauthenticated GET returned 403 once; the retained error response is not the PDF. The accessible text and this scoped reading note are derivative evidence. The URL upload directory is 2020; section I dates Executive Committee approval to 2016.\n\nSection A, PDF1, recites a June 2016 partial settlement providing approximately $631 million to participating states and $20 million to NAAG for consumer-protection oversight, training and enforcement, including automobile fraud, and investigation-cost reimbursement. These are the rule's description of settlement provisions, not newly acquired cash receipts.\n\nSections B-C, PDF1-2, vest administration in a three-AG special committee: NAAG's president appoints from the Executive Committee, which approves appointments. At most two may share a political party. Two affirmative votes approve grants, including principal; decisions are final without appeal. Money must be in a new segregated account, not commingled; additional deposits must lawfully fit the purpose. NAAG invests under the special committee's approved policy. If the special committee determines the fund purpose no longer exists, the Executive Committee may dissolve it and direct remaining money to NAAG's General Fund. This is a conditional authority, not an observed transfer.\n\nSection D, PDF3, requires recipients, to the extent applicable and permitted by law, to promise repayment when funded litigation succeeds, limited by available recovery. Conferences, education, training and travel-related events are exempt. Accounting is due by June 30 following disbursement. These are not uniformly repayable loans. Section E requires a temporary replacement when a special-committee member's office applies; implementation of that recusal was not inspected.\n\nSection F, PDF3, authorizes NAAG a fee of 1.0% of total fund value at inception and annually thereafter, covering administration and third-party costs. The current fee amount and any Cash App linkage are unknown. Section H, PDF4, includes the fund within NAAG's annual audit procedures; no resulting audit was acquired.\n\nSection B treats fund statements as public records and calls for balances, applications, deposits and withdrawals to be supplied at the annual summer meeting or on a NAAG member's request. Attachment 1, PDF5, requires written applications signed by an AG or, for NAAG itself, its executive director. Applications identify the purpose, action, amount and payee; payments may be directed to vendors, AG offices, NAAG or a designated multistate office. Bills and receipts go to NAAG finance for disbursement. The text does not establish an unrestricted public inspection procedure for confidential applications.\n\n## Financial Services Fund: MOU effective 31 October 2012, seven pages\n\nOriginal URL: http[local research file]\n\nThe entire official seven-page web-reader rendition, including appendices, was read. One ordinary GET returned 403; no original PDF bytes were retained. This is the Financial Services and Consumer Protection Enforcement, Education, and Training Fund.\n\nPDF1 describes all five 4 April 2012 consent judgments against Bank of America, JPMorgan Chase, Wells Fargo, Citi and Ally/GMAC as containing an Exhibit B provision directing $15 million to NAAG for the fund. Do not silently describe that as the aggregate fund, or reconstruct $75 million received: the underlying five judgments and aggregate receipts were not acquired. The MOU describes the State Monitoring Committee's development of the rules; that body is distinct from the grant special committee.\n\nSections A-B, PDF2, authorize investigation/prosecution expenses, experts, consultants, information exchange and education concerning fraud and illegal or deceptive financial/consumer practices. A five-AG special committee is appointed by the NAAG president and approved by its Executive Committee, with one-year terms and at most three from one party. Grants of $5,000 or more require three votes; smaller grants require two votes from different parties. Principal can be spent, and decisions are final without appeal. Public fund statements and member-request reporting are specified.\n\nSections C-D, PDF3, establish segregated purpose-limited money and committee-approved investment policy, conditional litigation-success repayment, June 30 accounting, and the education/travel exceptions. This MOU does not contain the VW instrument's dissolution-to-General-Fund clause; that authority must not be transplanted. Sections E-H, PDF4, prescribe temporary replacement of an applicant committee member, annual audit inclusion, and a 1.0% administration fee on total fund value at NAAG's fiscal-year close. Attachment 2, PDF7, identifies written application and NAAG finance bills/receipts as records of approval and expenditure. Neither present balances nor applied fees were acquired.\n\n## Identity and accounting boundary\n\nThe official GAO report GAO-26-107956, published 18 June 2026, reports that NAAG was not required to file Form 990 in the August 2025 IRS exempt-organization data it examined. It groups NAAG with organizations coded as governmental units for this filing analysis. This does not establish a broader legal conclusion that NAAG is a government department, nor transfer NAIC's detailed IRS history or financial transparency to NAAG. Scoped read: report date; section on filing requirements of similar organizations; footnotes 41-42. Original HTML retained: http[local research file]\n\nAn IRS-hosted original 2024 return separately identifies National Association of Attorneys General Mission Foundation Inc, EIN 30-0088843. PDF1 states a calendar 2024 period, 501(c)(3) status and education/research/training purpose benefiting NAAG members. Its return signature date is 14 November 2025, not an independently established public-release date. Schedule A, PDF16, identifies a Type I supporting organization, sole supported organization NAAG, EIN 52-1322260, with monetary support of $0 and other support of $5,192,939. These are the Foundation's reported support categories; other support is not a cash payment and neither figure is an association-wide budget. Schedule R, PDF44, also lists NAAG as a related exempt organization. The form's tax categories should not be repurposed as a complete operational-control map.\n\nOriginal: http[local research file]\n\nForty-eight original pages retained; visual substantive reads were PDF1,13-16,44,47-48 only. PDF47-48 were blank schedule fields; PDF13-15 described training, Supreme Court advocacy and member services. The original is image-only for the attempted text extractor; the empty derived text is not evidence of a read. Other rendered pages were not substantively inspected. No Cash App item was sought by pretending those unread pages were searched successfully.\n\n## Negative frontier and next discriminating record\n\nNo NAAG association audited financial statement, current segregated fund balance, annual fund statement, Cash App remittance advice, deposit ledger or cost invoice was acquired. The two official PDF byte routes each stopped after one 403, using the readable official web rendition as the credible substitute. The IRS original succeeded by ordinary public GET after web-parser failures. No private/account or paid route was used. Failure to locate a Form 990 is not suspicious by itself given GAO's filing analysis. Membership dues, settlement reimbursements and restricted grants must remain separate, and a Foundation return cannot substitute for association accounts.\n\nThe useful next records are the current consolidated fund committee charter/amendments; latest fund-specific statements and audit notes; and, for Cash App, remittance plus NAAG finance's underlying invoices, coding and allocation. These would distinguish reimbursement of past expenses, retention in general operating resources, and an authorized restricted-fund deposit. A fund application, committee approval and disbursement record would be needed to name a later investigation actually financed. The instruments establish the mechanism, not that any particular later case was purchased or directed by an earlier defendant.\n", "file": "research/naag-resources-wave-2026-09-09-finance.md" }, { "id": "naag-resources-wave-2026-09-09-governance", "title": "NAAG: who organizes, approves and supplies collective capacity", "description": "Observed9September2026. This root lane is closed at a public governance/procedure breakpoint. It follows the specific NAAG investigation-cost reimbursement in the July Cash App settlement. It does not establish that NAAG received that payment, supplied a particular investigator, or chose the Cash App target. Separate finance.md examines formal fund instrumen", "html": "Observed9September2026. This root lane is closed at a public governance/procedure breakpoint. It follows the specific NAAG investigation-cost reimbursement in the July Cash App settlement. It does not establish that NAAG received that payment, supplied a particular investigator, or chose the Cash App target. Separate finance.md examines formal fund instruments and financial identity. Current web descriptions are organization accounts; the May2026 policy procedure is a published operational rule, not proof of every historical use.
\nThe [leadership page](http[local research file] describes a13-attorney-general Executive Committee governing NAAG's funds, property and management. It comprises four officers, the Finance Committee chair, and eight appointees split between president and president-elect. The committee employs an executive director for day-to-day operations; finance and audit committees have distinct oversight functions. Current named president is William Tong; president-elect Marty Jackley; vice-president Brian Schwalb; immediate past president John Formella. Tim Griffin is Finance chair. This is a dated page observation, not a claim these people held every role when older settlements were negotiated. Dan Schweitzer is listed as interim executive director and Theresia Heller as CFO.
\nThe [Consolidated Consumer Protection and Charity Funds Committee](http[local research file] has a separate remit: funds for consumer-protection/charity enforcement and training. The current roster names co-chairs Phil Weiser and Derek Brown, with Keith Ellison, Letitia James, Gentner Drummond, Dave Sunday and Jonathan Skrmetti. Kate Donoven, Center for Consumer Protection director, is the staff contact. This short page identifies oversight and membership but supplies neither a specific Cash App grant nor a grant-approval threshold; the formal fund documents must carry those claims.
\nNeither body is the July Cash App agreement's eight-state Multistate AG Executive Committee. Nor is either the January money-transmitter order's five-state executive group. An official may sit in more than one, but the grant, corporate-governance, enforcement and fallback-plan approval rights attach to different instruments. NAAG is a specified reimbursement payee in Oregon48; the Oregon money-award table identifies the state as judgment creditor. NAAG is not a signatory to that inspected stipulation.
\nNAAG's member-services page.
\nThe [FAQ](http[local research file] identifies member service fees, registrations and investment proceeds originating in past settlements as operating sources. It says stewarded funds are held apart from operating accounts, constrained by settlement and executive-committee restrictions, and that nearly97% cannot fund NAAG operations. That percentage is an undated organizational claim observed today, with no audited denominator or annual-spending calculation acquired in this root lane. The FAQ says fund committees decide by majority. It also says NAAG does not accept corporate/lobbyist donations, memberships or sponsorships; that is a stated policy, not a transaction-by-transaction audit. Public-conference attendance and registration are separately described.
\nThe same FAQ says NAAG is not a federal agency, the USAG is not a member, and NAAG is not itself a party to lawsuits or settlements. A designated reimbursement recipient can be distinct from the parties, as in the Cash App document. Its McKinsey example describes15million dollars for administration, including7million reimbursing the Financial Services Fund, roughly215,000 reimbursing documented investigation expenses, and the remainder for a public documents repository. This is an organization-reported example of resources being replenished and used; the underlying settlement and receipts were not read in this lane. It is not proof of an equivalent Cash App allocation, nor a reason to open the opioid branch now.
\nThe [May2026 Policy Letter Process](http[local research file] both pages read, specifies a concrete route. AG offices submit a draft and an accompanying colleague letter, backed by at least two Democratic and two Republican AGs. Sponsors set a deadline at least10businessdays ahead and supply lead-contact details. NAAG compiles and circulates the packet; it assists with formatting on request. The published process does not identify who originally authored a given draft or whether outside input preceded the sponsors' request.
\nOnce circulated, any change requires a new circulation and voids the prior signatures. AGs or designated staff use an authenticated sign-on site to confirm or reject for their jurisdiction, recording name, email and title for an audit trail. At36signatory states, the letter receives NAAG letterhead; NAAG coordinates press releases with signatories and transmits it. Below36, it is returned with sponsoring-state seals and the acquired signatures for the states to disseminate, and NAAG's role ends. The threshold governs the association's endorsed distribution, not a state's ability to speak or a court's authority. Retained signature images are used only for initiatives a jurisdiction joins under this process; possession of an image does not establish blanket consent.
\nThis gives the investigation specific expected records: initiating office request, colleague letter, circulated draft/version, named lead contacts, circulation date and deadline, jurisdictional confirmation/rejection log, revised-circulation record if text changed, and final transmission/press coordination. These are much more discriminating than a shared logo. The process does not prove every record survives or is publicly releasable, nor that every historical coalition letter used theMay2026version. No authenticated site, staff account or correspondence was accessed.
\nThe [committee directory](http[local research file] lists AI/internet safety, youth safety and gaming bodies, alongside other topics; their titles are discovery cues only. The [policy index](http[local research file] lists AI preemption, KOSA and KIDS Act letters, potentially relevant to earlier campaign branches. Several listing descriptions visibly repeat a deepfake-payment-platform summary under unrelated titles. The actual letters must be read before deriving demands or recipient connections from those cards. This root run did not open them.
\nThe best-supported mechanism is a member-governed resource and coordination institution, with several decision gates: funds are supervised through defined bodies, expertise and communications are supplied to members, and policy endorsement has a documented sponsorship/sign-off route. Ordinary cooperation and a consequential concentration of capacity can both be true. Whether settlement replenishment changes case selection or policy priorities remains a different question requiring a grant application, decision, expense schedule or worked correspondence chain. No inference of unlawful or hidden control follows merely from funding capacity.
\nEight NAAG primary pages/documents were read through the working public web reader. reader-sources-governance.json records exact scopes and references. OrdinaryGET attempts returned403 for those eight plus the CCP-work link whose web reader gave a cache miss; the saved HTMLs are error responses only, not originals. Those original-download routes are closed. The policy PDF's complete web-reader output is retained in naag-policy-procedure-web-reader.json, a derived reader representation rather than originalPDFbytes. Its first save hit a PowerShell smart-apostrophe parser issue; one byte-encoded save succeeded, with no change to source content or permission boundary. Full bylaws were not located in the limited search; unrelated organizations' results were discarded.
\nRoot did not enter an account, submit a form, contact staff, pay for records, or edit an earlier case/ZIP. Acquisition has stopped while the separate finance lane completes, before a combined forest review selects another research run.
", "path": "naag-resources-wave-2026-09-09/governance.md", "sourceIds": [ "w24-ag-s-faq", "w24-ag-s-consolidated", "w24-ag-s-leadership", "w24-ag-s-procedure", "w24-ag-s-services", "w24-ag-s-ccp" ], "markdown": "Observed9September2026. This root lane is closed at a public governance/procedure breakpoint. It follows the specific NAAG investigation-cost reimbursement in the July Cash App settlement. It does not establish that NAAG received that payment, supplied a particular investigator, or chose the Cash App target. Separate finance.md examines formal fund instruments and financial identity. Current web descriptions are organization accounts; the May2026 policy procedure is a published operational rule, not proof of every historical use.\n\n## Different bodies, different rights\n\nThe leadership page (see source references) describes a13-attorney-general Executive Committee governing NAAG's funds, property and management. It comprises four officers, the Finance Committee chair, and eight appointees split between president and president-elect. The committee employs an executive director for day-to-day operations; finance and audit committees have distinct oversight functions. Current named president is William Tong; president-elect Marty Jackley; vice-president Brian Schwalb; immediate past president John Formella. Tim Griffin is Finance chair. This is a dated page observation, not a claim these people held every role when older settlements were negotiated. Dan Schweitzer is listed as interim executive director and Theresia Heller as CFO.\n\nThe Consolidated Consumer Protection and Charity Funds Committee (see source references) has a separate remit: funds for consumer-protection/charity enforcement and training. The current roster names co-chairs Phil Weiser and Derek Brown, with Keith Ellison, Letitia James, Gentner Drummond, Dave Sunday and Jonathan Skrmetti. Kate Donoven, Center for Consumer Protection director, is the staff contact. This short page identifies oversight and membership but supplies neither a specific Cash App grant nor a grant-approval threshold; the formal fund documents must carry those claims.\n\nNeither body is the July Cash App agreement's eight-state Multistate AG Executive Committee. Nor is either the January money-transmitter order's five-state executive group. An official may sit in more than one, but the grant, corporate-governance, enforcement and fallback-plan approval rights attach to different instruments. NAAG is a specified reimbursement payee in Oregon48; the Oregon money-award table identifies the state as judgment creditor. NAAG is not a signatory to that inspected stipulation.\n\n## Resource provision is explicit, allocation still needs its own record\n\nNAAG's member-services page.\n\nThe FAQ (see source references) identifies member service fees, registrations and investment proceeds originating in past settlements as operating sources. It says stewarded funds are held apart from operating accounts, constrained by settlement and executive-committee restrictions, and that nearly97% cannot fund NAAG operations. That percentage is an undated organizational claim observed today, with no audited denominator or annual-spending calculation acquired in this root lane. The FAQ says fund committees decide by majority. It also says NAAG does not accept corporate/lobbyist donations, memberships or sponsorships; that is a stated policy, not a transaction-by-transaction audit. Public-conference attendance and registration are separately described.\n\nThe same FAQ says NAAG is not a federal agency, the USAG is not a member, and NAAG is not itself a party to lawsuits or settlements. A designated reimbursement recipient can be distinct from the parties, as in the Cash App document. Its McKinsey example describes15million dollars for administration, including7million reimbursing the Financial Services Fund, roughly215,000 reimbursing documented investigation expenses, and the remainder for a public documents repository. This is an organization-reported example of resources being replenished and used; the underlying settlement and receipts were not read in this lane. It is not proof of an equivalent Cash App allocation, nor a reason to open the opioid branch now.\n\n## The policy-letter machinery leaves identifiable records\n\nThe May2026 Policy Letter Process (see source references) both pages read, specifies a concrete route. AG offices submit a draft and an accompanying colleague letter, backed by at least two Democratic and two Republican AGs. Sponsors set a deadline at least10businessdays ahead and supply lead-contact details. NAAG compiles and circulates the packet; it assists with formatting on request. The published process does not identify who originally authored a given draft or whether outside input preceded the sponsors' request.\n\nOnce circulated, any change requires a new circulation and voids the prior signatures. AGs or designated staff use an authenticated sign-on site to confirm or reject for their jurisdiction, recording name, email and title for an audit trail. At36signatory states, the letter receives NAAG letterhead; NAAG coordinates press releases with signatories and transmits it. Below36, it is returned with sponsoring-state seals and the acquired signatures for the states to disseminate, and NAAG's role ends. The threshold governs the association's endorsed distribution, not a state's ability to speak or a court's authority. Retained signature images are used only for initiatives a jurisdiction joins under this process; possession of an image does not establish blanket consent.\n\nThis gives the investigation specific expected records: initiating office request, colleague letter, circulated draft/version, named lead contacts, circulation date and deadline, jurisdictional confirmation/rejection log, revised-circulation record if text changed, and final transmission/press coordination. These are much more discriminating than a shared logo. The process does not prove every record survives or is publicly releasable, nor that every historical coalition letter used theMay2026version. No authenticated site, staff account or correspondence was accessed.\n\n## Discovery boundaries and next discriminator\n\nThe committee directory (see source references) lists AI/internet safety, youth safety and gaming bodies, alongside other topics; their titles are discovery cues only. The policy index (see source references) lists AI preemption, KOSA and KIDS Act letters, potentially relevant to earlier campaign branches. Several listing descriptions visibly repeat a deepfake-payment-platform summary under unrelated titles. The actual letters must be read before deriving demands or recipient connections from those cards. This root run did not open them.\n\nThe best-supported mechanism is a member-governed resource and coordination institution, with several decision gates: funds are supervised through defined bodies, expertise and communications are supplied to members, and policy endorsement has a documented sponsorship/sign-off route. Ordinary cooperation and a consequential concentration of capacity can both be true. Whether settlement replenishment changes case selection or policy priorities remains a different question requiring a grant application, decision, expense schedule or worked correspondence chain. No inference of unlawful or hidden control follows merely from funding capacity.\n\n## Source custody and limitations\n\nEight NAAG primary pages/documents were read through the working public web reader. reader-sources-governance.json records exact scopes and references. OrdinaryGET attempts returned403 for those eight plus the CCP-work link whose web reader gave a cache miss; the saved HTMLs are error responses only, not originals. Those original-download routes are closed. The policy PDF's complete web-reader output is retained in naag-policy-procedure-web-reader.json, a derived reader representation rather than originalPDFbytes. Its first save hit a PowerShell smart-apostrophe parser issue; one byte-encoded save succeeded, with no change to source content or permission boundary. Full bylaws were not located in the limited search; unrelated organizations' results were discarded.\n\nRoot did not enter an account, submit a form, contact staff, pay for records, or edit an earlier case/ZIP. Acquisition has stopped while the separate finance lane completes, before a combined forest review selects another research run.\n", "file": "research/naag-resources-wave-2026-09-09-governance.md" }, { "id": "ncmec-advertising-input-wave-2026-09-09-category-authority", "title": "DV category authority: a later adoption claim and an upstream correction boundary", "description": "9 September 2026. This bounded pass found later, dated company reporting that materially advances the February launch account, and a directly relevant upstream list-maintenance description. It did not acquire an applicable versioned DV instrument specifying override rights, correction deadlines or downstream recipients' duties. Acquisition is closed at that ", "html": "9 September 2026. This bounded pass found later, dated company reporting that materially advances the February launch account, and a directly relevant upstream list-maintenance description. It did not acquire an applicable versioned DV instrument specifying override rights, correction deadlines or downstream recipients' duties. Acquisition is closed at that distinction.
\nDV's [7 February 2025 response](http[local research file] before the category launch, describes traffic-triggered page classification, optional blocking of unclassified pages, and campaign settings governing placement. It also reports taking additional blocking measures during its review and announces a P2P category under development. This helps distinguish measurement, classification and action, but it is an interested party's response to criticism. No underlying disputed test data were inspected, and no image-hosting service was visited. Its general client-settings claim predates the named defaults and does not prove that every customer may disable them. Original-derived26–31; full substantive1–33 read. The displayed date is February7,2025.
\nThe [24 April 2025 progress release](http[local research file] says both categories are enabled by default. DV reports nearly30 platforms had implemented Highly Illicit, which it offers free to all platforms, including nonclients. This is stronger than the February statement of intended defaults or distribution to more than100 partners: it is a dated adoption report, not independently tested uptake. The release also announces an IWF partnership and says DV will begin incorporating IWF's URL List into Highly Illicit. Planned reciprocal flagging is not a disclosed transaction or completed referral. No executed data agreement, platform identities/configurations or onboarding evidence was acquired. Original-derived75–80; full substantive61–86 read, displayed date62/67; primary reader91–95.
\nIWF's [8 May 2025 counterpart announcement](http[local research file] confirms the membership/partnership relationship, but still says DV will integrate the URL list. It supports the relationship from the counterparty; repeated executive quotations are not independent validation of implementation, and the post is not an executed licence. Original-derived56–81, particularly63–67; primary reader53–68.
\nThe [IWF 2025 annual-report section on dynamic lists](http[local research file] describes an assessment-based URL list, including locations linking to or advertising identified imagery. IWF says it rechecks daily, removes entries the same day when the content is no longer present, and updates the list twice daily. A separately attributed DV member testimonial describes integration of IWF signals. These are institutional and participant accounts, not observed delivery logs. Most importantly, removing an obsolete input is not proof that DV changes a broader domain category or that an advertiser restores placements. No exact publication date was inspected;2025 is the report period, not an assigned deployment date. Original-derived52–71 and75–82; primary reader63–91, decisive82 and88–91. Hash/NPI/keyword services are outside this lane.
\nThe accepted February13 original remains the source for DV's description of its NCMEC input, its own site assessment and the large-platform exclusions. Its public NCMEC material is distinct from the later IWF member-access feed. The records describe different origins and access arrangements; neither an equivalent unit of assessment nor a shared schema is assumed. NCMEC data meaning is assigned to the other research lane.
\nThe accepted current product page and June2024 review statement remain the sources for broad customer choice and an internal publisher-review route. They do not disclose rights applicable to every named2025 category, an override priority hierarchy, review deadlines, all recipients who must receive a correction, whether changed classifications invalidate cached signals, or any obligation to revisit an advertiser's prior exclusion. The existing February25 URL-reporting statement concerns qualified advertiser access; it does not supply these rules.
\nConsequently, four actions must remain distinct in any integration: IWF updating a URL input; DV revising a category; a platform receiving and applying that revision; and a customer changing placement. The new records support the first as a published maintenance account and provide further reported integration, but do not establish the remaining steps in a worked correction. Conversely, the absence of that instrument here does not establish that DV or recipients ignore updates. An automatic, timely correction path and a separate campaign-level decision are both still plausible.
\nTen targeted official-site searches are preserved in category-search-scope.json. They sought the named categories with policy, opt-out, override, dispute, review and PDF terms, plus one directly relevant IWF counterpart search. Results returned the existing launch/review material, the two new DV statements and the two new IWF descriptions. Search-only historical page-exception and generic filing snippets were not promoted into current category rights. No version history, category-definition manual, applicable customer terms, correction API/change-log or service-level review instrument was acquired. This is a bounded public-search result, not an exhaustive assertion that such material does not exist.
Reopen this lane for an actual, dated DV category/default specification or applicable agreement that identifies who may disable or override these categories and how a corrected input/classification reaches recipients; alternatively, a publicly documented correction with recipient acknowledgment and action. A customer configuration or source-data deletion alone would answer only part of that question. No sales contact, account, form, publisher submission, live category lookup, NCMEC dataset, IWF list entry or harmful-content location was accessed.
\nFour ordinary GETs succeeded. Original hashes, dates, exact scopes and derivative hashes are in category-captures.json. Every saved original was read within the stated scope; neither capture nor search indexing establishes deployment. The baseline four originals were reused from the accepted FilmOn product packet and were not downloaded again.
| Retained original | \nSHA256 | \n
|---|---|
category-adalytics-response2025.html | \nd243a7439049bcfa7f8ce3611bdb12d4276187bd0c72d77bd5b1acd8943c9ba9 | \n
category-iwf-progress2025.html | \nc9ea0b9f1d8c6a2253382cb3cf037d95a94634e5326561b85690c02cc9765c83 | \n
category-iwf-partnership2025.html | \n64763eefe3e058b885e105dd3d9a9c538c4450a7f6c26bb9a735bdfdd3f178d2 | \n
category-iwf-dynamic2025.html | \n6a98771131af9c350f6f7a2cab2c0d868d9331a8193ac89758e125b6c6fad869 | \n
Useful directional additions would be DV → category-default/adoption report, IWF ↔ DV membership/planned feed integration, and IWF → URL-list maintenance account. Each requires its reported/planned date and source attribution. Do not create an NCMEC endorsement, FBI instruction, named platform-implementation, mandatory category override or restored-income edge from these records. No case or presentation was edited.
", "path": "ncmec-advertising-input-wave-2026-09-09/category-authority.md", "sourceIds": [ "w30-class-s-dv-progress", "w30-class-s-dv-response", "w30-class-s-iwf-partnership", "w30-class-s-iwf-dynamic" ], "markdown": "9 September 2026. This bounded pass found later, dated company reporting that materially advances the February launch account, and a directly relevant upstream list-maintenance description. It did not acquire an applicable versioned DV instrument specifying override rights, correction deadlines or downstream recipients' duties. Acquisition is closed at that distinction.\n\n## Dated records and their limits\n\nDV's 7 February 2025 response (see source references) before the category launch, describes traffic-triggered page classification, optional blocking of unclassified pages, and campaign settings governing placement. It also reports taking additional blocking measures during its review and announces a P2P category under development. This helps distinguish measurement, classification and action, but it is an interested party's response to criticism. No underlying disputed test data were inspected, and no image-hosting service was visited. Its general client-settings claim predates the named defaults and does not prove that every customer may disable them. Original-derived26–31; full substantive1–33 read. The displayed date is February7,2025.\n\nThe 24 April 2025 progress release (see source references) says both categories **are enabled by default**. DV reports nearly30 platforms had implemented Highly Illicit, which it offers free to all platforms, including nonclients. This is stronger than the February statement of intended defaults or distribution to more than100 partners: it is a dated adoption report, not independently tested uptake. The release also announces an IWF partnership and says DV **will begin incorporating** IWF's URL List into Highly Illicit. Planned reciprocal flagging is not a disclosed transaction or completed referral. No executed data agreement, platform identities/configurations or onboarding evidence was acquired. Original-derived75–80; full substantive61–86 read, displayed date62/67; primary reader91–95.\n\nIWF's 8 May 2025 counterpart announcement (see source references) confirms the membership/partnership relationship, but still says DV **will integrate** the URL list. It supports the relationship from the counterparty; repeated executive quotations are not independent validation of implementation, and the post is not an executed licence. Original-derived56–81, particularly63–67; primary reader53–68.\n\nThe IWF 2025 annual-report section on dynamic lists (see source references) describes an assessment-based URL list, including locations linking to or advertising identified imagery. IWF says it rechecks daily, removes entries the same day when the content is no longer present, and updates the list twice daily. A separately attributed DV member testimonial describes integration of IWF signals. These are institutional and participant accounts, not observed delivery logs. Most importantly, removing an obsolete input is not proof that DV changes a broader domain category or that an advertiser restores placements. No exact publication date was inspected;2025 is the report period, not an assigned deployment date. Original-derived52–71 and75–82; primary reader63–91, decisive82 and88–91. Hash/NPI/keyword services are outside this lane.\n\n## Authority remains split\n\nThe accepted February13 original remains the source for DV's description of its NCMEC input, its own site assessment and the large-platform exclusions. Its public NCMEC material is distinct from the later IWF member-access feed. The records describe different origins and access arrangements; neither an equivalent unit of assessment nor a shared schema is assumed. NCMEC data meaning is assigned to the other research lane.\n\nThe accepted current product page and June2024 review statement remain the sources for broad customer choice and an internal publisher-review route. They do not disclose rights applicable to every named2025 category, an override priority hierarchy, review deadlines, all recipients who must receive a correction, whether changed classifications invalidate cached signals, or any obligation to revisit an advertiser's prior exclusion. The existing February25 URL-reporting statement concerns qualified advertiser access; it does not supply these rules.\n\nConsequently, four actions must remain distinct in any integration: IWF updating a URL input; DV revising a category; a platform receiving and applying that revision; and a customer changing placement. The new records support the first as a published maintenance account and provide further reported integration, but do not establish the remaining steps in a worked correction. Conversely, the absence of that instrument here does not establish that DV or recipients ignore updates. An automatic, timely correction path and a separate campaign-level decision are both still plausible.\n\n## Finite search and exact reopening record\n\nTen targeted official-site searches are preserved in `category-search-scope.json`. They sought the named categories with policy, opt-out, override, dispute, review and PDF terms, plus one directly relevant IWF counterpart search. Results returned the existing launch/review material, the two new DV statements and the two new IWF descriptions. Search-only historical page-exception and generic filing snippets were not promoted into current category rights. No version history, category-definition manual, applicable customer terms, correction API/change-log or service-level review instrument was acquired. This is a bounded public-search result, not an exhaustive assertion that such material does not exist.\n\nReopen this lane for an actual, dated DV category/default specification or applicable agreement that identifies who may disable or override these categories and how a corrected input/classification reaches recipients; alternatively, a publicly documented correction with recipient acknowledgment and action. A customer configuration or source-data deletion alone would answer only part of that question. No sales contact, account, form, publisher submission, live category lookup, NCMEC dataset, IWF list entry or harmful-content location was accessed.\n\n## Source custody and integration candidates\n\nFour ordinary GETs succeeded. Original hashes, dates, exact scopes and derivative hashes are in `category-captures.json`. Every saved original was read within the stated scope; neither capture nor search indexing establishes deployment. The baseline four originals were reused from the accepted FilmOn product packet and were not downloaded again.\n\n| Retained original | SHA256 |\n| --- | --- |\n| `category-adalytics-response2025.html` | `d243a7439049bcfa7f8ce3611bdb12d4276187bd0c72d77bd5b1acd8943c9ba9` |\n| `category-iwf-progress2025.html` | `c9ea0b9f1d8c6a2253382cb3cf037d95a94634e5326561b85690c02cc9765c83` |\n| `category-iwf-partnership2025.html` | `64763eefe3e058b885e105dd3d9a9c538c4450a7f6c26bb9a735bdfdd3f178d2` |\n| `category-iwf-dynamic2025.html` | `6a98771131af9c350f6f7a2cab2c0d868d9331a8193ac89758e125b6c6fad869` |\n\nUseful directional additions would be DV → category-default/adoption report, IWF ↔ DV membership/planned feed integration, and IWF → URL-list maintenance account. Each requires its reported/planned date and source attribution. Do not create an NCMEC endorsement, FBI instruction, named platform-implementation, mandatory category override or restored-income edge from these records. No case or presentation was edited.\n", "file": "research/ncmec-advertising-input-wave-2026-09-09-category-authority.md" }, { "id": "ncmec-advertising-input-wave-2026-09-09-input-meaning", "title": "NCMEC public tables: input meaning before an advertising exclusion", "description": "Acquisition closed on9 September2026. The historical source family is identified; DoubleVerify's exact three years, files, filters and provider-to-domain mapping remain unestablished.", "html": "Acquisition closed on9 September2026. The historical source family is identified; DoubleVerify's exact three years, files, filters and provider-to-domain mapping remain unestablished.
\nThe retained13 February2025 DoubleVerify announcement says it reviewed three years of public NCMEC data as a starting point, analyzed over300 domains/electronic service providers that had received CSAM-related notices, cross-referenced existing classifications, and conducted site-by-site analysis for the remainder while excluding large, well-moderated social/search services. It does not identify those years or the files and columns used.
\nThose qualifications matter. This investigation has not established that DV equated every NCMEC row with illegal content, nor that it confined its inputs to the CSAM column. Its own claimed additional analysis is part of the transformation to be tested. NCMEC's publication of data does not establish endorsement of DV's category or a direct data-supply agreement.
\nThe [2023 Reports by ESP table](http[local research file] PDF1, counts reports providers submitted to the CyberTipline about suspected exploitation/apparent CSAM on their systems. It says volume and content vary because the reporting framework did not then require proactive detection or prescribe the information to include. An incoming report count is not a count of outbound NCMEC notices, unique victims, confirmed incidents or criminal domains.
\nThe currently hosted2023 annual report, PDF7, adds that higher reporting may reflect robust detection/removal, and that inadequate reports can add volume without helping an actionable investigation. Its current file was modified after DV's announcement, so this passage is supportive methodological explanation, not a proven byte-for-byte historical input.
\nThe [2021 outbound table](http[local research file] and [2022 outbound table](http[local research file] each PDF1, describe NCMEC visually reviewing imagery reported by victims, caregivers and, expressly in2022, INHOPE hotlines, then notifying relevant providers. These are more specific than unreviewed incoming report totals. They still record notification activity and response, not final court findings about whole domains.
\nBoth the [responsive](http[local research file] and [unresponsive](http[local research file] tables explain their three categories on PDF1: CSAM; exploitative content depicting identified child victims that does not meet the legal CSAM definition; and predatory text involving sexual comments or personal information about identified victims. Those categories must remain separate. A total across them cannot simply be labelled confirmed criminal CSAM.
\nNCMEC says staff review the imagery, notify the provider, manually track status and send additional notices until content is addressed. The responsive introduction says company terms may lead to imagery removal or user blocking. Its table reports notifications and average days to respond/takedown, by category and total. The unresponsive table reports notification counts without response-time columns. Neither introductory methodology defines the counting unit as a unique incident, file, victim or site; repeat notifications expressly prevent treating it as such.
\nThe combined response/takedown heading should not be silently rewritten as a pure time-to-removal metric. The2021 footnote, PDF10, says some removal times are unavailable after unsuccessful notification efforts and NCMEC seeks other options such as contacting upstream providers. The2022 footnote, PDF12, makes clear that missing times can concern some notices, even where a provider has a numerical average for others. The tables do not provide a universal deadline, a causal explanation for every nonresponse, or the underlying notifications.
\nThe same named provider can appear in both2023 tables. CloudFlare, Inc appears on responsivePDF4 with5,183 notifications across the three categories and an average of1.53 days, and on unresponsivePDF1 with7 CSAM notices. These are separate published aggregates, not a current rate or a count this inquiry independently audited. This example disproves a reading of the headings as permanently exclusive whole-company labels. Alibaba likewise appears in both (responsivePDF2, unresponsivePDF1). The notices' dates, subjects and reasons for differing outcomes were not acquired; no individual notice is inferred.
\nRows mix provider names, domains and grouped services. The explicit upstream-contact route cautions against converting a hosting/infrastructure recipient directly into the offending destination site. The relevant entity-to-domain and content-location mapping has to be established separately. No listed site was visited.
\nThe current NCMEC archive page retained in the earlier PhotoDNA wave links2021 and2022 notification tables and the2023 split tables. Those links establish the archive's present organization, not its exact state on13 February2025.
\n| Artifact acquired now | \nDate evidence in the original | \nWhat can be concluded | \n
|---|---|---|
| 2021 notification table | \nCopyright2022; PDF created/modified22 March2022 | \nA historical2021 table with pre-announcement file metadata; first public upload not independently established. | \n
| 2022 notification table | \nCopyright2023; PDF created/modified1 May2023 | \nSame limited chronology conclusion; not proof DV selected it. | \n
| 2023 responsive/unresponsive tables and Reports by ESP | \nCopyright2024; PDF created/modified12 April2024 | \nPlausible public-source candidates before DV's announcement; exact historical public versions/selection not established. | \n
| 2023 annual report currently hosted | \nCreated15 April2024; modified3 June2025 | \nCurrent bytes cannot be assumed identical to the version available in February2025. No specific content change identified. | \n
| NCMEC2024-data release post | \nDated8 May2025 | \nThis acquired release is later than DV's announcement and cannot establish what DV had on13 February. | \n
PDF metadata is document metadata, not an authenticated publication log. This packet does not declare2021-2023 to be DV's three years, or assume all2024 annual data were unavailable in February2025.
\nThe official 5 February2024 post.
\nThe separate hash-sharing programme is not the same resource as these public provider/notification tables. The previously acquired15 April2024 NCMEC audit announcement concerns reviewed image/video hashes. Its assurance about that resource cannot certify a separate advertising domain list or identify DV's input files.
\nOne reading consistent with DV's account is that it used historical outbound-notice tables, filtered relevant CSAM notices, resolved provider/domain identities and applied additional classification judgment. Another possibility is broader use of totals or provider membership across heterogeneous categories. The acquired originals establish why those choices matter, but do not establish which occurred. Nor does the existence of responsive notices erase the potential relevance of underlying harmful material; response performance and content risk are different questions.
\nThe high-value missing record is DV's dated source inventory and transformation specification: exact NCMEC URLs/versions/years; included columns and responsive/unresponsive treatment; repeat-notice handling; provider/domain resolution; exclusions; and the provenance for final category decisions. The next archival record, if needed, is a dated public snapshot of the selected NCMEC files. Neither requires access to illegal imagery or private CyberTipline reports.
\nEight new originals were acquired by ordinary unauthenticated GET, all200: six PDFs and two HTML posts. input-get.json and input-date-get.json retain transport/metadata; input-captures.json records SHA256 and reading scope. No acquisition route failed or was retried. A few focused discovery searches returned current NCMEC pages, unrelated material and public dates; no unrelated branch was opened.
\nActual table scope:2021 PDF1 and final-page10 footnote/totals;2022 PDF1 and final-page12 footnote/totals;2023 responsivePDF1, selected AlibabaPDF2 and CloudFlarePDF4 rows, and final-page22 totals;2023 unresponsivePDF1 and final-page4 totals, with intervening text visible in extraction but no claimed complete row audit;2023 Reports by ESP PDF1 introduction/header only. The key2023 headers/CloudFlare rows and2021/22 footnotes were visually checked in retained PNGs. No row totals were recalculated across years or treated as a prevalence denominator.
\nThe annual report was keyword-scanned; selected extractedPDF7,12-13,15-16 were read, with onlyPDF7 and12-13 material used for this question and the modification-date caveat retained. Both dated posts' complete substantive text was read; embedded videos/forms were not used. The prior DV announcement's whole substantive body and the prior NCMEC archive links were read locally. Acquisition now stops for combined review; no case, reader, ZIP or operational data was changed.
", "path": "ncmec-advertising-input-wave-2026-09-09/input-meaning.md", "sourceIds": [ "w30-class-s-ncmec2021", "w30-class-s-ncmec2022", "w30-class-s-ncmec-responsive", "w30-class-s-ncmec-unresponsive", "w30-class-s-ncmec-incoming" ], "markdown": "Acquisition closed on9 September2026. The historical source family is identified; DoubleVerify's exact three years, files, filters and provider-to-domain mapping remain unestablished.\n\n## What DoubleVerify actually said\n\nThe retained13 February2025 [DoubleVerify announcement](https://doubleverify.com/blog/omnichannel/verify/taking-action-strengthening-protections-against-high-risk-ad-supported-content-and-attacking-the-csam-issue-together) says it reviewed three years of public NCMEC data as a starting point, analyzed over300 domains/electronic service providers that had received CSAM-related notices, cross-referenced existing classifications, and conducted site-by-site analysis for the remainder while excluding large, well-moderated social/search services. It does not identify those years or the files and columns used.\n\nThose qualifications matter. This investigation has not established that DV equated every NCMEC row with illegal content, nor that it confined its inputs to the CSAM column. Its own claimed additional analysis is part of the transformation to be tested. NCMEC's publication of data does not establish endorsement of DV's category or a direct data-supply agreement.\n\n## Two opposite directions of information\n\nThe 2023 Reports by ESP table (see source references) PDF1, counts reports providers submitted to the CyberTipline about suspected exploitation/apparent CSAM on their systems. It says volume and content vary because the reporting framework did not then require proactive detection or prescribe the information to include. An incoming report count is not a count of outbound NCMEC notices, unique victims, confirmed incidents or criminal domains.\n\nThe currently hosted2023 annual report, PDF7, adds that higher reporting may reflect robust detection/removal, and that inadequate reports can add volume without helping an actionable investigation. Its current file was modified after DV's announcement, so this passage is supportive methodological explanation, not a proven byte-for-byte historical input.\n\nThe 2021 outbound table (see source references) and 2022 outbound table (see source references) each PDF1, describe NCMEC visually reviewing imagery reported by victims, caregivers and, expressly in2022, INHOPE hotlines, then notifying relevant providers. These are more specific than unreviewed incoming report totals. They still record notification activity and response, not final court findings about whole domains.\n\n## The2023 columns are consequential\n\nBoth the responsive (see source references) and unresponsive (see source references) tables explain their three categories on PDF1: CSAM; exploitative content depicting identified child victims that does not meet the legal CSAM definition; and predatory text involving sexual comments or personal information about identified victims. Those categories must remain separate. A total across them cannot simply be labelled confirmed criminal CSAM.\n\nNCMEC says staff review the imagery, notify the provider, manually track status and send additional notices until content is addressed. The responsive introduction says company terms may lead to imagery removal or user blocking. Its table reports notifications and average days to respond/takedown, by category and total. The unresponsive table reports notification counts without response-time columns. Neither introductory methodology defines the counting unit as a unique incident, file, victim or site; repeat notifications expressly prevent treating it as such.\n\nThe combined response/takedown heading should not be silently rewritten as a pure time-to-removal metric. The2021 footnote, PDF10, says some removal times are unavailable after unsuccessful notification efforts and NCMEC seeks other options such as contacting upstream providers. The2022 footnote, PDF12, makes clear that missing times can concern some notices, even where a provider has a numerical average for others. The tables do not provide a universal deadline, a causal explanation for every nonresponse, or the underlying notifications.\n\nThe same named provider can appear in both2023 tables. CloudFlare, Inc appears on responsivePDF4 with5,183 notifications across the three categories and an average of1.53 days, and on unresponsivePDF1 with7 CSAM notices. These are separate published aggregates, not a current rate or a count this inquiry independently audited. This example disproves a reading of the headings as permanently exclusive whole-company labels. Alibaba likewise appears in both (responsivePDF2, unresponsivePDF1). The notices' dates, subjects and reasons for differing outcomes were not acquired; no individual notice is inferred.\n\nRows mix provider names, domains and grouped services. The explicit upstream-contact route cautions against converting a hosting/infrastructure recipient directly into the offending destination site. The relevant entity-to-domain and content-location mapping has to be established separately. No listed site was visited.\n\n## Chronology: reporting year is not publication date\n\nThe current NCMEC archive page retained in the earlier PhotoDNA wave links2021 and2022 notification tables and the2023 split tables. Those links establish the archive's present organization, not its exact state on13 February2025.\n\n| Artifact acquired now | Date evidence in the original | What can be concluded |\n| --- | --- | --- |\n| 2021 notification table | Copyright2022; PDF created/modified22 March2022 | A historical2021 table with pre-announcement file metadata; first public upload not independently established. |\n| 2022 notification table | Copyright2023; PDF created/modified1 May2023 | Same limited chronology conclusion; not proof DV selected it. |\n| 2023 responsive/unresponsive tables and Reports by ESP | Copyright2024; PDF created/modified12 April2024 | Plausible public-source candidates before DV's announcement; exact historical public versions/selection not established. |\n| 2023 annual report currently hosted | Created15 April2024; modified3 June2025 | Current bytes cannot be assumed identical to the version available in February2025. No specific content change identified. |\n| NCMEC2024-data release post | Dated8 May2025 | This acquired release is later than DV's announcement and cannot establish what DV had on13 February. |\n\nPDF metadata is document metadata, not an authenticated publication log. This packet does not declare2021-2023 to be DV's three years, or assume all2024 annual data were unavailable in February2025.\n\nThe official 5 February2024 post.\n\nThe separate hash-sharing programme is not the same resource as these public provider/notification tables. The previously acquired15 April2024 NCMEC audit announcement concerns reviewed image/video hashes. Its assurance about that resource cannot certify a separate advertising domain list or identify DV's input files.\n\n## Competing interpretations and the record that would decide\n\nOne reading consistent with DV's account is that it used historical outbound-notice tables, filtered relevant CSAM notices, resolved provider/domain identities and applied additional classification judgment. Another possibility is broader use of totals or provider membership across heterogeneous categories. The acquired originals establish why those choices matter, but do not establish which occurred. Nor does the existence of responsive notices erase the potential relevance of underlying harmful material; response performance and content risk are different questions.\n\nThe high-value missing record is DV's dated source inventory and transformation specification: exact NCMEC URLs/versions/years; included columns and responsive/unresponsive treatment; repeat-notice handling; provider/domain resolution; exclusions; and the provenance for final category decisions. The next archival record, if needed, is a dated public snapshot of the selected NCMEC files. Neither requires access to illegal imagery or private CyberTipline reports.\n\n## Custody and actual scope\n\nEight new originals were acquired by ordinary unauthenticated GET, all200: six PDFs and two HTML posts. input-get.json and input-date-get.json retain transport/metadata; input-captures.json records SHA256 and reading scope. No acquisition route failed or was retried. A few focused discovery searches returned current NCMEC pages, unrelated material and public dates; no unrelated branch was opened.\n\nActual table scope:2021 PDF1 and final-page10 footnote/totals;2022 PDF1 and final-page12 footnote/totals;2023 responsivePDF1, selected AlibabaPDF2 and CloudFlarePDF4 rows, and final-page22 totals;2023 unresponsivePDF1 and final-page4 totals, with intervening text visible in extraction but no claimed complete row audit;2023 Reports by ESP PDF1 introduction/header only. The key2023 headers/CloudFlare rows and2021/22 footnotes were visually checked in retained PNGs. No row totals were recalculated across years or treated as a prevalence denominator.\n\nThe annual report was keyword-scanned; selected extractedPDF7,12-13,15-16 were read, with onlyPDF7 and12-13 material used for this question and the modification-date caveat retained. Both dated posts' complete substantive text was read; embedded videos/forms were not used. The prior DV announcement's whole substantive body and the prior NCMEC archive links were read locally. Acquisition now stops for combined review; no case, reader, ZIP or operational data was changed.\n", "file": "research/ncmec-advertising-input-wave-2026-09-09-input-meaning.md" }, { "id": "nominet-domain-authority-wave-2026-09-09-exercised-action", "title": "Nominet exercised domain action: reported requests and reversals", "description": "9 September 2026. Bounded acquisition CLOSED. Official Nominet reader/index records only; no original GET, domain operation, account, form, content/media access, outreach or Explorer action. Current standing/remedy rules belong to the parallel authority lane.", "html": "9 September 2026. Bounded acquisition CLOSED. Official Nominet reader/index records only; no original GET, domain operation, account, form, content/media access, outreach or Explorer action. Current standing/remedy rules belong to the parallel authority lane.
\nThe inspected reports document IWF suspension requests, but do not identify their domains or publish an individual decision trail. They also report reversals and requests that did not lead to suspension. These demonstrate reported operational activity and limits; they do not establish wrongful decisions, independent verification, or rights created by Nominet's funding of UKSIC.
\n| Reporting period | \nIWF-labelled suspension requests | \nCriminal Practices Policy domains suspended | \nRequests not resulting in suspension | \nSuspension reversals | \n
|---|---|---|---|---|
| 1 November 2019–31 October 2020 | \n0 | \n22,158 | \n47 | \n15 | \n
| 1 November 2022–31 October 2023 | \n1 | \n1,193 | \n3 | \n19 | \n
| 1 November 2023–31 October 2024 | \n1 | \n913 | \n0 | \n16 | \n
Sources are the official [2020 report](http[local research file] [2023 report](http[local research file] and [2025 Impact Issue](http[local research file] These are selected periods, not a continuous time series or September2026 snapshot. Publication/upload year is not activity year.
\nThe IWF metric is expressly a request. Neither a case-specific successful suspension nor an IWF reversal is established by the aggregate columns. Do not derive either from a zero total of unsuccessful requests. The source does not reconcile the IWF-labelled metric with the agency totals; it must not be added to the reported suspension total as an extra domain. No UKSIC consortium-labelled request or named IWF-origin domain case was located in this pass.
\nBoth one-page reports and the2025 spread depict agencies notifying Nominet, Nominet making administrative checks, and Nominet notifying the registrant and working with the registrar to suspend the domain. That is their summary of the law-enforcement process; it does not prove IWF is a law-enforcement agency or that its requests use every step of that route. Current eligibility, discretion and challenge rules must come from the parallel policy packet. The named examples in the reports belong to Domain Watch phishing categories, not the IWF case path.
\nThe2023 report explains that a non-suspension can reflect previous suspension, a court-ordered transfer, or subsequent compliance. It attributes reversal to cessation of the offending behaviour followed by confirmation from the enforcing agency. Its19 reversals are therefore not19 adjudicated false positives, independent appeals or reinstated IWF domains. It also reports1,058 proscribed-term flags and no domain suspensions, separating an automated flag from final action. No registrar names, individual certification, applicant correspondence or restoration trace accompanies the IWF statistic.
\nThe [1 December 2020 Nominet announcement](http[local research file] retained as an indexed-primary body after full-reader failure, explicitly assigns 5,006 suspensions pending checks and 558 completed registrations to Domain Watch. This prose corroborates the category assignment without relying on the one-page PDF's ambiguous extracted column order. That is positive Nominet-reported movement out of a hold, distinct from the 15 Criminal Practices Policy reversals. A separate paragraph reports 3,811 COVID-related holds and 1,568 completed checks as of 28 October 2020. The categories must not simply be added; overlap is not resolved. None proves that an initial flag was unlawful, that all false positives were remedied, or that an IWF request was reversed. This is the same publisher's corroborating account, not independent verification of individual outcomes.
\nNominet's [9 December2021 announcement](http[local research file] is available here only as an indexed-primary body after reader failure. It distinguishes539 separate law-enforcement requests from3,434 suspended domains for the year to31October2021. A request can therefore concern more than one domain; report vocabulary must be retained rather than treated as a uniform unit across all years. It also reports no IWF requests in that period. No full-page or original-byte read is claimed for this announcement.
\nAn indexed passage of [Nominet's retrospective on IWF's2020 report](http[local research file] says there were11 CSAM reports in .UK but no domain suspensions were necessary after investigation; it gives search-result deindexing as a more appropriate intervention in an example category. This is a Nominet account, not11 cases independently reviewed here. The full article returned429 and the route is closed. Its exact publication date was not established. The passage is useful because content reports, IWF domain-suspension requests and domain actions plainly need not be the same thing. Other numerical ratios in the indexed passage were not adopted or audited.
\nRead the complete substantive textual extraction of each one-page2020 and2023 report (including category descriptions and example labels); read the2025 Impact Issue's complete Criminal Activity spread, PDF9/printed16–17, reader lines971–1070. Surrounding pages appeared in the reader but no additional domain-case claims rely on them. No PDF facsimile was inspected in this lane. Four local .txt captures retain reader text, a clearly labelled scoped extraction, and indexed-primary article passages, including the subsequently reviewed 1 December 2020 category corroboration. Their SHA256 values are capture hashes, never original PDF/HTML hashes. exercised-action-captures.json records exact sources, dates, scopes and access failures.
\nSeven focused discovery queries tested IWF requests, recent reversals, UKSIC-labelled action and the older report-category contrast. No named IWF/UKSIC refusal or reinstatement was located; that is a scoped retrieval gap, not absence of such cases. Unrelated IWF acronym results and third-party commentary were excluded. The old2013 registration-review terms-of-reference lead was passed to the authority lane, not acquired here. No failed Nominet GET or FY2026 report was reopened.
\nThe next discriminating record is the administrative file behind one of the publicly reported IWF requests: original notification/certification, the domain-level decision, registrar implementation, and any review/restoration correspondence. A redacted case record could resolve standing and exercised discretion without disclosing abusive media or a harmful-content URL. Likely custodians are Nominet's compliance/legal function and IWF's operational referral team. No contact is authorised or attempted. Until such a record is available, the supported relation is reported IWF requests plus Nominet's separately described domain machinery, not a completed proof that a funding relationship compelled suspension.
", "path": "nominet-domain-authority-wave-2026-09-09/exercised-action.md", "sourceIds": [ "w29-enf-s-action2020", "w29-enf-s-action2023", "w29-enf-s-impact2025-actions" ], "markdown": "9 September 2026. Bounded acquisition CLOSED. Official Nominet reader/index records only; no original GET, domain operation, account, form, content/media access, outreach or Explorer action. Current standing/remedy rules belong to the parallel authority lane.\n\nThe inspected reports document IWF suspension requests, but do not identify their domains or publish an individual decision trail. They also report reversals and requests that did not lead to suspension. These demonstrate reported operational activity and limits; they do not establish wrongful decisions, independent verification, or rights created by Nominet's funding of UKSIC.\n\n## Comparable labels, bounded periods\n\n| Reporting period | IWF-labelled suspension requests | Criminal Practices Policy domains suspended | Requests not resulting in suspension | Suspension reversals |\n| --- | --- | --- | --- | --- |\n| 1 November 2019–31 October 2020 | 0 | 22,158 | 47 | 15 |\n| 1 November 2022–31 October 2023 | 1 | 1,193 | 3 | 19 |\n| 1 November 2023–31 October 2024 | 1 | 913 | 0 | 16 |\n\nSources are the official 2020 report (see source references) 2023 report (see source references) and 2025 Impact Issue (see source references) These are selected periods, not a continuous time series or September2026 snapshot. Publication/upload year is not activity year.\n\nThe IWF metric is expressly a request. Neither a case-specific successful suspension nor an IWF reversal is established by the aggregate columns. Do not derive either from a zero total of unsuccessful requests. The source does not reconcile the IWF-labelled metric with the agency totals; it must not be added to the reported suspension total as an extra domain. No UKSIC consortium-labelled request or named IWF-origin domain case was located in this pass.\n\n## Requests, decisions and implementation\n\nBoth one-page reports and the2025 spread depict agencies notifying Nominet, Nominet making administrative checks, and Nominet notifying the registrant and working with the registrar to suspend the domain. That is their summary of the law-enforcement process; it does not prove IWF is a law-enforcement agency or that its requests use every step of that route. Current eligibility, discretion and challenge rules must come from the parallel policy packet. The named examples in the reports belong to Domain Watch phishing categories, not the IWF case path.\n\nThe2023 report explains that a non-suspension can reflect previous suspension, a court-ordered transfer, or subsequent compliance. It attributes reversal to cessation of the offending behaviour followed by confirmation from the enforcing agency. Its19 reversals are therefore not19 adjudicated false positives, independent appeals or reinstated IWF domains. It also reports1,058 proscribed-term flags and no domain suspensions, separating an automated flag from final action. No registrar names, individual certification, applicant correspondence or restoration trace accompanies the IWF statistic.\n\n## A different positive correction account\n\nThe 1 December 2020 Nominet announcement (see source references) retained as an indexed-primary body after full-reader failure, explicitly assigns 5,006 suspensions pending checks and 558 completed registrations to Domain Watch. This prose corroborates the category assignment without relying on the one-page PDF's ambiguous extracted column order. That is positive Nominet-reported movement out of a hold, distinct from the 15 Criminal Practices Policy reversals. A separate paragraph reports 3,811 COVID-related holds and 1,568 completed checks as of 28 October 2020. The categories must not simply be added; overlap is not resolved. None proves that an initial flag was unlawful, that all false positives were remedied, or that an IWF request was reversed. This is the same publisher's corroborating account, not independent verification of individual outcomes.\n\nNominet's 9 December2021 announcement (see source references) is available here only as an indexed-primary body after reader failure. It distinguishes539 separate law-enforcement requests from3,434 suspended domains for the year to31October2021. A request can therefore concern more than one domain; report vocabulary must be retained rather than treated as a uniform unit across all years. It also reports no IWF requests in that period. No full-page or original-byte read is claimed for this announcement.\n\nAn indexed passage of Nominet's retrospective on IWF's2020 report (see source references) says there were11 CSAM reports in .UK but no domain suspensions were necessary after investigation; it gives search-result deindexing as a more appropriate intervention in an example category. This is a Nominet account, not11 cases independently reviewed here. The full article returned429 and the route is closed. Its exact publication date was not established. The passage is useful because content reports, IWF domain-suspension requests and domain actions plainly need not be the same thing. Other numerical ratios in the indexed passage were not adopted or audited.\n\n## Scope, limitations and next record\n\nRead the complete substantive textual extraction of each one-page2020 and2023 report (including category descriptions and example labels); read the2025 Impact Issue's complete Criminal Activity spread, PDF9/printed16–17, reader lines971–1070. Surrounding pages appeared in the reader but no additional domain-case claims rely on them. No PDF facsimile was inspected in this lane. Four local .txt captures retain reader text, a clearly labelled scoped extraction, and indexed-primary article passages, including the subsequently reviewed 1 December 2020 category corroboration. Their SHA256 values are capture hashes, never original PDF/HTML hashes. exercised-action-captures.json records exact sources, dates, scopes and access failures.\n\nSeven focused discovery queries tested IWF requests, recent reversals, UKSIC-labelled action and the older report-category contrast. No named IWF/UKSIC refusal or reinstatement was located; that is a scoped retrieval gap, not absence of such cases. Unrelated IWF acronym results and third-party commentary were excluded. The old2013 registration-review terms-of-reference lead was passed to the authority lane, not acquired here. No failed Nominet GET or FY2026 report was reopened.\n\nThe next discriminating record is the administrative file behind one of the publicly reported IWF requests: original notification/certification, the domain-level decision, registrar implementation, and any review/restoration correspondence. A redacted case record could resolve standing and exercised discretion without disclosing abusive media or a harmful-content URL. Likely custodians are Nominet's compliance/legal function and IWF's operational referral team. No contact is authorised or attempted. Until such a record is available, the supported relation is reported IWF requests plus Nominet's separately described domain machinery, not a completed proof that a funding relationship compelled suspension.\n", "file": "research/nominet-domain-authority-wave-2026-09-09-exercised-action.md" }, { "id": "nominet-domain-authority-wave-2026-09-09-policy-authority", "title": "Nominet domain-action authority: certification, private contract and review", "description": "Observed 9 September 2026. Acquisition CLOSED. This is a policy/standing packet; exercised requests and reversals belong to the parallel outcomes lane. No registry operation, account, form, outreach, domain-content retrieval or ordinary Nominet GET was used.", "html": "Observed 9 September 2026. Acquisition CLOSED. This is a policy/standing packet; exercised requests and reversals belong to the parallel outcomes lane. No registry operation, account, form, outreach, domain-content retrieval or ordinary Nominet GET was used.
\nIWF's documented cooperation with Nominet predates the 2021 UKSIC grant. The evidence does not establish that receiving the grant created a suspension entitlement. The current policy's named-agency list, a historical IWF cooperation statement and reported IWF requests are distinct evidence types. Their combination warrants investigating a separate referral arrangement; it does not supply its terms.
\nThe official [.UK Policy page](http[local research file] read at its current-policy section and future-policy distinction, links the Criminal Practices Policy dated 6 July 2026 and registrant terms dated 20 March 2024. It separately identifies revised agreements/policies taking effect on 9 February 2027, and says the Criminal Practices Policy will remain unchanged at that transition. The upload directories are not substituted for document dates. This establishes what Nominet currently publishes, not historical applicability to every reported case.
\nThe [6 July 2026 policy](http[local research file] supplies this mechanism:
\n| Question | \nPublished rule | \nLocator | \n
|---|---|---|
| Who establishes criminality? | \nA listed UK enforcement agency provides official written certification. Nominet describes its role as administrative checking, acknowledging limited criminal-assessment expertise. | \n§§2–3, PDF 3 | \n
| What follows? | \nFollowing checks and notice, Nominet places the domain in special status unless satisfactory action already occurred. | \n§3, PDF 3 | \n
| Notice and urgency | \nAt least 48 hours' notice; significant evidence of imminent public harm permits immediate suspension with senior-management approval. Agency confirmation that criminal use ended stops the process. | \n§3, PDF 4 | \n
| Duration/effect | \n.UK: minimum 24 months; .cymru/.wales: until expiry, then cancellation/re-registration. Suspension/redirection disrupts access without removing hosted content. | \n§§3,5, PDF 4,7 | \n
| Challenge | \nCriminality assessment: agency complaint/appeal routes, with ombudsman, police-conduct and judicial-review examples. Policy application: Nominet complaints. | \n§3, PDF 4 | \n
| Named standing | \nSection 4 lists 15 enforcement entries, including Gambling Commission. Neither IWF nor UKSIC appears. This is a list observation, not exclusion of every separate route. | \n§4, PDF 4–6 | \n
Nominet's current-linked [registrant terms](http[local research file] prohibit unlawful domain use (§6.1.5). Section10.1.2 reserves sole-discretion action for use, or high risk of use, likely to endanger DNS, users or Nominet systems; this differs from an agency certifying criminality. Section10.2.5 separately permits changes to implement a valid binding court order or avoid contempt. The private policy therefore must not be redescribed as a court order.
\nThe terms' proscribed-name definition requires both an apparent serious-sexual-offence meaning and no reasonably contemplated legitimate use (§1); §6.1.4 applies it to the domain's characters. That is not a prohibition of any unpopular website content. Section13.2 gives nonparties no enforcement rights under these terms; it does not negate rights arising under a separate agreement.
\nRemedies also have limits: §§12.3–12.5 exclude specified losses and cap liability at £5,000, with those clauses inapplicable to consumers under §12.6. Section13.8 selects England/Wales law and courts, accepting local law/courts for Scottish or Northern Irish consumers. These are contract provisions, not an enforceability opinion or case-specific ruling. Material locators: PDF 3,5–9.
\nThe historical [Domain registration policy review: Terms of Reference](http[local research file] PDF 2 §3, reports that Nominet is an IWF member and will act promptly to remove access to child sexual abuse images and criminally obscene adult content. It describes enforcement-agency working relationships in the preceding paragraph. The record is set in 2013 by its reference to the 4 August 2013 newspaper article and dated review timetable; an exact publication day was not established.
\nThis is explicit historical cooperation, not a signed IWF referral protocol. It does not identify evidential thresholds, a binding certification power, notice exceptions, an IWF-specific appeal process or UKSIC authority. It predates the 2021 consortium pledge examined in the accepted award packet. That chronology weighs against a claim that the later grant originated the relationship. It does not prove there was no earlier financing or that the relationship's scope never changed. Current IWF case counts belong to the outcomes packet and must retain their periods and request/action distinctions.
\nNominet's [3 April 2014 announcement](http[local research file] announced changes effective 4 May 2014. Its criminal-domain section said a suspension could be lifted when criminal conduct ended, when wrongly applied, or when the referring agency failed to respond within 10 working days. Its separate proscribed-name section described senior-manager checking, a 30-day appeal window and an independent external review of unresolved disagreements. These statements concern different decision tracks.
\nThose historical safeguards are useful, but the current CPP does not reproduce that 10-working-day nonresponse rule in its inspected complaints section. I have not established whether another operative document preserves it, nor identified the proscribed-name external reviewer. The current general complaints landing failed to load; an indexed excerpt was discovered but was not treated as the full procedure. The generic registrar complaint template surfaced in search is expressly a template, not a substitute for Nominet's own current procedure. No general right to automatic reinstatement has been inferred.
\nNominet's [16 July 2026 blog](http[local research file] reports Royal Assent to the Crime and Policing Act on 29 April 2026, describes court-ordered domain suspension as a backstop, and says commencement legislation was still needed, with an indicative end-November expectation. It reports government reassurance that voluntary CPP suspensions remain the first route. This packet does not verify the Act's provisions, its commencement as of 9 September, or any actual order. The linked legislation.gov.uk PDF returned 429 and was not read. The statutory instrument/commencement question is parked for a separately selected run; the blog is evidence of Nominet's stated implementation position.
\nAll successful source reads in this lane used the official web reader. No original HTML or PDF bytes were acquired. Retained policy-*.reader.txt files are exact tool-return derivatives; their hashes identify those derivatives, never the underlying originals. policy-captures.json records URLs, dates, ancestry, locators, failures and source states.
\nThe CPP's complete seven-page reader text was read, including every agency entry and the version table. Registrant terms: complete nine-page reader text, with the substantive authority/remedy analysis confined to the cited clauses. The historical review's complete four-page text was read; no contribution form was used. The 2014 announcement's substantive body, lines 43–73, and the July 2026 blog's substantive body, lines 43–102, were read. The current policy landing's lines 51–141 establish current/future links; unrelated council names and member-account material are not findings here. No screenshot or visual-original inspection is claimed.
\nThe old /policies link returned a non-retryable safe-open error. Targeted official search identified a genuinely different current publication page; the failed URL was not retried. The Security & Protection reader returned 429; the Complaints link returned an internal error; the linked Act PDF returned 429. All stayed closed, with no GET, URL-variant, mirror or support-layer recovery. Search-only hits and generic templates are recorded as discovery/exclusion rather than read operative authority. No new donor or grant-agreement hunt was undertaken.
\nThe finite next record is an IWF–Nominet referral/membership operating protocol or a public administrative case record explaining how an IWF request becomes a registry decision, who may contest the classification, and who can authorize reversal. Likely custodians are Nominet registry/legal operations and IWF. A second narrow gap is the actual current Nominet complaint procedure or identified external-review mandate. Until obtained, the defensible mechanism is separate contractual registry action, listed-agency certification and documented IWF cooperation. Expertise and earlier operational cooperation remain alternatives to a funding-created-right explanation.
", "path": "nominet-domain-authority-wave-2026-09-09/policy-authority.md", "sourceIds": [ "w29-enf-s-cpp2026", "w29-enf-s-terms2024", "w29-enf-s-iwf2013", "w29-enf-s-formal2014", "w29-enf-s-statutory-account2026" ], "markdown": "Observed 9 September 2026. Acquisition CLOSED. This is a policy/standing packet; exercised requests and reversals belong to the parallel outcomes lane. No registry operation, account, form, outreach, domain-content retrieval or ordinary Nominet GET was used.\n\nIWF's documented cooperation with Nominet predates the 2021 UKSIC grant. The evidence does not establish that receiving the grant created a suspension entitlement. The current policy's named-agency list, a historical IWF cooperation statement and reported IWF requests are distinct evidence types. Their combination warrants investigating a separate referral arrangement; it does not supply its terms.\n\n## Which instruments are current-linked?\n\nThe official .UK Policy page (see source references) read at its current-policy section and future-policy distinction, links the Criminal Practices Policy dated **6 July 2026** and registrant terms dated **20 March 2024**. It separately identifies revised agreements/policies taking effect on **9 February 2027**, and says the Criminal Practices Policy will remain unchanged at that transition. The upload directories are not substituted for document dates. This establishes what Nominet currently publishes, not historical applicability to every reported case.\n\n## Criminal Practices Policy: the operative allocation of decisions\n\nThe 6 July 2026 policy (see source references) supplies this mechanism:\n\n| Question | Published rule | Locator |\n| --- | --- | --- |\n| Who establishes criminality? | A listed UK enforcement agency provides official written certification. Nominet describes its role as administrative checking, acknowledging limited criminal-assessment expertise. | §§2–3, PDF 3 |\n| What follows? | Following checks and notice, Nominet places the domain in special status unless satisfactory action already occurred. | §3, PDF 3 |\n| Notice and urgency | At least 48 hours' notice; significant evidence of imminent public harm permits immediate suspension with senior-management approval. Agency confirmation that criminal use ended stops the process. | §3, PDF 4 |\n| Duration/effect | .UK: minimum 24 months; .cymru/.wales: until expiry, then cancellation/re-registration. Suspension/redirection disrupts access without removing hosted content. | §§3,5, PDF 4,7 |\n| Challenge | Criminality assessment: agency complaint/appeal routes, with ombudsman, police-conduct and judicial-review examples. Policy application: Nominet complaints. | §3, PDF 4 |\n| Named standing | Section 4 lists 15 enforcement entries, including Gambling Commission. Neither IWF nor UKSIC appears. This is a list observation, not exclusion of every separate route. | §4, PDF 4–6 |\n\n## Contractual authority is broader than the certification procedure\n\nNominet's current-linked registrant terms (see source references) prohibit unlawful domain use (§6.1.5). Section10.1.2 reserves sole-discretion action for use, or high risk of use, likely to endanger DNS, users or Nominet systems; this differs from an agency certifying criminality. Section10.2.5 separately permits changes to implement a valid binding court order or avoid contempt. The private policy therefore must not be redescribed as a court order.\n\nThe terms' proscribed-name definition requires both an apparent serious-sexual-offence meaning and no reasonably contemplated legitimate use (§1); §6.1.4 applies it to the domain's characters. That is not a prohibition of any unpopular website content. Section13.2 gives nonparties no enforcement rights under these terms; it does not negate rights arising under a separate agreement.\n\nRemedies also have limits: §§12.3–12.5 exclude specified losses and cap liability at £5,000, with those clauses inapplicable to consumers under §12.6. Section13.8 selects England/Wales law and courts, accepting local law/courts for Scottish or Northern Irish consumers. These are contract provisions, not an enforceability opinion or case-specific ruling. Material locators: PDF 3,5–9.\n\n## The IWF relationship before the consortium grant\n\nThe historical Domain registration policy review: Terms of Reference (see source references) PDF 2 §3, reports that Nominet is an IWF member and will act promptly to remove access to child sexual abuse images and criminally obscene adult content. It describes enforcement-agency working relationships in the preceding paragraph. The record is set in 2013 by its reference to the 4 August 2013 newspaper article and dated review timetable; an exact publication day was not established.\n\nThis is explicit historical cooperation, not a signed IWF referral protocol. It does not identify evidential thresholds, a binding certification power, notice exceptions, an IWF-specific appeal process or UKSIC authority. It predates the 2021 consortium pledge examined in the accepted [award packet](nominet-resources-wave-2026-09-09-uksic-awards.md). That chronology weighs against a claim that the later grant originated the relationship. It does not prove there was no earlier financing or that the relationship's scope never changed. Current IWF case counts belong to the outcomes packet and must retain their periods and request/action distinctions.\n\n## Historical remedy promises should not silently become current deadlines\n\nNominet's 3 April 2014 announcement (see source references) announced changes effective 4 May 2014. Its criminal-domain section said a suspension could be lifted when criminal conduct ended, when wrongly applied, or when the referring agency failed to respond within 10 working days. Its separate proscribed-name section described senior-manager checking, a 30-day appeal window and an independent external review of unresolved disagreements. These statements concern different decision tracks.\n\nThose historical safeguards are useful, but the current CPP does not reproduce that 10-working-day nonresponse rule in its inspected complaints section. I have not established whether another operative document preserves it, nor identified the proscribed-name external reviewer. The current general complaints landing failed to load; an indexed excerpt was discovered but was not treated as the full procedure. The generic registrar complaint template surfaced in search is expressly a template, not a substitute for Nominet's own current procedure. No general right to automatic reinstatement has been inferred.\n\n## Separate statutory development: reported status only\n\nNominet's 16 July 2026 blog (see source references) reports Royal Assent to the Crime and Policing Act on 29 April 2026, describes court-ordered domain suspension as a backstop, and says commencement legislation was still needed, with an indicative end-November expectation. It reports government reassurance that voluntary CPP suspensions remain the first route. This packet does not verify the Act's provisions, its commencement as of 9 September, or any actual order. The linked legislation.gov.uk PDF returned 429 and was not read. The statutory instrument/commencement question is parked for a separately selected run; the blog is evidence of Nominet's stated implementation position.\n\n## Custody, reading scope and closed routes\n\nAll successful source reads in this lane used the official web reader. **No original HTML or PDF bytes were acquired.** Retained policy-*.reader.txt files are exact tool-return derivatives; their hashes identify those derivatives, never the underlying originals. policy-captures.json records URLs, dates, ancestry, locators, failures and source states.\n\nThe CPP's complete seven-page reader text was read, including every agency entry and the version table. Registrant terms: complete nine-page reader text, with the substantive authority/remedy analysis confined to the cited clauses. The historical review's complete four-page text was read; no contribution form was used. The 2014 announcement's substantive body, lines 43–73, and the July 2026 blog's substantive body, lines 43–102, were read. The current policy landing's lines 51–141 establish current/future links; unrelated council names and member-account material are not findings here. No screenshot or visual-original inspection is claimed.\n\nThe old /policies link returned a non-retryable safe-open error. Targeted official search identified a genuinely different current publication page; the failed URL was not retried. The Security & Protection reader returned 429; the Complaints link returned an internal error; the linked Act PDF returned 429. All stayed closed, with no GET, URL-variant, mirror or support-layer recovery. Search-only hits and generic templates are recorded as discovery/exclusion rather than read operative authority. No new donor or grant-agreement hunt was undertaken.\n\nThe finite next record is an **IWF–Nominet referral/membership operating protocol or a public administrative case record** explaining how an IWF request becomes a registry decision, who may contest the classification, and who can authorize reversal. Likely custodians are Nominet registry/legal operations and IWF. A second narrow gap is the actual current Nominet complaint procedure or identified external-review mandate. Until obtained, the defensible mechanism is separate contractual registry action, listed-agency certification and documented IWF cooperation. Expertise and earlier operational cooperation remain alternatives to a funding-created-right explanation.\n", "file": "research/nominet-domain-authority-wave-2026-09-09-policy-authority.md" }, { "id": "nominet-resources-wave-2026-09-09-donor-resources", "title": "Nominet donor resources and decision rights", "description": "Bounded primary-source read, 9 September 2026. Acquisition CLOSED. This is a donor-side packet; uksic records in this wave belong to the separate recipient/award lane. No case, prior reader or ZIP changed.", "html": "Bounded primary-source read, 9 September 2026. Acquisition CLOSED. This is a donor-side packet; uksic records in this wave belong to the separate recipient/award lane. No case, prior reader or ZIP changed.
The premise is partly right: registry business helps finance public-benefit work. The acquired record does not describe a simple annual .uk surplus passed straight to charities. It describes a mixed business, accumulated investments/reserves, Board-agreed commitments and a further distribution step through a CAF account. This establishes resource provision, not donor direction of every recipient policy or classification.
\nThe current official governance footer identifies Nominet UK, company 03203859, a company limited by guarantee. Its public governance description reports more than 2,000 members; that is the organisation's current account, not an inspected membership register. Corporate members are not shareholders entitled to distributed profits. [Governance page](http[local research file] substantive body and footer, read 9 September 2026.
\nThe current-linked memorandum permits registry and related services and public-benefit contributions. Clause 3.12 authorizes donations, subscriptions, guarantees or free services for public, charitable or useful objects, with a political-party exclusion. Clause 6 directs income/capital to company objects and bars member profit distribution, subject to specified ordinary compensation/expense exceptions. This is a company constitution, not evidence that Nominet itself is a registered charity. [Memorandum](http[local research file] PDF pages 1–4 and targeted 5–7, clauses 3.12, 6, 8–9. Adoption date not established; the URL's upload directory is not that date.
\nThe current-linked articles, amended by special resolution 22 September 2020, give members election rights for up to four non-executive directors; the Board can appoint additional non-executives and executives under separate provisions. Board-appointed non-executives must seek member reappointment by ordinary resolution at the following AGM and periodically thereafter (article 30); executive appointments remain a separate category. Articles 35–42 put management of affairs/property with the Board, provide meeting/quorum/conflict rules, and permit delegation to committees or executives on conditions the Board can change or revoke. These powers support a Board-budget authority path, not an identified vote approving the UKSIC grant. [Articles](http[local research file] PDF pages 5–12, articles 19, 26–31 and 33–42. Registry-policy consultation under article 43 is not a general grant-approval consultation right.
\nThe current-linked dated voting by-law allocates 25% of poll votes equally among members and 75% by paid-for domains associated with member tags, with caps and connected-party rules. This makes commercial involvement relevant to constitutional/election voting; it does not establish a particular registrar's control of grant decisions. This is the linked 2012-file rule, not a reconstruction of every later ballot's applicable caps. No current member-weight register, decisive ballot or donor appointment instruction was acquired. [Voting by-law](http[local research file] PDF pages 1–3, clauses 4.1–4.3, scoped text through line 82.
\nGroup figures in the [FY2025 report](http[local research file] registry revenue £40.147m, DNS revenue £15.750m, total £55.897m (PDF32). Registry is not necessarily exclusively .uk; recognized revenue is not cash receipts.
\nThe 2021 commitment of up to £65m could use reserves or annual surplus. FY2025 trading operating profit was £5.463m; public-benefit expense was £10.979m. Investment assets were £88.6m. These are distinct flow/stock measures (PDF5–6,26).
\nPublic-benefit expense comprised programme £10.585m and operating costs £0.394m. Programme spending included £1.6m transferred to Nominet's CAF account for onward distribution in FY2026/27. The grant table excludes that transfer and reports UKSIC £1.647m. Do not add the transfer again or equate expenses with beneficiary bank receipts (PDF6).
\nThe former Public Benefit Committee's work moved to the main Board. Accounting policy recognizes public-benefit initiatives on an accrual basis against Board-agreed commitments; donations have a different paid-recipient recognition condition (PDF13,21). This establishes reported governance/accounting, not observed grant approval or completion.
\nThe supported path is business income plus accumulated reserves → Nominet Board resource decisions → programme commitments/distribution arrangements → recipient work. A second path is commercially weighted member votes → some director appointments → Board oversight. These should remain distinct: neither proves that large registrars ordered a specific grant or that receiving a grant surrendered editorial discretion.
\nThe public constitution lets the Board delegate; it does not reveal whether a particular programme award required a Board vote, delegated approval or later amendments. Financial review and reported impact monitoring are also different from a contractual right to clear recipients' advocacy. Recipient autonomy remains a live alternative unless the actual instrument assigns such rights.
\nThe annual-reports page now links FY2026. Its web reader rejected the PDF for size; the single ordinary public GET returned 403. It was not read. FY2025 is therefore a dated inspected financial snapshot, not a claim about the latest balance sheet or current budget.
\nSeven ordinary GETs returned identical 403 error bodies. They are retained with hashes as access observations only. The constitution and FY2025 claims above come from inspected official web-reader text; no successful PDF-original bytes were retained in this lane. donor-captures.json records exact scopes, reader-only status and error-body custody. The already-read current landing links establish which older instruments the site presented, not that no later instrument exists.
The next discriminating records are the current public-benefit budget/reserve policy and actual UKSIC award/renewal instrument or approval minute, including delegated limits, payment schedule, reporting, termination and any policy-clearance clauses. Board powers and accounting categories alone do not answer those questions. No wider donor genealogy or operational registry inquiry was opened.
", "path": "nominet-resources-wave-2026-09-09/donor-resources.md", "sourceIds": [ "w28-access-s-nominet-governance", "w28-access-s-nominet-articles2020", "w28-access-s-nominet-memorandum", "w28-access-s-nominet-voting2012", "w28-access-s-nominet-accounts2025", "w28-access-s-nominet-accounts2026-unacquired" ], "markdown": "Bounded primary-source read, 9 September 2026. Acquisition CLOSED. This is a donor-side packet; `uksic` records in this wave belong to the separate recipient/award lane. No case, prior reader or ZIP changed.\n\n## Finding\n\nThe premise is partly right: registry business helps finance public-benefit work. The acquired record does **not** describe a simple annual .uk surplus passed straight to charities. It describes a mixed business, accumulated investments/reserves, Board-agreed commitments and a further distribution step through a CAF account. This establishes resource provision, not donor direction of every recipient policy or classification.\n\n## Legal actor and corporate authority\n\nThe current official governance footer identifies **Nominet UK, company 03203859**, a company limited by guarantee. Its public governance description reports more than 2,000 members; that is the organisation's current account, not an inspected membership register. Corporate members are not shareholders entitled to distributed profits. Governance page (see source references) substantive body and footer, read 9 September 2026.\n\nThe current-linked memorandum permits registry and related services and public-benefit contributions. Clause 3.12 authorizes donations, subscriptions, guarantees or free services for public, charitable or useful objects, with a political-party exclusion. Clause 6 directs income/capital to company objects and bars member profit distribution, subject to specified ordinary compensation/expense exceptions. This is a company constitution, not evidence that Nominet itself is a registered charity. Memorandum (see source references) PDF pages 1–4 and targeted 5–7, clauses 3.12, 6, 8–9. Adoption date not established; the URL's upload directory is not that date.\n\nThe current-linked articles, amended by special resolution **22 September 2020**, give members election rights for up to four non-executive directors; the Board can appoint additional non-executives and executives under separate provisions. Board-appointed non-executives must seek member reappointment by ordinary resolution at the following AGM and periodically thereafter (article 30); executive appointments remain a separate category. Articles 35–42 put management of affairs/property with the Board, provide meeting/quorum/conflict rules, and permit delegation to committees or executives on conditions the Board can change or revoke. These powers support a Board-budget authority path, not an identified vote approving the UKSIC grant. Articles (see source references) PDF pages 5–12, articles 19, 26–31 and 33–42. Registry-policy consultation under article 43 is not a general grant-approval consultation right.\n\nThe current-linked dated voting by-law allocates **25% of poll votes equally among members and 75% by paid-for domains associated with member tags**, with caps and connected-party rules. This makes commercial involvement relevant to constitutional/election voting; it does not establish a particular registrar's control of grant decisions. This is the linked 2012-file rule, not a reconstruction of every later ballot's applicable caps. No current member-weight register, decisive ballot or donor appointment instruction was acquired. Voting by-law (see source references) PDF pages 1–3, clauses 4.1–4.3, scoped text through line 82.\n\n## Dated financial path: year ended 31 March 2025\n\nGroup figures in the FY2025 report (see source references) registry revenue **£40.147m**, DNS revenue **£15.750m**, total **£55.897m** (PDF32). Registry is not necessarily exclusively .uk; recognized revenue is not cash receipts.\n\nThe 2021 commitment of **up to £65m** could use reserves **or annual surplus**. FY2025 trading operating profit was **£5.463m**; public-benefit expense was **£10.979m**. Investment assets were **£88.6m**. These are distinct flow/stock measures (PDF5–6,26).\n\nPublic-benefit expense comprised programme **£10.585m** and operating costs **£0.394m**. Programme spending included **£1.6m transferred to Nominet's CAF account** for onward distribution in FY2026/27. The grant table excludes that transfer and reports **UKSIC £1.647m**. Do not add the transfer again or equate expenses with beneficiary bank receipts (PDF6).\n\nThe former Public Benefit Committee's work moved to the main Board. Accounting policy recognizes public-benefit initiatives on an accrual basis against Board-agreed commitments; donations have a different paid-recipient recognition condition (PDF13,21). This establishes reported governance/accounting, not observed grant approval or completion.\n\n## Mechanism and competing interpretation\n\nThe supported path is **business income plus accumulated reserves → Nominet Board resource decisions → programme commitments/distribution arrangements → recipient work**. A second path is **commercially weighted member votes → some director appointments → Board oversight**. These should remain distinct: neither proves that large registrars ordered a specific grant or that receiving a grant surrendered editorial discretion.\n\nThe public constitution lets the Board delegate; it does not reveal whether a particular programme award required a Board vote, delegated approval or later amendments. Financial review and reported impact monitoring are also different from a contractual right to clear recipients' advocacy. Recipient autonomy remains a live alternative unless the actual instrument assigns such rights.\n\n## Limits, custody and next record\n\nThe annual-reports page now links **FY2026**. Its web reader rejected the PDF for size; the single ordinary public GET returned 403. It was not read. FY2025 is therefore a dated inspected financial snapshot, **not a claim about the latest balance sheet or current budget**.\n\nSeven ordinary GETs returned identical 403 error bodies. They are retained with hashes as access observations only. The constitution and FY2025 claims above come from inspected official web-reader text; no successful PDF-original bytes were retained in this lane. `donor-captures.json` records exact scopes, reader-only status and error-body custody. The already-read current landing links establish which older instruments the site presented, not that no later instrument exists.\n\nThe next discriminating records are the **current public-benefit budget/reserve policy and actual UKSIC award/renewal instrument or approval minute**, including delegated limits, payment schedule, reporting, termination and any policy-clearance clauses. Board powers and accounting categories alone do not answer those questions. No wider donor genealogy or operational registry inquiry was opened.\n", "file": "research/nominet-resources-wave-2026-09-09-donor-resources.md" }, { "id": "nominet-resources-wave-2026-09-09-recipient-delivery", "title": "Nominet resources: recipient accounts and delivery", "description": "Observed 9 September 2026. Bounded acquisition closed after the two intended recipient accounts. All amounts below are GBP. This packet adds a recipient-side accounting and governance view; it does not establish bank receipts, a consortium allocation schedule or donor control over policy.", "html": "Observed 9 September 2026. Bounded acquisition closed after the two intended recipient accounts. All amounts below are GBP. This packet adds a recipient-side accounting and governance view; it does not establish bank receipts, a consortium allocation schedule or donor control over policy.
\nThe three charities' accounts all cover the year ended 31 March 2025. Their UKSIC-related restricted income rows sum to £1,779,208, compared with Nominet's £1,647,000 UKSIC grants-table amount for the same year end. The £132,208 difference is unreconciled. It is not missing or diverted cash. The retained records give no bridge establishing identical coverage and recognition bases.
\n| Recipient and exact row | \nFY2025 income | \nFY2025 expenditure | \nTransfer into fund | \nFY2024 income | \nOriginal locator | \n
|---|---|---|---|---|---|
| South West Grid for Learning Trust: Nominet Grant | \n£773,395 | \n£773,395 | \nNo transfer column; opening and closing balances zero | \n£842,780 | \nNote 19, PDF 34–36 / printed 32–34 | \n
| Internet Watch Foundation: Nominet UKSIC | \n£568,707 | \n£593,101 | \n£24,394 | \n£551,852 | \nNote 18a–c, PDF/printed 30–32; also Note 3, PDF 24 | \n
| Childnet: Nominet UKSIC 9 | \n£437,106 | \n£546,529 | \n£109,423 | \n£422,168 | \nNote 18, PDF 35–36 / printed 33–34 | \n
Each of the three FY2025 rows has zero opening and closing fund balances. Those balances and recorded expenditure are not evidence that every deliverable was accepted or every payable settled. IWF explicitly explains that unrestricted funds covered the UKSIC deficit. Childnet shows a transfer into its UKSIC fund; the same table includes transfers into and out of several other funds, so the individual transfer should not be confused with its net general-fund movement of £107,192.
\nThe [SWGfL original](http[local research file] PDF 36, expressly identifies its Nominet fund as funding from Nominet for operating UKSIC with consortium members. This is separate from the same page's Pershing/StopNCII purpose. Its PDF 22 recognition policy requires entitlement, performance conditions met, probable receipt and reliable measurement, with deferral where applicable.
\nThe [IWF original](http[local research file] Note 18c/PDF 32, explicitly describes its UKSIC fund as a grant from Nominet assisting Hotline running costs as part of UKSIC. Its attribution is therefore not inferred solely from a label. The same paragraph states that there are no unfulfilled conditions relating to amounts recognized that year. This is positive evidence at that recognition scope. It is not an inspected donor acceptance certificate, banking ledger or assurance that every obligation outside those recognized amounts was fulfilled. Notes 1d–e/PDF 21–22 recognize income when entitlement, relevant performance conditions, probable receipt and reliable measurement are satisfied; expenditure is accrued against obligations.
\nThe Childnet original.
\nThe donor comparison is taken from the parallel, reviewed donor-resources.md: Nominet's [FY2025 report](http[local research file] PDF 6 grants table, reports £1,647,000 for UKSIC and excludes the separate £1.6m transfer into Nominet's CAF account for later distribution. The donor lane read the official web-reader text; no successful original PDF bytes were retained. This recipient worker read that scoped packet, not the donor original. Donor commitments/accrual policy, recipient recognition timing, award-period spillover, other income within a recipient fund or the donor table's coverage are possible reconciliation questions, not established explanations. Matching year ends alone does not settle them.
\nOther Nominet programmes remain separate. IWF Note 3/PDF 24 reports £226,352 of technical-project grant income; Note 18a/PDF 30 reports £226,351 for the corresponding technical fund. Both original tables were visually checked: the £1 discrepancy is retained without correction or an inferred cause. Neither amount belongs in the UKSIC comparison. Childnet's separate Nominet Enhance fund has £11,783 income, £1,959 expenditure and £9,824 transferred out (PDF 35), also excluded from the UKSIC sum.
\nChildnet's trustees' report gives a concrete next instrument. PDF 12/printed 10 says it has a formal partnership agreement with SWGfL and IWF for UKSIC projects now part-funded by Nominet. The same page says trustees review and approve formal partnerships and major contractual relationships before signature. Day-to-day authority sits with the CEO and Business Manager under the annual business plan and budget, with payments above £3,000 additionally authorized by the Chair. These are reported internal controls and an agreement-existence disclosure, not the agreement's text, an observed approval minute or a Nominet reserved right.
\nThis complements the separate uksic-awards.md finding from the [IWF Project Officer recruitment pack](http[local research file] PDF 7/printed 6: financial returns and reports through the UKSIC Secretariat, liaison with Nominet-identified evaluators/subcontractors, and internal progress reporting. That earlier operational role specification is not proof that a named person was appointed, that an evaluator accepted these FY2025 outputs or that the donor had publication clearance. No signed consortium/grant agreement, acceptance report, bespoke editorial condition or renewal schedule was acquired in this recipient pass.
\nChildnet reports Safer Internet Day on 11 February 2025, over 1,700 registered supporting organisations, almost 160,000 educational-resource downloads and more than 86,000 quiz uses (PDF 8/printed 6). It explicitly credits Nominet support alongside other funders. This supports a reported delivery relationship, without attributing all outputs or costs exclusively to Nominet. The same page's survey-evaluation prose mixes 2024 and 2025 references; its percentage results are not relied on here. The auditors' report, PDF 16–17/printed 14–15, gives the financial-statements opinion and expressly distinguishes other annual-report information, so the delivery narrative is not independently audited impact evidence.
\nIWF's trustees' report, PDF 7, attributes the IntelliGrade Multichild feature to Nominet-funded technology work and reports 70,898 children recorded during 2024 who would otherwise have gone unrecorded. This is an operator-reported counting capability/output; it is not a separately verified number of removals, and the inspected text does not assign this feature to the UKSIC Hotline grant rather than the separately described technical programme. This packet preserves that funding-purpose boundary.
\nThe award packet separately reads Nominet's 2025 Impact Issue.
\nThe newly retained IWF original has 34 PDF pages; SHA256 0f96cd95f93b3155cb46163dc64fe04604148d8a57e6d9f5e99c5459d465d433. Text read: PDF 1,7,9,16,17,21,22,24,30–32; PDF 24 and30–32 visually checked. Capture is not a full-document read.
\nThe newly retained Childnet original has 38 scanned PDF pages; SHA256 64a68bdf33c2ecdb9c36da5ae5c11c912b1b3079e7eb979d2623d7abe6e8506e. Full-size images read: PDF 2,5,6,8,9,11–14,16–17,23–24,26,33–36. The whole-document contact sheet was navigation only. Text extraction produced no substantive content; its headings-only derivative is excluded as evidence. Local rendering of retained bytes succeeded. The public Companies House overview identifies Childnet, company 03958416; the original notes identify charity 1080173. No officer/residential details were researched or included.
\nThe reused SWGfL original has SHA256 3dce487d6ef022565efc4d10e947ee4427c401edb7a3b056ef98b37cdfac96c5 and remains in the prior stopncii-governance wave. This lane visually reread PDF 22 and34–36; prior wider scope remains in operator-resources.md. No reacquisition or prior sidecar mutation was made. IWF is company 03426366; SWGfL is company 05589479. Exact original URLs, successful GET observations, hashes, scopes and derivative ancestry are in recipient-captures.json.
\nAcquisition closed at the intended two new recipient records. Targeted primary searches and Companies House's actual filing links were used. The Companies House overview's failed web-reader method was replaced once with a successful ordinary public GET. Closed Nominet, oversized FY2026 and Charity Commission routes were not retried. No sign-in, form, outreach or paid route was used.
\nThe strongest next discriminator is the consortium partnership/grant agreement and the FY2025 donor-to-recipient reconciliation or financial return, held by the three recipient charities/UKSIC Secretariat and Nominet. These would distinguish allocation and timing from extra fund income and show actual reporting, acceptance or approval rights. The present record already establishes purpose restrictions, recipient expenditure, internal approval and reported delivery. It leaves donor direction of specific advocacy and the £132,208 accounting bridge unresolved. Recipient governance and contributions from other funds remain a defensible alternative to treating financing as an undifferentiated command relationship. No further branch was opened.
", "path": "nominet-resources-wave-2026-09-09/recipient-delivery.md", "sourceIds": [ "w28-access-s-childnet-accounts2025", "w28-access-s-childnet-identity", "w28-access-s-iwf-accounts2025" ], "markdown": "Observed 9 September 2026. Bounded acquisition closed after the two intended recipient accounts. All amounts below are GBP. This packet adds a recipient-side accounting and governance view; it does not establish bank receipts, a consortium allocation schedule or donor control over policy.\n\nThe three charities' accounts all cover the year ended **31 March 2025**. Their UKSIC-related restricted income rows sum to **£1,779,208**, compared with Nominet's **£1,647,000** UKSIC grants-table amount for the same year end. The **£132,208 difference is unreconciled**. It is not missing or diverted cash. The retained records give no bridge establishing identical coverage and recognition bases.\n\n## Restricted funds, attribution and the comparison\n\n| Recipient and exact row | FY2025 income | FY2025 expenditure | Transfer into fund | FY2024 income | Original locator |\n| --- | --- | --- | --- | --- | --- |\n| South West Grid for Learning Trust: Nominet Grant | £773,395 | £773,395 | No transfer column; opening and closing balances zero | £842,780 | Note 19, PDF 34–36 / printed 32–34 |\n| Internet Watch Foundation: Nominet UKSIC | £568,707 | £593,101 | £24,394 | £551,852 | Note 18a–c, PDF/printed 30–32; also Note 3, PDF 24 |\n| Childnet: Nominet UKSIC 9 | £437,106 | £546,529 | £109,423 | £422,168 | Note 18, PDF 35–36 / printed 33–34 |\n\nEach of the three FY2025 rows has zero opening and closing fund balances. Those balances and recorded expenditure are not evidence that every deliverable was accepted or every payable settled. IWF explicitly explains that unrestricted funds covered the UKSIC deficit. Childnet shows a transfer into its UKSIC fund; the same table includes transfers into and out of several other funds, so the individual transfer should not be confused with its net general-fund movement of £107,192.\n\nThe SWGfL original (see source references) PDF 36, expressly identifies its Nominet fund as funding from Nominet for operating UKSIC with consortium members. This is separate from the same page's Pershing/StopNCII purpose. Its PDF 22 recognition policy requires entitlement, performance conditions met, probable receipt and reliable measurement, with deferral where applicable.\n\nThe IWF original (see source references) Note 18c/PDF 32, explicitly describes its UKSIC fund as **a grant from Nominet** assisting Hotline running costs as part of UKSIC. Its attribution is therefore not inferred solely from a label. The same paragraph states that there are no unfulfilled conditions relating to amounts recognized that year. This is positive evidence at that recognition scope. It is not an inspected donor acceptance certificate, banking ledger or assurance that every obligation outside those recognized amounts was fulfilled. Notes 1d–e/PDF 21–22 recognize income when entitlement, relevant performance conditions, probable receipt and reliable measurement are satisfied; expenditure is accrued against obligations.\n\nThe Childnet original.\n\nThe donor comparison is taken from the parallel, reviewed [donor-resources.md](nominet-resources-wave-2026-09-09-donor-resources.md): Nominet's FY2025 report (see source references) PDF 6 grants table, reports £1,647,000 for UKSIC and excludes the separate £1.6m transfer into Nominet's CAF account for later distribution. The donor lane read the official web-reader text; no successful original PDF bytes were retained. This recipient worker read that scoped packet, not the donor original. Donor commitments/accrual policy, recipient recognition timing, award-period spillover, other income within a recipient fund or the donor table's coverage are possible reconciliation questions, **not established explanations**. Matching year ends alone does not settle them.\n\nOther Nominet programmes remain separate. IWF Note 3/PDF 24 reports £226,352 of technical-project grant income; Note 18a/PDF 30 reports £226,351 for the corresponding technical fund. Both original tables were visually checked: the £1 discrepancy is retained without correction or an inferred cause. Neither amount belongs in the UKSIC comparison. Childnet's separate Nominet Enhance fund has £11,783 income, £1,959 expenditure and £9,824 transferred out (PDF 35), also excluded from the UKSIC sum.\n\n## A disclosed agreement and an accountable internal route\n\nChildnet's trustees' report gives a concrete next instrument. PDF 12/printed 10 says it has a **formal partnership agreement with SWGfL and IWF** for UKSIC projects now part-funded by Nominet. The same page says trustees review and approve formal partnerships and major contractual relationships before signature. Day-to-day authority sits with the CEO and Business Manager under the annual business plan and budget, with payments above £3,000 additionally authorized by the Chair. These are reported internal controls and an agreement-existence disclosure, not the agreement's text, an observed approval minute or a Nominet reserved right.\n\nThis complements the separate [uksic-awards.md](nominet-resources-wave-2026-09-09-uksic-awards.md) finding from the IWF Project Officer recruitment pack (see source references) PDF 7/printed 6: financial returns and reports through the UKSIC Secretariat, liaison with Nominet-identified evaluators/subcontractors, and internal progress reporting. That earlier operational role specification is not proof that a named person was appointed, that an evaluator accepted these FY2025 outputs or that the donor had publication clearance. No signed consortium/grant agreement, acceptance report, bespoke editorial condition or renewal schedule was acquired in this recipient pass.\n\n## Delivery and the March 2025 boundary\n\nChildnet reports Safer Internet Day on **11 February 2025**, over 1,700 registered supporting organisations, almost 160,000 educational-resource downloads and more than 86,000 quiz uses (PDF 8/printed 6). It explicitly credits Nominet support alongside other funders. This supports a reported delivery relationship, without attributing all outputs or costs exclusively to Nominet. The same page's survey-evaluation prose mixes 2024 and 2025 references; its percentage results are not relied on here. The auditors' report, PDF 16–17/printed 14–15, gives the financial-statements opinion and expressly distinguishes other annual-report information, so the delivery narrative is not independently audited impact evidence.\n\nIWF's trustees' report, PDF 7, attributes the IntelliGrade Multichild feature to Nominet-funded technology work and reports 70,898 children recorded during 2024 who would otherwise have gone unrecorded. This is an operator-reported counting capability/output; it is not a separately verified number of removals, and the inspected text does not assign this feature to the UKSIC Hotline grant rather than the separately described technical programme. This packet preserves that funding-purpose boundary.\n\nThe [award packet](nominet-resources-wave-2026-09-09-uksic-awards.md) separately reads Nominet's 2025 Impact Issue.\n\n## Source custody and finite frontier\n\nThe newly retained IWF original has 34 PDF pages; SHA256 **0f96cd95f93b3155cb46163dc64fe04604148d8a57e6d9f5e99c5459d465d433**. Text read: PDF 1,7,9,16,17,21,22,24,30–32; PDF 24 and30–32 visually checked. Capture is not a full-document read.\n\nThe newly retained Childnet original has 38 scanned PDF pages; SHA256 **64a68bdf33c2ecdb9c36da5ae5c11c912b1b3079e7eb979d2623d7abe6e8506e**. Full-size images read: PDF 2,5,6,8,9,11–14,16–17,23–24,26,33–36. The whole-document contact sheet was navigation only. Text extraction produced no substantive content; its headings-only derivative is excluded as evidence. Local rendering of retained bytes succeeded. The public Companies House overview identifies Childnet, company 03958416; the original notes identify charity 1080173. No officer/residential details were researched or included.\n\nThe reused SWGfL original has SHA256 **3dce487d6ef022565efc4d10e947ee4427c401edb7a3b056ef98b37cdfac96c5** and remains in the prior stopncii-governance wave. This lane visually reread PDF 22 and34–36; prior wider scope remains in [operator-resources.md](stopncii-governance-wave-2026-09-09-operator-resources.md). No reacquisition or prior sidecar mutation was made. IWF is company 03426366; SWGfL is company 05589479. Exact original URLs, successful GET observations, hashes, scopes and derivative ancestry are in recipient-captures.json.\n\nAcquisition closed at the intended two new recipient records. Targeted primary searches and Companies House's actual filing links were used. The Companies House overview's failed web-reader method was replaced once with a successful ordinary public GET. Closed Nominet, oversized FY2026 and Charity Commission routes were not retried. No sign-in, form, outreach or paid route was used.\n\nThe strongest next discriminator is the **consortium partnership/grant agreement and the FY2025 donor-to-recipient reconciliation or financial return**, held by the three recipient charities/UKSIC Secretariat and Nominet. These would distinguish allocation and timing from extra fund income and show actual reporting, acceptance or approval rights. The present record already establishes purpose restrictions, recipient expenditure, internal approval and reported delivery. It leaves donor direction of specific advocacy and the £132,208 accounting bridge unresolved. Recipient governance and contributions from other funds remain a defensible alternative to treating financing as an undifferentiated command relationship. No further branch was opened.\n", "file": "research/nominet-resources-wave-2026-09-09-recipient-delivery.md" }, { "id": "nominet-resources-wave-2026-09-09-uksic-awards", "title": "Nominet → UK Safer Internet Centre: award, closure and later commissioned work", "description": "Observed 9 September 2026. Bounded award/conditions lane, now closed. This packet does not replace donor governance or recipient accounts work. No case, renderer, ZIP, current-work or Explorer changes.", "html": "Observed 9 September 2026. Bounded award/conditions lane, now closed. This packet does not replace donor governance or recipient accounts work. No case, renderer, ZIP, current-work or Explorer changes.
\nThe public record separates a three-year programme from later support. Nominet announced a £5.1m pledge in October 2021; its 2025 retrospective says that grant ended in March. A subsequent named research commission includes disclosed donor input into study design. An early-2022 IWF role description also establishes a financial-reporting and evaluator-liaison structure. None of these records supplies an executed UKSIC grant agreement or general donor editorial veto.
\n| Record | \nWhat it establishes | \nWhat it does not establish | \n
|---|---|---|
| Nominet announcement, 29 October 2021 | \nGBP5.1m pledged to help fund UKSIC over three years; intended to replace lost EU support. UKSIC comprises IWF, Childnet and SWGfL. | \nExecuted agreement, instalment schedule, precise commencement day, recipient allocation or completed payment. | \n
| Nominet announcement, 25 January 2022 | \nA GBP250,000 commitment for Safer Internet Day 2022, described as made before the October pledge. | \nPayment, precise commitment day or a consolidated accounting reconciliation. Keep this named commitment separate; do not mechanically add it to annual UKSIC amounts. | \n
| Nominet FY2024 summary, grants table | \nUKSIC GBP1,599,000 for year ended31March2024, in a table of grants over£250k. IWF GBP250,000 is a separate row. | \nSWGfL-only amount, three equal instalments, or a banking record showing the consortium allocation. | \n
| FY2025 donor table supplied by parallel lane | \nUKSIC GBP1,647,000; donor lane owns substantive accounts reading. This lane encountered the indexed table only. | \nIndependent full-original validation by this worker. Neither annual number should be added to the aggregate award as an additional award. | \n
| 2025 Impact Issue, PDF8/printed14–15 | \nNominet reports the three-year grant ended in March and the partnership ended in its existing form, while committing to continued work/support. March2025 follows the edition context; no exact termination day asserted. | \nAutomatic renewal, total withdrawal of support, default/breach, or exercise of a contractual termination right. | \n
| Safer Internet Day2026 report | \nA distinct Nominet-funded commission with disclosed design input and named research delivery. | \nAmount, executed contract, cash dates, or proof this commission was an extension/drawdown of the £5.1m award. | \n
The [2021 announcement](http[local research file] describes a pledge despite its headline saying the Centre receives funding. It also describes separate Childnet Digital Leaders+, SWGfL Project Evolve and IWF Countering Harm support. Those programmes cannot silently be included in the consortium allocation. The [January2022 announcement](http[local research file] distinguishes the earlier Safer Internet Day commitment. The [FY2024 summary](http[local research file] is a donor financial summary, not a recipient bank statement.
\nThe retained [IWF Project Officer recruitment pack](http[local research file] contains a dedicated “Nominet grant for UK Safer Internet Centre” section at PDF7/printed6. The role is to communicate with Nominet and subcontractors it identifies, including evaluators; liaise with UKSIC partners, particularly the Secretariat, on financial returns/reports to Nominet; and coordinate IWF compliance with its obligations. Quarterly progress reports expressly go to IWF’s Senior Leadership Team. The neighbouring grant-management section describes internal monitoring/evaluation and approval by the Deputy CEO before new applications; those generic duties must not become bespoke Nominet contractual rights.
\nThis is stronger evidence than a logo: IWF publicly designed a role around grant communication and reporting, including funder-identified evaluators. It is still a recruitment specification, not an executed agreement, appointed-person record, named evaluator engagement, submitted report or evidence that Nominet accepted an output. No publication-clearance, withdrawal, clawback or renewal clause appears in the inspected job-description passages. PDF10/printed9 gives a closing date of11February2022; PDF metadata creation/modification is20January2022. These do not establish an exact public publication date or current staffing arrangement.
\nThe [2025 retrospective](http[local research file] credited to Social Impact Lead Adam Groves, is explicit about the grant ending. It associates funding with Safer Internet Day, helplines, removal work and policy contributions, including the Online Safety Act and school-filtering standards. These are donor/recipient impact claims, not output-acceptance documents or independent evidence that funding caused particular legal changes. Their presence does not confer donor clearance authority.
\nThe [2026 report](http[local research file] says Nominet commissioned and funded the research; Childnet led study design with Nominet input; Opinium conducted surveys in November2025 of2,018 children and2,000 parents/carers. Childnet separately consulted youth groups in November–December2025 and ran school focus groups in May–June2025. This supports an attributed commissioning-and-design-input relation, with Childnet retaining the stated lead role. It does not establish which questions Nominet changed, a right of veto, or control over survey responses. The public report is a delivered publication, without a located contractual acceptance record. Its year is2026; the January URL directory is not an exact publication date. No survey-result validity audit was conducted.
\nA [20December2024 parliamentary answer](http[local research file] to a question about support after March2025, promises continued government engagement toward sustainable funding. It is not a government funding award or underwriting commitment.
\nAll acquisition ended after nine focused query strings and the decisive public records. Nominet original GET routes were already closed in the parallel donor lane; this worker made no Nominet GET or URL-variant retry. Nominet sources here are retained web-reader/search responses, not original HTML/PDF bytes, and their hashes must never be labelled original hashes. The 2025 Impact Issue screenshot returned an internal error. The2026 screenshot returned a reference without an image available for this worker to inspect; no visual verification claimed for either. The IWF ordinary public GET succeeded after its web reader failed; original bytes, derived text and the visually checked PDF7 image are retained.
\nActual source scopes:2021 announcement reader lines43–84 (complete article including notes);2022 announcement lines43–90 (complete article);2025 Impact Issue PDF8/printed14–15 complete two-page spread, plus foreword context PDF2;FY2024 summary PDF5 funding table and adjacent funding-account context in reader;2026 report cover, PDF2 commissioning/method paragraph and accompanying sample/context, PDF3–4 forewords; parliamentary question/answer lines17–39. IWF PDF1 and5–11 text read, decisive PDF7 visually checked; no claim to have read PDF2–4. Current Our Impact page was only partially inspected and supports no additional substantive proposition here. The February2024 Nominet blog reader failed; its search excerpt is retained as a lead only. Scope metadata and exact hashes are in uksic-awards-captures.json.
\nNo executed grant, allocation schedule, amendment/renewal agreement, bespoke approval/clawback clause, named evaluator contract or delivered financial return was located in this pass. That is a bounded retrieval result, not proof no such conditions exist. Current DNS Fund application rules are a different instrument and were not imported into UKSIC grants.
\nThe most useful next record is the2021/2022 Nominet–UKSIC grant agreement and financial/reporting schedule, including any Secretariat administration arrangement and evaluator terms. Likely custodians are Nominet’s public-benefit team and the UKSIC Secretariat/three consortium charities. A second precise record is the2025/2026 Nominet–Childnet research brief or commissioning agreement: it would show whether disclosed design input was advisory, jointly agreed or backed by final approval. No outreach is authorised or undertaken. The local forest implication is a demonstrable funding-to-reporting/design route with a dated programme end, not an undifferentiated continuing donor command.
", "path": "nominet-resources-wave-2026-09-09/uksic-awards.md", "sourceIds": [ "w28-access-s-nominet-pledge2021", "w28-access-s-nominet-sid2022", "w28-access-s-nominet-summary2024", "w28-access-s-nominet-impact2025", "w28-access-s-sid-research2026", "w28-access-s-iwf-role2022" ], "markdown": "Observed 9 September 2026. Bounded award/conditions lane, now closed. This packet does not replace donor governance or recipient accounts work. No case, renderer, ZIP, current-work or Explorer changes.\n\nThe public record separates a three-year programme from later support. Nominet announced a £5.1m pledge in October 2021; its 2025 retrospective says that grant ended in March. A subsequent named research commission includes disclosed donor input into study design. An early-2022 IWF role description also establishes a financial-reporting and evaluator-liaison structure. None of these records supplies an executed UKSIC grant agreement or general donor editorial veto.\n\n## Amounts and instrument boundaries\n\n| Record | What it establishes | What it does not establish |\n| --- | --- | --- |\n| Nominet announcement, 29 October 2021 | GBP5.1m pledged to help fund UKSIC over three years; intended to replace lost EU support. UKSIC comprises IWF, Childnet and SWGfL. | Executed agreement, instalment schedule, precise commencement day, recipient allocation or completed payment. |\n| Nominet announcement, 25 January 2022 | A GBP250,000 commitment for Safer Internet Day 2022, described as made before the October pledge. | Payment, precise commitment day or a consolidated accounting reconciliation. Keep this named commitment separate; do not mechanically add it to annual UKSIC amounts. |\n| Nominet FY2024 summary, grants table | UKSIC GBP1,599,000 for year ended31March2024, in a table of grants over£250k. IWF GBP250,000 is a separate row. | SWGfL-only amount, three equal instalments, or a banking record showing the consortium allocation. |\n| FY2025 donor table supplied by parallel lane | UKSIC GBP1,647,000; donor lane owns substantive accounts reading. This lane encountered the indexed table only. | Independent full-original validation by this worker. Neither annual number should be added to the aggregate award as an additional award. |\n| 2025 Impact Issue, PDF8/printed14–15 | Nominet reports the three-year grant ended in March and the partnership ended in its existing form, while committing to continued work/support. March2025 follows the edition context; no exact termination day asserted. | Automatic renewal, total withdrawal of support, default/breach, or exercise of a contractual termination right. |\n| Safer Internet Day2026 report | A distinct Nominet-funded commission with disclosed design input and named research delivery. | Amount, executed contract, cash dates, or proof this commission was an extension/drawdown of the £5.1m award. |\n\nThe 2021 announcement (see source references) describes a pledge despite its headline saying the Centre receives funding. It also describes separate Childnet Digital Leaders+, SWGfL Project Evolve and IWF Countering Harm support. Those programmes cannot silently be included in the consortium allocation. The January2022 announcement (see source references) distinguishes the earlier Safer Internet Day commitment. The FY2024 summary (see source references) is a donor financial summary, not a recipient bank statement.\n\n## Reporting and selection of evaluators: a public operational record\n\nThe retained IWF Project Officer recruitment pack (see source references) contains a dedicated “Nominet grant for UK Safer Internet Centre” section at PDF7/printed6. The role is to communicate with Nominet and subcontractors it identifies, including evaluators; liaise with UKSIC partners, particularly the Secretariat, on financial returns/reports to Nominet; and coordinate IWF compliance with its obligations. Quarterly progress reports expressly go to IWF’s Senior Leadership Team. The neighbouring grant-management section describes internal monitoring/evaluation and approval by the Deputy CEO before new applications; those generic duties must not become bespoke Nominet contractual rights.\n\nThis is stronger evidence than a logo: IWF publicly designed a role around grant communication and reporting, including funder-identified evaluators. It is still a recruitment specification, not an executed agreement, appointed-person record, named evaluator engagement, submitted report or evidence that Nominet accepted an output. No publication-clearance, withdrawal, clawback or renewal clause appears in the inspected job-description passages. PDF10/printed9 gives a closing date of11February2022; PDF metadata creation/modification is20January2022. These do not establish an exact public publication date or current staffing arrangement.\n\n## Programme end, continuing support and disclosed research input\n\nThe 2025 retrospective (see source references) credited to Social Impact Lead Adam Groves, is explicit about the grant ending. It associates funding with Safer Internet Day, helplines, removal work and policy contributions, including the Online Safety Act and school-filtering standards. These are donor/recipient impact claims, not output-acceptance documents or independent evidence that funding caused particular legal changes. Their presence does not confer donor clearance authority.\n\nThe 2026 report (see source references) says Nominet commissioned and funded the research; Childnet led study design with Nominet input; Opinium conducted surveys in November2025 of2,018 children and2,000 parents/carers. Childnet separately consulted youth groups in November–December2025 and ran school focus groups in May–June2025. This supports an attributed commissioning-and-design-input relation, with Childnet retaining the stated lead role. It does not establish which questions Nominet changed, a right of veto, or control over survey responses. The public report is a delivered publication, without a located contractual acceptance record. Its year is2026; the January URL directory is not an exact publication date. No survey-result validity audit was conducted.\n\nA 20December2024 parliamentary answer (see source references) to a question about support after March2025, promises continued government engagement toward sustainable funding. It is not a government funding award or underwriting commitment.\n\n## Reading scope and custody\n\nAll acquisition ended after nine focused query strings and the decisive public records. Nominet original GET routes were already closed in the parallel donor lane; this worker made no Nominet GET or URL-variant retry. Nominet sources here are retained web-reader/search responses, not original HTML/PDF bytes, and their hashes must never be labelled original hashes. The 2025 Impact Issue screenshot returned an internal error. The2026 screenshot returned a reference without an image available for this worker to inspect; no visual verification claimed for either. The IWF ordinary public GET succeeded after its web reader failed; original bytes, derived text and the visually checked PDF7 image are retained.\n\nActual source scopes:2021 announcement reader lines43–84 (complete article including notes);2022 announcement lines43–90 (complete article);2025 Impact Issue PDF8/printed14–15 complete two-page spread, plus foreword context PDF2;FY2024 summary PDF5 funding table and adjacent funding-account context in reader;2026 report cover, PDF2 commissioning/method paragraph and accompanying sample/context, PDF3–4 forewords; parliamentary question/answer lines17–39. IWF PDF1 and5–11 text read, decisive PDF7 visually checked; no claim to have read PDF2–4. Current Our Impact page was only partially inspected and supports no additional substantive proposition here. The February2024 Nominet blog reader failed; its search excerpt is retained as a lead only. Scope metadata and exact hashes are in uksic-awards-captures.json.\n\n## Finite gap and next discriminator\n\nNo executed grant, allocation schedule, amendment/renewal agreement, bespoke approval/clawback clause, named evaluator contract or delivered financial return was located in this pass. That is a bounded retrieval result, not proof no such conditions exist. Current DNS Fund application rules are a different instrument and were not imported into UKSIC grants.\n\nThe most useful next record is the2021/2022 Nominet–UKSIC grant agreement and financial/reporting schedule, including any Secretariat administration arrangement and evaluator terms. Likely custodians are Nominet’s public-benefit team and the UKSIC Secretariat/three consortium charities. A second precise record is the2025/2026 Nominet–Childnet research brief or commissioning agreement: it would show whether disclosed design input was advisory, jointly agreed or backed by final approval. No outreach is authorised or undertaken. The local forest implication is a demonstrable funding-to-reporting/design route with a dated programme end, not an undifferentiated continuing donor command.\n", "file": "research/nominet-resources-wave-2026-09-09-uksic-awards.md" }, { "id": "oak-cma-support-wave-2026-09-09-oak-support", "title": "Oak support for Check My Ads: a verified grant, bounded terms", "description": "Closed acquisition: 9 September 2026. This packet uses donor records and one already recipient record. It does not establish an executed CMA agreement or a payment receipt.", "html": "Closed acquisition: 9 September 2026. This packet uses donor records and one already retained recipient record. It does not establish an executed CMA agreement or a payment receipt.
\nOak's own 2025 Annual Report, PDF/printed page 24, lists Check My Ads — USD 600,000 (3 years). The purpose is core support for greater transparency and fairness in digital advertising. It appears in the International Human Rights programme. The annual report presents its list as 2025 grant-making; the amount is the listed three-year grant, not an annual amount or evidence that $600,000 reached CMA in 2025. Exact approval, start, end and instalment dates are absent from this entry. [Current donor original](http[local research file] retained as oak-annual2025-current.pdf; visual check oak-annual2025-current-p24.png.
| Field | \nSupported reading | \n
|---|---|
| Public grantor attribution | \nOak Foundation. The report's page 4 describes a group of philanthropic organisations; it does not identify which legal entity executed this grant. | \n
| Recipient as printed | \nCheck My Ads. This row supplies no legal suffix, EIN or registered address. Do not silently expand it to a verified contracting entity. | \n
| Amount and duration | \nUSD 600,000 over a reported three-year grant. No per-year allocation or payment schedule is given. | \n
| Resource classification | \nDonor-reported core-support grant in its 2025 report. Cash disbursement and recipient recognition remain unverified. | \n
| Specific campaign allocation | \nNone disclosed in the grant entry. No link to the February 2025 Senate letters or the Ofcom submission is established by this record. | \n
The currently linked report has production metadata dated 20 March 2026 and modification metadata dated 3 June 2026. These are file-production dates, not the exact date it became public or the date of grant approval. Its page 6 reports grants made; pages 6–7 do not supply a CMA payment reconciliation. Pages 8–9 discuss general support, evaluation and reporting, including selected pilots; they do not disclose CMA-specific reporting terms.
\nSource reuse: root checked the current case and identified this exact application guide as existing source w3-oak-grant-process, with the same URL and SHA-256. Its senior-decision role (w3-oak-senior-decision) and generic decision/letter relation already exist. This capture adds a current-link observation and scoped reread, not an independent corroborating source or a newly discovered approval mechanism. The new extension is the CMA grant.
The general process reveals where decisions occur. In the currently linked [application guide](http[local research file] PDF pages 2–3, programme officers screen strategic fit and work with applicants. Staff, leadership and trustees may assess an early concept; a complete application is reviewed by the programme officer, finance team, programme director and possibly an external reviewer. The programme officer recommends a funding decision to senior foundation members. On approval, the guide describes a grant letter reviewed and signed by both parties, followed by a bank wire. Its timelines are approximate. This is a public description of a process, not proof that this particular letter was signed, what it required, or when a payment occurred. The guide's metadata dates to 28 November 2019; its continuing public link was observed on 9 September 2026. Original: oak-application-guide.pdf, all three pages read; page 3 visually checked in oak-application-guide-p3.png.
The live [grant-making page](http[local research file] describes recipients setting their priorities and Oak matching existing goals to its programmes, using flexible project funding or core support. Renewals are described as separate grants. The current [FAQ](http[local research file] especially “About” questions 4–9, “Grant-making” questions 1–2, “Due diligence and review,” and “What Oak Foundation does not fund” questions 4 and 6, says core support is unrestricted and recipients direct their work. It also identifies financial, safeguarding and internal-control review, continued discussion during grants, and exclusions for election campaigns or party-affiliated political activity. These are current public policies, not recovered historical CMA contractual clauses. The general 20% organisational / 50% project support norms allow exceptions and cannot establish CMA's budget, Oak's share of it, or a three-year grant's annual allocation.
\nThe strongest defensible mechanism is resource selection and sustained organisational capacity, with donor screening and an expected agreement, alongside an express public policy of recipient independence. That is stronger than a mere donor-roster association. It still does not show Oak chose CMA's targets, approved particular submissions, or purchased outcomes. The rival explanation supported by the public policies is that Oak selected an independently directed organisation already pursuing aligned goals. The decisive evidence needed to distinguish more specific conditions is the CMA grant letter and incorporated application, budget, reporting schedule and amendments, together with payment/recognition records. These were not acquired; the generic workflow only identifies the likely instrument.
\nFor historical context, CMA's retained 2025 annual report, PDF35, labels foundation grants as 82.8% of funding. That chart is not a dollar denominator for this three-year Oak amount and does not reconcile an Oak payment. Reused original: ../cma-danish-support-wave-2026-09-09/cma-annual2025.pdf, page 35 text read in this lane; root previously visually checked the chart. [Recipient report](http[local research file]
Source custody and finite frontier: the initially indexed March-path PDF returned 404 through the official reader and one ordinary GET; that route is closed. Its error body is preserved separately. Oak's actual homepage then supplied a different current June-path PDF link, which was captured successfully. This was an observed publisher link, not a guessed filename retry. The [latest-grants page](http[local research file] directs readers to the annual report while a new database is being prepared; no database or application system was queried. The FAQ, grant-making page and linked guide were read for the applicable public framework; no extra branding, safeguarding-policy or family-history branch was pursued.
\noak-captures.json records exact URLs, acquisition times where retained, SHA-256 hashes, ancestry and reading scope. The 60-page report was captured and text-extracted in full, but substantive reading was scoped to PDF1, 3–9, 22 and 24, with page 5 partial and page 24 visual. Other pages were not silently promoted to read evidence. Originals and derivative files remain distinct. No new graph, current-work, forest, presentation or package changes were made. Acquisition is closed pending root review.
Bounded acquisition closed on 9 September 2026 after two DFF primary records. Both are DFF's own accounts, not independent court records or a recipient payment receipt.
\nDFF's [18 August 2026 grant announcement](http[local research file] names Liberty's challenge to the UK government forcing Apple to break encryption among grants approved in June 2026. It attributes application review and grant selection to an external Community Peer Group, which it says has made these decisions since 2025. This establishes a reported intermediate allocation decision: Oak's support for DFF does not become direct Oak authorship of Liberty's litigation. The announcement gives a round total of EUR 500,000 for 16 grants; that total must not be assigned to Liberty. Retained original: dff-grants2026.html; full substantive text read in dff-grants2026.derived.txt. Decisive official-reader locators: lines 9–18 and 37–41.
The announcement's actual Liberty link opens [DFF's encryption case study](http[local research file] Its principal organisation field names Liberty, with EUR 36,000, Litigation Track Support, and Ongoing status. DFF describes Liberty seeking a ruling that the Investigatory Powers Act 2016 Technical Capability Notice regime breaches rights, preventing similar future notices. Its stated strategic objective is to protect end-to-end encryption and resist government attempts to undermine it. This is a concrete challenge to a coercive power, rather than generic privacy language. It is Liberty's described legal action, funded by DFF; neither DFF nor Oak should be substituted as the claimant or author. Retained original: dff-liberty-tcn.html; full substantive text read in dff-liberty-tcn.derived.txt. Decisive reader locators: lines 3–9, 12–26, 30–46.
The source's current label is a reported ongoing case, not a verified court result. Its account of Apple's notices and proceedings explicitly uses reported or alleged descriptions. No notice, pleading, judgment or payment record was read in this lane. The grant amount is not a cash-transfer finding. The dated announcement places the linked project in a June 2026 approval round; the case page does not disclose a precise award date, term, instalments, reporting conditions or signed agreement.
\nThere is one source defect worth preserving: after its Liberty description and related links, the case page repeats an “Organisation Name” label followed by “Women’s Link Worldwide.” This appears in both the original HTML and the reader (lines 52–54). The main organisation field, repeated body and separate grant announcement all identify Liberty. The trailing inconsistent label is not evidence of an additional recipient or co-claimant, and its cause was not established. It remains an explicit metadata anomaly rather than a silently repaired field.
\nThe evidence supplies a specific counterweight to an inference that shared foundation funding entails uniform support for expanding state access: DFF reports supporting litigation aimed at disabling a UK power to compel interference with encryption. Oak's separately reviewed core support for DFF does not establish an earmark for this grant, a traced payment chain, or the purchase of opposition.
\nIt does not yet prove a direct disagreement with 5Rights, WeProtect or Check My Ads over the same instrument. The challenged power is an IPA Technical Capability Notice, not the Online Safety Act's age-assurance rules or CMA's monetisation proposal. Protection of encryption can coexist with other online-safety duties; the case does not establish a position against all age verification. To claim a direct policy clash, the next discriminating evidence would be the actual positions of both sides on the same power. No such additional branch was opened here.
\nLikewise, the reported Community Peer Group decision should be kept separate from a claim that it possesses final legal execution authority or that DFF/Oak have no reserved rights. Its charter, appointment rules and the Liberty grant agreement were not read. The source supports the stated selection role, not an invented complete governance settlement.
\nFor later integration, useful relationships are DFF → Liberty (reported EUR 36,000 litigation support), Community Peer Group → the June 2026 grant selection (DFF-attributed decision), and Liberty → the UK TCN regime (reported rights challenge). None is a payment-to-Apple edge or proof of a successful judgment. The exact Oak → DFF donor instrument remains with the root's donor lane.
\nThe two HTML originals were obtained with one successful ordinary GET each on 9 September 2026, after substantive official-reader access. No failure workaround or gated system was needed. dff-captures.json contains exact URLs, timestamps, hashes, query scope and limitations. Source texts were retained in full; the notes rely on the relevant passages identified above. The Community Peer Group governance page, linked Liberty article and other projects were discovered but not opened. This finite pass did not establish a direct DFF intervention against UK age-assurance duties; that is a limit of this selected reading, not proof none exists.
The most useful missing record is Liberty's actual TCN pleading or a published judgment, to verify the pleaded scope and adjudicated result; a specific grant agreement would separately test allocation and reporting conditions. No further acquisition, graph, forest, current-work, visual or package change was undertaken.
", "path": "oak-digital-rights-divergence-wave-2026-09-09/dff-divergence.md", "sourceIds": [ "w31-fa-s-dffround", "w31-fa-s-dffliberty" ], "markdown": "Bounded acquisition closed on 9 September 2026 after two DFF primary records. Both are DFF's own accounts, not independent court records or a recipient payment receipt.\n\nDFF's 18 August 2026 grant announcement (see source references) names Liberty's challenge to the UK government forcing Apple to break encryption among grants approved in June 2026. It attributes application review and grant selection to an external Community Peer Group, which it says has made these decisions since 2025. This establishes a reported intermediate allocation decision: Oak's support for DFF does not become direct Oak authorship of Liberty's litigation. The announcement gives a round total of EUR 500,000 for 16 grants; that total must not be assigned to Liberty. Retained original: `dff-grants2026.html`; full substantive text read in `dff-grants2026.derived.txt`. Decisive official-reader locators: lines 9–18 and 37–41.\n\nThe announcement's actual Liberty link opens DFF's encryption case study (see source references) Its principal organisation field names Liberty, with **EUR 36,000**, **Litigation Track Support**, and **Ongoing** status. DFF describes Liberty seeking a ruling that the Investigatory Powers Act 2016 Technical Capability Notice regime breaches rights, preventing similar future notices. Its stated strategic objective is to protect end-to-end encryption and resist government attempts to undermine it. This is a concrete challenge to a coercive power, rather than generic privacy language. It is Liberty's described legal action, funded by DFF; neither DFF nor Oak should be substituted as the claimant or author. Retained original: `dff-liberty-tcn.html`; full substantive text read in `dff-liberty-tcn.derived.txt`. Decisive reader locators: lines 3–9, 12–26, 30–46.\n\nThe source's current label is a reported ongoing case, not a verified court result. Its account of Apple's notices and proceedings explicitly uses reported or alleged descriptions. No notice, pleading, judgment or payment record was read in this lane. The grant amount is not a cash-transfer finding. The dated announcement places the linked project in a June 2026 approval round; the case page does not disclose a precise award date, term, instalments, reporting conditions or signed agreement.\n\nThere is one source defect worth preserving: after its Liberty description and related links, the case page repeats an “Organisation Name” label followed by “Women’s Link Worldwide.” This appears in both the original HTML and the reader (lines 52–54). The main organisation field, repeated body and separate grant announcement all identify Liberty. The trailing inconsistent label is not evidence of an additional recipient or co-claimant, and its cause was not established. It remains an explicit metadata anomaly rather than a silently repaired field.\n\n## What this changes in the network interpretation\n\nThe evidence supplies a specific counterweight to an inference that shared foundation funding entails uniform support for expanding state access: DFF reports supporting litigation aimed at disabling a UK power to compel interference with encryption. Oak's separately reviewed core support for DFF does not establish an earmark for this grant, a traced payment chain, or the purchase of opposition.\n\nIt does **not** yet prove a direct disagreement with 5Rights, WeProtect or Check My Ads over the same instrument. The challenged power is an IPA Technical Capability Notice, not the Online Safety Act's age-assurance rules or CMA's monetisation proposal. Protection of encryption can coexist with other online-safety duties; the case does not establish a position against all age verification. To claim a direct policy clash, the next discriminating evidence would be the actual positions of both sides on the same power. No such additional branch was opened here.\n\nLikewise, the reported Community Peer Group decision should be kept separate from a claim that it possesses final legal execution authority or that DFF/Oak have no reserved rights. Its charter, appointment rules and the Liberty grant agreement were not read. The source supports the stated selection role, not an invented complete governance settlement.\n\nFor later integration, useful relationships are DFF → Liberty (reported EUR 36,000 litigation support), Community Peer Group → the June 2026 grant selection (DFF-attributed decision), and Liberty → the UK TCN regime (reported rights challenge). None is a payment-to-Apple edge or proof of a successful judgment. The exact Oak → DFF donor instrument remains with the root's donor lane.\n\n## Custody and stopping point\n\nThe two HTML originals were obtained with one successful ordinary GET each on 9 September 2026, after substantive official-reader access. No failure workaround or gated system was needed. `dff-captures.json` contains exact URLs, timestamps, hashes, query scope and limitations. Source texts were retained in full; the notes rely on the relevant passages identified above. The Community Peer Group governance page, linked Liberty article and other projects were discovered but not opened. This finite pass did not establish a direct DFF intervention against UK age-assurance duties; that is a limit of this selected reading, not proof none exists.\n\nThe most useful missing record is Liberty's actual TCN pleading or a published judgment, to verify the pleaded scope and adjudicated result; a specific grant agreement would separately test allocation and reporting conditions. No further acquisition, graph, forest, current-work, visual or package change was undertaken.\n", "file": "research/oak-digital-rights-divergence-wave-2026-09-09-dff-divergence.md" }, { "id": "oak-digital-rights-divergence-wave-2026-09-09-statewatch-divergence", "title": "Oak support and Statewatch: a specific limit on the common-funder inference", "description": "9 September 2026. Root acquisition closed. This lane compares named donor grants with actions by their recipients; it does not claim a grant paid for a particular action. The Digital Freedom Fund lane is separate.", "html": "9 September 2026. Root acquisition closed. This lane compares named donor grants with actions by their recipients; it does not claim a grant paid for a particular action. The Digital Freedom Fund lane is separate.
\nOak's retained 2025 Annual Report, PDF/printed25, lists Statewatch USD303622 over three years for news, information and analysis on EU spending and justice/home-affairs practice. It lists Digital Freedom Fund USD631579 over three years in core support for the European digital-rights community. Root read PDF22–26 and visually checked the complete page25 columns. These are reported grant amounts and durations; exact contracting entities, start/end dates, payment timing and conditions are absent. The Statewatch entry does not say unrestricted core support. [Donor report](http[local research file] is reused from the preceding Oak packet, not a second confirmation.
\nThe same donor's current International Human Rights programme sets priorities that include reliable information, media pluralism, an open and accountable digital sphere, and protection for human-rights defenders. It explicitly allows intermediary organisations in some geographies. This provides a programme-level rationale that could accommodate watchdogs and opponents of excessive surveillance. It supplies no specific grant-approval chain or direction to Statewatch. Root read the complete substantive programme section in the official reader and retained original HTML. [Programme](http[local research file]
\nStatewatch's dated 13 September 2023 statement identifies itself as a signatory to an EDRi-coordinated call for EU governments to reject the proposed Child Sexual Abuse Regulation until it protects rights and security. The statement objects to blanket scanning of private communications and potential ID checks, on privacy, encryption and exclusion grounds. Those are the signatories' claims about the then proposal; this comparison does not independently resolve every legal or technical assertion or establish the current 2026 text. [Signed statement](http[local research file] Root read the substantive statement and signatory list, including Statewatch. It predates the 2025 grant: the later grant cannot explain the origin of this 2023 position.
\nFor the comparator, root refreshed existing case source src-12639f108d96 and read all four pages of the WeProtect Secretariat's original EC consultation submission. It strongly supports the legislative initiative, mandatory detection obligations and a new EU centre, alongside safeguards, constrained information extraction, oversight and classification standards. It also expresses interest in collaborating with the proposed centre. The undated document gives membership figures as of 12 September 2022; this is not an exact submission date. PDF4 explicitly separates Secretariat views from members' positions. [Submission](http[local research file]
\nThese are different positions on the same legislative initiative at different dates, with potentially different draft versions. Both invoke rights and safeguards; they disagree over whether the proposed detection regime merits support. Do not turn this into a vote on identical final clauses, a position held by every WeProtect member, or proof Oak financed opposition and support simultaneously. Current case relationship n25 already records the supporting submission; the fresh original improves its source custody. The imported source's access_state had remained discovered despite the relationship's historical read note; this new scoped reread supplies explicit current evidence for later integration.
\nStatewatch also signed an Open Rights Group-coordinated letter to Ofcom, Meta, Alphabet, X and ByteDance, dated 4 August 2025 on the signatory publisher's page. The retained five-page PDF identifies ORG on its letterhead and Statewatch on PDF5. It seeks clarification and safeguards against over-removing lawful political expression under the OSA following Palestine Action's proscription: clear guidance, transparency, impact assessment, appeal/restoration processes, conditional independent dispute resolution, and clarity on global versus UK removal. These are requested protections, not demonstrated remedies or acquired regulatory commitments. [Publisher's letter page](http[local research file] [attached PDF](http[local research file]
\nRoot read the whole PDF and substantive HTML, and visually checked PDF5. Its characterisations of legal powers, particular detentions and the regulator's bypass strategy remain attributed assertions here; no independent legal/current-proscription audit was performed. The PDF itself has no extracted date; the date comes from the publisher. The 2026/04 URL path is a hosting path, not the letter date. The PDF lists D3 Portugal, absent from the current page's signatory text, so do not claim exact signatory-list identity. Statewatch appears in both. No sent-mail evidence, recipient reply or implemented correction was acquired. The date coincides with the separately evidenced DSIT/Ofcom meeting; a shared date alone supplies no connection between that meeting and this letter.
\nA common Oak funding node connects organisations that have publicly advocated different boundaries on digital power. This supports a model with recipient agency, differing programme objectives and contest over implementation. It weakens a uniform-policy inference from shared funding alone. It neither erases selection/renewal leverage nor proves all recipients independent in every decision. Statewatch's UK letter is critical engagement seeking constrained enforcement and remedies, not evidence it rejects all online-safety regulation or every form of age assurance. The sharper questions are which decisions a funder can approve, which decisions intermediaries allocate elsewhere, and whose proposed controls are actually adopted.
\nFinite acquisition: two Statewatch intervention records and one already-cited WeProtect comparator. The Statewatch month-index reader returned a challenge; an indexed direct article supplied a valid source and original. PDF reader cache fetch failed once; ordinary publisher GET succeeded. No bypass, account, outreach or broad genealogy. Original HTML/PDF bytes, their derived text, rendered pages and exact hashes are preserved in root-captures.json. Root has not searched for a response to this letter in this run.
", "path": "oak-digital-rights-divergence-wave-2026-09-09/statewatch-divergence.md", "sourceIds": [ "src-12639f108d96", "w31-fa-s-swcsa", "w31-fa-s-swletterpage", "w31-fa-s-swletter", "w31-fa-s-oakihr" ], "markdown": "9 September 2026. Root acquisition closed. This lane compares named donor grants with actions by their recipients; it does not claim a grant paid for a particular action. The Digital Freedom Fund lane is separate.\n\nOak's retained 2025 Annual Report, PDF/printed25, lists Statewatch USD303622 over three years for news, information and analysis on EU spending and justice/home-affairs practice. It lists Digital Freedom Fund USD631579 over three years in core support for the European digital-rights community. Root read PDF22–26 and visually checked the complete page25 columns. These are reported grant amounts and durations; exact contracting entities, start/end dates, payment timing and conditions are absent. The Statewatch entry does not say unrestricted core support. Donor report (see source references) is reused from the preceding Oak packet, not a second confirmation.\n\nThe same donor's current International Human Rights programme sets priorities that include reliable information, media pluralism, an open and accountable digital sphere, and protection for human-rights defenders. It explicitly allows intermediary organisations in some geographies. This provides a programme-level rationale that could accommodate watchdogs and opponents of excessive surveillance. It supplies no specific grant-approval chain or direction to Statewatch. Root read the complete substantive programme section in the official reader and retained original HTML. Programme (see source references)\n\n## A documented opposing position on the EU proposal\n\nStatewatch's dated 13 September 2023 statement identifies itself as a signatory to an EDRi-coordinated call for EU governments to reject the proposed Child Sexual Abuse Regulation until it protects rights and security. The statement objects to blanket scanning of private communications and potential ID checks, on privacy, encryption and exclusion grounds. Those are the signatories' claims about the then proposal; this comparison does not independently resolve every legal or technical assertion or establish the current 2026 text. Signed statement (see source references) Root read the substantive statement and signatory list, including Statewatch. It predates the 2025 grant: the later grant cannot explain the origin of this 2023 position.\n\nFor the comparator, root refreshed existing case source src-12639f108d96 and read all four pages of the WeProtect Secretariat's original EC consultation submission. It strongly supports the legislative initiative, mandatory detection obligations and a new EU centre, alongside safeguards, constrained information extraction, oversight and classification standards. It also expresses interest in collaborating with the proposed centre. The undated document gives membership figures as of 12 September 2022; this is not an exact submission date. PDF4 explicitly separates Secretariat views from members' positions. Submission (see source references)\n\nThese are different positions on the same legislative initiative at different dates, with potentially different draft versions. Both invoke rights and safeguards; they disagree over whether the proposed detection regime merits support. Do not turn this into a vote on identical final clauses, a position held by every WeProtect member, or proof Oak financed opposition and support simultaneously. Current case relationship n25 already records the supporting submission; the fresh original improves its source custody. The imported source's access_state had remained discovered despite the relationship's historical read note; this new scoped reread supplies explicit current evidence for later integration.\n\n## A direct UK intervention seeking limits and remedies\n\nStatewatch also signed an Open Rights Group-coordinated letter to Ofcom, Meta, Alphabet, X and ByteDance, dated 4 August 2025 on the signatory publisher's page. The retained five-page PDF identifies ORG on its letterhead and Statewatch on PDF5. It seeks clarification and safeguards against over-removing lawful political expression under the OSA following Palestine Action's proscription: clear guidance, transparency, impact assessment, appeal/restoration processes, conditional independent dispute resolution, and clarity on global versus UK removal. These are requested protections, not demonstrated remedies or acquired regulatory commitments. Publisher's letter page (see source references) attached PDF (see source references)\n\nRoot read the whole PDF and substantive HTML, and visually checked PDF5. Its characterisations of legal powers, particular detentions and the regulator's bypass strategy remain attributed assertions here; no independent legal/current-proscription audit was performed. The PDF itself has no extracted date; the date comes from the publisher. The 2026/04 URL path is a hosting path, not the letter date. The PDF lists D3 Portugal, absent from the current page's signatory text, so do not claim exact signatory-list identity. Statewatch appears in both. No sent-mail evidence, recipient reply or implemented correction was acquired. The date coincides with the separately evidenced DSIT/Ofcom meeting; a shared date alone supplies no connection between that meeting and this letter.\n\n## What the comparison can establish\n\nA common Oak funding node connects organisations that have publicly advocated different boundaries on digital power. This supports a model with recipient agency, differing programme objectives and contest over implementation. It weakens a uniform-policy inference from shared funding alone. It neither erases selection/renewal leverage nor proves all recipients independent in every decision. Statewatch's UK letter is critical engagement seeking constrained enforcement and remedies, not evidence it rejects all online-safety regulation or every form of age assurance. The sharper questions are which decisions a funder can approve, which decisions intermediaries allocate elsewhere, and whose proposed controls are actually adopted.\n\nFinite acquisition: two Statewatch intervention records and one already-cited WeProtect comparator. The Statewatch month-index reader returned a challenge; an indexed direct article supplied a valid source and original. PDF reader cache fetch failed once; ordinary publisher GET succeeded. No bypass, account, outreach or broad genealogy. Original HTML/PDF bytes, their derived text, rendered pages and exact hashes are preserved in root-captures.json. Root has not searched for a response to this letter in this run.\n", "file": "research/oak-digital-rights-divergence-wave-2026-09-09-statewatch-divergence.md" }, { "id": "ofcom-crisis-authority-wave-2026-09-09-crisis-authority", "title": "Crisis authority: inputs, decisions and the record left behind", "description": "Bounded root acquisition closed 9 September 2026. June's decision is distinct from current commencement, addressed in the companion status lane. Principal source is Ofcom's 9 June 2026 crisis statement, as scoped official-reader derivatives, not original PDF bytes: httplocal research file . One ordinary GET returned403 and is closed. Page numbers below are o", "html": "Bounded root acquisition closed 9 September 2026. June's decision is distinct from current commencement, addressed in the companion status lane. Principal source is Ofcom's 9 June 2026 crisis statement, retained as scoped official-reader derivatives, not original PDF bytes: http[local research file] . One ordinary GET returned403 and is closed. Page numbers below are one-based PDF pages and also match the statement's printed pages in these passages. No images were inspected.
\nParagraphs3.18-3.23 (PDF10-12) place crisis determination with the provider using its indicators. Local, regional or overseas events can qualify if linked to a serious UK public-safety threat and the specified increase in relevant illegal content or content harmful to children. Examples of external indicators include law enforcement/intelligence agencies, NGOs without trusted-flagger status, civil contingency bodies, news, other services and trusted flaggers; internal complaints/moderation signals can also count. These are permitted examples, not automatic activation orders or proof a named NGO has been chosen. The provider assembles a sufficiently senior cross-functional team, chooses systems and acts as soon as reasonably practicable after determination (3.20,3.26).
\nThere is a specifically added government input. Paragraphs3.11 and3.23 say an Ofcom public statement notice issued under a Secretary of State section175 direction should be considered alongside the provider's indicators. Paragraphs3.46-47 expressly connect this amendment to two stakeholders' questions about the interaction. Footnote29, PDF17, identifies Full Fact's response p17 and an Ofcom/DSIT meeting on4August2025. This is a regulator-attributed link between consultation/departmental engagement and an actual drafting change. It does not disclose what either participant asked or prove they proposed or controlled the resulting wording. Their underlying records are the discriminating evidence still to obtain.
\nSection175 itself has different powers. The retained original enacted Online Safety Act, PDF169-170/printed155-156, was read for the whole section: http[local research file] . On reasonable grounds for a public-health/safety or national-security threat, the Secretary can direct Ofcom's media-literacy priorities for a specified period, or require Ofcom to issue public statement notices to named or general providers. A notice requires a public statement by a specified date; subsection5 additionally permits Ofcom to require information for responding to the threat. Public-health/safety reasons must be published under subsection6; the same subsection does not impose that publication duty on the national-security ground. Direction and consequent notice can be varied/revoked. This is enacted-text custody, not a fresh current-consolidation or commencement audit; Ofcom's June statement independently treats the power as relevant. It is not, on this text alone, a direct power to order removal of any specified lawful post. A compelled explanation/information demand can nevertheless create practical pressure; exercised effects require a worked notice and response.
\nThe dedicated law-enforcement channel applies to large in-scope services once they determine a crisis is occurring or likely (3.27-29, PDF13). Its stated purpose is fast crisis information into the service. A1.61-64, PDF34-35, expressly preserves existing legal processes for routine data requests and places retention/preservation arrangements under existing law. Ofcom disclaims a general power in these measures to compel disclosure to police. Smaller high-risk services can be in protocol scope without being in dedicated-channel scope.
\nA1.48-59, PDF33-34, records requests for wider formal NGO/fact-checker contact and real-time data access. Ofcom encourages relevant collaboration but declines to extend formal contact obligations to those bodies or become a central cross-service coordinator. Existing voluntary industry cooperation is not made a recommended component here. Its separate discussion of a possible researcher-access framework is not an enacted data-access entitlement created by this crisis measure. This distinguishes an NGO input from a police channel, a chosen partnership and a compulsory shared database.
\nParagraphs3.30-34, PDF13-14, recommend ending the crisis protocol when criteria no longer apply or at90days, whichever is sooner, then recording key decisions in a post-crisis analysis. Continuing conditions move to ordinary operations and resilience planning. The90-day limit is therefore not a promised restoration of removed content, accounts or former moderation rules. Further detailed logs, rationale, partner engagement and responsible teams are suggested analysis contents, not a prescribed universal public ledger.
\nA1.40-47, PDF31-32, declines automatic publication or submission of every analysis. Providers retain records; Ofcom can seek them for supervision/enforcement and ask about activation. The public may consequently see less of the operating evidence than the provider/regulator. That is a documented visibility asymmetry, not proof that any particular record is suppressed or that no disclosure mechanism exists elsewhere.
\nRights safeguards coexist with an explicit risk judgment. A1.127-138, PDF44-45, preserves existing appeals, expression, privacy and data protection. But3.65-67, PDF19-20, acknowledges that provider implementation may increase false positives affecting political, journalistic or religious expression. Ofcom assesses the higher risk as likely proportionate during the defined rare crises, while saying speed-over-accuracy is not what its measures recommend. The record neither guarantees error-free operation nor prescribes indiscriminate removal. The relevant test is a real activation's decision log and correction results.
\nThe framework is a structured escalation process with several input sources and separated decision rights. A named regulatory drafting amendment can now be traced to Full Fact and DSIT engagement at the attribution level. Ministerial compulsion to explain/provide information, operational platform choice and police communication are different levers; combining them into a single takedown order would erase the mechanism we need to investigate. Conversely, calling the whole process merely voluntary would conceal the statutory duties and compelled-notice route.
\nNo particular crisis activation, section175 direction, platform notice, decision log or repaired false positive was acquired. No actual Full Fact p17 or4August2025meeting record was acquired in this lane. Those are concrete, likely existing records, with the statement itself identifying two of them. Root source-reading scope covers the named paragraphs and immediate context, not all70pages or their cited studies. Current status of remaining CMA proposals and crisis commencement is delegated to the companion lane. No new unrelated branch, outreach or transport retry.
\nRoot subsequently read the full operative C11/C12 text in the original GOV.UK18June HTML and the current official reader, plus HCWS124. The draft was submitted12June and laid18June; its preamble retains blank signatory/date and bracketed commencement. Its future-perfect template language is not a completed issue/commencement record. HCWS124 independently establishes laying and the conditional40-day scrutiny/21-days-after-issue process, not a final effective date. See companion status note for the dated public-status search.
\nThe draft makes an important precision visible: C11.7 expressly requires the analysis when the provider determines the crisis has ended or90days after it began, if earlier, then improvements to the protocol. It does not itself repeat the statement's separate end-the-protocol-at90days sentence. The statement supplies that explanatory recommendation. Neither text promises that90days restores removed posts or pre-crisis ordinary procedures. C11.8 cross-refers to named existing moderation/reporting/appeal safeguards rather than creating a freestanding new crisis appeal procedure. C12.3 remains conditional on the provider's determination. This comparison uses the children's draft; no unacquired final illegal-content code is silently substituted.
", "path": "ofcom-crisis-authority-wave-2026-09-09/crisis-authority.md", "sourceIds": [ "w31-fa-s-crisis" ], "markdown": "Bounded root acquisition closed 9 September 2026. June's decision is distinct from current commencement, addressed in the companion status lane. Principal source is Ofcom's 9 June 2026 crisis statement, retained as scoped official-reader derivatives, not original PDF bytes: http[local research file] . One ordinary GET returned403 and is closed. Page numbers below are one-based PDF pages and also match the statement's printed pages in these passages. No images were inspected.\n\n## Who can initiate and who decides\n\nParagraphs3.18-3.23 (PDF10-12) place crisis determination with the provider using its indicators. Local, regional or overseas events can qualify if linked to a serious UK public-safety threat and the specified increase in relevant illegal content or content harmful to children. Examples of external indicators include law enforcement/intelligence agencies, NGOs without trusted-flagger status, civil contingency bodies, news, other services and trusted flaggers; internal complaints/moderation signals can also count. These are permitted examples, not automatic activation orders or proof a named NGO has been chosen. The provider assembles a sufficiently senior cross-functional team, chooses systems and acts as soon as reasonably practicable after determination (3.20,3.26).\n\nThere is a specifically added government input. Paragraphs3.11 and3.23 say an Ofcom public statement notice issued under a Secretary of State section175 direction should be considered alongside the provider's indicators. Paragraphs3.46-47 expressly connect this amendment to two stakeholders' questions about the interaction. Footnote29, PDF17, identifies Full Fact's response p17 and an Ofcom/DSIT meeting on4August2025. This is a regulator-attributed link between consultation/departmental engagement and an actual drafting change. It does not disclose what either participant asked or prove they proposed or controlled the resulting wording. Their underlying records are the discriminating evidence still to obtain.\n\nSection175 itself has different powers. The retained original enacted Online Safety Act, PDF169-170/printed155-156, was read for the whole section: http[local research file] . On reasonable grounds for a public-health/safety or national-security threat, the Secretary can direct Ofcom's media-literacy priorities for a specified period, or require Ofcom to issue public statement notices to named or general providers. A notice requires a public statement by a specified date; subsection5 additionally permits Ofcom to require information for responding to the threat. Public-health/safety reasons must be published under subsection6; the same subsection does not impose that publication duty on the national-security ground. Direction and consequent notice can be varied/revoked. This is enacted-text custody, not a fresh current-consolidation or commencement audit; Ofcom's June statement independently treats the power as relevant. It is not, on this text alone, a direct power to order removal of any specified lawful post. A compelled explanation/information demand can nevertheless create practical pressure; exercised effects require a worked notice and response.\n\n## Communication and the limits of the channel\n\nThe dedicated law-enforcement channel applies to large in-scope services once they determine a crisis is occurring or likely (3.27-29, PDF13). Its stated purpose is fast crisis information into the service. A1.61-64, PDF34-35, expressly preserves existing legal processes for routine data requests and places retention/preservation arrangements under existing law. Ofcom disclaims a general power in these measures to compel disclosure to police. Smaller high-risk services can be in protocol scope without being in dedicated-channel scope.\n\nA1.48-59, PDF33-34, records requests for wider formal NGO/fact-checker contact and real-time data access. Ofcom encourages relevant collaboration but declines to extend formal contact obligations to those bodies or become a central cross-service coordinator. Existing voluntary industry cooperation is not made a recommended component here. Its separate discussion of a possible researcher-access framework is not an enacted data-access entitlement created by this crisis measure. This distinguishes an NGO input from a police channel, a chosen partnership and a compulsory shared database.\n\n## Persistence, accountability and acknowledged error\n\nParagraphs3.30-34, PDF13-14, recommend ending the crisis protocol when criteria no longer apply or at90days, whichever is sooner, then recording key decisions in a post-crisis analysis. Continuing conditions move to ordinary operations and resilience planning. The90-day limit is therefore not a promised restoration of removed content, accounts or former moderation rules. Further detailed logs, rationale, partner engagement and responsible teams are suggested analysis contents, not a prescribed universal public ledger.\n\nA1.40-47, PDF31-32, declines automatic publication or submission of every analysis. Providers retain records; Ofcom can seek them for supervision/enforcement and ask about activation. The public may consequently see less of the operating evidence than the provider/regulator. That is a documented visibility asymmetry, not proof that any particular record is suppressed or that no disclosure mechanism exists elsewhere.\n\nRights safeguards coexist with an explicit risk judgment. A1.127-138, PDF44-45, preserves existing appeals, expression, privacy and data protection. But3.65-67, PDF19-20, acknowledges that provider implementation may increase false positives affecting political, journalistic or religious expression. Ofcom assesses the higher risk as likely proportionate during the defined rare crises, while saying speed-over-accuracy is not what its measures recommend. The record neither guarantees error-free operation nor prescribes indiscriminate removal. The relevant test is a real activation's decision log and correction results.\n\n## Meaning for the case and finite boundary\n\nThe framework is a structured escalation process with several input sources and separated decision rights. A named regulatory drafting amendment can now be traced to Full Fact and DSIT engagement at the attribution level. Ministerial compulsion to explain/provide information, operational platform choice and police communication are different levers; combining them into a single takedown order would erase the mechanism we need to investigate. Conversely, calling the whole process merely voluntary would conceal the statutory duties and compelled-notice route.\n\nNo particular crisis activation, section175 direction, platform notice, decision log or repaired false positive was acquired. No actual Full Fact p17 or4August2025meeting record was acquired in this lane. Those are concrete, likely existing records, with the statement itself identifying two of them. Root source-reading scope covers the named paragraphs and immediate context, not all70pages or their cited studies. Current status of remaining CMA proposals and crisis commencement is delegated to the companion lane. No new unrelated branch, outreach or transport retry.\n\n## Root comparison with the laid children's draft\n\nRoot subsequently read the full operative C11/C12 text in the original GOV.UK18June HTML and the current official reader, plus HCWS124. The draft was submitted12June and laid18June; its preamble retains blank signatory/date and bracketed commencement. Its future-perfect template language is not a completed issue/commencement record. HCWS124 independently establishes laying and the conditional40-day scrutiny/21-days-after-issue process, not a final effective date. See companion status note for the dated public-status search.\n\nThe draft makes an important precision visible: C11.7 expressly requires the analysis when the provider determines the crisis has ended or90days after it began, if earlier, then improvements to the protocol. It does not itself repeat the statement's separate end-the-protocol-at90days sentence. The statement supplies that explanatory recommendation. Neither text promises that90days restores removed posts or pre-crisis ordinary procedures. C11.8 cross-refers to named existing moderation/reporting/appeal safeguards rather than creating a freestanding new crisis appeal procedure. C12.3 remains conditional on the provider's determination. This comparison uses the children's draft; no unacquired final illegal-content code is silently substituted.\n", "file": "research/ofcom-crisis-authority-wave-2026-09-09-crisis-authority.md" }, { "id": "ofcom-crisis-authority-wave-2026-09-09-disposition-status", "title": "Status of remaining Additional Safety Measures", "description": "Bounded official-record check closed on 9 September 2026. This is a status packet, not a fresh interpretation of the crisis protocol's operational powers.", "html": "Bounded official-record check closed on 9 September 2026. This is a status packet, not a fresh interpretation of the crisis protocol's operational powers.
\nThe strongest current public evidence located still schedules the wider Additional Safety Measures statement for autumn 2026. No final disposition of CMA's creator/advertiser checks or pre-monetization review was located in the finite official search. That is a qualified acquisition result, not proof that no decision exists or that every proposal remains substantively undecided.
\n| Item | \nStatus established by inspected records | \nWhat is not established | \n
|---|---|---|
| Remaining Additional Safety Measures | \nOfcom's roadmap, updated 28 July, says it is considering responses and will publish its statement by autumn. Its compliance milestones, updated 24 August, still list autumn publication with the statement link to be confirmed. | \nA final decision accepting/rejecting CMA's identity checks or pre-monetization review; an implementation date for those proposals. | \n
| Crisis-response amendments | \nJune decision; draft submitted to the Secretary of State on 12 June and laid on 18 June. Parliamentary scrutiny and subsequent Ofcom issue precede the stated 21-day commencement period. | \nA completed issue notice and exact date these crisis amendments entered, or will enter, force. | \n
| Intimate-image hash matching | \nThe 9 September official enforcement announcement expressly sets 30 September 2026 as the new measures' start date. The current regulatory index lists issued but not-yet-in-force illegal-content code amendments. | \nThat this separate commencement also applies to crisis measures or the wider CMA proposals. | \n
The [Ofcom implementation roadmap](http[local research file] displayed updated 28 July 2026, places the wider Additional Safety Measures statement in autumn (reader lines200–202). The [compliance milestones](http[local research file] displayed updated 24 August 2026, repeats that schedule (line197) and warns that dates may change (line141). These are newer affirmative scheduling records than the June crisis statement's forecast or the consultation's stale aggregate pending label.
\nThe roadmap separately reports two consultations opened on 10 July: one on Category 1 duties including identity verification, and another on fraudulent advertising for Categories 1/2a (lines188–197). This does not establish adoption of CMA's monetizing-creator/advertiser proposal. This lane did not read those separate consultation instruments and does not collapse their identity provisions into CMA's request.
\nThe accepted CMA packet remains the source for the proposal itself: submission-findings.md. Its verification subjects include monetizing creators and advertisers; that can affect pseudonymity without constituting universal viewer ID. The June crisis response explicitly declining specific monetization practices remains a disposition of that part of CMA's request, not a blanket rejection of its other proposals.
\n[HCWS124, 18 June 2026](http[local research file] by the Minister for AI and Online Safety, states that draft crisis amendments are being laid that day and describes the 40-day parliamentary period followed by force 21 days after code issue (reader lines27–33). The statement also distinguishes existing duties and voluntary early implementation from the amended code's start. It does not certify that the scrutiny and issue steps have already finished.
\nThe [GOV.UK children's-code draft](http[local research file] published 18 June, records submission to the Secretary of State on 12 June. Its preamble 1.1–1.6 contains anticipated completed-process wording, but the document is expressly a draft, its commencement remains bracketed and its signature name, position and date are placeholders. Those features prevent using preamble 1.4 as proof that scrutiny concluded or 1.5 as a dated commencement notice. Original HTML retained as status-draft-childrens-amendments.html; locally extracted text is separately available.
\nThe accepted statement's cover date is 9 June; the current Ofcom roadmap and regulatory index call the crisis announcement/publication 18 June, while government records establish laying 18 June. These dates are retained as source-specific metadata and procedural events, not silently normalized into one date. This packet does not establish the reason for the publication-label difference.
\nThe [current regulatory-document index](http[local research file] updated 9 September, lists upcoming user-to-user and search Illegal Content Codes as issued 9 September and not yet in force, with amendments taking effect 21 days after issue (lines129–141). Opening the new user-to-user code link 44 through the reader returned Internal Error. A subsequent ordinary GET of the index returned 403 and could not supply the target URL for a direct-code substitute. The route was closed; the new PDF's contents were not inspected. Consequently, this packet does not infer from the index whether that PDF contains crisis amendments.
\nThe [9 September NCII enforcement record](http[local research file] lines129–156, independently names 30 September for the hash-matching measures. Its [same-day news account](http[local research file] lines131–138, also preserves an alternative route using equally effective systems/processes. This establishes a specific upcoming measure and enforcement programme, not that every platform must purchase one named tool. The news also announces an SWGfL information/expertise/evidence partnership and recommends StopNCII; no partnership terms or procurement were acquired in this status lane.
\nstatus-reader-scopes.txt retains selected successful official-reader passages. These are derivatives, not original HTML or PDF bytes. disposition-captures.json records exact scopes, queries, files, hashes and failures. The only substantive new original acquired here is the GOV.UK draft HTML. The ordinary parliamentary GET returned 403, but the successful official reader supplied the full substantive HCWS124 statement. Earlier failed submission and crisis-PDF transports were not retried.
\nThe next discriminating record is a completed, dated Ofcom issue notice/consolidated crisis code and a final statement specifically disposing of CMA's verification/review proposals. The bounded check did not obtain either. Do not calculate a crisis start date by simply adding 40 and 21 calendar days to 18 June, or treat absence from search results as proof of non-publication. No further branch was opened.
", "path": "ofcom-crisis-authority-wave-2026-09-09/disposition-status.md", "sourceIds": [ "w31-fa-s-roadmap", "w31-fa-s-statusindex", "w31-fa-s-nciinews", "w31-fa-s-nciistatus", "w31-fa-s-laying", "w31-fa-s-childdraft" ], "markdown": "Bounded official-record check closed on 9 September 2026. This is a status packet, not a fresh interpretation of the crisis protocol's operational powers.\n\nThe strongest current public evidence located still schedules the wider Additional Safety Measures statement for autumn 2026. No final disposition of CMA's creator/advertiser checks or pre-monetization review was located in the finite official search. That is a qualified acquisition result, not proof that no decision exists or that every proposal remains substantively undecided.\n\n| Item | Status established by inspected records | What is not established |\n| --- | --- | --- |\n| Remaining Additional Safety Measures | Ofcom's roadmap, updated 28 July, says it is considering responses and will publish its statement by autumn. Its compliance milestones, updated 24 August, still list autumn publication with the statement link to be confirmed. | A final decision accepting/rejecting CMA's identity checks or pre-monetization review; an implementation date for those proposals. |\n| Crisis-response amendments | June decision; draft submitted to the Secretary of State on 12 June and laid on 18 June. Parliamentary scrutiny and subsequent Ofcom issue precede the stated 21-day commencement period. | A completed issue notice and exact date these crisis amendments entered, or will enter, force. |\n| Intimate-image hash matching | The 9 September official enforcement announcement expressly sets 30 September 2026 as the new measures' start date. The current regulatory index lists issued but not-yet-in-force illegal-content code amendments. | That this separate commencement also applies to crisis measures or the wider CMA proposals. |\n\n## Remaining measures: a current schedule, not a completed decision\n\nThe Ofcom implementation roadmap (see source references) displayed updated **28 July 2026**, places the wider Additional Safety Measures statement in autumn (reader lines200–202). The compliance milestones (see source references) displayed updated **24 August 2026**, repeats that schedule (line197) and warns that dates may change (line141). These are newer affirmative scheduling records than the June crisis statement's forecast or the consultation's stale aggregate pending label.\n\nThe roadmap separately reports two consultations opened on 10 July: one on Category 1 duties including identity verification, and another on fraudulent advertising for Categories 1/2a (lines188–197). This does not establish adoption of CMA's monetizing-creator/advertiser proposal. This lane did not read those separate consultation instruments and does not collapse their identity provisions into CMA's request.\n\nThe accepted CMA packet remains the source for the proposal itself: [submission-findings.md](cma-ofcom-monetization-wave-2026-09-09-submission-findings.md). Its verification subjects include monetizing creators and advertisers; that can affect pseudonymity without constituting universal viewer ID. The June crisis response explicitly declining specific monetization practices remains a disposition of that part of CMA's request, not a blanket rejection of its other proposals.\n\n## Crisis: actual parliamentary step located, commencement still unverified\n\nHCWS124, 18 June 2026 (see source references) by the Minister for AI and Online Safety, states that draft crisis amendments are being laid that day and describes the 40-day parliamentary period followed by force 21 days after code issue (reader lines27–33). The statement also distinguishes existing duties and voluntary early implementation from the amended code's start. It does not certify that the scrutiny and issue steps have already finished.\n\nThe GOV.UK children's-code draft (see source references) published 18 June, records submission to the Secretary of State on 12 June. Its preamble 1.1–1.6 contains anticipated completed-process wording, but the document is expressly a draft, its commencement remains bracketed and its signature name, position and date are placeholders. Those features prevent using preamble 1.4 as proof that scrutiny concluded or 1.5 as a dated commencement notice. Original HTML retained as [status-draft-childrens-amendments.html](https://www.gov.uk/government/publications/online-safety-act-draft-amendments-to-the-protection-of-children-code-of-practice-for-user-to-user-services/draft-amendments-to-the-protection-of-children-code-of-practice-for-user-to-user-services); locally extracted text is separately available.\n\nThe accepted statement's cover date is 9 June; the current Ofcom roadmap and regulatory index call the crisis announcement/publication 18 June, while government records establish laying 18 June. These dates are retained as source-specific metadata and procedural events, not silently normalized into one date. This packet does not establish the reason for the publication-label difference.\n\n## What changed on 9 September\n\nThe current regulatory-document index (see source references) updated 9 September, lists upcoming user-to-user and search Illegal Content Codes as issued 9 September and not yet in force, with amendments taking effect 21 days after issue (lines129–141). Opening the new user-to-user code link 44 through the reader returned Internal Error. A subsequent ordinary GET of the index returned 403 and could not supply the target URL for a direct-code substitute. The route was closed; the new PDF's contents were not inspected. Consequently, this packet does not infer from the index whether that PDF contains crisis amendments.\n\nThe 9 September NCII enforcement record (see source references) lines129–156, independently names 30 September for the hash-matching measures. Its same-day news account (see source references) lines131–138, also preserves an alternative route using equally effective systems/processes. This establishes a specific upcoming measure and enforcement programme, not that every platform must purchase one named tool. The news also announces an SWGfL information/expertise/evidence partnership and recommends StopNCII; no partnership terms or procurement were acquired in this status lane.\n\n## Reading and closure boundary\n\nstatus-reader-scopes.txt retains selected successful official-reader passages. These are derivatives, not original HTML or PDF bytes. disposition-captures.json records exact scopes, queries, files, hashes and failures. The only substantive new original acquired here is the GOV.UK draft HTML. The ordinary parliamentary GET returned 403, but the successful official reader supplied the full substantive HCWS124 statement. Earlier failed submission and crisis-PDF transports were not retried.\n\nThe next discriminating record is a completed, dated Ofcom issue notice/consolidated crisis code and a final statement specifically disposing of CMA's verification/review proposals. The bounded check did not obtain either. Do not calculate a crisis start date by simply adding 40 and 21 calendar days to 18 June, or treat absence from search results as proof of non-publication. No further branch was opened.\n", "file": "research/ofcom-crisis-authority-wave-2026-09-09-disposition-status.md" }, { "id": "osdi-comparison-wave-2026-09-09-osdi-evaluation-output", "title": "OSDI evaluation output: bounded public-source pass", "description": "Observed 9 September 2026. Five search queries completed, followed by inspection of one specifically identified primary PDF. Acquisition stops here for the forest review.", "html": "Observed 9 September 2026. Five search queries completed, followed by inspection of one specifically identified primary PDF. Acquisition stops here for the forest review.
\nNo PA Consulting OSDI final assessment, inception report, detailed agreed evaluation scope, or original OSDI product-evaluation framework was located in this pass. This is a search-coverage result, not evidence that evaluation did not occur. The public record found does include a later, substantive technical evaluation that expressly attributes part of its approach to OSDI. That supplies a documented methodological reuse route while leaving the original OSDI assessment gap intact.
\nThe [official appointment account](http[local research file] dated 23 September 2021 and authored by project delivery lead Fiona Hallett, identifies PA Consulting as OSDI's external evaluator. Its sections on the evaluator's role and working relationship describe assurance of delivery processes, resource use, spending and quality; recommendations on governance; and communication of findings to senior civil servants. DCMS says it agreed responsibilities and priorities with PA in collaboration with the Faculty/PUBLIC supplier consortium. This is a public description of the agreed relationship, not the agreement or an evaluation result. No report attachment or detailed scope instrument was linked in the inspected article body. Reading scope: entire article body, web-reader rendition; no original HTML-byte capture by this worker.
\nThe previously retained public-gcloud-sandbox.pdf, final PDF page 13, is a separate supplier account of PUBLIC's product-evaluation sandbox. Its framework, dataset, pilot and final-report claims do not establish that PA conducted those product tests. Root inspected that original; this worker did not independently audit its full contents. Neither OSDI role is the separate CDDP commission.
[Evaluation of Safety Tech Challenge Fund Round 2](http[local research file] cover dated October 2024, is a PUBLIC-authored report hosted by government. PDF page 3 says DSIT commissioned this separate, small-scale evaluation in February 2024. PDF page 10 and PDF page 33 explicitly credit OSDI practices alongside other inputs. Page 33 sets out performance, robustness, scalability, data privacy/security and user feedback criteria, with supplier surveys, interviews and quarterly reporting as evidence sources. Footnote 17 links only to the OSDI blog homepage; it does not expose the original framework.
\nThis report contains actual assessments: PDF page 41 assigns C4FF TRL4, CameraForensics TRL5 and Vistalworks TRL5, subject to further independent testing. PDF pages 9–10 describe a late-starting, high-level evaluation and limited testing data. Page 38 labels performance figures self-reported, cautions against unsupported cross-company comparison and calls for independent verification. These are PUBLIC's bounded findings, not independently replicated scores or PA's OSDI conclusions. The visible reuse establishes an attributed methodological contribution, not adoption of an OSDI standard into binding regulation.
\nThe acquired original is evaluation-stcf2-original.pdf: 774,350 bytes, 54 PDF pages; SHA-256 1f44c2e0c818976571274eaa5b22c3b4ceafb85b7d4c4e42e83573d2e671f0aa. One ordinary public GET succeeded. Text reading was scoped to PDF pages 1–5, 9–10, 31–34, 38, 41, 47 and 53, plus targeted web-reader findings. PDF pages 33 and 41 were rendered and visually inspected as evaluation-stcf2-pdf33.png and evaluation-stcf2-pdf41.png; the images confirm table alignment and printed-page numbering. Printed numbers are one less than physical PDF page numbers: the key approach is printed 32 / PDF 33, and the TRL table is printed 40 / PDF 41. The whole report was not comprehensively read. Embedded annotation inspection confirmed the exact footnote-17 URL http[local research file].
Queries used: "Online Safety Data Initiative" "PA Consulting" evaluation report; "Online Safety Data Initiative" "evaluation framework"; site:gov.uk "Online Safety Data Initiative" evaluation results; "OSDI" "evaluation" "PA Consulting" report; and "Online Safety Data Initiative" "evaluation" "final" report. These returned the evaluator announcement, programme planning/longlist material, the retained supplier service definition, sector/data reviews and the STCF2 report. Search-result references to other work do not establish that work was read. Root owns the newer Data Landscape Review; no parallel reread or new branch was started here.
The next useful record family is precise: PA's OSDI inception/evaluation plan containing the jointly agreed responsibilities and priorities; its final or periodic assurance report to the responsible senior civil servants; and the departmental acceptance or management response. The expected government custodian is the department holding the former DCMS online-safety technology programme records, with PA as the evaluator counterpart. For the product strand, the original framework version, pilot-testing protocol/results and final recommendations described by PUBLIC would distinguish a commissioned prototype from demonstrated effectiveness. None was requested or contacted, and no exact undiscovered filename is invented.
\nThe breakpoint is therefore an identifiable assurance-output gap plus a directly attributed reuse example. Do not replace the scoped gap with “no evaluation,” merge the three programmes, equate supplier self-report with replicated performance, or infer that the later STCF2 criteria exactly reproduce OSDI's missing original instrument.
", "path": "osdi-comparison-wave-2026-09-09/osdi-evaluation-output.md", "sourceIds": [ "w33-con-s-pa", "w33-con-s-stcf" ], "markdown": "Observed 9 September 2026. Five search queries completed, followed by inspection of one specifically identified primary PDF. Acquisition stops here for the forest review.\n\nNo PA Consulting OSDI final assessment, inception report, detailed agreed evaluation scope, or original OSDI product-evaluation framework was located in this pass. This is a search-coverage result, not evidence that evaluation did not occur. The public record found does include a later, substantive technical evaluation that expressly attributes part of its approach to OSDI. That supplies a documented methodological reuse route while leaving the original OSDI assessment gap intact.\n\n## Programme assurance and product evaluation are distinct\n\nThe official appointment account (see source references) dated 23 September 2021 and authored by project delivery lead Fiona Hallett, identifies PA Consulting as OSDI's external evaluator. Its sections on the evaluator's role and working relationship describe assurance of delivery processes, resource use, spending and quality; recommendations on governance; and communication of findings to senior civil servants. DCMS says it agreed responsibilities and priorities with PA in collaboration with the Faculty/PUBLIC supplier consortium. This is a public description of the agreed relationship, not the agreement or an evaluation result. No report attachment or detailed scope instrument was linked in the inspected article body. Reading scope: entire article body, web-reader rendition; no original HTML-byte capture by this worker.\n\nThe previously retained `public-gcloud-sandbox.pdf`, final PDF page 13, is a separate supplier account of PUBLIC's product-evaluation sandbox. Its framework, dataset, pilot and final-report claims do not establish that PA conducted those product tests. Root inspected that original; this worker did not independently audit its full contents. Neither OSDI role is the separate CDDP commission.\n\n## A published reuse record, with evaluation limits\n\nEvaluation of Safety Tech Challenge Fund Round 2 (see source references) cover dated October 2024, is a PUBLIC-authored report hosted by government. PDF page 3 says DSIT commissioned this separate, small-scale evaluation in February 2024. PDF page 10 and PDF page 33 explicitly credit OSDI practices alongside other inputs. Page 33 sets out performance, robustness, scalability, data privacy/security and user feedback criteria, with supplier surveys, interviews and quarterly reporting as evidence sources. Footnote 17 links only to the OSDI blog homepage; it does not expose the original framework.\n\nThis report contains actual assessments: PDF page 41 assigns C4FF TRL4, CameraForensics TRL5 and Vistalworks TRL5, subject to further independent testing. PDF pages 9–10 describe a late-starting, high-level evaluation and limited testing data. Page 38 labels performance figures self-reported, cautions against unsupported cross-company comparison and calls for independent verification. These are PUBLIC's bounded findings, not independently replicated scores or PA's OSDI conclusions. The visible reuse establishes an attributed methodological contribution, not adoption of an OSDI standard into binding regulation.\n\nThe acquired original is `evaluation-stcf2-original.pdf`: 774,350 bytes, 54 PDF pages; SHA-256 `1f44c2e0c818976571274eaa5b22c3b4ceafb85b7d4c4e42e83573d2e671f0aa`. One ordinary public GET succeeded. Text reading was scoped to PDF pages 1–5, 9–10, 31–34, 38, 41, 47 and 53, plus targeted web-reader findings. PDF pages 33 and 41 were rendered and visually inspected as `evaluation-stcf2-pdf33.png` and `evaluation-stcf2-pdf41.png`; the images confirm table alignment and printed-page numbering. Printed numbers are one less than physical PDF page numbers: the key approach is printed 32 / PDF 33, and the TRL table is printed 40 / PDF 41. The whole report was not comprehensively read. Embedded annotation inspection confirmed the exact footnote-17 URL `http[local research file]`.\n\n## Search coverage and discriminating next record\n\nQueries used: `\"Online Safety Data Initiative\" \"PA Consulting\" evaluation report`; `\"Online Safety Data Initiative\" \"evaluation framework\"`; `site:gov.uk \"Online Safety Data Initiative\" evaluation results`; `\"OSDI\" \"evaluation\" \"PA Consulting\" report`; and `\"Online Safety Data Initiative\" \"evaluation\" \"final\" report`. These returned the evaluator announcement, programme planning/longlist material, the retained supplier service definition, sector/data reviews and the STCF2 report. Search-result references to other work do not establish that work was read. Root owns the newer Data Landscape Review; no parallel reread or new branch was started here.\n\nThe next useful record family is precise: PA's OSDI inception/evaluation plan containing the jointly agreed responsibilities and priorities; its final or periodic assurance report to the responsible senior civil servants; and the departmental acceptance or management response. The expected government custodian is the department holding the former DCMS online-safety technology programme records, with PA as the evaluator counterpart. For the product strand, the original framework version, pilot-testing protocol/results and final recommendations described by PUBLIC would distinguish a commissioned prototype from demonstrated effectiveness. None was requested or contacted, and no exact undiscovered filename is invented.\n\nThe breakpoint is therefore an identifiable assurance-output gap plus a directly attributed reuse example. Do not replace the scoped gap with “no evaluation,” merge the three programmes, equate supplier self-report with replicated performance, or infer that the later STCF2 criteria exactly reproduce OSDI's missing original instrument.\n", "file": "research/osdi-comparison-wave-2026-09-09-osdi-evaluation-output.md" }, { "id": "osdi-comparison-wave-2026-09-09-osdi-governance-market", "title": "The advisory route into a government-funded safety-technology programme", "description": "9 September 2026. Completed root lane of the OSDI comparison. These findings supplement the r21 graph; they are not yet graph-integrated. OSDI, CDDP and the Safety Tech Challenge Fund are distinct programmes. Advisory access, supplier delivery accounts and actual evaluated outputs retain different evidence states.", "html": "9 September 2026. Completed root lane of the OSDI comparison. These findings supplement the r21 graph; they are not yet graph-integrated. OSDI, CDDP and the Safety Tech Challenge Fund are distinct programmes. Advisory access, supplier delivery accounts and actual evaluated outputs retain different evidence states.
\nThe government's 2 July 2021 account identifies a ten-person Online Harms Expert Group convened and chaired by the Centre for Data Ethics and Innovation (CDEI) for OSDI. Its role included advising on project scope and direction, privacy-preserving and legally robust data sharing, and data-stewardship practice. CDEI says the group had already encouraged attention to users' online experience alongside the technology industry's needs. That is an official account of advice and input, not acquired meeting minutes or an approval instrument.
\nThe roster includes Poppy Wood, then Senior Advisor at Reset Tech, alongside ICO technology-policy head Ali Shah, Ofcom technology-policy lead Fred Langford, NSPCC policy head Andy Burrows, Open Data Institute programme lead Jack Hardinges, BBC product manager Jon Howard, Web Foundation adviser Azmina Dhrodia, Bristol researchers Asma Vranaki and Claudia Peersman, and Oxford researcher Jun Zhao. Roles are as listed in July2021, not claims about current employment. Poppy Wood's later Reset UK directorship/member-control records are separate dated relationships already in the case.
\nThis is a concrete government-programme advisory route for a named Reset representative. It does not grant Reset a procurement vote, veto over deployment, ownership of government data, or approval of any later content decision. Those would require their own evidence.
\nSource: [CDEI's expert-group account](http[local research file] Full article and roster read; original retained as cdei-governance.html, 53,671 bytes, SHA256 5e0d8421c0f20775de28d1db2ed837351e9fee559d2b38029a8017c6c66660ed.
\nThe [official OSDI description](http[local research file] describes a fifteen-month project to improve access to data for technologies identifying and removing illegal and harmful content. It identifies government, OSTIA, Faculty Science and PUBLIC as participants. Shared Outcomes Fund financing is named; the fund-wide GBP200m is not OSDI's own allocation.
\nThe [June2021 longlist](http[local research file] describes nearly seventy discovery interviews, workshops involving OSTIA members, and a Whitehall roundtable. Its candidate responses include data repositories, privacy-preserving access, common taxonomies, benchmarking/accreditation and procurement/marketplace support. These were twenty-three candidate solutions awaiting shortlisting, not twenty-three delivered products or enacted requirements. The separate [security workstream account](http[local research file] describes early security, privacy and legal design, including ICO guidance. It is an account of approach, not an independent compliance audit.
\nThe advisory group, discovery participants and government team therefore had a route to shape which industry problems became publicly funded projects. This locates influence in priority formation, before later procurement or deployment. The minutes, shortlisting matrix and specific feedback dispositions remain unacquired.
\nPUBLIC's G-Cloud14 service definition uses OSDI as a case study. It says it delivered a TRL4 evaluation sandbox to DCMS for pilot review/testing, focused on text models moderating suicide and self-harm. It describes a labelled dataset, technical and qualitative evaluation frameworks, three pilot firms, a DPIA, user guidance and scaling recommendations. Its final stage proposes progression to live operation; this is not proof of a live national service.
\nThis is first-party evidence of a claimed delivery in a commercial service description, not an inspected sandbox, replicated classifier test, named pilot roster or government acceptance certificate. Its reported two hundred-plus interviews and three roundtables describe a broader programme than the earlier seventy-interview discovery account; differing scope or timing is possible. Do not treat these figures as contradictory counts of the same measured event or sum them.
\n[Original service definition](http[local research file] final PDF page13. Retained public-gcloud-sandbox.pdf: 3,142,085 bytes; SHA256 db602fec0ede3076b488f78be0bcd5698a16cf3c0cfb4306d8eb3ad641be4174. Root read the final case study and targeted surrounding methodology; not every service offering was evaluated.
\nThe separate evaluation-output note distinguishes PA Consulting's programme-assurance role from PUBLIC's product-evaluation role. It also identifies a later PUBLIC-authored STCF2 report explicitly reusing OSDI practices, with self-reporting and testing limitations. No original PA final assessment or original OSDI framework was located in that bounded search.
\nA DSIT-commissioned, PUBLIC-produced [Data Landscape Review 2024–2025](http[local research file] published26March2026, gives a later account. It explicitly says PUBLIC produced earlier internal data-access reviews for DCMS. Its research comprises thirteen interviews between28January and5March2025, nine survey responses (six relevant to this workstream), and more than fifty-five sources. The appendix describes a targeted stakeholder tracker with selection reasons, tailored interview scripts and an affinity map; these underlying artifacts were not acquired. This small, purposive sector study is not a representative population survey or a direct causal evaluation of OSDI.
\nThe report's acknowledgements include AVPA, Ofcom, OSTIA, Resolver, Yoti and other suppliers. That establishes participation or research input at the stated combined level, not that each answered every question, endorsed all conclusions or belonged to the earlier CDEI group. Its contextual citations include a Paladin/PUBLIC/Perspective Economics market study and a Tony Blair Institute paper. Citation does not establish commissioning by either organization.
\nIts findings complicate any picture of a seamless apparatus. Providers report persistent data costs, commercial/IP barriers, platform reluctance and inconsistent classification practices. Market demand and customer awareness are described as especially pressing. Some interviewees report moving toward brand protection where demand is stronger. Government programmes are appreciated, yet respondents describe slower momentum and uncertainty about enforcement. The report treats regulation as a demand driver; that is attributed market analysis, not proof that any firm secretly authored legislation.
\nOne consequential reported development is sharing processed signals or classification outputs instead of raw datasets. That can preserve data custody while spreading judgments between services. The report gives Checkstep/Sightengine as one example and discusses other collaborative models. These are research-report claims requiring provider agreements, interface specifications and actual downstream decisions before promotion into an exercised cross-platform control claim. No WEF reference was found in the retained report's text search; no exclusive taxonomy lineage is inferred.
\nRetained report HTML: 336,941 bytes, SHA256 71fd0eb7b10593d398668bda52d09c45b66b414dada9640d3ca6274e2ae29e1f. Substantive reading: introduction/objectives and methodology; selected market, barrier and classification passages; intervention-impact section6.2; conclusion; appendix/acknowledgements and selected citations. A full HTML capture is not evidence every case study was independently checked. The link labelled as the full2022 Data Landscape Review resolves to an unrelated RDI organizational review; that page was not treated as the missing original.
\nWe can now connect a named NGO representative to a specified government advisory group, an industry-informed selection process, commercial prototype development and later reuse of evaluation practices. That is more explanatory than a generic membership map. The field also contains legal constraints, resource shortages, competing interests and explicit limitations on measured performance.
\nThe next discriminating records are ordinary project artifacts: group minutes and advice dispositions; the shortlisting/selection matrix; original evaluation instruments and government acceptance; and, for the signal-sharing branch, contractual permission and downstream action rules. PUBLIC's exact CDDP commission remains unresolved in the separate contract review. The OSDI consortium cannot fill that gap by association.
", "path": "osdi-comparison-wave-2026-09-09/osdi-governance-market.md", "sourceIds": [ "w22-sig-s-landscape-case4", "w33-con-s-cdei", "w33-con-s-osdiabout", "w33-con-s-longlist", "w33-con-s-sandbox" ], "markdown": "9 September 2026. Completed root lane of the OSDI comparison. These findings supplement the r21 graph; they are not yet graph-integrated. OSDI, CDDP and the Safety Tech Challenge Fund are distinct programmes. Advisory access, supplier delivery accounts and actual evaluated outputs retain different evidence states.\n\n## A direct Reset connection, with a defined remit\n\nThe government's 2 July 2021 account identifies a ten-person Online Harms Expert Group convened and chaired by the Centre for Data Ethics and Innovation (CDEI) for OSDI. Its role included advising on project scope and direction, privacy-preserving and legally robust data sharing, and data-stewardship practice. CDEI says the group had already encouraged attention to users' online experience alongside the technology industry's needs. That is an official account of advice and input, not acquired meeting minutes or an approval instrument.\n\nThe roster includes **Poppy Wood, then Senior Advisor at Reset Tech**, alongside ICO technology-policy head Ali Shah, Ofcom technology-policy lead Fred Langford, NSPCC policy head Andy Burrows, Open Data Institute programme lead Jack Hardinges, BBC product manager Jon Howard, Web Foundation adviser Azmina Dhrodia, Bristol researchers Asma Vranaki and Claudia Peersman, and Oxford researcher Jun Zhao. Roles are as listed in July2021, not claims about current employment. Poppy Wood's later Reset UK directorship/member-control records are separate dated relationships already in the case.\n\nThis is a concrete government-programme advisory route for a named Reset representative. It does not grant Reset a procurement vote, veto over deployment, ownership of government data, or approval of any later content decision. Those would require their own evidence.\n\nSource: CDEI's expert-group account (see source references) Full article and roster read; original retained as cdei-governance.html, 53,671 bytes, SHA256 5e0d8421c0f20775de28d1db2ed837351e9fee559d2b38029a8017c6c66660ed.\n\n## What the programme was building\n\nThe official OSDI description (see source references) describes a fifteen-month project to improve access to data for technologies identifying and removing illegal and harmful content. It identifies government, OSTIA, Faculty Science and PUBLIC as participants. Shared Outcomes Fund financing is named; the fund-wide GBP200m is not OSDI's own allocation.\n\nThe June2021 longlist (see source references) describes nearly seventy discovery interviews, workshops involving OSTIA members, and a Whitehall roundtable. Its candidate responses include data repositories, privacy-preserving access, common taxonomies, benchmarking/accreditation and procurement/marketplace support. These were twenty-three candidate solutions awaiting shortlisting, not twenty-three delivered products or enacted requirements. The separate security workstream account (see source references) describes early security, privacy and legal design, including ICO guidance. It is an account of approach, not an independent compliance audit.\n\nThe advisory group, discovery participants and government team therefore had a route to shape which industry problems became publicly funded projects. This locates influence in priority formation, before later procurement or deployment. The minutes, shortlisting matrix and specific feedback dispositions remain unacquired.\n\n## A delivered prototype is claimed, but production performance is not established\n\nPUBLIC's G-Cloud14 service definition uses OSDI as a case study. It says it delivered a TRL4 evaluation sandbox to DCMS for pilot review/testing, focused on text models moderating suicide and self-harm. It describes a labelled dataset, technical and qualitative evaluation frameworks, three pilot firms, a DPIA, user guidance and scaling recommendations. Its final stage proposes progression to live operation; this is not proof of a live national service.\n\nThis is first-party evidence of a claimed delivery in a commercial service description, not an inspected sandbox, replicated classifier test, named pilot roster or government acceptance certificate. Its reported two hundred-plus interviews and three roundtables describe a broader programme than the earlier seventy-interview discovery account; differing scope or timing is possible. Do not treat these figures as contradictory counts of the same measured event or sum them.\n\nOriginal service definition (see source references) final PDF page13. Retained public-gcloud-sandbox.pdf: 3,142,085 bytes; SHA256 db602fec0ede3076b488f78be0bcd5698a16cf3c0cfb4306d8eb3ad641be4174. Root read the final case study and targeted surrounding methodology; not every service offering was evaluated.\n\nThe [separate evaluation-output note](osdi-comparison-wave-2026-09-09-osdi-evaluation-output.md) distinguishes PA Consulting's programme-assurance role from PUBLIC's product-evaluation role. It also identifies a later PUBLIC-authored STCF2 report explicitly reusing OSDI practices, with self-reporting and testing limitations. No original PA final assessment or original OSDI framework was located in that bounded search.\n\n## The same market reports persistent friction\n\nA DSIT-commissioned, PUBLIC-produced Data Landscape Review 2024–2025 (see source references) published26March2026, gives a later account. It explicitly says PUBLIC produced earlier internal data-access reviews for DCMS. Its research comprises thirteen interviews between28January and5March2025, nine survey responses (six relevant to this workstream), and more than fifty-five sources. The appendix describes a targeted stakeholder tracker with selection reasons, tailored interview scripts and an affinity map; these underlying artifacts were not acquired. This small, purposive sector study is not a representative population survey or a direct causal evaluation of OSDI.\n\nThe report's acknowledgements include AVPA, Ofcom, OSTIA, Resolver, Yoti and other suppliers. That establishes participation or research input at the stated combined level, not that each answered every question, endorsed all conclusions or belonged to the earlier CDEI group. Its contextual citations include a Paladin/PUBLIC/Perspective Economics market study and a Tony Blair Institute paper. Citation does not establish commissioning by either organization.\n\nIts findings complicate any picture of a seamless apparatus. Providers report persistent data costs, commercial/IP barriers, platform reluctance and inconsistent classification practices. Market demand and customer awareness are described as especially pressing. Some interviewees report moving toward brand protection where demand is stronger. Government programmes are appreciated, yet respondents describe slower momentum and uncertainty about enforcement. The report treats regulation as a demand driver; that is attributed market analysis, not proof that any firm secretly authored legislation.\n\nOne consequential reported development is sharing processed signals or classification outputs instead of raw datasets. That can preserve data custody while spreading judgments between services. The report gives Checkstep/Sightengine as one example and discusses other collaborative models. These are research-report claims requiring provider agreements, interface specifications and actual downstream decisions before promotion into an exercised cross-platform control claim. No WEF reference was found in the retained report's text search; no exclusive taxonomy lineage is inferred.\n\nRetained report HTML: 336,941 bytes, SHA256 71fd0eb7b10593d398668bda52d09c45b66b414dada9640d3ca6274e2ae29e1f. Substantive reading: introduction/objectives and methodology; selected market, barrier and classification passages; intervention-impact section6.2; conclusion; appendix/acknowledgements and selected citations. A full HTML capture is not evidence every case study was independently checked. The link labelled as the full2022 Data Landscape Review resolves to an unrelated RDI organizational review; that page was not treated as the missing original.\n\n## What this changes\n\nWe can now connect a named NGO representative to a specified government advisory group, an industry-informed selection process, commercial prototype development and later reuse of evaluation practices. That is more explanatory than a generic membership map. The field also contains legal constraints, resource shortages, competing interests and explicit limitations on measured performance.\n\nThe next discriminating records are ordinary project artifacts: group minutes and advice dispositions; the shortlisting/selection matrix; original evaluation instruments and government acceptance; and, for the signal-sharing branch, contractual permission and downstream action rules. PUBLIC's exact CDDP commission remains unresolved in the [separate contract review](osdi-comparison-wave-2026-09-09-public-cddp.md). The OSDI consortium cannot fill that gap by association.\n", "file": "research/osdi-comparison-wave-2026-09-09-osdi-governance-market.md" }, { "id": "osdi-comparison-wave-2026-09-09-public-cddp", "title": "PUBLIC's CDDP role: an attributed delivery claim, with the subcontract identity withheld", "description": "Bounded review, 9 September 2026. PUBLIC/CDDP only; OSDI governance and its separate consortium remain in the root lane. No new procurement award or PUBLIC-specific CDDP payment was established. Acquisition stops here for the forest review.", "html": "Bounded review, 9 September 2026. PUBLIC/CDDP only; OSDI governance and its separate consortium remain in the root lane. No new procurement award or PUBLIC-specific CDDP payment was established. Acquisition stops here for the forest review.
\nPUBLIC itself says it helped deliver three Shared Outcomes Fund projects, including the Counter-Disinformation Data Platform. That supports an attributed claim of participation. It does not specify what PUBLIC delivered for CDDP, whether it held a direct commission or a subcontract, the relevant legal entity, a fee, a work period or an accepted deliverable. The original CDDP procurement link now redirects to the marketplace homepage. The retained official contract, however, exposes a precise missing record: the buyer's approved key-subcontractor identities in the project-management call-off are redacted. That is a credible location where PUBLIC's role might be resolved, without identifying PUBLIC as an obscured entry.
\nThe current PUBLIC website identifies its controller as PUBLIC Group International Ltd; the Companies House register supplies the exact legal name PUBLIC GROUP INTERNATIONAL LIMITED, number 10608507. This establishes a current brand/entity association. It does not prove that this company was the party performing the historical CDDP work.
\n| Record and exact scope | \nSupported result | \nLimit | \n
|---|---|---|
| [PUBLIC, Better collaboration, bigger impact](http[local research file] substantive article body from its first digital-transformation paragraph through its proposed roundtable; author Mahlet Yared | \nPUBLIC describes participation in three Shared Outcomes Fund projects over the preceding 24 months, working with more than eight government departments in aggregate. Its CDDP paragraph identifies DCMS as lead and says delivery involved seven departments. The stated purpose is improved cross-government sharing and analysis of disinformation data. | \nNo article publication date appears in the read page or retained date metadata. The current footer's 2025 copyright date and the relative 24-month period cannot supply a historical start/end date. The proposed roundtable is an announcement, without observed delivery or attendance. The aggregate departmental count is not PUBLIC's CDDP-specific customer count. | \n
| [PUBLIC privacy policy](http[local research file] heading “As of January 2026” and “Identity and contact details of the data controller” | \nPUBLIC Group International Ltd, at 1 Horse Guards Avenue, is the named controller for the site. | \nA website-controller notice is not a CDDP contract or historical commissioning record. | \n
| [Companies House company 10608507](http[local research file] overview identity, status, type and incorporation fields | \nPUBLIC GROUP INTERNATIONAL LIMITED, active private limited company, incorporated 8 February 2017; registered office at PUBLIC Hall, 1 Horse Guards Avenue. | \nNo CDDP activity, payment or subcontract appears in these overview fields. No inference from directors, ownership or other clients was used. | \n
| [Full Fact Report 2022](http[local research file] Chapter 3 paragraph listing CDDP and footnotes 36–38, especially footnote 38 | \nFull Fact identifies an original DCMS procurement titled Counter Disinformation Data Platform Digital / Technical Project Management Team, dated 26 August 2021, with opportunity 15487. It distinguishes CDDP from the OSDI entry and its consortium in the preceding paragraph. | \nThis is primary evidence of Full Fact's citation, secondary evidence about an original notice that is no longer readable at that route. It does not cite PUBLIC's delivery article or establish an article date. | \n
| [Disclosed CDDP bundle](http[local research file] PDF 114–120 / order-form printed 2–8, call-off 102548 | \nDCMS is buyer and Faculty Science Limited, 08873131, is supplier for the Digital/Technical Project Management Team. PDF 115 dates the order form10 January 2022; PDF 116 states call-off start24 November 2021, expiry31 March 2023. The service encompasses discovery, development and delivery of a prototype. The read schedule includes project/product management, business analysis, service design, data science, legal/security and disinformation advisory roles, plus reports and sprint governance. | \nScope breadth makes a specialist contribution possible; it does not assign any role to PUBLIC. The initial phase dates and funding-gated later phase must not be flattened into proof that every listed service was performed. Execution, acceptance and PUBLIC-specific work were not established in this lane. | \n
| Same bundle, PDF 120 / printed 8, “Key Subcontractor(s)” | \nThe form says entry into it provides buyer written approval for the listed key subcontractors. The identifying entries immediately below are blacked out. This was visually checked in the original, rather than inferred from blank extracted text. | \nDo not identify an obscured entry as PUBLIC, count the black bars as a proven count of entities, or call them empty fields. | \n
| Same bundle, PDF 180–181 and186, distinct call-off 102887 | \nThis is the CDDP Technical Build Team G-Cloud 12 order, effective11 April 2022, with Faculty Science Ltd as supplier. Its supplier-information field states the subcontractors/partners list is “Not Applicable”. | \nThis is contract-specific. It neither resolves the redacted 102548 subcontractor list nor excludes a separate commission, earlier role or later variation involving PUBLIC. | \n
The official [FOI release's document inventory](http[local research file] lists the disclosed Faculty, NewsWhip and Oliver Wyman instruments, while withholding the Discovery Report and Product Operating Model. It is not evidence that PUBLIC never worked on the project. The release page itself has inconsistent displayed chronology: published 28 June 2024, request 25 August 2023, response 25 September 2024. Those labels are preserved rather than silently reconciled. This lane relies on the retained contract text for its dates, not a reconstructed release date.
\nA PUBLIC contribution as one of the project-management supplier's specialists or approved subcontractors is compatible with both the supplier's claim and the redacted subcontract field. A separate consultancy commission or contribution at another phase is also compatible. A broad marketing description that compresses a narrower contribution remains possible. No obtained record discriminates between these explanations.
\nThe OSDI Faculty/PUBLIC consortium is evidence about OSDI. It cannot supply the missing CDDP contractual edge. Similarly, the 2022 G-Cloud build order's subcontract field cannot override the distinct project-management order's redacted list. The combined record supports a PUBLIC claim of CDDP involvement alongside an independently documented Faculty commission; it does not yet support a named PUBLIC subcontract, direct DCMS award, CDDP fee, software ownership or government acceptance of PUBLIC work.
\n| From → to | \nType / stage | \nScope and confidence | \n
|---|---|---|
| PUBLIC brand → CDDP | \nreports participation in delivery / supplier self-report | \nDirectly read claim; historical period undated and exact contribution unresolved. | \n
| PUBLIC website → PUBLIC GROUP INTERNATIONAL LIMITED 10608507 | \nidentifies current website controller / website-controller disclosure plus register identity | \nStrong current association; no backdated CDDP contract identity. | \n
| DCMS → Faculty Science Limited 08873131 | \ncommissions through disclosed call-off 102548 / contractual instrument | \nDigital/Technical Project Management Team; order-form date10 January 2022 and stated start24 November 2021. Separate from performance or acceptance. | \n
| DCMS → undisclosed key subcontractor identities | \nprovides approval through102548 order form / redacted identification | \nMechanism observed at PDF 120, identities unreadable; keep this as a record-gap relation, not a PUBLIC edge. | \n
| DCMS → Faculty Science Ltd | \ncommissions through separate call-off 102887 / contractual instrument | \nTechnical Build Team effective11 April 2022; no listed subcontractor/partner in the inspected form. | \n
The supplier article's original HTML is retained as public-claim.html (58,137 bytes; SHA256 1edec8751a3944880d52ffc927caad6e160d73f706313906be556dbe15fb4e36) with a derived plain-text copy. Status: read. Read scope is the substantive article, author and relevant footer links, not the unrelated recommended articles.
The original opportunity 15487 reader returned a non-retryable safe-opening error. One ordinary GET substitute returned HTTP 200 after redirecting to http[local research file]. Retained public-procurement-15487.html is therefore a generic homepage capture, not a successful procurement capture (22,990 bytes; SHA256 adb1e4071e457420c1e5abb82c830e0ebaea4890a88bc8779303f581b215caa1). Requested-record status: inaccessible by this route; returned page inspected and excluded as substantive procurement evidence. No retry or alternate host was attempted after that substitute.
The Companies House overview, PUBLIC privacy identity section, Full Fact Chapter 3/footnotes and official FOI inventory were read through the public web reader. Their original response bytes were not retained in this lane. No hashes are claimed for those web-reader-only scopes.
\nThe existing 262-page CDDP bundle was reused, not downloaded again: ../supplier-platform-wave-2026-09-09/cddp-foi-bundle.pdf, SHA256 71bd6a522a765d53f9aa79b239134226be64891ea3c67711abc9d788fb87748f. New directly read contract scope: PDF 114–122,180–181,186, with PDF 120 visually inspected. The extract is retained as public-contract-extract.txt; the PDF 120 render as public-page-120.png. Targeted name searches across extracted text found no PUBLIC legal-name variant. This search is a discovery aid, not a full reading of every page and not evidence of absence; redactions and possible names outside the searched variants matter. The unrelated retained public-gcloud-sandbox.pdf was checked only by first page and targeted CDDP/name search; it was not used to support a CDDP relation.
The discriminating records are the unredacted key-subcontractor entries or buyer approval record for call-off 102548, and the relevant signed PUBLIC statement of work/subcontract, role allocation or accepted deliverable. Likely custodians are DSIT as successor department, Faculty as prime contractor, and the PUBLIC entity that performed the work. These records were not requested from anyone. A genuinely accessible official award/commission record naming PUBLIC or a public delivery artifact explicitly allocating CDDP work to it would also reopen the question. Further general PUBLIC profiles or the separate OSDI consortium would not resolve it.
", "path": "osdi-comparison-wave-2026-09-09/public-cddp.md", "sourceIds": [ "w33-con-s-publicclaim", "w33-con-s-publicprivacy", "w33-con-s-publiccompany" ], "markdown": "Bounded review, 9 September 2026. PUBLIC/CDDP only; OSDI governance and its separate consortium remain in the root lane. No new procurement award or PUBLIC-specific CDDP payment was established. Acquisition stops here for the forest review.\n\n## Finding\n\nPUBLIC itself says it helped deliver three Shared Outcomes Fund projects, including the Counter-Disinformation Data Platform. That supports an attributed claim of participation. It does not specify what PUBLIC delivered for CDDP, whether it held a direct commission or a subcontract, the relevant legal entity, a fee, a work period or an accepted deliverable. The original CDDP procurement link now redirects to the marketplace homepage. The retained official contract, however, exposes a precise missing record: the buyer's approved key-subcontractor identities in the project-management call-off are redacted. That is a credible location where PUBLIC's role might be resolved, without identifying PUBLIC as an obscured entry.\n\nThe current PUBLIC website identifies its controller as PUBLIC Group International Ltd; the Companies House register supplies the exact legal name PUBLIC GROUP INTERNATIONAL LIMITED, number 10608507. This establishes a current brand/entity association. It does not prove that this company was the party performing the historical CDDP work.\n\n## What the records actually establish\n\n| Record and exact scope | Supported result | Limit |\n| --- | --- | --- |\n| PUBLIC, Better collaboration, bigger impact (see source references) substantive article body from its first digital-transformation paragraph through its proposed roundtable; author Mahlet Yared | PUBLIC describes participation in three Shared Outcomes Fund projects over the preceding 24 months, working with more than eight government departments in aggregate. Its CDDP paragraph identifies DCMS as lead and says delivery involved seven departments. The stated purpose is improved cross-government sharing and analysis of disinformation data. | No article publication date appears in the read page or retained date metadata. The current footer's 2025 copyright date and the relative 24-month period cannot supply a historical start/end date. The proposed roundtable is an announcement, without observed delivery or attendance. The aggregate departmental count is not PUBLIC's CDDP-specific customer count. |\n| PUBLIC privacy policy (see source references) heading “As of January 2026” and “Identity and contact details of the data controller” | PUBLIC Group International Ltd, at 1 Horse Guards Avenue, is the named controller for the site. | A website-controller notice is not a CDDP contract or historical commissioning record. |\n| Companies House company 10608507 (see source references) overview identity, status, type and incorporation fields | PUBLIC GROUP INTERNATIONAL LIMITED, active private limited company, incorporated 8 February 2017; registered office at PUBLIC Hall, 1 Horse Guards Avenue. | No CDDP activity, payment or subcontract appears in these overview fields. No inference from directors, ownership or other clients was used. |\n| Full Fact Report 2022 (see source references) Chapter 3 paragraph listing CDDP and footnotes 36–38, especially footnote 38 | Full Fact identifies an original DCMS procurement titled Counter Disinformation Data Platform Digital / Technical Project Management Team, dated 26 August 2021, with opportunity 15487. It distinguishes CDDP from the OSDI entry and its consortium in the preceding paragraph. | This is primary evidence of Full Fact's citation, secondary evidence about an original notice that is no longer readable at that route. It does not cite PUBLIC's delivery article or establish an article date. |\n| Disclosed CDDP bundle (see source references) PDF 114–120 / order-form printed 2–8, call-off 102548 | DCMS is buyer and Faculty Science Limited, 08873131, is supplier for the Digital/Technical Project Management Team. PDF 115 dates the order form10 January 2022; PDF 116 states call-off start24 November 2021, expiry31 March 2023. The service encompasses discovery, development and delivery of a prototype. The read schedule includes project/product management, business analysis, service design, data science, legal/security and disinformation advisory roles, plus reports and sprint governance. | Scope breadth makes a specialist contribution possible; it does not assign any role to PUBLIC. The initial phase dates and funding-gated later phase must not be flattened into proof that every listed service was performed. Execution, acceptance and PUBLIC-specific work were not established in this lane. |\n| Same bundle, PDF 120 / printed 8, “Key Subcontractor(s)” | The form says entry into it provides buyer written approval for the listed key subcontractors. The identifying entries immediately below are blacked out. This was visually checked in the original, rather than inferred from blank extracted text. | Do not identify an obscured entry as PUBLIC, count the black bars as a proven count of entities, or call them empty fields. |\n| Same bundle, PDF 180–181 and186, distinct call-off 102887 | This is the CDDP Technical Build Team G-Cloud 12 order, effective11 April 2022, with Faculty Science Ltd as supplier. Its supplier-information field states the subcontractors/partners list is “Not Applicable”. | This is contract-specific. It neither resolves the redacted 102548 subcontractor list nor excludes a separate commission, earlier role or later variation involving PUBLIC. |\n\nThe official FOI release's document inventory (see source references) lists the disclosed Faculty, NewsWhip and Oliver Wyman instruments, while withholding the Discovery Report and Product Operating Model. It is not evidence that PUBLIC never worked on the project. The release page itself has inconsistent displayed chronology: published 28 June 2024, request 25 August 2023, response 25 September 2024. Those labels are preserved rather than silently reconciled. This lane relies on the retained contract text for its dates, not a reconstructed release date.\n\n## Rival explanations and what must not be merged\n\nA PUBLIC contribution as one of the project-management supplier's specialists or approved subcontractors is compatible with both the supplier's claim and the redacted subcontract field. A separate consultancy commission or contribution at another phase is also compatible. A broad marketing description that compresses a narrower contribution remains possible. No obtained record discriminates between these explanations.\n\nThe OSDI Faculty/PUBLIC consortium is evidence about OSDI. It cannot supply the missing CDDP contractual edge. Similarly, the 2022 G-Cloud build order's subcontract field cannot override the distinct project-management order's redacted list. The combined record supports a PUBLIC claim of CDDP involvement alongside an independently documented Faculty commission; it does not yet support a named PUBLIC subcontract, direct DCMS award, CDDP fee, software ownership or government acceptance of PUBLIC work.\n\n## Candidate directional relations\n\n| From → to | Type / stage | Scope and confidence |\n| --- | --- | --- |\n| PUBLIC brand → CDDP | reports participation in delivery / supplier self-report | Directly read claim; historical period undated and exact contribution unresolved. |\n| PUBLIC website → PUBLIC GROUP INTERNATIONAL LIMITED 10608507 | identifies current website controller / website-controller disclosure plus register identity | Strong current association; no backdated CDDP contract identity. |\n| DCMS → Faculty Science Limited 08873131 | commissions through disclosed call-off 102548 / contractual instrument | Digital/Technical Project Management Team; order-form date10 January 2022 and stated start24 November 2021. Separate from performance or acceptance. |\n| DCMS → undisclosed key subcontractor identities | provides approval through102548 order form / redacted identification | Mechanism observed at PDF 120, identities unreadable; keep this as a record-gap relation, not a PUBLIC edge. |\n| DCMS → Faculty Science Ltd | commissions through separate call-off 102887 / contractual instrument | Technical Build Team effective11 April 2022; no listed subcontractor/partner in the inspected form. |\n\n## Custody, access and reading extent\n\nThe supplier article's original HTML is retained as `public-claim.html` (58,137 bytes; SHA256 `1edec8751a3944880d52ffc927caad6e160d73f706313906be556dbe15fb4e36`) with a derived plain-text copy. Status: read. Read scope is the substantive article, author and relevant footer links, not the unrelated recommended articles.\n\nThe original opportunity 15487 reader returned a non-retryable safe-opening error. One ordinary GET substitute returned HTTP 200 after redirecting to `http[local research file]`. Retained `public-procurement-15487.html` is therefore a generic homepage capture, not a successful procurement capture (22,990 bytes; SHA256 `adb1e4071e457420c1e5abb82c830e0ebaea4890a88bc8779303f581b215caa1`). Requested-record status: inaccessible by this route; returned page inspected and excluded as substantive procurement evidence. No retry or alternate host was attempted after that substitute.\n\nThe Companies House overview, PUBLIC privacy identity section, Full Fact Chapter 3/footnotes and official FOI inventory were read through the public web reader. Their original response bytes were not retained in this lane. No hashes are claimed for those web-reader-only scopes.\n\nThe existing 262-page CDDP bundle was reused, not downloaded again: `../supplier-platform-wave-2026-09-09/cddp-foi-bundle.pdf`, SHA256 `71bd6a522a765d53f9aa79b239134226be64891ea3c67711abc9d788fb87748f`. New directly read contract scope: PDF 114–122,180–181,186, with PDF 120 visually inspected. The extract is retained as `public-contract-extract.txt`; the PDF 120 render as `public-page-120.png`. Targeted name searches across extracted text found no PUBLIC legal-name variant. This search is a discovery aid, not a full reading of every page and not evidence of absence; redactions and possible names outside the searched variants matter. The unrelated retained `public-gcloud-sandbox.pdf` was checked only by first page and targeted CDDP/name search; it was not used to support a CDDP relation.\n\n## Precise reopening condition\n\nThe discriminating records are the unredacted key-subcontractor entries or buyer approval record for call-off 102548, and the relevant signed PUBLIC statement of work/subcontract, role allocation or accepted deliverable. Likely custodians are DSIT as successor department, Faculty as prime contractor, and the PUBLIC entity that performed the work. These records were not requested from anyone. A genuinely accessible official award/commission record naming PUBLIC or a public delivery artifact explicitly allocating CDDP work to it would also reopen the question. Further general PUBLIC profiles or the separate OSDI consortium would not resolve it.\n", "file": "research/osdi-comparison-wave-2026-09-09-public-cddp.md" }, { "id": "payment-termination-wave-2026-09-09-acquirer-implementation", "title": "Stripe implementation — MATCH screening and processing access", "description": "9 September 2026. The public implementation counterpart establishes a substantial onward effect: Stripe generally declines MATCH-listed merchants while reserving consideration for documented exceptional circumstances. This is evidence of one provider's published practice. It does not establish a Mastercard rule that every provider must reject every listed bu", "html": "9 September 2026. The public implementation counterpart establishes a substantial onward effect: Stripe generally declines MATCH-listed merchants while reserving consideration for documented exceptional circumstances. This is evidence of one provider's published practice. It does not establish a Mastercard rule that every provider must reject every listed business, or loss of access to all banking.
\n[Stripe, Terminated merchant files](http[local research file] also labelled High risk merchant lists in its navigation. Original HTML retained as acquirer-stripe-match.html, SHA256 b867b5801678b18afc54e0a9b0ea684ca788ac70132239ff19cb97f6979773c7; HTTP 200, 825,132 bytes, captured 2026-09-09 at 17:12:10 UTC. No revision or effective date was displayed in the substantive text. This is an observed current-page account, not a dated historic policy. The locale is English (United Kingdom), but the captured country selector reads United States; it does not identify a UK acquiring-bank contract.
The complete substantive original text was read, including the separate Visa section to identify the boundary. The operative MATCH scope is acquirer-stripe-match.html.derived.txt lines 88–170; the Visa section is not adopted into the MATCH findings. The primary web reader independently exposed the same text, lines 10–80 in reference turn1482view0. The original HTML is the retained source, and the parsed text is a derivative, not a second publication.
| Locator in the retained text | \nScoped finding | \n
|---|---|
| Opening, lines 88–89; qualification, 92–93 | \nStripe describes preapproval database screening and qualifying-termination reporting as acquirer duties. Its broader assertion that most processors reject listings is an attributed industry claim, not an independently established frequency. | \n
| Information added, 149–162 | \nThe described record includes business and principal-owner identifiers, dates and reason code. Stripe says it does not disclose database information to users. This run obtained no actual merchant record. | \n
| Entry removal, 163–168 | \nStripe describes removal for erroneous entry or cured PCI noncompliance, through the listing acquirer; it directs merchants unable to identify that acquirer to Mastercard. Its general retention summary is five years. These are Stripe's summaries, subject to the current network instrument. | \n
| Processing limitations, 169–170 | \nA listing generally disqualifies Stripe processing. Consideration requires evidence of an exceptional circumstance, with verified identity theft given as an example. This promises consideration, not approval. Resolving excessive-chargeback issues does not by itself permit Stripe to remove the qualifying record. | \n
The example transaction arithmetic elsewhere on the page is expressly hypothetical. It supplies neither a merchant case nor evidence that a particular referral was acted upon. No quantitative eligibility threshold is independently adopted from this page where the current Mastercard manual is the governing source.
\nPinpoint reading of root's shared [Security Rules and Procedures—Merchant Edition, 4 August 2026](http[local research file] sharpens the distinction. Section 11.2, PDF/printed page 148, expressly permits acquirers to onboard listed merchants and payment facilitators to onboard listed sponsored merchants. Stripe's general disqualification is therefore its stated acceptance practice, not an unavoidable consequence of mandatory database use.
\nThe summaries also differ. Section 11.5, page 151, specifies five calendar days under its enumerated termination/awareness triggers, versus Stripe's one working day. Table 11.4, page 157, uses previous-three-month language and a 1.5% threshold for code 04, versus Stripe's monthly 1%; its code 05 period also differs. The manual's awkward period wording is not reconstructed into a new formula here. Section 11.13, pages 155–156, places removal with Mastercard and allows the merchant itself to request removal for cured PCI noncompliance if the acquirer is unwilling or unable, following the stated proof process. That alternative is absent from Stripe's opening acquirer-only summary.
\nThese are source/version mismatches, not proof of when rules changed or that Stripe applied the wrong rule in any case. The dated network instrument controls the network-rule description; Stripe remains the primary account of its own published acceptance practice. No additional provider was acquired to manufacture agreement.
\nThe source is a primary-reader rendition with no original PDF retained. Independently inspected relevant reader passages in the shared rules-chapter11.reader.txt and rules-chapter11-gaps.reader.txt, and pinpoint working-reader references turn1493view1 and turn1494view0: section 11.2 final paragraph; sections 11.5 and 11.6; section 11.10; section 11.13; and table 11.4 codes 04–05. This is not a claim to have read the whole manual or independently rendered its pages. Root owns full Chapter 11 review and lifecycle synthesis.
Reused, without reacquisition, the [Mastercard Anti-Piracy Policy](http[local research file] sections 1–2, retained in ../wipo-payment-wave-2026-09-09/participant-policies.reader.txt, reader lines 161–179. It separately requires investigation and cessation of payment acceptance for a qualifying infringing product, then conditionally requires applicable MATCH listing if the acquirer terminates the merchant. That source remains a retained primary-reader derivative; its ordinary original GET previously failed with 403 and remains closed.
The supported relationship chain is therefore conditional: an applicable termination can generate a shared record; screening exposes the record to later underwriting; Stripe's stated acceptance policy gives that record a further restrictive consequence. This does not prove an ALERT-PAY-specific route into MATCH, a Stripe listing or decision concerning any identified merchant, or that Stripe itself was the acquiring bank which entered a record. A listing, termination of one processing relationship, denial of a new application, and closure of a deposit account are distinct events.
\nFor integration, useful typed relations are Mastercard MATCH record → Stripe underwriting consideration (provider_published_acceptance_policy), and listing acquirer → merchant record (reported_correction_custodian, qualified by the current Mastercard rules). Do not add WIPO → Stripe, universal bank-denial, or exercised-removal edges from this evidence.
No second implementation source was needed. There were no live database queries, account access, merchant data, operational API calls or submitted forms. The current SPME manual is root's separate lane, with the pinpoint comparison above; its ordinary PDF GET failed and a working reader identifies the 4 August 2026 version. No February version date or original-PDF custody is imported here.
\nThe strongest rival to a universal-ban account is differentiated underwriting: the record is shared, but acceptance and exceptional consideration remain provider decisions under the applicable rules. The highest-value missing record is a public, reasoned acquiring or processor decision showing how a MATCH record was assessed, challenged, corrected and reconsidered. A public policy supplies the mechanism; it does not establish its frequency, error rate or a particular customer's restoration. Acquisition closes here for the combined forest review.
", "path": "payment-termination-wave-2026-09-09/acquirer-implementation.md", "sourceIds": [ "w30-class-s-stripe-match" ], "markdown": "9 September 2026. The public implementation counterpart establishes a substantial onward effect: Stripe generally declines MATCH-listed merchants while reserving consideration for documented exceptional circumstances. This is evidence of one provider's published practice. It does not establish a Mastercard rule that every provider must reject every listed business, or loss of access to all banking.\n\n## Source, version and reading scope\n\nStripe, Terminated merchant files (see source references) also labelled High risk merchant lists in its navigation. Original HTML retained as `acquirer-stripe-match.html`, SHA256 `b867b5801678b18afc54e0a9b0ea684ca788ac70132239ff19cb97f6979773c7`; HTTP 200, 825,132 bytes, captured 2026-09-09 at 17:12:10 UTC. No revision or effective date was displayed in the substantive text. This is an observed current-page account, not a dated historic policy. The locale is English (United Kingdom), but the captured country selector reads United States; it does not identify a UK acquiring-bank contract.\n\nThe complete substantive original text was read, including the separate Visa section to identify the boundary. The operative MATCH scope is `acquirer-stripe-match.html.derived.txt` lines 88–170; the Visa section is not adopted into the MATCH findings. The primary web reader independently exposed the same text, lines 10–80 in reference `turn1482view0`. The original HTML is the retained source, and the parsed text is a derivative, not a second publication.\n\n## Decisive provisions\n\n| Locator in the retained text | Scoped finding |\n| --- | --- |\n| Opening, lines 88–89; qualification, 92–93 | Stripe describes preapproval database screening and qualifying-termination reporting as acquirer duties. Its broader assertion that most processors reject listings is an attributed industry claim, not an independently established frequency. |\n| Information added, 149–162 | The described record includes business and principal-owner identifiers, dates and reason code. Stripe says it does not disclose database information to users. This run obtained no actual merchant record. |\n| Entry removal, 163–168 | Stripe describes removal for erroneous entry or cured PCI noncompliance, through the listing acquirer; it directs merchants unable to identify that acquirer to Mastercard. Its general retention summary is five years. These are Stripe's summaries, subject to the current network instrument. |\n| Processing limitations, 169–170 | A listing generally disqualifies Stripe processing. Consideration requires evidence of an exceptional circumstance, with verified identity theft given as an example. This promises consideration, not approval. Resolving excessive-chargeback issues does not by itself permit Stripe to remove the qualifying record. |\n\nThe example transaction arithmetic elsewhere on the page is expressly hypothetical. It supplies neither a merchant case nor evidence that a particular referral was acted upon. No quantitative eligibility threshold is independently adopted from this page where the current Mastercard manual is the governing source.\n\n## Comparison with the current network instrument\n\nPinpoint reading of root's shared Security Rules and Procedures—Merchant Edition, 4 August 2026 (see source references) sharpens the distinction. Section 11.2, PDF/printed page 148, expressly permits acquirers to onboard listed merchants and payment facilitators to onboard listed sponsored merchants. Stripe's general disqualification is therefore its stated acceptance practice, not an unavoidable consequence of mandatory database use.\n\nThe summaries also differ. Section 11.5, page 151, specifies five calendar days under its enumerated termination/awareness triggers, versus Stripe's one working day. Table 11.4, page 157, uses previous-three-month language and a 1.5% threshold for code 04, versus Stripe's monthly 1%; its code 05 period also differs. The manual's awkward period wording is not reconstructed into a new formula here. Section 11.13, pages 155–156, places removal with Mastercard and allows the merchant itself to request removal for cured PCI noncompliance if the acquirer is unwilling or unable, following the stated proof process. That alternative is absent from Stripe's opening acquirer-only summary.\n\nThese are source/version mismatches, not proof of when rules changed or that Stripe applied the wrong rule in any case. The dated network instrument controls the network-rule description; Stripe remains the primary account of its own published acceptance practice. No additional provider was acquired to manufacture agreement.\n\nThe source is a primary-reader rendition with no original PDF retained. Independently inspected relevant reader passages in the shared `rules-chapter11.reader.txt` and `rules-chapter11-gaps.reader.txt`, and pinpoint working-reader references `turn1493view1` and `turn1494view0`: section 11.2 final paragraph; sections 11.5 and 11.6; section 11.10; section 11.13; and table 11.4 codes 04–05. This is not a claim to have read the whole manual or independently rendered its pages. Root owns full Chapter 11 review and lifecycle synthesis.\n\n## Connection to the earlier policy\n\nReused, without reacquisition, the Mastercard Anti-Piracy Policy (see source references) sections 1–2, retained in `../wipo-payment-wave-2026-09-09/participant-policies.reader.txt`, reader lines 161–179. It separately requires investigation and cessation of payment acceptance for a qualifying infringing product, then conditionally requires applicable MATCH listing if the acquirer terminates the merchant. That source remains a retained primary-reader derivative; its ordinary original GET previously failed with 403 and remains closed.\n\nThe supported relationship chain is therefore conditional: an applicable termination can generate a shared record; screening exposes the record to later underwriting; Stripe's stated acceptance policy gives that record a further restrictive consequence. This does not prove an ALERT-PAY-specific route into MATCH, a Stripe listing or decision concerning any identified merchant, or that Stripe itself was the acquiring bank which entered a record. A listing, termination of one processing relationship, denial of a new application, and closure of a deposit account are distinct events.\n\nFor integration, useful typed relations are Mastercard MATCH record → Stripe underwriting consideration (`provider_published_acceptance_policy`), and listing acquirer → merchant record (`reported_correction_custodian`, qualified by the current Mastercard rules). Do not add WIPO → Stripe, universal bank-denial, or exercised-removal edges from this evidence.\n\n## Boundary and next discriminator\n\nNo second implementation source was needed. There were no live database queries, account access, merchant data, operational API calls or submitted forms. The current SPME manual is root's separate lane, with the pinpoint comparison above; its ordinary PDF GET failed and a working reader identifies the 4 August 2026 version. No February version date or original-PDF custody is imported here.\n\nThe strongest rival to a universal-ban account is differentiated underwriting: the record is shared, but acceptance and exceptional consideration remain provider decisions under the applicable rules. The highest-value missing record is a public, reasoned acquiring or processor decision showing how a MATCH record was assessed, challenged, corrected and reconsidered. A public policy supplies the mechanism; it does not establish its frequency, error rate or a particular customer's restoration. Acquisition closes here for the combined forest review.\n", "file": "research/payment-termination-wave-2026-09-09-acquirer-implementation.md" }, { "id": "payment-termination-wave-2026-09-09-correction-remedies", "title": "MATCH correction, privacy and a worked judicial challenge", "description": "Bounded acquisition closed 9 September 2026. This lane inspected the current official privacy notice and one historical court opinion. The root owns the current Security Rules and Procedures; no manual was rehere. Neither MATCH nor a merchant account was accessed.", "html": "Bounded acquisition closed 9 September 2026. This lane inspected the current official privacy notice and one historical court opinion. The root owns the current Security Rules and Procedures; no manual was reacquired here. Neither MATCH nor a merchant account was accessed.
\nThe notice effective 16 March 2026 assigns storage/distribution to Mastercard and entry, maintenance and accuracy to financial institutions. Institutions may use a match for additional diligence, safeguards or refusing a contract. Depending on applicable law, individuals can seek access, correction, restriction, objection or deletion and complain to a supervisory authority. California requests go to the relevant institution; elsewhere Mastercard may receive and redirect requests. The notice describes automatic deletion after five years, subject to its legal-retention qualification. It does not promise erasure of independent recipient records or restoration of services. Its final contact paragraph promises a written reply within one month unless law supplies another period; that is not a correction deadline. These rights concern personal information, not an unconditional entitlement to merchant processing. The notice's introductory fraud framing should not replace the manual's complete eligibility rules. [Official notice](http[local research file] reader lines146-159,171-174,198-206,217-218,257-258; full substantive notice lines140-262 read. Exact paragraphs, including surrounding jurisdiction language, are retained in correction-match-privacy.reader.txt.
In the unpublished 17 May 2023 opinion eGumball v. Merrick Bank, G060853, the court describes Visa's investigation, Merrick's termination/reporting and eGumball's disputed explanation. On 3 November 2021 the trial court ordered Merrick, Paysafe and Visa to request removal. The mandatory provision was stayed on appeal; the appellate court reversed the TRO because the evidence did not establish the required interim harm before a preliminary-injunction hearing. It expressly reserved that hearing's merits and other contested issues. Its record showed only Merrick could request this merchant's addition/removal. It also explained that delisting would not itself restore processing: another acquirer still had to approve the merchant. Thus neither the provisional order nor its reversal proves that the listing was false, justified, finally corrected or followed by restored service. [Republished court opinion](http[local research file] factual sectionsIII-VIII; DiscussionIII-IV; Disposition. Decisive reader lines238-244,285-289,303-328, especially311-325. This is a primary judicial text accessed through a republication, not a court-hosted original or a current docket check.
\nThere are at least three different questions: accuracy of shared personal information, power to change the merchant record, and willingness of another institution to offer service. A remedy at one stage need not resolve the others. The case gives a concrete example of court scrutiny directed at those different roles, while also showing how the requested emergency remedy can fail without determining the underlying dispute.
\nNo notice-to-prior-recipient or automatic restoration promise was located in the inspected privacy instrument. Applicable law or other rules may impose additional obligations; this lane did not audit those regimes. A current rule allowing correction is also not evidence that the correction channel worked in a particular case. Conversely, the unsuccessful TRO does not establish that meaningful correction is impossible.
\nThe next useful evidence would be an attributable completed correction: the responsible institution's decision, resulting MATCH status and notice to earlier recipients, separately from a subsequent onboarding decision. For eGumball specifically, a later final disposition or court-filed removal confirmation would settle much more than another summary of the 2023 procedural ruling. No ALERT-PAY, adult-content or NGO involvement in this dispute was established or inferred.
\nSeven discovery queries were used: MATCH Mastercard terminated merchant list court removal acquirer decision privacy; site.ftc.gov Mastercard MATCH terminated merchant list; site.courtlistener.com Mastercard MATCH removal merchant; "Mastercard" "MATCH" "removal" court injunction merchant; "MATCH list" "removed" court judgment Mastercard; "eGumball" "G062863" site:courts.ca.gov; "eGumball" "G060853" opinion court PDF. Other search returns were not treated as acquired findings. In particular, anonymous accounts of removal were not used. The later same-case G062863M anti-SLAPP decision appeared as an unread search lead; no final merits/removal outcome was inferred from it. Other FTC and merchant cases were not opened into new branches.
No court-hosted eGumball original was located by the finite official-domain search. The one selected public opinion republication was read as a credible substitute. Ordinary GETs of both the Mastercard notice and Casemine page returned403; those error bytes are retained and clearly labeled. Their successful public web-reader results were retained instead, with source URL and exact reader locators. There was no alternate-domain or language attempt after failure. The privacy reader contains the complete substantive notice; the opinion capture combines actually returned lines116-161 and173-330, leaving the unreturned162-172 gap explicit. This is scoped reading, not an original court docket, underlying-exhibit inspection, audio/video review or proof of present merchant status.
\ncorrection-captures.json identifies the error responses, derivative captures, hashes, dates and read scope. No canonical case, reader, ZIP or other lane's packet was changed.
9 September 2026. Root acquisition is closed for review. These are public rules, not operational records, an individual merchant's status or an ALERT-PAY agreement. The specific route connecting WIPO's programme to Mastercard remains unacquired.
\nThe successful official reader now displays Security Rules and Procedures, Merchant Edition, dated 4 August 2026, with 236 pages: http[local research file] . The initial search extract described 3 February 2026; that older indexed date must not govern the following passages. The ordinary GET returned 403 HTML, correctly named rules-spme2026.error.html. No original PDF was retained and that route is closed.
\nRoot read the title and continuous Chapter 11 text, PDF pages 147–158, reader lines 5380–5822, across the two saved overlapping reader files. The coverage check proves retained text continuity only. No page image, live system or complete 236-page manual was inspected. Tables are reader-supported, without a facsimile check.
\nThe public rules index also failed ordinary GET but succeeded in the official reader: http[local research file] . Root read its complete substantive rules, disclaimer and programme sections. It warns that public documents may not be entirely current and that Mastercard's authoritative standards prevail. This is therefore a dated public edition, not a certified complete current contract set. Chapter 11 ends with a privacy heading and a short introductory sentence; Chapter 12 is expressly omitted. This does not establish why material is omitted or that no other privacy rules exist. The separate public privacy notice is covered by the remedy lane.
\nSections 11.1–3 describe a mandatory acquirer system carrying reports about certain terminated merchants and their owners. Authorized roles include approved acquirers and appropriately registered processors, payment facilitators and data-storage providers. Service-provider access requires acquirer approval and appropriate registration/provisioning. No actual authorization was inspected.
\nSections 11.2 and 11.6.4 explicitly permit an acquirer to onboard a listed merchant; section 11.2 also permits a payment facilitator to onboard a listed sponsored merchant. Section 11.6 requires an inquiry before signing an agreement or enabling acceptance. Mandatory inquiry is not mandatory rejection. The chapter supplies noncompliance consequences for failures to use, report or inquire. A provider may adopt a stricter admission policy, as Stripe's separate guidance illustrates. A processor service, an acquiring relationship, card acceptance and ordinary bank-account access remain different things.
\nSection 11.5 combines an act to terminate by either the acquirer or merchant with the acquirer's reason to believe a listed condition exists. It specifies five calendar days from the earlier of its enumerated events: the acquirer's termination decision, receipt of the merchant's termination decision, or awareness of an issue meeting a reason code. The third item must not be detached from the opening condition. The effective termination date is explicitly not controlling. A merchant-initiated exit can still qualify; every exit does not.
\nSection 11.4 leaves accuracy responsibility with the acquirer even when a service provider handles work. It requires good faith and reasonable conduct, prohibits using MATCH as a collection threat for minor discretionary activity, and requires a defined reason to be met or suspected at the termination decision. Failure to correct erroneous data within a reasonable period can attract assessments. It also requires supplying the listing acquirer identifier and reason code to the merchant or another acquirer, and maintaining working contacts. These are duties, not observed compliance.
\nRemoval requests require a response within 30 calendar days; questions about a listing require a response within seven. A response is not a removal guarantee. Section 11.5.1 expressly permits a merchant to request removal without legal counsel and identifies the required information. No request was prepared or sent.
\nThe system can associate up to five principal owners with a merchant and search using exact or phonetic/spelling-variant matches. Those are possible identity matches, not findings of new misconduct. Section 11.4 requires checking relevance, with special care for identity-theft reason code 14: a legitimate owner may not know someone previously used their identity. No person's record was queried.
\nSection 11.6.4 describes retroactive alerts when another authorized user adds information after an initial inquiry, within the following 365 days. The alert remains visible for 30 days and is reported once in batch results. Recipients assess appropriate next steps; the text again allows onboarding. This is a documented route for later information to reach other institutions, not a demonstrated cascade of closures. No API, testing facility or production service was operated. The chapter's test-environment rules do not establish that this investigation performed any test.
\nSection 11.10 provides five-year retention and automatic purge for merchant listings. Separately, authorized users must retain their own relevant MATCH records for at least two years after the underlying agreement ends; system inquiry records last 365 days. Database purge therefore cannot by itself prove deletion of every recipient's records, reversal of its assessment or restored processing.
\nSection 11.13 permits early removal when an authorized user reports an erroneous addition, or when reason-code-12 PCI security noncompliance has been remedied with specified confirmation and validation. For the PCI route, the merchant can apply directly if the acquirer is unwilling or unable, with the same proof requirements. This is a narrow alternative, not a general bypass for every dispute. The chapter does not expressly make later business reform a removal ground for every other reason code; it also does not establish that external privacy, contractual or judicial remedies are unavailable.
\nCorrecting a report, removing a listing and obtaining new processing are distinct outcomes. The public case and privacy lane tests the latter stages separately.
\nTable 11.4 covers several kinds of risk and misconduct, including compromised data, transaction laundering, chargebacks/fraud, coercion, insolvency, standards violations, PCI compliance, illegal transactions and identity theft. Some descriptions expressly address American Express reports. That is evidence of a cross-network reporting category, not proof of Visa sharing or every brand's rules. The table does not identify which code an ALERT-PAY or IP case would actually use.
\nStripe's undated public guidance summarizes a one-working-day reporting requirement and different chargeback/fraud periods or thresholds; the August manual specifies five calendar days under its listed triggers and different table wording. This is a documented source mismatch. No amendment notice or exact change date was acquired. A stricter provider deadline could coexist with a network limit; each statement needs its own attribution. The dated manual is stronger evidence for the public network rule than a provider's general summary, subject to the manual's own currency caveat.
\nThe shared system combines mandatory reporting and inquiry with recipient discretion, possible identity matches, conditional removal and separate retention rules. It can affect access across institutions without a universal instruction to refuse every listed business. Its explicit duties also provide concrete standards against which practice can be tested. The revealing missing evidence is a worked report, recipient assessment, challenge and outcome, including whether correction reached the recipient. There is no established universal loss of banking, automatic account closure or intent to target a particular group in this packet.
", "path": "payment-termination-wave-2026-09-09/rules-lifecycle.md", "sourceIds": [ "w30-class-s-match-rules" ], "markdown": "9 September 2026. Root acquisition is closed for review. These are public rules, not operational records, an individual merchant's status or an ALERT-PAY agreement. The specific route connecting WIPO's programme to Mastercard remains unacquired.\n\n## Source and version\n\nThe successful official reader now displays Security Rules and Procedures, Merchant Edition, dated 4 August 2026, with 236 pages: http[local research file] . The initial search extract described 3 February 2026; that older indexed date must not govern the following passages. The ordinary GET returned 403 HTML, correctly named rules-spme2026.error.html. No original PDF was retained and that route is closed.\n\nRoot read the title and continuous Chapter 11 text, PDF pages 147–158, reader lines 5380–5822, across the two saved overlapping reader files. The coverage check proves retained text continuity only. No page image, live system or complete 236-page manual was inspected. Tables are reader-supported, without a facsimile check.\n\nThe public rules index also failed ordinary GET but succeeded in the official reader: http[local research file] . Root read its complete substantive rules, disclaimer and programme sections. It warns that public documents may not be entirely current and that Mastercard's authoritative standards prevail. This is therefore a dated public edition, not a certified complete current contract set. Chapter 11 ends with a privacy heading and a short introductory sentence; Chapter 12 is expressly omitted. This does not establish why material is omitted or that no other privacy rules exist. The separate public privacy notice is covered by the remedy lane.\n\n## Who reports, and who admits the next merchant\n\nSections 11.1–3 describe a mandatory acquirer system carrying reports about certain terminated merchants and their owners. Authorized roles include approved acquirers and appropriately registered processors, payment facilitators and data-storage providers. Service-provider access requires acquirer approval and appropriate registration/provisioning. No actual authorization was inspected.\n\nSections 11.2 and 11.6.4 explicitly permit an acquirer to onboard a listed merchant; section 11.2 also permits a payment facilitator to onboard a listed sponsored merchant. Section 11.6 requires an inquiry before signing an agreement or enabling acceptance. Mandatory inquiry is not mandatory rejection. The chapter supplies noncompliance consequences for failures to use, report or inquire. A provider may adopt a stricter admission policy, as Stripe's separate guidance illustrates. A processor service, an acquiring relationship, card acceptance and ordinary bank-account access remain different things.\n\n## Entry, timing and responsibility\n\nSection 11.5 combines an act to terminate by either the acquirer or merchant with the acquirer's reason to believe a listed condition exists. It specifies five calendar days from the earlier of its enumerated events: the acquirer's termination decision, receipt of the merchant's termination decision, or awareness of an issue meeting a reason code. The third item must not be detached from the opening condition. The effective termination date is explicitly not controlling. A merchant-initiated exit can still qualify; every exit does not.\n\nSection 11.4 leaves accuracy responsibility with the acquirer even when a service provider handles work. It requires good faith and reasonable conduct, prohibits using MATCH as a collection threat for minor discretionary activity, and requires a defined reason to be met or suspected at the termination decision. Failure to correct erroneous data within a reasonable period can attract assessments. It also requires supplying the listing acquirer identifier and reason code to the merchant or another acquirer, and maintaining working contacts. These are duties, not observed compliance.\n\nRemoval requests require a response within 30 calendar days; questions about a listing require a response within seven. A response is not a removal guarantee. Section 11.5.1 expressly permits a merchant to request removal without legal counsel and identifies the required information. No request was prepared or sent.\n\n## How the record travels\n\nThe system can associate up to five principal owners with a merchant and search using exact or phonetic/spelling-variant matches. Those are possible identity matches, not findings of new misconduct. Section 11.4 requires checking relevance, with special care for identity-theft reason code 14: a legitimate owner may not know someone previously used their identity. No person's record was queried.\n\nSection 11.6.4 describes retroactive alerts when another authorized user adds information after an initial inquiry, within the following 365 days. The alert remains visible for 30 days and is reported once in batch results. Recipients assess appropriate next steps; the text again allows onboarding. This is a documented route for later information to reach other institutions, not a demonstrated cascade of closures. No API, testing facility or production service was operated. The chapter's test-environment rules do not establish that this investigation performed any test.\n\n## Several different retention and correction clocks\n\nSection 11.10 provides five-year retention and automatic purge for merchant listings. Separately, authorized users must retain their own relevant MATCH records for at least two years after the underlying agreement ends; system inquiry records last 365 days. Database purge therefore cannot by itself prove deletion of every recipient's records, reversal of its assessment or restored processing.\n\nSection 11.13 permits early removal when an authorized user reports an erroneous addition, or when reason-code-12 PCI security noncompliance has been remedied with specified confirmation and validation. For the PCI route, the merchant can apply directly if the acquirer is unwilling or unable, with the same proof requirements. This is a narrow alternative, not a general bypass for every dispute. The chapter does not expressly make later business reform a removal ground for every other reason code; it also does not establish that external privacy, contractual or judicial remedies are unavailable.\n\nCorrecting a report, removing a listing and obtaining new processing are distinct outcomes. The public case and privacy lane tests the latter stages separately.\n\n## Scope and source differences\n\nTable 11.4 covers several kinds of risk and misconduct, including compromised data, transaction laundering, chargebacks/fraud, coercion, insolvency, standards violations, PCI compliance, illegal transactions and identity theft. Some descriptions expressly address American Express reports. That is evidence of a cross-network reporting category, not proof of Visa sharing or every brand's rules. The table does not identify which code an ALERT-PAY or IP case would actually use.\n\nStripe's undated public guidance summarizes a one-working-day reporting requirement and different chargeback/fraud periods or thresholds; the August manual specifies five calendar days under its listed triggers and different table wording. This is a documented source mismatch. No amendment notice or exact change date was acquired. A stricter provider deadline could coexist with a network limit; each statement needs its own attribution. The dated manual is stronger evidence for the public network rule than a provider's general summary, subject to the manual's own currency caveat.\n\n## What this changes\n\nThe shared system combines mandatory reporting and inquiry with recipient discretion, possible identity matches, conditional removal and separate retention rules. It can affect access across institutions without a universal instruction to refuse every listed business. Its explicit duties also provide concrete standards against which practice can be tested. The revealing missing evidence is a worked report, recipient assessment, challenge and outcome, including whether correction reached the recipient. There is no established universal loss of banking, automatic account closure or intent to target a particular group in this packet.\n", "file": "research/payment-termination-wave-2026-09-09-rules-lifecycle.md" }, { "id": "photodna-access-wave-2026-09-09-coalition-sublicensing", "title": "Tech Coalition PhotoDNA sublicensing: delegated access, undisclosed decision terms", "description": "9 September 2026. Bounded acquisition closed. This packet owns the Coalition selection/authority lane. It does not determine Microsoft SDK/cloud terms or NCMEC hash access, which are separate lanes. No application, account, software execution, media, hash list, outreach or payment was attempted.", "html": "9 September 2026. Bounded acquisition closed. This packet owns the Coalition selection/authority lane. It does not determine Microsoft SDK/cloud terms or NCMEC hash access, which are separate lanes. No application, account, software execution, media, hash list, outreach or payment was attempted.
\nThe Coalition is a publicly acknowledged licensing intermediary. Its [2023 annual report](http[local research file] reproduces a statement attributed to Microsoft describing provision of PhotoDNA for the Coalition to license to its members. This is a counterparty-attributed statement in the recipient's report, not the original interparty agreement.
\nThe [2024 annual report](http[local research file] dates becoming a sublicensor to 2023 and extending sublicenses to qualifying non-members to 2024. PDF13/printed13 separately reports integrating PhotoDNA with open-source Hasher-Matcher-Actioner for already licensed companies, alongside a Microsoft statement attributed to Liz Thomas supporting the licensing arrangement. Thus non-membership access predates the January 2025 announcement below. These are retrospective programme accounts; no individual executed sublicense or exact grant date appears in the acquired passages. Page28 also describes companies acting on content under their own policies and procedures. Neither integration nor access establishes a common removal instruction.
\nThe [27 January 2025 announcement](http[local research file] says the Coalition expanded its agreement for Microsoft's updated technology. It reports licensing scope extended beyond CSAM to terrorist content and adults' non-consensual intimate imagery. Its implementation section invites member and non-member licensing inquiries, offers the code without charge and describes responses as typically arriving within a week. That is a response estimate, not a guaranteed licence award or observed turnaround. It offers implementation assistance and separately ongoing support for members. Compatibility with NCMEC, GIFCT and StopNCII is claimed; access to those bodies' data is not granted by that sentence. This page publishes neither the Microsoft–Coalition agreement nor an applicant contract or approval rubric. Its promotional privacy and effectiveness statements are not treated here as independent technical findings.
\nThe [current Pathways page](http[local research file] retained earlier on 9 September 2026 and reread in full, permits employees of technology companies hosting or providing access to user content or communications. It excludes vendors/service providers without that function and staff of existing Coalition members. Staff review submitted information for eligibility. Participation is free and does not require or guarantee later membership. A separate opportunity allows companies to apply for a PhotoDNA sublicense. These are Pathways admission conditions; the acquired page does not establish that every participant receives a licence or that its exclusions govern the separate member licensing route. Intake labels include sector, products, region, company size and confidentiality consent; no form was submitted.
\nThe [nine published non-member account terms](http[local research file] govern the site and mentoring. Clauses1–4 cover discretionary updates, reasonable-effort information responses, accurate registration and a revocable, nontransferable site licence with limited copying permissions. Clauses5–6 cover mentoring confidentiality. Clause7 preserves voluntary use and the participant's decision whether/how to use materials, disclaims creation of a legal/community standard, and recommends the participant's own legal review. Clauses8–9 address severability and dispute rules. These are actual published conditions on resources. Their site-licence revocability cannot be transplanted into PhotoDNA's unknown sublicense withdrawal terms. The published document is not evidence that any particular company assented or was sanctioned. Its current capture has no established publication/revision date.
\n| Typed relationship | \nSupported state | \nWhat it does not establish | \n
|---|---|---|
| Microsoft → Tech Coalition: sublicensing delegation | \nCoalition's retrospective programme accounts, including attributed Microsoft statements, and January 2025 expansion announcement | \nFull delegated powers, Microsoft reserved approval, revocation, audit or downstream-use clauses | \n
| Tech Coalition → applicants: licensing route | \nMember/non-member inquiries, free code, administrative response estimate | \nNamed decision-maker, licence guaranteed by eligibility, a particular granted/denied licence | \n
| Tech Coalition → Pathways participants: selective resource access | \nPublished admission criteria, staff review and site/mentoring terms | \nThese conditions are the PhotoDNA contract or admission confers membership/hash access | \n
| Tech Coalition → licensed implementers: integration support | \n2024 HMA integration and 2025 assistance announcement | \nPhotoDNA itself is open source, automatic access to reference hashes, or mandated production configuration | \n
| Provider → its content decisions: retained implementation role | \nReported use under provider policies; voluntary-materials clause | \nUnqualified freedom from separate contracts or applicable law | \n
A concrete intermediary can reduce transaction and engineering costs while retaining an administrative gate. Both features are visible. The currently acquired public evidence cannot distinguish routine eligibility/safety screening from broader discretion in actual licence decisions. It also cannot establish donor or member influence over an individual grant or refusal. The published non-member route is evidence against treating paid membership as universally necessary for access.
\nSix distinct discovery queries were attempted: site:technologycoalition.org "PhotoDNA" license; the same domain with "PhotoDNA" sublicense; the same domain with "PhotoDNA" eligibility; "PhotoDNA" "Tech Coalition" "sublicense" -site:technologycoalition.org; "PhotoDNA" "Tech Coalition" eligibility approval license denied agreement; and "PhotoDNA" "Tech Coalition" ("license granted" OR "sublicensed" OR "denied" OR "rejected" OR "licensing agreement"). The initial three-query response was wholly truncated and not used as evidence; the sublicense query was rerun. Later visible results identified the announcement and annual reports; the last exact decision query returned no results. Not every returned snippet/link was opened. A proxy-hosted OpenAI result and third-party/LinkedIn snippets were not acquired or treated as original records. No licence recipient was inferred from an adoption count, HMA use or membership testimonial.
No public selection rubric, named issued/denied Coalition sublicense, executed applicant agreement, or licence review/revocation procedure was located within this finite scope. This is a bounded search result, not proof that none exists. The decisive next records are the Microsoft–Coalition delegation and current applicant sublicense/template, with any eligibility schedule; these would locate reserved approvals, permitted purposes, term, termination, derivative/code distribution, data access and reporting conditions. A dated executed licence or refusal/appeal correspondence would then establish how those powers were exercised. Likely custodians are Microsoft, the Coalition and the relevant applicant. No request was made.
\ncoalition-captures.json records three new successful official originals, plus two reused Pathways originals with verified hashes and prior-manifest ancestry. The 2025 announcement's full substantive body was read through the primary web reader and retained HTML-derived text. Both reports were text-searched in full for PhotoDNA/sublicensing; complete extracted text of 2023 PDF10/20 and 2024 PDF13/28/29/35 was read. Only the licensing/provider-policy passages identified above support this packet; chart graphics and other report pages were not substantively inspected. The Pathways substantive programme/FAQ/intake labels and all nine account terms were reread from retained derivatives. Original HTML/PDF bytes remain distinct from derived text. All three new ordinary public GETs succeeded; no failed access route or known closed source was retried. Acquisition stops here for the combined forest assessment.
Observed 9 September 2026. Acquisition closed after six focused queries and the named primary pages below. This is an administrative-document review, not a service test or a complete licence audit.
\nThe evidence supports conditional access through different instruments. It does not establish a public entitlement to PhotoDNA, a licence refusal against a named applicant, or permission to obtain every reference-hash collection. The dates and product distinction matter: the posted Cloud terms carry a May 2018 update date; that text cannot establish the terms of a later Tech Coalition software sublicense.
\nMicrosoft describes PhotoDNA as software that computes an image signature for comparison with other signatures; a reference database supplies previously identified material for matching. Its landing page reports a donation of the technology to NCMEC and free provision to qualified organizations and law-enforcement tool providers. These statements do not themselves convey an open-source licence or a recipient's independent rights to every comparison database. [Microsoft PhotoDNA](http[local research file] substantive headings “How does PhotoDNA technology work?” and “Who uses PhotoDNA technology?”, retained derivative lines97–104.
\nThe FAQ distinguishes installed technology requiring integration from the hosted REST service. It describes Cloud applicants as trusted online providers/businesses hosting user content, approved individually following an unnamed third party's evaluation; keys follow vetting. It directs law enforcement toward forensic-tool providers or the on-premise licensing process rather than this Cloud offering. The posted FAQ says Cloud handles images, while a separate PhotoDNA technique can handle video. It advertises a benign-image integration environment and says customer images are not stored. These are provider descriptions, not independently observed operating guarantees. [Microsoft FAQ](http[local research file] retained derivative lines87–116.
\nThe documentation landing places portal access, sample code and test transactions after onboarding and describes subscription-key REST calls comparing a supplied image or URL against a hash set. That establishes a described matching interface; it does not promise a raw reference-set download. The linked portal was not opened. [Documentation](http[local research file] retained derivative lines86–91.
\nThe live-posted terms, visibly updated May 2018, incorporate MOSA/Online Services Terms and give these specific terms precedence. Microsoft retains eligibility discretion, reverification and access suspension. Customers must supply accurate information, secure accounts and report account-security breaches. Use is restricted to preventing CSAM distribution and supporting related investigations. Customers authorize usage audits and customer-identified aggregate match-count reports to NCMEC; those reports do not replace customers' own legal reporting duties.
\nThe text specifies internal use, excludes managed-service provision, gives a 30-day subscription term, allows throttling, and describes limited free transactions or Content Moderator charges for high volume. It supplies neither support nor an SLA and contractually excludes Microsoft Azure Service status. That designation differs from the FAQ's hosting statement that the service runs on Azure. The dated pricing references are not a verified September 2026 commercial offer. No refusal-review procedure was located in these seven sections; incorporated agreement versions remain unexamined. [Cloud terms](http[local research file] sections1–7, retained derivative lines86–115.
\nMicrosoft's 19 March 2012 announcement says the “Netherlands Forensics Institute” (publisher's wording) and New Zealand Department of Internal Affairs had licensed PhotoDNA source code. It separately describes NetClean and CETS integration routes and direct licensing for agencies able to integrate the code. This is a dated licensor account of named software access, not an executed agreement, current licence status, cash grant or proof of a particular enforcement action. [Microsoft announcement](http[local research file] reader lines14–18. Original GET returned403; the complete substantive public-reader extraction is retained instead.
\nThe parallel Tech Coalition lane owns its expanded sublicense announcement and application rules. None of the Microsoft pages read here identifies Tech Coalition as the unnamed Cloud vetter, supplies its master sublicense agreement, or establishes that the Cloud terms govern its software distribution. Keep these as separate instruments unless a contract cross-reference joins them. Likewise, Microsoft's NCMEC technology-donation account is not a finding about who selects or corrects a particular reference entry; the root NCMEC lane examines those administrative rules separately.
\nNo named Microsoft rejection or revocation decision, current signed on-premise licence, or TC master licence was located in this finite pass. That is a search limitation, not proof of universal approval or absence of restrictions. A current dated software licence/sublicence form and a public executed approval or reasoned refusal would discriminate between advertised availability and the precise approval, downstream-use, termination and reconsideration machinery. Likely custodians are Microsoft licensing and the relevant sublicensor; no contact was made.
\nThe local forest implication is that technical availability passes through separate approval and data-access decisions. A historical software grant supplies an actual access outcome; the posted Cloud terms supply express continuing control. Neither alone establishes authority over a recipient's individual moderation decision or a working correction chain after an underlying reference changes.
\nmicrosoft-captures.json records URL, observation time, original/derived hashes, HTTP state and exact substantive reading scopes. Four original HTML responses were retained and hash-checked: home, FAQ, Cloud terms and documentation. Their complete substantive bodies were read; navigation and footer material were not treated as evidence.
The 2012 source has only microsoft-licensees-2012.reader.json: its SHA identifies the reader extraction, not original HTML. The ordinary403 route is closed. No incorporation-by-link terms were presumed read; no application, account, authenticated portal, endpoint transaction, media or reference hash set was accessed. Six recorded focused queries close this instrument-level pass. The court-text leads Rosales and Bebris were relayed to the root lane as secondary leads only and supply no findings in this packet.
9September2026. Root acquisition closed. Public administrative documentation separates NCMEC's access administration, submitting members' rights, clients' local datasets and platform actions. No live hash/media data, credentials, endpoint, application or account was used.
\nThe [NCMEC Hash Sharing v2 documentation](http[local research file] visibly last updated5November2025, requires NCMEC-issued credentials and separate read-only/read-write permissions in each environment. Its table distinguishes industry, nonprofit and law-enforcement services, with separate CSAM/exploitative and generative-AI entries; a label is not a legal finding about every entry. Submitting members can limit which other members see their entries. No explicit grants means access for all members; deleting the last grant restores that default. Self-only access is separately described. These are documented functions, not an observed denial or a public dataset.
\nOnly the submitting member can retract its entries through the described API. Recipients' queries can return a deletion marker identifying submitter, entry and timestamp. Other members can provide affirmative/negative fingerprint feedback; example reason IDs are expressly synthetic. This is not an adjudicated appeal or a promise to remove an account restriction. Read scopes: sections2,4.6,7.5,7.9,7.10 completely; other sections were discovery/search context only. No operational examples were executed.
\nThe [Usage Recommendations](http[local research file] make the client side consequential. Clients maintain their own datasets and choose polling frequency; the API supplies the latest state using update times, rather than a historical event log. Owners can withdraw visibility, and the guide expressly says the API has no explicit mechanism for detecting others' visibility changes. New grants can require historical backfill. This differs from a published retraction marker: loss of access is not automatically evidence an entry was judged wrong. The guide does not establish a particular client's cache deletion, duty to erase, or reversal of past moderation. Full substantive page read; date matches the main document footer.
\nNCMEC's [2025 CyberTipline data page](http[local research file] Hash Sharing and Removing Content, describes repeated staff review and reports12.1million-plus hashes shared with78 voluntary ESP participants as of31December2025. Its notice programme distinguishes CSAM, exploitative imagery of identified victims that may fall below the CSAM threshold, and predatory text. These notification categories must not all become criminal-content findings or be assigned to every reference environment. Removal/blocking is described as based on company terms. Counts and response times are NCMEC's reported outcomes; no individual notification or client log was acquired. Root read those sections and adjacent response context, not the entire page or downloadable outcome tables.
\nThe [12April2024 Concentrix attestation](http[local research file] all3pages read, identifies an April2023 RFP and13July2023 audit contract. It covers538,922 unique files, with two reviews each against the stated federal definition, conducted16October–8December2023. It reports59 exceptions:50 judged not to meet that definition;5 cartoons/anime also not meeting it;2 unplayable videos;2 with uncertain depicted age. These are the auditor's classifications, not our inspection of content or59 identical kinds of error.
\nPage3 says Concentrix notified NCMEC and was told that NCMEC removed all59 in less than five business days. This is a concrete auditor-reported correction notification, not an independently observed deletion log or proof of recipient-wide correction. Reported99.99% agreement concerns that reviewed list/version; it is not all NCMEC environments or future datasets. The entire audit contract, workpapers and downstream acknowledgements remain unacquired.
\nNCMEC's [15April2024 announcement](http[local research file] full substantive body read, says the audit was donated, valued at$300,000. This is a reported in-kind contribution and recipient valuation, not a cash payment. The announcement also describes systems/training supplied by NCMEC and the auditor's independent content decisions; its own promotional conclusion is not additional independent assurance. No new corporate genealogy or moderator/subject identification was undertaken.
\nA [25November2025 Rosales opinion reproduced by FindLaw](http[local research file] cases01-23-00876-CR/00877-CR, describes separate Microsoft→NCMEC→Dropbox software sublicensing, database-access agreements and a signatures MOU. The reproduced text says the MOU grants restricted royalty-free access and an agreement to attempt detection; database clauses retain voluntary access and disclaim NCMEC direction over use and responsibility for other nonprofits' entries. The court rejected government-agency attribution on that record, while assuming NCMEC's status for argument. It also records that actual use of that software in the contested search was not proved. This is not a universal judgment that collaborations cannot exercise influence.
\nCustody matters: root read the reproduced analysis123–196, relevant footnotes and conclusion through the reader. It is not the official original, full contracts, or current Dropbox terms. The available official link was for a distinct July2025 opinion; reader failure followed by one GET403 closed that route. FindLaw GET likewise returned an error while its reader worked. No original or signed agreement is claimed. A later attempt to retain serialized reader output hit Windows command-length limits before creating a file; tool history/read locators, rather than a fabricated original, preserve this scope. No further recovery route was opened.
\nThere are several administrators, not just one database owner: NCMEC grants environment access; a submitter controls the visibility/retraction of its entries; a recipient chooses its local subset, polling and effects. Review can produce a documented correction, yet feed withdrawal and a restored item remain different steps. The reproduced litigation shows why signed access rights, general detection commitments and case-specific direction need separate evidence. Main licence/TC sublicence findings are in the sibling lanes and are not inferred here.
\nncmec-captures.json records five successful primary originals, the FindLaw error response and the closed official-opinion access observation. HTML derivatives/section extracts and PDF text are labelled. Root checked disk hashes after capture and after final scope edits. Source-provided examples are not real entries. Four focused NCMEC discovery queries and three court-identity queries were used; unhelpful same-name cases were discarded. No data endpoint, status check, application, private system, outreach or paid source was accessed. Acquisition is stopped for combined review and forest; the case and ZIP remain unchanged.
", "path": "photodna-access-wave-2026-09-09/ncmec-access.md", "sourceIds": [ "w28-access-s-ncmec-api", "w28-access-s-ncmec-usage", "w28-access-s-ncmec-audit", "w28-access-s-ncmec-donation", "w28-access-s-ncmec-provider" ], "markdown": "9September2026. Root acquisition closed. Public administrative documentation separates NCMEC's access administration, submitting members' rights, clients' local datasets and platform actions. No live hash/media data, credentials, endpoint, application or account was used.\n\n## Software rights do not supply reference-data rights\n\nThe NCMEC Hash Sharing v2 documentation (see source references) visibly last updated5November2025, requires NCMEC-issued credentials and separate read-only/read-write permissions in each environment. Its table distinguishes industry, nonprofit and law-enforcement services, with separate CSAM/exploitative and generative-AI entries; a label is not a legal finding about every entry. Submitting members can limit which other members see their entries. No explicit grants means access for all members; deleting the last grant restores that default. Self-only access is separately described. These are documented functions, not an observed denial or a public dataset.\n\nOnly the submitting member can retract its entries through the described API. Recipients' queries can return a deletion marker identifying submitter, entry and timestamp. Other members can provide affirmative/negative fingerprint feedback; example reason IDs are expressly synthetic. This is not an adjudicated appeal or a promise to remove an account restriction. Read scopes: sections2,4.6,7.5,7.9,7.10 completely; other sections were discovery/search context only. No operational examples were executed.\n\nThe Usage Recommendations (see source references) make the client side consequential. Clients maintain their own datasets and choose polling frequency; the API supplies the latest state using update times, rather than a historical event log. Owners can withdraw visibility, and the guide expressly says the API has no explicit mechanism for detecting others' visibility changes. New grants can require historical backfill. This differs from a published retraction marker: loss of access is not automatically evidence an entry was judged wrong. The guide does not establish a particular client's cache deletion, duty to erase, or reversal of past moderation. Full substantive page read; date matches the main document footer.\n\n## Review and removal: a specific public record\n\nNCMEC's 2025 CyberTipline data page (see source references) Hash Sharing and Removing Content, describes repeated staff review and reports12.1million-plus hashes shared with78 voluntary ESP participants as of31December2025. Its notice programme distinguishes CSAM, exploitative imagery of identified victims that may fall below the CSAM threshold, and predatory text. These notification categories must not all become criminal-content findings or be assigned to every reference environment. Removal/blocking is described as based on company terms. Counts and response times are NCMEC's reported outcomes; no individual notification or client log was acquired. Root read those sections and adjacent response context, not the entire page or downloadable outcome tables.\n\nThe 12April2024 Concentrix attestation (see source references) all3pages read, identifies an April2023 RFP and13July2023 audit contract. It covers538,922 unique files, with two reviews each against the stated federal definition, conducted16October–8December2023. It reports59 exceptions:50 judged not to meet that definition;5 cartoons/anime also not meeting it;2 unplayable videos;2 with uncertain depicted age. These are the auditor's classifications, not our inspection of content or59 identical kinds of error.\n\nPage3 says Concentrix notified NCMEC and was told that NCMEC removed all59 in less than five business days. This is a concrete auditor-reported correction notification, not an independently observed deletion log or proof of recipient-wide correction. Reported99.99% agreement concerns that reviewed list/version; it is not all NCMEC environments or future datasets. The entire audit contract, workpapers and downstream acknowledgements remain unacquired.\n\nNCMEC's 15April2024 announcement (see source references) full substantive body read, says the audit was donated, valued at$300,000. This is a reported in-kind contribution and recipient valuation, not a cash payment. The announcement also describes systems/training supplied by NCMEC and the auditor's independent content decisions; its own promotional conclusion is not additional independent assurance. No new corporate genealogy or moderator/subject identification was undertaken.\n\n## A named agreement trail, through reproduced judicial text\n\nA 25November2025 Rosales opinion reproduced by FindLaw (see source references) cases01-23-00876-CR/00877-CR, describes separate Microsoft→NCMEC→Dropbox software sublicensing, database-access agreements and a signatures MOU. The reproduced text says the MOU grants restricted royalty-free access and an agreement to attempt detection; database clauses retain voluntary access and disclaim NCMEC direction over use and responsibility for other nonprofits' entries. The court rejected government-agency attribution on that record, while assuming NCMEC's status for argument. It also records that actual use of that software in the contested search was not proved. This is not a universal judgment that collaborations cannot exercise influence.\n\nCustody matters: root read the reproduced analysis123–196, relevant footnotes and conclusion through the reader. It is not the official original, full contracts, or current Dropbox terms. The available official link was for a distinct July2025 opinion; reader failure followed by one GET403 closed that route. FindLaw GET likewise returned an error while its reader worked. No original or signed agreement is claimed. A later attempt to retain serialized reader output hit Windows command-length limits before creating a file; tool history/read locators, rather than a fabricated original, preserve this scope. No further recovery route was opened.\n\n## Synthesis and exact remaining gap\n\nThere are several administrators, not just one database owner: NCMEC grants environment access; a submitter controls the visibility/retraction of its entries; a recipient chooses its local subset, polling and effects. Review can produce a documented correction, yet feed withdrawal and a restored item remain different steps. The reproduced litigation shows why signed access rights, general detection commitments and case-specific direction need separate evidence. Main licence/TC sublicence findings are in the sibling lanes and are not inferred here.\n\nncmec-captures.json records five successful primary originals, the FindLaw error response and the closed official-opinion access observation. HTML derivatives/section extracts and PDF text are labelled. Root checked disk hashes after capture and after final scope edits. Source-provided examples are not real entries. Four focused NCMEC discovery queries and three court-identity queries were used; unhelpful same-name cases were discarded. No data endpoint, status check, application, private system, outreach or paid source was accessed. Acquisition is stopped for combined review and forest; the case and ZIP remain unchanged.\n", "file": "research/photodna-access-wave-2026-09-09-ncmec-access.md" }, { "id": "pipcu-priorities-wave-2026-09-09-funding-design", "title": "PIPCU funding design: reported annual signing, unacquired pilot conditions", "description": "Closed 9 September 2026 after the authorised public-source pass. This packet adds an author-attributed report of a 2026–27 Service Agreement signing and a meeting involving the Alliance for Intellectual Property and Sky. It does not acquire the agreement, a pilot specification, an industry contribution or a private right to select police work.", "html": "Closed 9 September 2026 after the authorised public-source pass. This packet adds an author-attributed report of a 2026–27 Service Agreement signing and a meeting involving the Alliance for Intellectual Property and Sky. It does not acquire the agreement, a pilot specification, an industry contribution or a private right to select police work.
\nThe three periods remain separate: the government announced public baseline funding for April 2026–March 2029; a later public LinkedIn post reports signing the first annual Service Agreement for 2026–27; the industry co-funding pilot remains planned from April 2027. The annual document must not be relabelled as an acquired three-year contract or a signed industry-funding agreement.
\nThe retained [IPO/DSIT announcement of 9 March 2026](http[local research file] opening paragraph and Background 3–4, specifies approximately £3 million per year of baseline funding, provided by DSIT via the IPO, through March 2029. Its pilot paragraph describes testing additional enforcement activity with partnership funding. The paragraph immediately after Adam Williams's quotation assigns development of the pilot's priorities and contribution level to the IPO and PIPCU together. It anticipates stakeholder engagement; it gives neither contributor admission criteria nor a method for accepting or rejecting contributions. The Commissioner describes public-sector enforcement powers and strategic oversight alongside industry expertise and intelligence. That is a public account of roles, not an acquired independence or conflicts clause.
\n[Parliamentary answer 119446](http[local research file] asked 10 March and answered by Ian Murray for DCMS on 19 March 2026, confirms funding through 2029 in an answer about illegal streaming and broadcaster sustainability. The complete question and answer were read; the decisive funding passage is the second answer paragraph (reader lines 17–18). This corroborates the announced duration without supplying an assurance schedule, payment condition or pilot rule. No actual police receipt is established by either announcement.
\nThe new [public post displayed as IPO Counter-Infringement](http[local research file] says its enforcement team and CEO Adam Williams met to sign the 2026–27 Service Agreement and met representatives of the Alliance for Intellectual Property and Sky. Its embedded post, displayed as Intellectual Property Office UK, says the annual agreement was signed as the first step after the three-year commitment. It names Daniel Guthrie, Director General of the Alliance, and Matthew Hibbert, Director of Anti-Piracy at Sky, among attendees. This establishes an attributed report of an actual meeting, a narrower advance beyond the March intention to engage. Neither body says that Sky or the Alliance paid, committed a contribution, approved a priority, obtained a seat, or signed the agreement.
\nThe main post's retained JSON-LD gives datePublished as 2026-05-29T15:21:33.883Z. The displayed date is 3mo / Edited. The embedded body says the visit occurred that week; an exact meeting date, the embedded post's publication date and the date of any edit are not independently established. Its image caption/alt text names Adam Williams and Pete O'Doherty as signing. That caption is readable source text, not our visual verification of signatures: the image pixels, signature page and agreement text were not acquired or inspected.
Author identity also has a limit. Root's separate check found that the official GOV.UK IPO landing links a LinkedIn profile at /company/uk-ipo/, whereas this retained page identifies its embedded author at /company/intellectual-property-office-uk. The gov-linked profile returned 999 when root tried it; that route is closed. The account alias has not been independently reconciled. Accordingly the signing and attendance findings are attributed to the displayed authors of the public posts, not presented as a separately verified GOV.UK publication. This is distinct from the directly attributable government funding announcement.
The March government announcement attributes to Dan Guthrie a welcome for continued public funding and an intention to engage with IPO/PIPCU about expanding the unit's work. The [Alliance's own About page](http[local research file] first two substantive paragraphs and Staff and board section, describes a UK coalition and identifies Dan Guthrie as Director General. This supports counterpart identification; it supplies no corporate contracting identity, pilot mandate or authority to commit every member.
\nThe retained signing-page HTML also includes a separate related-post card displayed as Alliance for Intellectual Property welcoming the security of three years' funding. Its exact link is [the Alliance's related post](http[local research file] That text was read within the retained page; this separate URL was not fetched. It confirms the displayed counterpart's supportive public position at that scope. It does not promise to co-fund.
\nThe Alliance's [Consumer Protection page](http[local research file] links PIPCU but publishes no pilot conditions. Its [Online Accountability page](http[local research file] including the Know Your Business Customer section, advocates verification protocols by hosting providers, online advertising companies, proxies, payment gateways and marketplaces, with loss of business-critical hosting/payment access as a consequence for illicit activity. This is a business-customer verification demand, not a consumer age-identification requirement. This is a policy position of a named participant in the reported meeting. No inspected record connects that position to an adopted PIPCU pilot priority or policing condition. The page's precise publication/update date is unknown; its copyright year is not treated as that date. No linked policy genealogy was followed.
\n| Decision or resource boundary | \nBest public evidence in this packet | \nUnacquired discriminator | \n
|---|---|---|
| Baseline and additional capacity | \nMarch announcement separates the public baseline from proposed additional pilot work. | \nExecuted 2026–27 agreement, financial schedules, baseline definition and any substitution restriction. | \n
| Who develops pilot priorities | \nMarch announcement identifies IPO and PIPCU jointly. | \nPriority-selection procedure, committee membership, approval rights and minutes. | \n
| Who may contribute, and how much | \nGeneral invitation to interested industry partners; contribution level under development. | \nEligibility/exclusion rules, contribution bands, recipient and acceptance instrument. | \n
| Industry access | \nPublic post reports Alliance and Sky attendance; displayed Alliance card welcomes funding. | \nAgenda, minutes or consultation output establishing what was proposed and what was decided. | \n
| Conflict controls and operational independence | \nCommissioner describes public strategic oversight; no current pilot clause acquired. | \nConflicts, recusal, earmarking, non-contributor access, reporting, audit, withdrawal and repayment provisions. | \n
The defensible additions are government confirmation of baseline duration, displayed-author report of annual agreement signing, reported meeting participation by Alliance and Sky, and attributed Alliance support for public funding. Their stages are respectively an official public commitment, reported execution, reported engagement, and a public position. There is no supported contribution-to-priority entitlement edge.
\nA capacity-expansion explanation remains strongest: public funding continues, public bodies develop a supplementary pilot, and industry contributes expertise and discussion. A competing explanation is that future earmarked resources could shape the queue or preferred sectors without formally transferring police powers. The meeting record makes that mechanism more concrete as a question, but it does not resolve it. The most discriminating records are the annual agreement and a published pilot participation/priority protocol, followed by records showing the disposition of industry proposals. None was acquired in this pass.
\nfunding-design-captures.json records five newly retained original HTML files, one Parliamentary error response, read scopes and three reused original hashes. funding-design-post-scoped.json preserves the main post, embedded IPO text, relevant related cards, precise observed links, metadata and image-caption scope. funding-design-parliament-reader-note.json is a scoped reading note, not an original Parliamentary document. The successful public web reader supplied the substantive answer; the ordinary preservation GET returned 403 and was not repeated. The LinkedIn reader returned Cache miss; its one ordinary GET substitute succeeded. No login, account, form, image fetch or operational interaction was used.
The full original Alliance news index was read. Its first entry is 30 June 2026, an AI-licensing event, whereas the web-reader rendition began at July 2025. The ordinary original supersedes the earlier index-date observation. Neither version supplies pilot conditions; that is only an observation about the inspected index, not proof that engagement did not occur. The successful signing post is positive evidence that engagement was reported elsewhere.
\nfunding-design-search-routes.json records the exact queries and inspected public navigation routes across government, Parliament, police/governance and Alliance surfaces. These yielded no executed agreement, assurance schedule, pilot specification, contributor list, contribution table or conflict protocol. Search absence does not establish nonexistence, concealment or a lack of safeguards. The prior non-public police budget appendix and all closed routes remain untouched. Prior accounts record an IPO-to-DSIT cost-covering dividend; that remains distinct from the new baseline commitment, a police receipt, or future industry money.
Acquisition is closed at this instrument/engagement boundary for root review and a forest assessment. No case, atlas, ZIP, current-work or forest file was changed.
", "path": "pipcu-priorities-wave-2026-09-09/funding-design.md", "sourceIds": [ "w29-enf-s-signing-post2026", "w29-enf-s-alliance-about", "w29-enf-s-alliance-online", "w29-enf-s-funding-answer2026" ], "markdown": "Closed 9 September 2026 after the authorised public-source pass. This packet adds an author-attributed report of a 2026–27 Service Agreement signing and a meeting involving the Alliance for Intellectual Property and Sky. It does not acquire the agreement, a pilot specification, an industry contribution or a private right to select police work.\n\nThe three periods remain separate: the government announced public baseline funding for April 2026–March 2029; a later public LinkedIn post reports signing the first annual Service Agreement for 2026–27; the industry co-funding pilot remains planned from April 2027. The annual document must not be relabelled as an acquired three-year contract or a signed industry-funding agreement.\n\n## What the public records establish\n\nThe retained IPO/DSIT announcement of 9 March 2026 (see source references) opening paragraph and Background 3–4, specifies approximately £3 million per year of baseline funding, provided by DSIT via the IPO, through March 2029. Its pilot paragraph describes testing additional enforcement activity with partnership funding. The paragraph immediately after Adam Williams's quotation assigns development of the pilot's priorities and contribution level to the IPO and PIPCU together. It anticipates stakeholder engagement; it gives neither contributor admission criteria nor a method for accepting or rejecting contributions. The Commissioner describes public-sector enforcement powers and strategic oversight alongside industry expertise and intelligence. That is a public account of roles, not an acquired independence or conflicts clause.\n\nParliamentary answer 119446 (see source references) asked 10 March and answered by Ian Murray for DCMS on 19 March 2026, confirms funding through 2029 in an answer about illegal streaming and broadcaster sustainability. The complete question and answer were read; the decisive funding passage is the second answer paragraph (reader lines 17–18). This corroborates the announced duration without supplying an assurance schedule, payment condition or pilot rule. No actual police receipt is established by either announcement.\n\nThe new public post displayed as IPO Counter-Infringement (see source references) says its enforcement team and CEO Adam Williams met to sign the 2026–27 Service Agreement and met representatives of the Alliance for Intellectual Property and Sky. Its embedded post, displayed as **Intellectual Property Office UK**, says the annual agreement was signed as the first step after the three-year commitment. It names Daniel Guthrie, Director General of the Alliance, and Matthew Hibbert, Director of Anti-Piracy at Sky, among attendees. This establishes an attributed report of an actual meeting, a narrower advance beyond the March intention to engage. Neither body says that Sky or the Alliance paid, committed a contribution, approved a priority, obtained a seat, or signed the agreement.\n\nThe main post's retained JSON-LD gives `datePublished` as **2026-05-29T15:21:33.883Z**. The displayed date is **3mo / Edited**. The embedded body says the visit occurred that week; an exact meeting date, the embedded post's publication date and the date of any edit are not independently established. Its image caption/alt text names Adam Williams and Pete O'Doherty as signing. That caption is readable source text, **not our visual verification of signatures**: the image pixels, signature page and agreement text were not acquired or inspected.\n\nAuthor identity also has a limit. Root's separate check found that the official GOV.UK IPO landing links a LinkedIn profile at `/company/uk-ipo/`, whereas this retained page identifies its embedded author at `/company/intellectual-property-office-uk`. The gov-linked profile returned 999 when root tried it; that route is closed. The account alias has not been independently reconciled. Accordingly the signing and attendance findings are attributed to the displayed authors of the public posts, not presented as a separately verified GOV.UK publication. This is distinct from the directly attributable government funding announcement.\n\n## The named industry's position and the missing decision rights\n\nThe March government announcement attributes to Dan Guthrie a welcome for continued public funding and an intention to engage with IPO/PIPCU about expanding the unit's work. The Alliance's own About page (see source references) first two substantive paragraphs and Staff and board section, describes a UK coalition and identifies Dan Guthrie as Director General. This supports counterpart identification; it supplies no corporate contracting identity, pilot mandate or authority to commit every member.\n\nThe retained signing-page HTML also includes a **separate related-post card displayed as Alliance for Intellectual Property** welcoming the security of three years' funding. Its exact link is the Alliance's related post (see source references) That text was read within the retained page; this separate URL was not fetched. It confirms the displayed counterpart's supportive public position at that scope. It does not promise to co-fund.\n\nThe Alliance's Consumer Protection page (see source references) links PIPCU but publishes no pilot conditions. Its Online Accountability page (see source references) including the Know Your Business Customer section, advocates verification protocols by hosting providers, online advertising companies, proxies, payment gateways and marketplaces, with loss of business-critical hosting/payment access as a consequence for illicit activity. This is a business-customer verification demand, not a consumer age-identification requirement. This is a policy position of a named participant in the reported meeting. **No inspected record connects that position to an adopted PIPCU pilot priority or policing condition.** The page's precise publication/update date is unknown; its copyright year is not treated as that date. No linked policy genealogy was followed.\n\n| Decision or resource boundary | Best public evidence in this packet | Unacquired discriminator |\n| --- | --- | --- |\n| Baseline and additional capacity | March announcement separates the public baseline from proposed additional pilot work. | Executed 2026–27 agreement, financial schedules, baseline definition and any substitution restriction. |\n| Who develops pilot priorities | March announcement identifies IPO and PIPCU jointly. | Priority-selection procedure, committee membership, approval rights and minutes. |\n| Who may contribute, and how much | General invitation to interested industry partners; contribution level under development. | Eligibility/exclusion rules, contribution bands, recipient and acceptance instrument. |\n| Industry access | Public post reports Alliance and Sky attendance; displayed Alliance card welcomes funding. | Agenda, minutes or consultation output establishing what was proposed and what was decided. |\n| Conflict controls and operational independence | Commissioner describes public strategic oversight; no current pilot clause acquired. | Conflicts, recusal, earmarking, non-contributor access, reporting, audit, withdrawal and repayment provisions. |\n\n## Reusable relations and strongest rival\n\nThe defensible additions are **government confirmation of baseline duration**, **displayed-author report of annual agreement signing**, **reported meeting participation by Alliance and Sky**, and **attributed Alliance support for public funding**. Their stages are respectively an official public commitment, reported execution, reported engagement, and a public position. There is no supported contribution-to-priority entitlement edge.\n\nA capacity-expansion explanation remains strongest: public funding continues, public bodies develop a supplementary pilot, and industry contributes expertise and discussion. A competing explanation is that future earmarked resources could shape the queue or preferred sectors without formally transferring police powers. The meeting record makes that mechanism more concrete as a question, but it does not resolve it. The most discriminating records are the annual agreement and a published pilot participation/priority protocol, followed by records showing the disposition of industry proposals. None was acquired in this pass.\n\n## Custody, reading scope and finite frontier\n\n`funding-design-captures.json` records five newly retained original HTML files, one Parliamentary error response, read scopes and three reused original hashes. `funding-design-post-scoped.json` preserves the main post, embedded IPO text, relevant related cards, precise observed links, metadata and image-caption scope. `funding-design-parliament-reader-note.json` is a scoped reading note, **not an original Parliamentary document**. The successful public web reader supplied the substantive answer; the ordinary preservation GET returned 403 and was not repeated. The LinkedIn reader returned Cache miss; its one ordinary GET substitute succeeded. No login, account, form, image fetch or operational interaction was used.\n\nThe full original Alliance news index was read. Its first entry is **30 June 2026**, an AI-licensing event, whereas the web-reader rendition began at July 2025. The ordinary original supersedes the earlier index-date observation. Neither version supplies pilot conditions; that is only an observation about the inspected index, not proof that engagement did not occur. The successful signing post is positive evidence that engagement was reported elsewhere.\n\n`funding-design-search-routes.json` records the exact queries and inspected public navigation routes across government, Parliament, police/governance and Alliance surfaces. These yielded no executed agreement, assurance schedule, pilot specification, contributor list, contribution table or conflict protocol. Search absence does not establish nonexistence, concealment or a lack of safeguards. The prior non-public police budget appendix and all closed routes remain untouched. Prior accounts record an IPO-to-DSIT cost-covering dividend; that remains distinct from the new baseline commitment, a police receipt, or future industry money.\n\nAcquisition is closed at this instrument/engagement boundary for root review and a forest assessment. No case, atlas, ZIP, current-work or forest file was changed.\n", "file": "research/pipcu-priorities-wave-2026-09-09-funding-design.md" }, { "id": "pipcu-priorities-wave-2026-09-09-priority-rules", "title": "PIPCU and IPO priority rules: an actual historical industry role, incomplete current continuity", "description": "Observed 9 September 2026. Acquisition CLOSED. One new police-guide official-reader record and one new government annual-report original, compared with the accepted August2025 Gateway announcement and posted2020 Ashiko policy. No referral form, account, contact, operational endpoint, non-public appendix or visible UI was accessed.", "html": "Observed 9 September 2026. Acquisition CLOSED. One new police-guide official-reader record and one new government annual-report original, compared with the accepted August2025 Gateway announcement and posted2020 Ashiko policy. No referral form, account, contact, operational endpoint, non-public appendix or visible UI was accessed.
\nThe posted [PIPCU Referral Guide](http[local research file] PDF4, says a Steering Group comprising IPO, City of London Police, rights holders and industry bodies agreed broad priorities: physical goods with an online connection and digital content. This is an explicit shared priority-setting role, not just consultation attendance.
\nPDF5 lists six acceptance considerations: remit; organised activity; potential public-safety danger and the extent of actual or anticipated financial loss or reputational harm to the rights holder/referrer; continuing offending; realistic asset recovery; and prospects of identifying suspects and prosecuting successfully. Historical offending is exceptional. It excludes certain product categories and normally requires the original agency to refer previously investigated matters.
\nPDF6–8 expect investigative groundwork, original evidence and explained inferences from referrers. PDF9 describes DI refusal letters giving reasons, alternatives and a ten-working-day decision aim, with contact if complexity/information needs prevent it. Accepted referrers must support evidential proceedings. PDF10 permits proportionate focus/disruption, CPS cooperation and preservation of relevant material.
\nThe guide is undated in the inspected text. PDF2 mentions October2014 funding; it describes an old referral form. The search engine's age estimate is not a publication date. These are historical posted rules, not established current Gateway procedure.
\nThe accepted [12 August2025 announcement](http[local research file] expressly replaced the PIPCU form from 11August. Initial assessment and prioritisation now sit with IPO's Gateway team, which may develop intelligence or refer onward. It supplies no weighted scoring scheme, commitment to the old ten-day target, industry vote, complaint escalation or adoption of the guide's exact criteria. The police guide's case-acceptance provisions also should not be imposed on every Ashiko bulk suspension: the separately posted2020 policy supplies its own DCI suspension-authorization rule.
\nA later official implementation account is available in the [IPO annual report for 2022/23](http[local research file] Counter infringement, reader427–441. It describes reformed IPCG structures as existing: a mixed enforcement/industry delivery group, an industry advisory committee, and an executive board of senior industry/enforcement decisionmakers. The report links these arrangements to tactical cooperation and strategic coordination. This advances beyond the earlier strategy's promise to build coordination, but is the institution's report of implementation, not an acquired constitution, minutes or observed case decision.
\nThe same section says an internal Strategic Threat Assessment was completed in November2022 while a public version was being developed; an external monitoring/evaluation commission aimed for May2023. Neither the assessment nor the evaluation output was acquired here. The report period is 1April2022–31March2023, with publication identified as July2023; the inspected HTML says updated9January2025. Its present-tense descriptions are not September2026 currency guarantees.
\nThe IPCG executive board, PIPCU Steering Group and IPO Gateway are not identified as one body. No acquired instrument assigns the IPCG industry representatives authority to score individual Gateway referrals or sign police suspension requests. Nor is an industry participant's police-database access implied. Root separately owns the already-retained2022 strategy and FY2025/26 enforcement-capability comparison; their promises and reported implementation remain source-specific.
\nIndustry participation in choosing a broad remit and a harm factor explicitly including referrer interests are substantive agenda-setting mechanisms. The guide also expects evidence, police explanations, proportionate action and prosecution-service involvement. Those are concrete counterweights on paper, not independent proof of each decision's quality. Neither industry command nor complete immunity from industry influence follows.
\nThe acquired public texts do not provide a current Gateway conflict/recusal protocol, weights, escalation rights, performance assessment or an executed allocation of authority between these bodies. The historical guide does not state committee quorum, voting shares or recusal rules. This bounded failure to locate them is not proof they do not exist, were withheld, or were ignored. No general policing or corporate governance rule was substituted for a Gateway-specific instrument.
\nThe next discriminating records are the current Gateway assessment/priority SOP and reasoned allocation or refusal template, and the PIPCU Steering Group/IPCG terms of reference, continuity decision and conflict register or recusal provision. Likely holders are IPO Intelligence and Law Enforcement and City of London Police/PIPCU. A redacted tasking decision could distinguish industry-informed harm criteria from preferential attention to contributors. No approach to those holders is authorized or attempted here.
\nThe new guide was read in the official reader at cover/PDF1; introduction/history/PDF2–3; broad priorities/PDF4; all six acceptance factors/PDF5; referral evidential instructions/PDF6–8; response/accountability/PDF9; resource/proportionality/preservation provisions/PDF10; and feedback statement/PDF12. Academy promotional material/PDF11 is excluded. This is not an executed agreement or a facsimile read. One ordinary public GET returned403; the response bytes are retained as an error, never the guide's original. The successful reader is retained in priority-referral-guide.reader.txt. That route is closed, with no variant or retry.
\nThe annual-report ordinary GET succeeded. Retained priority-ipo2023.html SHA256 is b7f1c94f0df2269776c32aeff112d2f7af7a3ee9da51c1f2046164b40486ab16. I read its complete Counter infringement section and period/publication/update blocks directly from the original. International visits and unrelated operations appearing in that section are not new findings. No whole annual-report read is claimed. The original, scoped extraction and scoped official-reader capture are distinct in priority-captures.json.
The August2025 original and2020 police-reader passages were reused locally from the accepted resources wave, not reacquired. Ten focused discovery queries covered exact Gateway/criteria, PIPCU priorities/conflicts/tasking and the exact Steering Group name, followed by the one permitted implementation paragraph. Search results for other police units, Creative, generic prioritisation frameworks, other local-authority intelligence gateways, corporate schemes and the operational form were not opened as substitute criteria. A police search result describing a2025 medicines case is discovery only, not a new case finding or proof the old exclusions were formally revoked. No current guide-adoption or Gateway scoring instrument was located in this finite pass. No branch follows before the combined forest.
", "path": "pipcu-priorities-wave-2026-09-09/priority-rules.md", "sourceIds": [ "w29-enf-s-referral-guide", "w29-enf-s-ipo-report2023" ], "markdown": "Observed 9 September 2026. Acquisition CLOSED. One new police-guide official-reader record and one new government annual-report original, compared with the accepted August2025 Gateway announcement and posted2020 Ashiko policy. No referral form, account, contact, operational endpoint, non-public appendix or visible UI was accessed.\n\n## The concrete new finding\n\nThe posted PIPCU Referral Guide (see source references) PDF4, says a Steering Group comprising IPO, City of London Police, rights holders and industry bodies agreed broad priorities: physical goods with an online connection and digital content. This is an explicit shared priority-setting role, not just consultation attendance.\n\nPDF5 lists six acceptance considerations: remit; organised activity; potential public-safety danger and the extent of actual or anticipated financial loss or reputational harm to the rights holder/referrer; continuing offending; realistic asset recovery; and prospects of identifying suspects and prosecuting successfully. Historical offending is exceptional. It excludes certain product categories and normally requires the original agency to refer previously investigated matters.\n\nPDF6–8 expect investigative groundwork, original evidence and explained inferences from referrers. PDF9 describes DI refusal letters giving reasons, alternatives and a ten-working-day decision aim, with contact if complexity/information needs prevent it. Accepted referrers must support evidential proceedings. PDF10 permits proportionate focus/disruption, CPS cooperation and preservation of relevant material.\n\nThe guide is undated in the inspected text. PDF2 mentions October2014 funding; it describes an old referral form. The search engine's age estimate is not a publication date. These are historical posted rules, not established current Gateway procedure.\n\n## What the later evidence does and does not establish\n\nThe accepted 12 August2025 announcement (see source references) expressly replaced the PIPCU form from 11August. Initial assessment and prioritisation now sit with IPO's Gateway team, which may develop intelligence or refer onward. It supplies no weighted scoring scheme, commitment to the old ten-day target, industry vote, complaint escalation or adoption of the guide's exact criteria. The police guide's case-acceptance provisions also should not be imposed on every Ashiko bulk suspension: the separately posted2020 policy supplies its own DCI suspension-authorization rule.\n\nA later official implementation account is available in the IPO annual report for 2022/23 (see source references) Counter infringement, reader427–441. It describes reformed IPCG structures as existing: a mixed enforcement/industry delivery group, an industry advisory committee, and an executive board of senior industry/enforcement decisionmakers. The report links these arrangements to tactical cooperation and strategic coordination. This advances beyond the earlier strategy's promise to build coordination, but is the institution's report of implementation, not an acquired constitution, minutes or observed case decision.\n\nThe same section says an internal Strategic Threat Assessment was completed in November2022 while a public version was being developed; an external monitoring/evaluation commission aimed for May2023. Neither the assessment nor the evaluation output was acquired here. The report period is 1April2022–31March2023, with publication identified as July2023; the inspected HTML says updated9January2025. Its present-tense descriptions are not September2026 currency guarantees.\n\nThe IPCG executive board, PIPCU Steering Group and IPO Gateway are **not identified as one body**. No acquired instrument assigns the IPCG industry representatives authority to score individual Gateway referrals or sign police suspension requests. Nor is an industry participant's police-database access implied. Root separately owns the already-retained2022 strategy and FY2025/26 enforcement-capability comparison; their promises and reported implementation remain source-specific.\n\n## Accountability and competing interpretations\n\nIndustry participation in choosing a broad remit and a harm factor explicitly including referrer interests are substantive agenda-setting mechanisms. The guide also expects evidence, police explanations, proportionate action and prosecution-service involvement. Those are concrete counterweights on paper, not independent proof of each decision's quality. Neither industry command nor complete immunity from industry influence follows.\n\nThe acquired public texts do not provide a current Gateway conflict/recusal protocol, weights, escalation rights, performance assessment or an executed allocation of authority between these bodies. The historical guide does not state committee quorum, voting shares or recusal rules. This bounded failure to locate them is not proof they do not exist, were withheld, or were ignored. No general policing or corporate governance rule was substituted for a Gateway-specific instrument.\n\nThe next discriminating records are the **current Gateway assessment/priority SOP and reasoned allocation or refusal template**, and the **PIPCU Steering Group/IPCG terms of reference, continuity decision and conflict register or recusal provision**. Likely holders are IPO Intelligence and Law Enforcement and City of London Police/PIPCU. A redacted tasking decision could distinguish industry-informed harm criteria from preferential attention to contributors. No approach to those holders is authorized or attempted here.\n\n## Custody and finite scope\n\nThe new guide was read in the official reader at cover/PDF1; introduction/history/PDF2–3; broad priorities/PDF4; all six acceptance factors/PDF5; referral evidential instructions/PDF6–8; response/accountability/PDF9; resource/proportionality/preservation provisions/PDF10; and feedback statement/PDF12. Academy promotional material/PDF11 is excluded. This is not an executed agreement or a facsimile read. One ordinary public GET returned403; the response bytes are retained as an error, never the guide's original. The successful reader is retained in priority-referral-guide.reader.txt. That route is closed, with no variant or retry.\n\nThe annual-report ordinary GET succeeded. Retained priority-ipo2023.html SHA256 is `b7f1c94f0df2269776c32aeff112d2f7af7a3ee9da51c1f2046164b40486ab16`. I read its complete Counter infringement section and period/publication/update blocks directly from the original. International visits and unrelated operations appearing in that section are not new findings. No whole annual-report read is claimed. The original, scoped extraction and scoped official-reader capture are distinct in priority-captures.json.\n\nThe August2025 original and2020 police-reader passages were reused locally from the accepted resources wave, not reacquired. Ten focused discovery queries covered exact Gateway/criteria, PIPCU priorities/conflicts/tasking and the exact Steering Group name, followed by the one permitted implementation paragraph. Search results for other police units, Creative, generic prioritisation frameworks, other local-authority intelligence gateways, corporate schemes and the operational form were not opened as substitute criteria. A police search result describing a2025 medicines case is discovery only, not a new case finding or proof the old exclusions were formally revoked. No current guide-adoption or Gateway scoring instrument was located in this finite pass. No branch follows before the combined forest.\n", "file": "research/pipcu-priorities-wave-2026-09-09-priority-rules.md" }, { "id": "pipcu-resources-wave-2026-09-09-funding-resources", "title": "PIPCU funding, staffing and financial conditions", "description": "Observed 9 September 2026. Acquisition CLOSED after five useful primary originals and a scoped read of the parallel lane's strategy. Amounts are GBP. PIPCU is treated as a City of London Police unit, not an independently incorporated recipient or a proxy for every City-funded organisation.", "html": "Observed 9 September 2026. Acquisition CLOSED after five useful primary originals and a scoped read of the parallel lane's retained strategy. Amounts are GBP. PIPCU is treated as a City of London Police unit, not an independently incorporated recipient or a proxy for every City-funded organisation.
\nThe strongest finding is a documented public financing route, alongside a future industry co-funding proposal and a historical industry secondment. The records establish resources and their intended purpose. They do not establish that brands buy a right to direct police investigations, or that the announced pilot is already operating.
\n| Record and stage | \nPayer/recipient path stated by source | \nAmount and period | \nBoundary | \n
|---|---|---|---|
| IPO FY2025/26 accounts, Note 7: reported payment | \nIPO additional dividend to DSIT, used to cover PIPCU costs | \n£2,533,000 paid in FY2025/26; £2,624,000 comparative in FY2024/25 | \nThe cost coverage is on an actuals basis since the preceding additional dividend. This is not a PIPCU bank-receipt record or an Ashiko allocation. | \n
| 9 March 2026 joint IPO/DSIT announcement: confirmed baseline funding commitment | \nDSIT via IPO for PIPCU | \nApproximately £3,000,000 per year, April 2026–March 2029 | \nThree-year funding agreement is reported by the Commissioner; its executed terms and payment schedule were not acquired. | \n
| Same announcement: planned pilot | \nIPO and PIPCU to develop collaborative industry co-funding | \nPlanned from April 2027; contribution amounts and participants not specified | \nAdditional enforcement activity is proposed. No pilot payment, participation agreement or exercised donor instruction established. | \n
The [IPO accounts](http[local research file] published 9 July 2026 and covering 1 April 2025–31 March 2026, distinguish the additional dividend from the ordinary dividend. Note 1(o) explains fees/cost recovery and DSIT's ability to withdraw surplus through additional dividends. Note 7 and its cashflow paragraph explicitly identify the £2.533m payment. This is evidence of a public financial transaction funded through the IPO trading-fund structure; it is not a reconciliation of subsequent DSIT/IPO/police transfers. Do not substitute the £5.470m ordinary dividend due for the additional PIPCU-related payment, or treat cost coverage since the last dividend as an independently established unit expenditure total for precisely the same financial year.
\nThe [9 March 2026 announcement](http[local research file] substantive body and Background 3–4, says IPO and PIPCU are working together on pilot priorities and contribution levels. The Commissioner describes public enforcement powers/strategic oversight alongside industry expertise/intelligence; the Alliance for IP expresses willingness to engage. That is a published allocation of proposed roles, not an operational approval clause or proof a private contribution has been accepted. The baseline and pilot also must not be added to the earlier cost-dividend as though all were separate disbursed awards.
\nThe [Revenue and Capital Budget 2025/26 paper](http[local research file] is dated 10 February 2025 for the Resources, Risks & Estates Committee and 12 February 2025 for the Police Authority Board. PDF 1 marks the Board item for decision; the acquired document is a proposed budget, not the approving minute.
\nPDF 9 paragraph 12 combines a £0.5m increase in contributions for three funded units—PIPCU, IFED and DCPCU—to meet rising costs. It says unrealised income would require corresponding funded-unit pay/non-pay reductions. No PIPCU-specific share is supplied. This is a documented resource constraint, not a direction to select a particular brand's case.
\nPDF 17 paragraph 31 expressly locates detailed income in non-public Appendix 3, citing sensitive funding streams. PDF 18/Table 9 aggregates staffing by broad business area; it is not a PIPCU establishment. No attempt was made to access the non-public appendix. The paragraph 12 expansion of PIPCU as Police Intellectual Property Office is a source naming anomaly; it does not create a second institution. None of the other units' resources is attributed to PIPCU here.
\nA [24 May 2018 draft financial-plan refresh](http[local research file] PDF 1 and Appendix B/PDF 11, separately forecasts PIPCU–IPO income of £1.613m in each year 2018/19–2022/23, marked unconfirmed. Its accompanying statement says sponsorship/grant agreements should rise with pay awards and that funding arrangements are under review. These are historical planning assumptions and a review statement, not proof those annual amounts were received or that any executed agreement contained automatic indexation. The table was visually checked.
\nThe [23 October 2014 government announcement](http[local research file] Notes to Editors 1–4, reports IPO's initial £2.56m over two years from September 2013 and a further £3m to 2017. It calls PIPCU operationally independent and reports a 21-person team, with added expertise from two secondees: an IPO Senior Intelligence Officer and a BPI Internet Investigator. I have not calculated an unqualified 23-person establishment from that wording.
\nThe BPI secondment is a concrete industry-resource connection, stronger than attendance at an event or a logo. Its duration, employer/payroll allocation, access conditions, reporting line and case-selection rights are absent from the acquired source. It must not become a quantified private donation or a current staffing claim. The historical ministerial announcement of additional money is also distinct from a payment ledger.
\nThe parallel lane's retained [2022–2027 counter-infringement strategy](http[local research file] Operation Ashiko paragraph, says one detective managed that operation under PIPCU management oversight, with no charge to referrers or brands. This is a dated operational resource account, not the unit's total or current staffing. It also does not exclude every possible resource contribution by industry. Its reported output counts and referral mechanics are owned by the parallel operational packet.
\nNo executed IPO/DSIT–police grant agreement, PIPCU-specific payment/receipt schedule, performance-return acceptance, audit/clawback clause, secondment agreement or pilot participation instrument was acquired. The accounts' actual-cost basis should not be promoted into an inspected invoice-approval procedure. Likewise, the police paper's conditional budget reduction establishes financial dependence without identifying a funder veto over individual investigations.
\nThe most useful next documents are the announced 2026–2029 funding agreement and financial/assurance schedule, held by IPO/DSIT and City of London Police; the BPI secondment instrument, held by BPI and the police; and, if publicly issued, the 2027 pilot specification, including eligibility, additionality, priority setting, conflict controls and who accepts contributions. The public budget specifically identifies a non-public detailed schedule; this is a precise record boundary, not proof all grant conditions are secret. No pilot-design or other-unit investigation was opened in this run.
\nThe defensible competing interpretation is publicly funded policing using private expertise and proposed additional capacity, with case approval governed separately. That interpretation does not remove influence through agenda-setting or inputs; it requires tracing those mechanisms instead of inferring command from money or a secondment alone.
\nFive original files are retained. All ordinary GETs succeeded; the 2018 financial-plan reader failure was resolved by its one ordinary public GET. No closed Nominet or legislation route was reopened.
\n| Retained original | \nActually read | \nSHA256 | \n
|---|---|---|
| funding-ipo2026.html | \nOriginal HTML Notes 1(o), 7 in full; official reader period/title, agency status and matching notes. No full annual-report read. | \n273ff627e0c2c5b987af622c44aed602c1f7dbf6eb536d12ff573369764c61a4 | \n
| funding-commitment2026.html | \nComplete substantive article, quotes and Background; matched retained original after official-reader read. | \nf6278214cffadb0133aaf7bfec157d7fe120b87adfcb92b824b9d8340ceb927c | \n
| funding-announcement2014.html | \nComplete substantive article and Notes to Editors; matched retained original. | \nb99f979b89bd948659ab33e05fab5880099162386c0eae8281e6692098ce36ca | \n
| funding-colp-budget2025.pdf | \nPDF 1–2,9,17–18; PDF 9 and17 visually checked. 30 pages captured, not fully read. | \n48b1d02f81b917f72fbb01fc85be24d00e1e177b5be1a31ea4c0ba866931a34a | \n
| funding-colp-mtfp.pdf | \nPDF 1,11–12; PDF 11 visually checked. 13 pages captured, not fully read. | \nfe56ce21f21683d1aed7d5e3701a11f2cd9456c3862917add77e0476d7395c55 | \n
The shared strategy original, referral-strategy2022.html, SHA256 9dcb6d69ad37942b0b4005543be05e9f4561eb854659d149e0f95510067e0654, was reused without reacquisition or modification. This lane read the complete Ashiko paragraph and neighbouring transition context only. A search result's old CAPTCHA-token PIPCU URL and 19-person claim were not opened or used as current staffing. Search-only earlier accounts and unrelated unit records are discovery, not findings. funding-captures.json contains exact URLs, GET observations, reading states, derivative ancestry and source hashes. Acquisition is closed pending the combined forest; no case, current-work, forest, atlas or ZIP edits were made.
", "path": "pipcu-resources-wave-2026-09-09/funding-resources.md", "sourceIds": [ "w29-enf-s-ipo-accounts2026", "w29-enf-s-police-budget2025", "w29-enf-s-police-forecast2018", "w29-enf-s-funding2026", "w29-enf-s-funding2014" ], "markdown": "Observed 9 September 2026. Acquisition CLOSED after five useful primary originals and a scoped read of the parallel lane's retained strategy. Amounts are GBP. PIPCU is treated as a City of London Police unit, not an independently incorporated recipient or a proxy for every City-funded organisation.\n\nThe strongest finding is a documented public financing route, alongside a future industry co-funding proposal and a historical industry secondment. The records establish resources and their intended purpose. They do not establish that brands buy a right to direct police investigations, or that the announced pilot is already operating.\n\n## Current commitment and actual financial transaction\n\n| Record and stage | Payer/recipient path stated by source | Amount and period | Boundary |\n| --- | --- | --- | --- |\n| IPO FY2025/26 accounts, Note 7: reported payment | IPO additional dividend to DSIT, used to cover PIPCU costs | £2,533,000 paid in FY2025/26; £2,624,000 comparative in FY2024/25 | The cost coverage is on an actuals basis since the preceding additional dividend. This is not a PIPCU bank-receipt record or an Ashiko allocation. |\n| 9 March 2026 joint IPO/DSIT announcement: confirmed baseline funding commitment | DSIT via IPO for PIPCU | Approximately £3,000,000 per year, April 2026–March 2029 | Three-year funding agreement is reported by the Commissioner; its executed terms and payment schedule were not acquired. |\n| Same announcement: planned pilot | IPO and PIPCU to develop collaborative industry co-funding | Planned from April 2027; contribution amounts and participants not specified | Additional enforcement activity is proposed. No pilot payment, participation agreement or exercised donor instruction established. |\n\nThe IPO accounts (see source references) published 9 July 2026 and covering 1 April 2025–31 March 2026, distinguish the additional dividend from the ordinary dividend. Note 1(o) explains fees/cost recovery and DSIT's ability to withdraw surplus through additional dividends. Note 7 and its cashflow paragraph explicitly identify the £2.533m payment. This is evidence of a public financial transaction funded through the IPO trading-fund structure; it is not a reconciliation of subsequent DSIT/IPO/police transfers. Do not substitute the £5.470m ordinary dividend due for the additional PIPCU-related payment, or treat cost coverage since the last dividend as an independently established unit expenditure total for precisely the same financial year.\n\nThe 9 March 2026 announcement (see source references) substantive body and Background 3–4, says IPO and PIPCU are working together on pilot priorities and contribution levels. The Commissioner describes public enforcement powers/strategic oversight alongside industry expertise/intelligence; the Alliance for IP expresses willingness to engage. That is a published allocation of proposed roles, not an operational approval clause or proof a private contribution has been accepted. The baseline and pilot also must not be added to the earlier cost-dividend as though all were separate disbursed awards.\n\n## Police budget: resource dependence and a specific non-public schedule\n\nThe Revenue and Capital Budget 2025/26 paper (see source references) is dated 10 February 2025 for the Resources, Risks & Estates Committee and 12 February 2025 for the Police Authority Board. PDF 1 marks the Board item for decision; the acquired document is a proposed budget, not the approving minute.\n\nPDF 9 paragraph 12 combines a £0.5m increase in contributions for three funded units—PIPCU, IFED and DCPCU—to meet rising costs. It says unrealised income would require corresponding funded-unit pay/non-pay reductions. No PIPCU-specific share is supplied. This is a documented resource constraint, not a direction to select a particular brand's case.\n\nPDF 17 paragraph 31 expressly locates detailed income in **non-public Appendix 3**, citing sensitive funding streams. PDF 18/Table 9 aggregates staffing by broad business area; it is not a PIPCU establishment. No attempt was made to access the non-public appendix. The paragraph 12 expansion of PIPCU as Police Intellectual Property Office is a source naming anomaly; it does not create a second institution. None of the other units' resources is attributed to PIPCU here.\n\nA 24 May 2018 draft financial-plan refresh (see source references) PDF 1 and Appendix B/PDF 11, separately forecasts PIPCU–IPO income of £1.613m in each year 2018/19–2022/23, marked unconfirmed. Its accompanying statement says sponsorship/grant agreements should rise with pay awards and that funding arrangements are under review. These are historical planning assumptions and a review statement, not proof those annual amounts were received or that any executed agreement contained automatic indexation. The table was visually checked.\n\n## Staff resources: named historical contribution, no current headcount inference\n\nThe 23 October 2014 government announcement (see source references) Notes to Editors 1–4, reports IPO's initial £2.56m over two years from September 2013 and a further £3m to 2017. It calls PIPCU operationally independent and reports a 21-person team, with added expertise from two secondees: an IPO Senior Intelligence Officer and a BPI Internet Investigator. I have not calculated an unqualified 23-person establishment from that wording.\n\nThe BPI secondment is a concrete industry-resource connection, stronger than attendance at an event or a logo. Its duration, employer/payroll allocation, access conditions, reporting line and case-selection rights are absent from the acquired source. It must not become a quantified private donation or a current staffing claim. The historical ministerial announcement of additional money is also distinct from a payment ledger.\n\nThe parallel lane's retained 2022–2027 counter-infringement strategy (see source references) Operation Ashiko paragraph, says one detective managed that operation under PIPCU management oversight, with no charge to referrers or brands. This is a dated operational resource account, not the unit's total or current staffing. It also does not exclude every possible resource contribution by industry. Its reported output counts and referral mechanics are owned by the parallel operational packet.\n\n## Acceptance rights and the remaining discriminator\n\nNo executed IPO/DSIT–police grant agreement, PIPCU-specific payment/receipt schedule, performance-return acceptance, audit/clawback clause, secondment agreement or pilot participation instrument was acquired. The accounts' actual-cost basis should not be promoted into an inspected invoice-approval procedure. Likewise, the police paper's conditional budget reduction establishes financial dependence without identifying a funder veto over individual investigations.\n\nThe most useful next documents are the announced **2026–2029 funding agreement and financial/assurance schedule**, held by IPO/DSIT and City of London Police; the **BPI secondment instrument**, held by BPI and the police; and, if publicly issued, the **2027 pilot specification**, including eligibility, additionality, priority setting, conflict controls and who accepts contributions. The public budget specifically identifies a non-public detailed schedule; this is a precise record boundary, not proof all grant conditions are secret. No pilot-design or other-unit investigation was opened in this run.\n\nThe defensible competing interpretation is publicly funded policing using private expertise and proposed additional capacity, with case approval governed separately. That interpretation does not remove influence through agenda-setting or inputs; it requires tracing those mechanisms instead of inferring command from money or a secondment alone.\n\n## Reading scope and custody\n\nFive original files are retained. All ordinary GETs succeeded; the 2018 financial-plan reader failure was resolved by its one ordinary public GET. No closed Nominet or legislation route was reopened.\n\n| Retained original | Actually read | SHA256 |\n| --- | --- | --- |\n| funding-ipo2026.html | Original HTML Notes 1(o), 7 in full; official reader period/title, agency status and matching notes. No full annual-report read. | 273ff627e0c2c5b987af622c44aed602c1f7dbf6eb536d12ff573369764c61a4 |\n| funding-commitment2026.html | Complete substantive article, quotes and Background; matched retained original after official-reader read. | f6278214cffadb0133aaf7bfec157d7fe120b87adfcb92b824b9d8340ceb927c |\n| funding-announcement2014.html | Complete substantive article and Notes to Editors; matched retained original. | b99f979b89bd948659ab33e05fab5880099162386c0eae8281e6692098ce36ca |\n| funding-colp-budget2025.pdf | PDF 1–2,9,17–18; PDF 9 and17 visually checked. 30 pages captured, not fully read. | 48b1d02f81b917f72fbb01fc85be24d00e1e177b5be1a31ea4c0ba866931a34a |\n| funding-colp-mtfp.pdf | PDF 1,11–12; PDF 11 visually checked. 13 pages captured, not fully read. | fe56ce21f21683d1aed7d5e3701a11f2cd9456c3862917add77e0476d7395c55 |\n\nThe shared strategy original, referral-strategy2022.html, SHA256 **9dcb6d69ad37942b0b4005543be05e9f4561eb854659d149e0f95510067e0654**, was reused without reacquisition or modification. This lane read the complete Ashiko paragraph and neighbouring transition context only. A search result's old CAPTCHA-token PIPCU URL and 19-person claim were not opened or used as current staffing. Search-only earlier accounts and unrelated unit records are discovery, not findings. funding-captures.json contains exact URLs, GET observations, reading states, derivative ancestry and source hashes. Acquisition is closed pending the combined forest; no case, current-work, forest, atlas or ZIP edits were made.\n", "file": "research/pipcu-resources-wave-2026-09-09-funding-resources.md" }, { "id": "pipcu-resources-wave-2026-09-09-referral-decision", "title": "Operation Ashiko: private referrals, police approval and registry action", "description": "Observed 9 September 2026. Acquisition CLOSED. Four primary substantive records, plus the previously accepted registry-policy packet. Local overlap was completed by root before acquisition; no existing case-r27 Ashiko/PIPCU coverage was found. No complaint form, domain, harmful material, account, paid route or operational endpoint was used.", "html": "Observed 9 September 2026. Acquisition CLOSED. Four primary substantive records, plus the previously accepted registry-policy packet. Local overlap was completed by root before acquisition; no existing case-r27 Ashiko/PIPCU coverage was found. No complaint form, domain, harmful material, account, paid route or operational endpoint was used.
\nThe public chain has several decision points. Rights holders supply intelligence and evidence; a police policy assigns approval to a senior police officer; registrars and Nominet act under their own terms. A later IPO announcement adds initial assessment and prioritisation before possible law-enforcement referral. These records establish a process and reported activity, not a private complainant's right to compel suspension or proof that the safeguards worked in an individual case.
\nThe official [Operation Ashiko Suspension Policy](http[local research file] dated 22 June 2020, version 1.2–20, identifies individual brands and enforcement agencies as intelligence/evidence sources (§2, PDF4). It describes a Nominet agreement for action without a court order, contingent on police investigation and due diligence (§4, PDF6–7). If necessary, brands, representatives or trade bodies confirm infringement (§8, PDF9). Every submitted suspension requires PIPCU Detective Chief Inspector authorisation (§9, PDF9).
\nThe notice asks the registrar to investigate terms breaches and act; after 48 hours without registrar action, PIPCU requests Nominet investigate and act if it finds a breach (§7, PDF8). Its requested duration is a minimum of 24 months, or the domain expiry if sooner. A registrant alleging error should contact PIPCU within the first 48 hours; police reassess and consider remediation, responding only to the WHOIS registrant or documented representative (§10, PDF9). Section11, PDF10 points to police complaints. The policy does not document an actual reversal here.
\nThis is the posted 2020 instrument, not an executed inter-party agreement or proof every clause still governs after later changes. The current-linked 2026 CPP is separately analysed in policy-authority.md. In particular, the older expiry wording should not silently replace the current CPP's .UK minimum-duration rule.
\nThe IPO's [12 August 2025 announcement](http[local research file] says that from 11 August 2025 the IPO Intelligence Gateway replaces the previous PIPCU referral form for IP rights holders and enforcement partners. Its team first assesses and prioritises submissions. It can develop intelligence itself or refer to a law-enforcement agency; it liaises with PIPCU, Trading Standards and other agencies about possible disruption. College of Policing-trained intelligence officers/managers work with IPO analysts, financial investigators and regional IP Crime Coordinators. Public reporting is separately directed to Trading Standards or Crimestoppers.
\nThus a current description that sends every brand directly to the old PIPCU form would be stale. The announcement does not amend Ashiko's named suspension approver, guarantee referral, publish priority scores or establish that every possible intelligence channel must use this form. The form link was observed but neither opened nor used. Read the complete substantive body and publication/update date, reader lines53–75; the retained HTML corroborates them.
\nThe PIPCU contribution to the official [IP Crime and Enforcement Report 2018–19](http[local research file] PDF59/printed55, describes brand-owner and supplier referrals supported by counterfeit-trading evidence, imaging of websites for evidence, then details sent to Nominet. It reports 32,995 websites suspended for 2018/19. This is an attributed police contribution in a government publication, not independent domain-by-domain verification; exact start/end dates beyond its period label are not supplied in that passage.
\nThe adjacent 32,927 requests / 32,669 from PIPCU cite a November 2018 Nominet report in footnote35. The same volume's Nominet spread, PDF56/printed52, explicitly covers 1 November 2017–31 October 2018. Therefore the figures must not be treated as matched inputs and outputs for one period, or subtracted to produce an error/refusal rate. Both pages were visually inspected from retained original bytes. The neighbouring Operation Creative/IWL and Operation Methone material concerns separate interventions and is excluded from the Ashiko mechanism.
\nThe IPO's [4 February 2022 strategy](http[local research file] the Operation Ashiko paragraph under future challenges, reports over115,000 sites removed from .UK, Nominet administrative checks and registrar implementation. It describes one detective with PIPCU management oversight and no charge to brands/referrers. These are a dated account of resources and delivery, not current headcount, a unit budget, proof of negligible total social cost or a cumulative total to September2026. Its preceding industry-intelligence/IWL paragraph describes Operation Creative; that stronger wording is not imported into Ashiko. The later all-TLD cumulative account root read in the 2025 Impact Issue is likewise not an annual Nominet-only count.
\nThe evidence separates three questions: who supplies allegations, who approves a police request, and who implements a contractual registry decision. An investigator's reliance on private expertise can influence which sites are examined without granting that supplier approval power. Conversely, the existence of a police sign-off rule does not establish the independence or quality of each assessment. No original brand dossier, completed DCI approval, particular registry determination, named error reversal or reinstatement trace was acquired.
\nThe finite next discriminator is a redacted end-to-end administrative file: referral and supporting evidence, IPO triage if applicable, police due-diligence/sign-off record, registrar/Nominet determination and any challenge/remediation. The underlying police–Nominet agreement and the current Gateway prioritisation criteria would resolve a narrower rules-versus-practice gap. Likely custodians are PIPCU/City of London Police, IPO Intelligence Gateway and Nominet registry/legal operations. No contact was attempted or authorised. Funding, staffing and financial conditions remain the parallel finance packet, not inferred from brand participation.
\nreferral-captures.json records source URLs, dates, exact reading scopes, original and derivative hashes. Original official HTML was retained for the Gateway and strategy; the government-published 136-page 2018–19 report was retained, but only its cover, PDF56, PDF58 attribution context and PDF59 were text-inspected, with PDF56 and59 visually inspected. No whole-report read is claimed. The full Gateway substantive body and the strategy's Ashiko paragraph/publication date were read from originals; other returned strategy passages do not support new findings.
\nThe police PDF's ordinary GET returned403; that route is closed. Its substantive policy sections2–11, PDF4–10, were read through the successful official reader, plus cover/date. No original PDF hash or facsimile inspection is claimed for that policy. Later generic privacy pages were incidentally returned but were not audited or treated as an Ashiko-specific correction protocol. Existing Nominet failures were not retried. Four focused discovery queries were used; third-party forum allegations, biographies and unrelated Ashiko/IWF names were excluded. No new branch follows this packet before root's forest review.
", "path": "pipcu-resources-wave-2026-09-09/referral-decision.md", "sourceIds": [ "w29-enf-s-ashiko-policy2020", "w29-enf-s-gateway2025", "w29-enf-s-delivery2019", "w29-enf-s-strategy2022" ], "markdown": "Observed 9 September 2026. Acquisition CLOSED. Four primary substantive records, plus the previously accepted registry-policy packet. Local overlap was completed by root before acquisition; no existing case-r27 Ashiko/PIPCU coverage was found. No complaint form, domain, harmful material, account, paid route or operational endpoint was used.\n\nThe public chain has several decision points. Rights holders supply intelligence and evidence; a police policy assigns approval to a senior police officer; registrars and Nominet act under their own terms. A later IPO announcement adds initial assessment and prioritisation before possible law-enforcement referral. These records establish a process and reported activity, not a private complainant's right to compel suspension or proof that the safeguards worked in an individual case.\n\n## The actual police administrative instrument\n\nThe official Operation Ashiko Suspension Policy (see source references) dated 22 June 2020, version 1.2–20, identifies individual brands and enforcement agencies as intelligence/evidence sources (§2, PDF4). It describes a Nominet agreement for action without a court order, contingent on police investigation and due diligence (§4, PDF6–7). If necessary, brands, representatives or trade bodies confirm infringement (§8, PDF9). Every submitted suspension requires PIPCU Detective Chief Inspector authorisation (§9, PDF9).\n\nThe notice asks the registrar to investigate terms breaches and act; after 48 hours without registrar action, PIPCU requests Nominet investigate and act if it finds a breach (§7, PDF8). Its requested duration is a minimum of 24 months, or the domain expiry if sooner. A registrant alleging error should contact PIPCU within the first 48 hours; police reassess and consider remediation, responding only to the WHOIS registrant or documented representative (§10, PDF9). Section11, PDF10 points to police complaints. The policy does not document an actual reversal here.\n\nThis is the posted 2020 instrument, not an executed inter-party agreement or proof every clause still governs after later changes. The current-linked 2026 CPP is separately analysed in [policy-authority.md](nominet-domain-authority-wave-2026-09-09-policy-authority.md). In particular, the older expiry wording should not silently replace the current CPP's .UK minimum-duration rule.\n\n## Initial intake changed in August 2025\n\nThe IPO's 12 August 2025 announcement (see source references) says that from **11 August 2025** the IPO Intelligence Gateway replaces the previous PIPCU referral form for IP rights holders and enforcement partners. Its team first assesses and prioritises submissions. It can develop intelligence itself or refer to a law-enforcement agency; it liaises with PIPCU, Trading Standards and other agencies about possible disruption. College of Policing-trained intelligence officers/managers work with IPO analysts, financial investigators and regional IP Crime Coordinators. Public reporting is separately directed to Trading Standards or Crimestoppers.\n\nThus a current description that sends every brand directly to the old PIPCU form would be stale. The announcement does not amend Ashiko's named suspension approver, guarantee referral, publish priority scores or establish that every possible intelligence channel must use this form. The form link was observed but neither opened nor used. Read the complete substantive body and publication/update date, reader lines53–75; the retained HTML corroborates them.\n\n## A dated account of delivery and its limits\n\nThe PIPCU contribution to the official IP Crime and Enforcement Report 2018–19 (see source references) **PDF59/printed55**, describes brand-owner and supplier referrals supported by counterfeit-trading evidence, imaging of websites for evidence, then details sent to Nominet. It reports **32,995 websites suspended for 2018/19**. This is an attributed police contribution in a government publication, not independent domain-by-domain verification; exact start/end dates beyond its period label are not supplied in that passage.\n\nThe adjacent **32,927 requests / 32,669 from PIPCU** cite a November 2018 Nominet report in footnote35. The same volume's Nominet spread, **PDF56/printed52**, explicitly covers **1 November 2017–31 October 2018**. Therefore the figures must not be treated as matched inputs and outputs for one period, or subtracted to produce an error/refusal rate. Both pages were visually inspected from retained original bytes. The neighbouring Operation Creative/IWL and Operation Methone material concerns separate interventions and is excluded from the Ashiko mechanism.\n\nThe IPO's 4 February 2022 strategy (see source references) the Operation Ashiko paragraph under future challenges, reports over115,000 sites removed from .UK, Nominet administrative checks and registrar implementation. It describes one detective with PIPCU management oversight and no charge to brands/referrers. These are a dated account of resources and delivery, not current headcount, a unit budget, proof of negligible total social cost or a cumulative total to September2026. Its preceding industry-intelligence/IWL paragraph describes Operation Creative; that stronger wording is not imported into Ashiko. The later all-TLD cumulative account root read in the 2025 Impact Issue is likewise not an annual Nominet-only count.\n\n## What remains unproved\n\nThe evidence separates three questions: who supplies allegations, who approves a police request, and who implements a contractual registry decision. An investigator's reliance on private expertise can influence which sites are examined without granting that supplier approval power. Conversely, the existence of a police sign-off rule does not establish the independence or quality of each assessment. No original brand dossier, completed DCI approval, particular registry determination, named error reversal or reinstatement trace was acquired.\n\nThe finite next discriminator is a redacted end-to-end administrative file: referral and supporting evidence, IPO triage if applicable, police due-diligence/sign-off record, registrar/Nominet determination and any challenge/remediation. The underlying police–Nominet agreement and the current Gateway prioritisation criteria would resolve a narrower rules-versus-practice gap. Likely custodians are PIPCU/City of London Police, IPO Intelligence Gateway and Nominet registry/legal operations. No contact was attempted or authorised. Funding, staffing and financial conditions remain the parallel finance packet, not inferred from brand participation.\n\n## Source custody and closed routes\n\nreferral-captures.json records source URLs, dates, exact reading scopes, original and derivative hashes. Original official HTML was retained for the Gateway and strategy; the government-published 136-page 2018–19 report was retained, but only its cover, PDF56, PDF58 attribution context and PDF59 were text-inspected, with PDF56 and59 visually inspected. No whole-report read is claimed. The full Gateway substantive body and the strategy's Ashiko paragraph/publication date were read from originals; other returned strategy passages do not support new findings.\n\nThe police PDF's ordinary GET returned403; that route is closed. Its substantive policy sections2–11, PDF4–10, were read through the successful official reader, plus cover/date. No original PDF hash or facsimile inspection is claimed for that policy. Later generic privacy pages were incidentally returned but were not audited or treated as an Ashiko-specific correction protocol. Existing Nominet failures were not retried. Four focused discovery queries were used; third-party forum allegations, biographies and unrelated Ashiko/IWF names were excluded. No new branch follows this packet before root's forest review.\n", "file": "research/pipcu-resources-wave-2026-09-09-referral-decision.md" }, { "id": "private-verification-wave-2026-09-09-gifct", "title": "GIFCT: organizational authority, financing and signal correction", "description": "Acquisition closed 9 September 2026. Bounded primary-source run; no database access, harmful-content inspection, account, outreach or payment. No canonical case changes. This packet separates current public descriptions, dated instruments and reported operation. It does not decide the legal effect of Australia's DIS section6; root owns that statutory connect", "html": "Acquisition closed 9 September 2026. Bounded primary-source run; no database access, harmful-content inspection, account, outreach or payment. No canonical case changes. This packet separates current public descriptions, dated instruments and reported operation. It does not decide the legal effect of Australia's DIS section6; root owns that statutory connection and the BSR assessment.
\nGIFCT has a company-led governing body, membership conditions and a disciplinary mechanism controlling access to shared infrastructure. The providers retain responsibility for their own moderation decisions. A shared hash can therefore transmit an upstream classification without itself constituting a universal removal order. Correction inside that infrastructure is documented; downstream restoration remains a separate evidentiary question.
\nCurrent organizational structure. The official governance page identifies Meta, Microsoft and YouTube as founding-company Operating Board members, with Discord and Twitch the elected at-large members for 2026; Meta chairs in 2026. The general membership elects at-large members annually. The Board supplies strategic guidance and oversight. The current IAC roster includes Australia's Department of Home Affairs in an advisory role; that role is not database access. The IAC advises on priorities and performance, has a civil-society majority and a nongovernment chair. The current page states the Operating Board periodically reviews IAC composition. Source: [current governance](http[local research file] substantive lines65–151, observed9September 2026. Current roster is not the 2021 roster.
\nThe official FAQ describes GIFCT as a501(c)(3) nonprofit, daily operations led by Executive Director Naureen Chowdhury Fink, member funding since inception and nonfounder contributions from 2021. This is an organizational self-description; it is not an independently acquired incorporation document. [Explainers](http[local research file] lines69–121 and footer. A bounded secondary tax-index search identifies the probable legal entity Global Internet Forum to Counter Terrorism Inc., EIN 84-4586487. The original return content was not acquired, so that identifier remains a retrieval lead, not independently established primary legal identity here.
\nDated financing and membership changes. The 2025 annual report records $3,760,921 in combined member/external contributions and $3,239,223 expenses, of which $2,296,459 is programming/delivery. Within that category it allocates $215,588 to incident response/HSDB, $1,226,367 member support, $391,669 GNET, $347,193 engagement/events and $115,642 working groups. These are report-level figures, not audited accounts, bank receipts or donor-specific allocations; the table uses the dollar symbol without separately labeling currency. Mandatory membership dues began in 2025. The report says membership support moved from a partner-based model to GIFCT's own MAP; its diagram places confirmation of full membership with GIFCT, members and Operating Board, without publishing a precise voting threshold. X ended membership in 2025, attributed to its focus on internal trust/safety; the source does not report expulsion. Source: [2025 Annual and Transparency Report](http[local research file] PDF 7–9,28,30; printed9–14,46,49. Do not reconstruct a donor's content veto from financial support.
\nThe access gate. The current membership page requires annual tiered dues and six substantive criteria, including a public terrorism/violent-extremism prohibition, a functioning reports/appeals route, relevant technical interest, transparency and human-rights commitments. The public resource guide expressly says GIFCT assists policy drafting but does not prescribe exact wording. Sources: [membership](http[local research file] lines65–87,171–187; resource guide.
\nThe current HSDB page distinguishes exact cryptographic matching from perceptual similarity and describes member-supplied hashes and labels, with agreement/disagreement feedback. It says GIFCT does not dictate enforcement. [HSDB](http[local research file] substantive lines65–89. The 2025 report adds that GIFCT itself can create hashes and limits access to GIFCT plus member technology companies completing the information-sharing agreement. It expressly separates matching from each platform's independent action; its listed companies are integrated or integrating, not all proved operational. Same annual report, PDF 11/printed17.
\nThe currently linked December 2022 v1.0 taxonomy supplies distinct qualifying routes: UN-sanctioned terrorist entities, qualifying perpetrator/incident content, or a behavior framework. The behavioral route requires more than a general ideological label: the document combines nongovernment status, a distinguishable extremist identifier, a hate-based mission and advocacy of civilian/vigilante violence toward that mission. Its URL provision names TAT's TCAP but says GIFCT includes associated hashes only when GIFCT's own inclusion parameters also hold. This is a filter between systems, not automatic import of every TCAP designation. [Taxonomy](http[local research file] PDF 3–5. Its incident terminology is historically dated; do not silently carry 2022 IRF details across the 2025 revision.
\nActual correction, with a meaningful limit. The 2022 report describes a member sampling/review exercise, corrected labels and removed hashes. It says another member sees a contributor's removal the next time it accesses the database. Separately, feedback on two hashes led the contributor to revisit a nonviolent music video and remove the hashes; some hashes were also removed cautiously when originals were no longer retained. Source:2022 report PDF 32–35. Its feedback sample is explicitly nonrandom and cannot establish a whole-database false-positive rate. The 2025 report still describes member review/removal, varying original-content retention and feedback on about2% of hashes. Same annual report PDF 13/printed21. Neither establishes that a removed hash compelled a member to restore a post or account. No universal signal-expiry period or downstream correction service-level agreement was acquired.
\nThe 2025 annual report's PDF 11 link points directly to Tech Solutions Code of Conduct v1.0,2022. Its public version omits detailed acceptable-use examples for operational security. Leadership enforces after consulting the affected organization and in consultation with the Operating Board. Available sanctions run from correction/warning to temporary technology bans, permanent bans or membership revocation, with graduated Board/IAC/public disclosure. An accused organization has10business days after notification to appeal; the Board has30business days to affirm, change or reverse and may consult the IAC. Its determination is final within this process. This is an organizational access appeal, not a platform user's content remedy. Source: [Code](http[local research file] all6pages, especiallyPDF 2–5. The existence of authority is established; no specific disciplinary exercise was proved in this run.
\nAdvisory independence has defined dependencies. The current governance page links a January 2026-path IAC TOR whose first page calls the text draft. Its7pages provide a nongovernment majority, IAC election of its chair, that chair's nonvoting Board participation and remuneration by GIFCT. They provide Board endorsement of appointments, budgeted secretariat/participation support, reciprocal summarized minutes and formal Board responses to IAC recommendations. Conflict provisions require disclosure and recorded recusal for direct interests, while allowing specified exceptions. The chair/secretariat are therefore structurally supported by the organization being advised; this is disclosed architecture, not proof of compromised judgment. Signed/adopted status and actual minutes/recusals were not acquired. [Current-linked TOR](http[local research file] PDF 1–7, especially1–5. The upload directory is not an established enactment date.
\nThe strongest next discriminating record is the information-sharing agreement's correction clause plus one paired correction receipt: originating classification withdrawal, delivery to a receiving platform, its acknowledged review and actual disposition. This would separate interoperable correction from reversible enforcement. GIFCT/member custodians hold those records; the public code does not supply them.
\nFor money and authority, the missing record is current audited/IRS financial detail joined to an actual donor agreement or board instrument identifying budget approval and any earmarked decision rights. Current reports establish resources and membership dues, not per-funder operational commands. The EIN lead above is useful for a later selected finance run; it is not a reason to expand this one.
\nCandidate relations are typed as: Operating Board→GIFCT strategic/governance oversight; member→GIFCT dues obligation; GIFCT→eligible member shared-tool access; member/GIFCT→HSDB taxonomy-qualified signal contribution; HSDB→provider identification input; provider→its content independent policy action; contributor→HSDB reported correction; Board→disciplined organization final internal appeal decision. The IAC→Board relation is advice, not a veto; Home Affairs→IAC is participation, not database access. These distinctions preserve both shared upstream influence and independent recipient discretion. No new nodes or edges were written into the case.
\nAll nine originals listed below were acquired by an ordinary unauthenticated GET and hash-checked again before closure. Their complete URLs, timestamps, original/derivative state and exact reading scopes are in gifct-captures.json. Original retention does not imply a full read. Annual-report page numbers here are PDF pages; some PDF pages contain printed spreads. The 2025 report was reached from GIFCT's transparency page; taxonomy from the current HSDB page; TOR from current governance; Code from the report's exact PDF 11 link annotation.
\nResource-guide web-reader text was read only through lines80–131 for the policy and appeals statements. The original page's full body was not retained. The initial Code-link text extraction returned no link; reading the report's own PDF link annotation resolved the document. A large first PDF extraction was truncated; decisive omitted TOR and annual pages were reread in compact page-specific form. The saved gifct-scoped-extracts.txt is a normalized derivative of listed PDF ranges, not an original or a transcript of a full-report read (SHA256 a0bd63ee596773e9e901425dd3bbda85367b2f62e58ec3a7035d0a9c42c58535).
\nLegal-identity route: [ProPublica's IRS-derived index](http[local research file] and [2023 return wrapper](http[local research file] identify the entity/EIN and iframe locations. The two nested IRS990PF/ScheduleB readers returned nonretryable unsafe-open errors. Those routes were closed without alternative retrieval; no tax-return contents or donor names from that unacquired filing are relied upon. The legal identifier is therefore expressly secondary in this packet. No other funder or organization branch was opened. Root's BSR 2021/2026-policy lane and Hubble's TAT lane remain separate.
\n| Original file | \nSHA256 | \n
|---|---|
| gifct-explainers.html | \neeae05820d80252a603dc8e3c7eb34e32bcb137615eda3835c805845eb37b9e5 | \n
| gifct-governance.html | \ncc2fe3e4666a9d9b6ed376047cabb3de743e5f3bbe94325a79fc7b6e2719ce26 | \n
| gifct-membership.html | \nc4a3ff9e97fc38e829bda7116869efe27f9e99c8d6e1b7a60039620bdb7a9f3f | \n
| gifct-hsdb.html | \n00e76f2317df4ab6498c0dbf438b6fec0d472606d11afe23fd3676c40bd89c7d | \n
| gifct-annual2025.pdf | \ndc22bea68d28284bb40b201da8a7e4a227f4eed7ea48eae1d6080383f4ec7e00 | \n
| gifct-taxonomy2022.pdf | \nd4c084c9ac1b330fd72a2c607446fcdd50f0fab93631d2e09b8fef2479a13571 | \n
| gifct-iac-tor2026.pdf | \ne77733eb7bf9fe96453108686c55a9e2708d64df7d0acc7ed6cf4db05f344af8 | \n
| gifct-transparency2022.pdf | \n1b7c2847247fe7388eefe3b0f70d237d6923082f22f01407d32ced2ee2bb8422 | \n
| gifct-code2022.pdf | \n0366c204ea1be5b547ccdd179e0c09cc8b16e6b28512aec6b739bc8c23a94863 | \n
Root lane closed 9 September 2026. This is a bounded comparison of the commissioned 2021 assessment, GIFCT's response and its May 2026 policy. It does not audit individual moderation outcomes.
\n[BSR's 2021 assessment](http[local research file] was paid for by GIFCT with BSR retaining editorial control; its December 2020-May 2021 work primarily assessed future organizational risks, not member-company outcomes. Stakeholders raised risks of government pressure and large-platform policy export. The report also records that some small platforms relied on human review because they lacked automation. It recommended stronger appeals, dispute holds, external database review and transparency about downstream use. Its then-current description assigned hash removal to the contributor and content remedy primarily to each platform. These observations and recommendations cannot establish today's implementation or a particular wrongful removal. Its governance discussion distinguished participation from voting power and recommended a later governance review; it did not itself perform a full governance audit.
\nRead scope: PDF3-10 and start of PDF11 via reader; targeted PDF38-39, PDF41-42, PDF44-48 and PDF63-65 passages, not each whole page or the full73-page report. Web references turn1189view0, turn1191view1/3/4, turn1192view0/1/2. Ordinary original GET returned403; no original PDF retained or hash. Two screenshot attempts (PDF44/45) returned cache misses. Reader-only evidence; visual verification unperformed. No retries.
\n[GIFCT's 20 July 2021 response](http[local research file] grouped its commitments by timeframe. It said membership criteria had already been revised, with appeals required, and identified further remedy capability as work underway. Human review before contribution, third-party review, research access and database governance changes were longer-term commitments or matters for exploration. Review of a multi-stakeholder Operating Board was proposed for two years later. A promise in this letter is not proof of completion. Root read the complete main body, reader lines75-173; retained original HTML in root-hria-captures.json.
\nThe [currently linked policy](http[local research file] retains a 2022 URL but explicitly states that GIFCT's team updated it in May2026. It commits to assessment, action, tracking and annual communication, with an annex assigning those activities across technical tools, incident response, research, working groups and transparency. The annex specifies metrics and a risk register; it does not supply their measurements or a worked user-remedy record. Root read all5 extracted pages and visually inspected the page5 matrix from the retained original. Web screenshot failed; local rendering succeeded. Original65316bytes, SHAdd704b12fc0ccfaa7f2584d9419b58f74aed2cb3a707b6a16706f8e4ae109e7d.
\nThe [18 May 2026 update](http[local research file] says a five-year report against the BSR assessment was being developed. It announces that previously BSR-delivered human-rights services will move into GIFCT's Membership Advisory Program. The statement is forward-looking; completion of that specific transition is not proved here. It also does not establish that every safeguard improved or deteriorated. Root read the full main body, lines75-101; original retained. Two scoped searches for the five-year review returned this announcement, not a completed report. This bounded failure to locate it does not prove nonpublication.
\nThere are identifiable safeguards to inspect and an explicit distinction between the organization and each recipient platform. The revealing test is the chain from disputed source entry to correction notification to reconsideration of the recipient's action. Current policy text alone cannot close that chain. Governance, actual reported corrections and member discipline are separately covered by the GIFCT lane. All acquisition in this lane has stopped before the combined forest.
", "path": "private-verification-wave-2026-09-09/human-rights-followthrough.md", "sourceIds": [ "w26-ver-s-bsr-hria2021", "w26-ver-s-gifct-response2021", "w26-ver-s-gifct-hr-policy2026", "w26-ver-s-gifct-hr-update2026" ], "markdown": "Root lane closed 9 September 2026. This is a bounded comparison of the commissioned 2021 assessment, GIFCT's response and its May 2026 policy. It does not audit individual moderation outcomes.\n\n## The assessment\n\nBSR's 2021 assessment (see source references) was paid for by GIFCT with BSR retaining editorial control; its December 2020-May 2021 work primarily assessed future organizational risks, not member-company outcomes. Stakeholders raised risks of government pressure and large-platform policy export. The report also records that some small platforms relied on human review because they lacked automation. It recommended stronger appeals, dispute holds, external database review and transparency about downstream use. Its then-current description assigned hash removal to the contributor and content remedy primarily to each platform. These observations and recommendations cannot establish today's implementation or a particular wrongful removal. Its governance discussion distinguished participation from voting power and recommended a later governance review; it did not itself perform a full governance audit.\n\nRead scope: PDF3-10 and start of PDF11 via reader; targeted PDF38-39, PDF41-42, PDF44-48 and PDF63-65 passages, not each whole page or the full73-page report. Web references turn1189view0, turn1191view1/3/4, turn1192view0/1/2. Ordinary original GET returned403; no original PDF retained or hash. Two screenshot attempts (PDF44/45) returned cache misses. Reader-only evidence; visual verification unperformed. No retries.\n\n## The 2021 response\n\nGIFCT's 20 July 2021 response (see source references) grouped its commitments by timeframe. It said membership criteria had already been revised, with appeals required, and identified further remedy capability as work underway. Human review before contribution, third-party review, research access and database governance changes were longer-term commitments or matters for exploration. Review of a multi-stakeholder Operating Board was proposed for two years later. A promise in this letter is not proof of completion. Root read the complete main body, reader lines75-173; retained original HTML in root-hria-captures.json.\n\n## The current policy and announced follow-through\n\nThe currently linked policy (see source references) retains a 2022 URL but explicitly states that GIFCT's team updated it in May2026. It commits to assessment, action, tracking and annual communication, with an annex assigning those activities across technical tools, incident response, research, working groups and transparency. The annex specifies metrics and a risk register; it does not supply their measurements or a worked user-remedy record. Root read all5 extracted pages and visually inspected the page5 matrix from the retained original. Web screenshot failed; local rendering succeeded. Original65316bytes, SHAdd704b12fc0ccfaa7f2584d9419b58f74aed2cb3a707b6a16706f8e4ae109e7d.\n\nThe 18 May 2026 update (see source references) says a five-year report against the BSR assessment was being developed. It announces that previously BSR-delivered human-rights services will move into GIFCT's Membership Advisory Program. The statement is forward-looking; completion of that specific transition is not proved here. It also does not establish that every safeguard improved or deteriorated. Root read the full main body, lines75-101; original retained. Two scoped searches for the five-year review returned this announcement, not a completed report. This bounded failure to locate it does not prove nonpublication.\n\n## What this lane contributes\n\nThere are identifiable safeguards to inspect and an explicit distinction between the organization and each recipient platform. The revealing test is the chain from disputed source entry to correction notification to reconsideration of the recipient's action. Current policy text alone cannot close that chain. Governance, actual reported corrections and member discipline are separately covered by the GIFCT lane. All acquisition in this lane has stopped before the combined forest.\n", "file": "research/private-verification-wave-2026-09-09-human-rights-followthrough.md" }, { "id": "private-verification-wave-2026-09-09-tat", "title": "Tech Against Terrorism: the verifier, its legal controller and its operating rules", "description": "9 September 2026. Bounded acquisition complete. No harmful-content database was accessed, no underlying propaganda was downloaded, and no registration, outreach or payment was made. GIFCT's own rules and the comparative rights/evaluation literature are separate lanes.", "html": "9 September 2026. Bounded acquisition complete. No harmful-content database was accessed, no underlying propaganda was downloaded, and no registration, outreach or payment was made. GIFCT's own rules and the comparative rights/evaluation literature are separate lanes.
\nThe strongest new relationship is concrete corporate and financial dependence: Tech Against Terrorism (TAT) is implemented by Online Harms Foundation (OHF); OHF's filed controlling entity is QuantSpark Limited; OHF also reports receiving loans from, being invoiced by, and owing money to QuantSpark. A separate official Canadian award directly funds development of the Terrorist Content Analytics Platform (TCAP). Neither record gives a donor a demonstrated right to order an individual content classification.
\nThe Australian DIS Standard2024 section6 definition of known pro-terror material, Note2, names TAT and GIFCT as examples of independent counterterrorism-expert organisations whose tools may be used to verify material. This supplies the observed legal connection. It is not a requirement to buy from TAT, a finding about every TCAP record, or incorporation of the entirety of TCAP's evolving taxonomy into Australian law. Retained ../esafety-service-wave-2026-09-09/legal-dis-standard.pdf, PDF13/printed9, text independently reread in this pass; original SHA256 afe30304b7094509de84a1845534b346a48c4fc2f2f07e163fd47fd040ce487d.
\nTAT's [About page](http[local research file] Governance and Funding sections, identifies OHF as implementer and describes TAT as independent and not-for-profit. Its privacy policy also names OHF and company11656320, although the page retains a QuantSpark label in its introductory wording. The legal-identity inference is therefore tied to the company number, not just a similar name.
\n[Companies House's OHF overview](http[local research file] lists an active private company limited by guarantee without share capital, using the Limited exemption; incorporation2November2018; previous name QuantSpark Foundation until11May2021. This pass establishes a company, not registered-charity status. The project website's earlier launch account is a different event from this legal entity's incorporation.
\nThe [current PSC record](http[local research file] names QuantSpark Limited, company10304694, notified2November2018, with 75% or more of voting rights and the right to appoint or remove directors. This is voting/director-appointment control in a guarantee company; it is not ownership of75% of nonexistent share capital. The public filing is a company-reported legal record, and Companies House states it does not check the accuracy of filed information. No donor voting right or instruction about a particular TCAP item is established by it.
\nOHF's unaudited financial statements for the year ended31July2025 were approved and filed28July2026. All8 retained PDF pages were visually read. The directors elected not to file the income-and-expenditure account. Consequently, this filing does not provide total annual income or a donor-by-donor allocation against which to verify TAT's stated balance of government and industry funding.
\n| Relationship | \nExact disclosed state | \nLimit | \n
|---|---|---|
| QuantSpark → OHF consultancy | \nFY2025 invoices GBP913,015; comparative FY2024 GBP2,160,943 | \nInvoiced services, not proved cash payment or a TCAP-only expense. | \n
| OHF → QuantSpark outstanding amount | \nGBP1,554,927 owed at31July2025; comparative GBP1,287,711 | \nRelated-party balance. The note does not disaggregate consultancy payable versus loans; do not label the entire balance a loan. | \n
| QuantSpark → OHF finance | \nGoing-concern note says OHF relies on loans received from QuantSpark, a related party. QuantSpark's director confirmed amounts would not be recalled in the next12months. | \nDated support statement, not perpetual funding, an item-level veto or evidence that all liabilities were forgiven. | \n
| Public Safety Canada → TCAP Phase2 | \n20September2022 announcement: commitment of up to CAD1.9million over three years, through the Community Resilience Fund. | \nProject-specific commitment ceiling; not proof all money was paid. | \n
| Public Safety Canada → TCAP creation | \nSame announcement retrospectively says approximately CAD1million was provided in June2019 through the fund to create TCAP. | \nOfficial reported provision, not independently obtained bank receipt or the underlying executed agreement. | \n
Accounts source: [filed FY2025 accounts](http[local research file] retained tat-accounts-2025.pdf. PDF3/printed1 identifies unaudited small-company treatment, omitted income/expenditure statement and approval date; PDF4/printed2 note1.2 gives going-concern support; PDF7/printed5 note7 gives related-party invoices and outstanding amounts. Note7 expressly calls QuantSpark a corporate director and related party through common control and a common director. It does not identify the particular services, underlying subcontract terms or project cost centre. These financial statements also use accrual accounting, not a cash-receipts presentation.
\nCanadian source: [Public Safety Canada20September2022 announcement](http[local research file] complete substantive news-release body read. Phase2 is described as expanding content types/platform coverage and helping develop a moderation tool for smaller companies. CAD reflects the Canadian public-programme currency context; no conversion is performed. The award names TAT, while the recipient legal-entity bridge comes from the separate TAT/Companies House records; the executed agreement naming its legal signatory was not acquired.
\nTAT's About page says bespoke services to technology companies generate funding and GIFCT contributes to its work. It also says funding comes equally from democratic governments and industry. This is an organisational self-description with no accounting period or reconciliation on the page, not an independently verified50:50 financial ratio. It does not establish that TCAP alerts themselves are paid, that a particular government funds a particular classification, or that the UN owns OHF.
\nThe public [classification-and-verification page](http[local research file] says in-house experts assess both source and content against the inclusion policy, using source-channel assessment and an intelligence assessment of likely organisational production. The full policy requires a form supplying name, work email, organisation and position. This is an explicit access boundary. The form was not submitted, and no alternate access path was attempted.
\nThe [tiered inclusion policy](http[local research file] is version-sensitive: it announces a July2023 expansion. It separates threat-to-life, crisis, designation and promotional tiers, with law-enforcement alerting confined to threat-to-life and live crisis content. Promotional content goes beyond officially produced material. A fifth tier for material not necessarily illegal remains labelled forthcoming on the page read; that is not proof it is operational now. Its references to an Oversight Board describe a safeguard prospectively, without a charter, appointment or adjudication record in this source. Footer2026 does not establish a2026 policy revision.
\n[Tier3's designation policy](http[local research file] follows selected UN/EU and national designation authorities, then applies TAT's consensus and offline/online threat assessments. It permits an expert assessment of splinter-group continuity when a designation effort fails, and includes supporter-produced material using official branding under an imitation rule. Thus the published process contains private judgement as well as legal-list inputs. The stated entity review interval is six months. Removal from all followed lists triggers removal from the inclusion policy; a major governmental-status change can also justify delisting. Historical material predating delisting continues to be alerted. This is an entity-level policy, not a demonstrated correction rule for a falsely labelled item. The displayed designation matrices were not visually audited and are not adopted here as current legal lists.
\nThe [Year2 transparency report](http[local research file] covering December2021–November2022, describes human verification followed by automated alerts. Platforms retain exclusive moderation discretion and can decline further notifications. Registered companies can dispute classification; TAT says it responds within seven working days. It reports36 disputes, all concerning automated online/offline-status errors:35 URL statuses were manually changed and one remained active. These are not36 terrorist-label reversals. An Academic Advisory Board was still being established at report-writing. Locators: PDF28–33, especially29–32;31–32 visually checked.
\nThe same report records18,995 submitted URLs,10,174 alerts to57 companies and82% subsequently offline (PDF5). Its monitor checks URL availability (PDF29). This is TAT's reported operational observation, not independently demonstrated causation, a measure of adjudicated illegality or proof that each disappearance was a provider removal caused by the alert.
\nThe [legacy FAQ](http[local research file] describes vetted company, academic and civil-society access, free smaller-platform/academic use and no content downloading. It excludes government database access, while allowing threat-to-life escalation to police. However, its future references to2021/2022 development make it unsuitable as sole evidence of current access rights. Database access, receipt of an emergency alert and financial sponsorship remain different relationships.
\nThe [current TAT product page](http[local research file] describes human analysts, automated email alerts and ongoing URL-status monitoring. Its marketing language about automated removal does not establish an independent power to delete another provider's content or override the narrower workflow evidence. No live service, API or customer configuration was inspected.
\nThe evidence establishes corporate voting/appointment control, related-party financial reliance, a project-specific government commitment, private classification judgement and an operating provider-dispute process. It does not establish donor-to-item commands, paid inclusion incentives or proven arbitrary classification. Nor does the presence of appeals language demonstrate independent, implemented adjudication.
\nThe decisive missing records are: OHF–QuantSpark consultancy and loan agreements plus project allocation ledgers; the executed Canadian contribution agreements and disbursement reports; the current TCAP verification policy and oversight charter/appointment decisions; a redacted item-classification appeal and its downstream correction notices, including hash withdrawal, recipient notification and provider reconsideration. OHF/QuantSpark hold the first family, Public Safety Canada and the grant recipient the second, and TAT plus participating providers the third. The filed balance cannot supply those terms. No further pursuit of these records starts before the root's forest review.
\nAll captures belong to this wave; tat-captures.json retains exactURLs, acquisition timestamps, byte sizes and SHA256. Website/company captures are original response HTML, not screenshots or reconstructions. The Companies House web-reader account link failed on a redirected cached URL; one ordinary public GET of the official filing link succeeded. Page2's web reader likewise missed cache; ordinary GET succeeded. These were transport/reader failures, not bypassed account gates.
\nCore original hashes: accounts2025 134815173e7eec20b1be9171b23da097d0db5b3eb9d5ffa7fe71e5b0c8c38ab9; PSC HTML 3ff455a4ac892cd12666f06583e52d9a7a5cc9e7c61075fa4055eeaeffbc5147; Canada announcement 4f830f539c2839ae245a4c04e29f2b8a41f072f58e2feee0a6b27faea4cfb619; Year2report 8e41db80e8a9619aa5fccf252b7a831508e34c82c02c924f58c68e79ee5c8d40. Other hashes are in the manifest.
\nRead scope: Companies House overview/PSC records complete substantive sections; filing-history entries identifying latest accounts/incorporation only. About page Governance/Funding and identity passages; privacy-page identity passage in indexed-primary excerpt only. Canada2022 announcement complete substantive body. FAQ substantive questions, with explicit legacy-date caveat. Inclusion policy substantive body; designation policy narrative criteria, imitation, splinter and delisting sections, excluding matrix accuracy. Verification public summary and form boundary only. Current TCAP product-page narrative. Accounts2025 all8pages visually inspected; its embedded text layer was empty. Year2report text PDF5–6,17,28–33; visual31–32 only; remaining40-page contents unreviewed. The incorporation bundle was captured as a58-page original but is image-only and not substantively read; no articles-of-association claim relies on it. Original capture does not imply content knowledge.
\nNo private database or underlying extremist media was accessed, copied or tested. This packet is frozen at the funding/control-to-verification breakpoint for the combined forest.
", "path": "private-verification-wave-2026-09-09/tat.md", "sourceIds": [ "w26-ver-s-tat-about", "w26-ver-s-ohf-company", "w26-ver-s-ohf-control", "w26-ver-s-ohf-accounts2025", "w26-ver-s-tat-verification", "w26-ver-s-tat-inclusion", "w26-ver-s-tat-designation", "w26-ver-s-tat-report2022", "w26-ver-s-dis-verifier-examples" ], "markdown": "9 September 2026. Bounded acquisition complete. No harmful-content database was accessed, no underlying propaganda was downloaded, and no registration, outreach or payment was made. GIFCT's own rules and the comparative rights/evaluation literature are separate lanes.\n\nThe strongest new relationship is concrete corporate and financial dependence: **Tech Against Terrorism (TAT) is implemented by Online Harms Foundation (OHF); OHF's filed controlling entity is QuantSpark Limited; OHF also reports receiving loans from, being invoiced by, and owing money to QuantSpark.** A separate official Canadian award directly funds development of the Terrorist Content Analytics Platform (TCAP). Neither record gives a donor a demonstrated right to order an individual content classification.\n\n## Why this actor belongs in the case\n\nThe Australian DIS Standard2024 section6 definition of known pro-terror material, Note2, names TAT and GIFCT as examples of independent counterterrorism-expert organisations whose tools may be used to verify material. This supplies the observed legal connection. It is not a requirement to buy from TAT, a finding about every TCAP record, or incorporation of the entirety of TCAP's evolving taxonomy into Australian law. Retained ../esafety-service-wave-2026-09-09/legal-dis-standard.pdf, PDF13/printed9, text independently reread in this pass; original SHA256 afe30304b7094509de84a1845534b346a48c4fc2f2f07e163fd47fd040ce487d.\n\n## Legal identity and control\n\nTAT's About page (see source references) Governance and Funding sections, identifies OHF as implementer and describes TAT as independent and not-for-profit. Its privacy policy also names OHF and company11656320, although the page retains a QuantSpark label in its introductory wording. The legal-identity inference is therefore tied to the company number, not just a similar name.\n\nCompanies House's OHF overview (see source references) lists an active private company limited by guarantee without share capital, using the Limited exemption; incorporation2November2018; previous name QuantSpark Foundation until11May2021. This pass establishes a company, not registered-charity status. The project website's earlier launch account is a different event from this legal entity's incorporation.\n\nThe current PSC record (see source references) names **QuantSpark Limited, company10304694**, notified2November2018, with **75% or more of voting rights and the right to appoint or remove directors**. This is voting/director-appointment control in a guarantee company; it is not ownership of75% of nonexistent share capital. The public filing is a company-reported legal record, and Companies House states it does not check the accuracy of filed information. No donor voting right or instruction about a particular TCAP item is established by it.\n\n## Money with distinct legal and accounting states\n\nOHF's unaudited financial statements for the year ended31July2025 were approved and filed28July2026. All8 retained PDF pages were visually read. The directors elected not to file the income-and-expenditure account. Consequently, this filing does not provide total annual income or a donor-by-donor allocation against which to verify TAT's stated balance of government and industry funding.\n\n| Relationship | Exact disclosed state | Limit |\n| --- | --- | --- |\n| QuantSpark → OHF consultancy | FY2025 invoices **GBP913,015**; comparative FY2024 **GBP2,160,943** | Invoiced services, not proved cash payment or a TCAP-only expense. |\n| OHF → QuantSpark outstanding amount | **GBP1,554,927** owed at31July2025; comparative **GBP1,287,711** | Related-party balance. The note does not disaggregate consultancy payable versus loans; do not label the entire balance a loan. |\n| QuantSpark → OHF finance | Going-concern note says OHF relies on loans received from QuantSpark, a related party. QuantSpark's director confirmed amounts would not be recalled in the next12months. | Dated support statement, not perpetual funding, an item-level veto or evidence that all liabilities were forgiven. |\n| Public Safety Canada → TCAP Phase2 | 20September2022 announcement: commitment of **up to CAD1.9million over three years**, through the Community Resilience Fund. | Project-specific commitment ceiling; not proof all money was paid. |\n| Public Safety Canada → TCAP creation | Same announcement retrospectively says approximately **CAD1million was provided in June2019** through the fund to create TCAP. | Official reported provision, not independently obtained bank receipt or the underlying executed agreement. |\n\nAccounts source: filed FY2025 accounts (see source references) retained tat-accounts-2025.pdf. PDF3/printed1 identifies unaudited small-company treatment, omitted income/expenditure statement and approval date; PDF4/printed2 note1.2 gives going-concern support; PDF7/printed5 note7 gives related-party invoices and outstanding amounts. Note7 expressly calls QuantSpark a corporate director and related party through common control and a common director. It does not identify the particular services, underlying subcontract terms or project cost centre. These financial statements also use accrual accounting, not a cash-receipts presentation.\n\nCanadian source: Public Safety Canada20September2022 announcement (see source references) complete substantive news-release body read. Phase2 is described as expanding content types/platform coverage and helping develop a moderation tool for smaller companies. CAD reflects the Canadian public-programme currency context; no conversion is performed. The award names TAT, while the recipient legal-entity bridge comes from the separate TAT/Companies House records; the executed agreement naming its legal signatory was not acquired.\n\nTAT's About page says bespoke services to technology companies generate funding and GIFCT contributes to its work. It also says funding comes equally from democratic governments and industry. This is an organisational self-description with no accounting period or reconciliation on the page, not an independently verified50:50 financial ratio. It does not establish that TCAP alerts themselves are paid, that a particular government funds a particular classification, or that the UN owns OHF.\n\n## What is verified, by whom, and under which published rules\n\nThe public classification-and-verification page (see source references) says in-house experts assess both source and content against the inclusion policy, using source-channel assessment and an intelligence assessment of likely organisational production. The **full policy requires a form supplying name, work email, organisation and position**. This is an explicit access boundary. The form was not submitted, and no alternate access path was attempted.\n\nThe tiered inclusion policy (see source references) is version-sensitive: it announces a July2023 expansion. It separates threat-to-life, crisis, designation and promotional tiers, with law-enforcement alerting confined to threat-to-life and live crisis content. Promotional content goes beyond officially produced material. A fifth tier for material not necessarily illegal remains labelled forthcoming on the page read; that is not proof it is operational now. Its references to an Oversight Board describe a safeguard prospectively, without a charter, appointment or adjudication record in this source. Footer2026 does not establish a2026 policy revision.\n\nTier3's designation policy (see source references) follows selected UN/EU and national designation authorities, then applies TAT's consensus and offline/online threat assessments. It permits an expert assessment of splinter-group continuity when a designation effort fails, and includes supporter-produced material using official branding under an imitation rule. Thus the published process contains private judgement as well as legal-list inputs. The stated entity review interval is six months. Removal from all followed lists triggers removal from the inclusion policy; a major governmental-status change can also justify delisting. **Historical material predating delisting continues to be alerted.** This is an entity-level policy, not a demonstrated correction rule for a falsely labelled item. The displayed designation matrices were not visually audited and are not adopted here as current legal lists.\n\n## Provider choice and an actual reported correction\n\nThe Year2 transparency report (see source references) covering December2021–November2022, describes human verification followed by automated alerts. Platforms retain exclusive moderation discretion and can decline further notifications. Registered companies can dispute classification; TAT says it responds within seven working days. It reports36 disputes, **all concerning automated online/offline-status errors**:35 URL statuses were manually changed and one remained active. These are not36 terrorist-label reversals. An Academic Advisory Board was still being established at report-writing. Locators: PDF28–33, especially29–32;31–32 visually checked.\n\nThe same report records18,995 submitted URLs,10,174 alerts to57 companies and82% subsequently offline (PDF5). Its monitor checks URL availability (PDF29). This is TAT's reported operational observation, not independently demonstrated causation, a measure of adjudicated illegality or proof that each disappearance was a provider removal caused by the alert.\n\nThe legacy FAQ (see source references) describes vetted company, academic and civil-society access, free smaller-platform/academic use and no content downloading. It excludes government database access, while allowing threat-to-life escalation to police. However, its future references to2021/2022 development make it unsuitable as sole evidence of current access rights. Database access, receipt of an emergency alert and financial sponsorship remain different relationships.\n\nThe current TAT product page (see source references) describes human analysts, automated email alerts and ongoing URL-status monitoring. Its marketing language about automated removal does not establish an independent power to delete another provider's content or override the narrower workflow evidence. No live service, API or customer configuration was inspected.\n\n## Limits and concrete next records\n\nThe evidence establishes corporate voting/appointment control, related-party financial reliance, a project-specific government commitment, private classification judgement and an operating provider-dispute process. It does not establish donor-to-item commands, paid inclusion incentives or proven arbitrary classification. Nor does the presence of appeals language demonstrate independent, implemented adjudication.\n\nThe decisive missing records are: OHF–QuantSpark consultancy and loan agreements plus project allocation ledgers; the executed Canadian contribution agreements and disbursement reports; the current TCAP verification policy and oversight charter/appointment decisions; a redacted item-classification appeal and its downstream correction notices, including hash withdrawal, recipient notification and provider reconsideration. OHF/QuantSpark hold the first family, Public Safety Canada and the grant recipient the second, and TAT plus participating providers the third. The filed balance cannot supply those terms. No further pursuit of these records starts before the root's forest review.\n\n## Capture and reading scope\n\nAll captures belong to this wave; tat-captures.json retains exactURLs, acquisition timestamps, byte sizes and SHA256. Website/company captures are original response HTML, not screenshots or reconstructions. The Companies House web-reader account link failed on a redirected cached URL; one ordinary public GET of the official filing link succeeded. Page2's web reader likewise missed cache; ordinary GET succeeded. These were transport/reader failures, not bypassed account gates.\n\nCore original hashes: accounts2025 **134815173e7eec20b1be9171b23da097d0db5b3eb9d5ffa7fe71e5b0c8c38ab9**; PSC HTML **3ff455a4ac892cd12666f06583e52d9a7a5cc9e7c61075fa4055eeaeffbc5147**; Canada announcement **4f830f539c2839ae245a4c04e29f2b8a41f072f58e2feee0a6b27faea4cfb619**; Year2report **8e41db80e8a9619aa5fccf252b7a831508e34c82c02c924f58c68e79ee5c8d40**. Other hashes are in the manifest.\n\nRead scope: Companies House overview/PSC records complete substantive sections; filing-history entries identifying latest accounts/incorporation only. About page Governance/Funding and identity passages; privacy-page identity passage in indexed-primary excerpt only. Canada2022 announcement complete substantive body. FAQ substantive questions, with explicit legacy-date caveat. Inclusion policy substantive body; designation policy narrative criteria, imitation, splinter and delisting sections, excluding matrix accuracy. Verification public summary and form boundary only. Current TCAP product-page narrative. Accounts2025 all8pages visually inspected; its embedded text layer was empty. Year2report text PDF5–6,17,28–33; visual31–32 only; remaining40-page contents unreviewed. The incorporation bundle was captured as a58-page original but is image-only and **not substantively read**; no articles-of-association claim relies on it. Original capture does not imply content knowledge.\n\nNo private database or underlying extremist media was accessed, copied or tested. This packet is frozen at the funding/control-to-verification breakpoint for the combined forest.\n", "file": "research/private-verification-wave-2026-09-09-tat.md" }, { "id": "quantspark-finance-wave-2026-09-09-canadian-awards", "title": "Canadian TCAP contributions: named recipient, amendment and agreement scope", "description": "9 September 2026. Bounded public acquisition closed. No account, outreach, paid request, harmful-content access, new case record or canonical mutation. This lane read actual proactive-disclosure record pages, not just indexed snippets. Nine original public responses and their hashes/scopes are in canadian-award-captures.json. Executed agreements and payment ", "html": "9 September 2026. Bounded public acquisition closed. No account, outreach, paid request, harmful-content access, new case record or canonical mutation. This lane read actual proactive-disclosure record pages, not just indexed snippets. Nine original public responses and their hashes/scopes are retained in canadian-award-captures.json. Executed agreements and payment evidence were not acquired.
\nThe Canadian disclosures supply the recipient/project connection that the earlier announcement left indirect. They name QuantSpark Foundation for TCAP agreement 21415, and Online Harms Foundation for Phase 2 agreement 22841. Their fields explicitly identify these as recipient legal names. The current official data schema says that field should come from the agreement, and that agreement value means the total stated in the agreement, in CAD, not an amendment increment. These are government-published disclosure assertions about agreements, not the signed instruments themselves.
\n| Public disclosure | \nNamed recipient and amount state | \nDates and project scope | \n
|---|---|---|
| [Original TCAP record](http[local research file] reference 214-2019-2020-Q4-0048; agreement 21415 | \nQuantSpark Foundation; CAD 1,155,622.44 agreement value; contribution, international nongovernment recipient | \nDisplayed agreement period26February 2020–30 September 2021. Purpose: knowledge development and rapid identification/removal of terrorist content, disrupting its online distribution and limiting radicalization. | \n
| [Amended TCAP record](http[local research file] reference 088-2021-2022-Q3-00035; same agreement 21415 | \nQuantSpark Foundation; CAD 1,256,688.09 revised total. Additional Information explicitly says the amendment increases funding from1,155,622.44 to1,256,688.09. | \nAmendment date 22 November 2021; displayed agreement period25February 2020–31 March 2022. Same project purpose. [Amendment panel](http[local research file] displays one Current entry with that amendment date and total. | \n
| [TCAP Phase 2 record](http[local research file] reference 088-2022-2023-Q4-00199; agreement 22841 | \nOnline Harms Foundation; CAD 1,945,904 agreement value; contribution, international nongovernment recipient | \nDisplayed agreement period 1 April 2023–31 March 2026. Note says signed in Q4 with start in Q1 of 2023–24; exact signature date is not given. Two workstreams: archiving/classifying terrorist content and constructing a content-moderation workflow tool for technology platforms. | \n
All three identify Public Safety Canada and (CRF) Community Resilience Fund. Full substantive record fields were read, including expected outcomes. Those expected outcomes concern identifying emerging issues and research/programming gaps, applying relevant initiatives, best practices and partner preparedness; they are not measurable project-specific milestone schedules or proof of outcomes delivered.
\nThe amendment's derived increase is CAD 101,065.65. Adding the two displayed 21415 totals would double-count the pre-amendment commitment. The start dates differ by one day—26 February in the original entry versus25February in the amended entry—and the end date changes. Preserve those source values; no signed original/amendment was acquired to resolve the one-day discrepancy. The panel is an amendment disclosure, not evidence of a third payment or award.
\nThe retained [20 September 2022 Public Safety release](http[local research file] announces up toCAD 1.9million over three years to TAT for Phase 2 and retrospectively says approximatelyCAD 1million was provided in June 2019 to create TCAP. Its original was acquired and fully read in the preceding private-verification lane, not reacquired here; SHA256 4f830f539c2839ae245a4c04e29f2b8a41f072f58e2feee0a6b27faea4cfb619. Prior scope: ../private-verification-wave-2026-09-09/tat.md and its capture manifest.
\nThe newly acquired records add precise totals, an explicit amendment, legal-name fields, agreement identifiers and later periods. The 2022 announcement's1.9million formulation is not exactly the1,945,904 later disclosed total. It may be announcement rounding or later agreement setting, but no acquired record explains the difference. Likewise, the 2019 statement and the 2020 agreement entry are not automatically separate funding: same project language is a strong reconciliation lead, but does not prove the timing or allocation of the reported 2019 provision. Do not add announcement figures to agreement totals or recast agreement value as paid cash.
\nIdentity boundary. The accepted Companies House packet identifies company 11656320 as Online Harms Foundation, previously QuantSpark Foundation until11May 2021. This supports reading these labels as the historical names of the same company, distinct from QuantSpark Limited company 10304694. That legal-name history is retained prior primary research, not a fresh registry inspection in this lane. The November 2021 Canadian amendment continues to display the former Foundation name after the reported UK name change; this may be a legacy agreement label. It is not evidence of a second foundation, an undisclosed recipient or the Limited company itself receiving the Canadian award. No UK company number appears in the inspected Canadian records, so the registry-number crosswalk remains separately sourced.
\nThe [CRF Terms and Conditions](http[local research file] page modified 25 May 2026, were read completely, sections1–15.6. They provide department-approved project budgets (§8), Steering Committee proposal/cost decisions (§12), and conditional, auditable contributions tied to milestones (§13). Contributions require financial/nonfinancial reporting and final expenditure accounting; frequency depends on risk (§15.2). Progress payments require departmental acceptance of interim financial/project reports, with risk/cash-flow-based advances possible (§15.4).
\nRecipient copyright remains; a royalty-free, permanent, nonexclusive departmental licence may be given (§10), not automatic government ownership. Onward redistribution requires prior departmental approval, while recipients independently choose beneficiaries with minimal guidance and do not act as government agents; selection records may be reviewed (§15.6). This is approval of redistribution plus recipient selection discretion, not unconstrained pass-through or automatic control of every subcontract. No actual redistribution, licence or content-classification instruction is proved.
\nThese currently published terms show a concrete programme architecture. They do not establish which version or bespoke schedules governed 21415/22841, when a particular milestone was accepted, whether money was advanced/reimbursed, or whether a donor exercised any selection right. The precise executed agreements, incorporated terms, approved budgets, reporting schedules, amendment and acceptance/payment records remain the decisive missing documents.
\nThe original announcement led to a narrowly targeted search of official grants disclosures. Indexed results surfaced TCAP identifiers but their very large paginated URLs had changed contents when opened; those stale snippets were discovery only. The live filtered public search returned the exact record hrefs subsequently read and retained. A web-reader failure on the TCAP search, a403 on the dataset landing and cache miss on the Phase 2 detail were resolved by one ordinary unauthenticated GET of each public page; there was no proxy, account, authentication workaround or repeated failed route. Other acquisition GETs succeeded on first attempt.
\nThe [official dataset landing](http[local research file] links the [public schema](http[local research file] Only the agreement-value, start/end-date, amendment-date/number and recipient-legal-name definitions were read. The schema distinguishes projected end dates from completed performance. Its current guidance helps interpret displayed records but does not retroactively certify historical data quality. The million-record bulk CSV was not downloaded. No broad recipient roster was harvested.
\nTen targeted search queries covered TCAP/TAT/OHF in official grants records, agreement 21415/22841, the dataset, CRF reporting and programme terms. A 2019–20 departmental-results-report hit and general programme-plan snippets were discovery only; they were not acquired or added as evidence. The old and current award record bodies and amendment panel were read in full. The four-record extract is a derivative for local navigation; no original text is altered. Full metadata and SHA256 appear in the adjacent manifest.
\nForest implication: the useful path is now Public Safety Canada→identified contribution→named foundation→specified TCAP workstreams. This does not prove Canada→QuantSpark consultancy payment, classification command, individual removal or Australian deployment. Those would require their own allocation/implementation records. The most discriminating next record is the executed 21415 amendment or 22841 agreement with its approved budget and payment/milestone annex, then a corresponding acceptance/payment record. Public Safety Canada's CRF grants/contributions administrators and the recipient hold them. No request or follow-up branch was initiated.
\n| Retained original | \nSHA256 | \n
|---|---|
| canada-grants-search.html | \n90d5434e9d8b815c60fe38710fa8f08f01b62596c474b63ffca4aeaa4c850554 | \n
| canada-grants-dataset.html | \n1c1d0b427330b7dee079325124c2c828b147f2d48317c58ceb80a8705f5fcece | \n
| canada-phase2-record.html | \nbd99ed33d85637435c4436f8f9e1ccd8466f16ff8c2850f476b67650ddd65aab | \n
| canada-quantspark-search.html | \n3003297c743b611f5d3f2598d7640c19830e355947b972201bcdd1f4892dd4a4 | \n
| canada-phase1-original-record.html | \n96ff7704f383ffd60d6e63d464369cc2cb0211397f2caa030bd700fbf02361ae | \n
| canada-phase1-amended-record.html | \naf134718afb4fab14bf577303bcc47d27dd73d50be594e54e21467ce64ee4a4d | \n
| canada-phase1-amendments.html | \n26ed0dfcd910e3fc7ec7683cf91d3d5438e7869600eedfb3a33ce9431d609f29 | \n
| canada-crf-terms-current.html | \n67410f69545551534785721c81ff75ac07d248b0ddb3a5e6e5cbbccaac50a0c0 | \n
| canada-grants-schema.json | \na9b230e6c1750dc5f555ee12464568330ae0882f606e772a64b4fcd933bf919a | \n
Four-record normalized extract (derivative) SHA256: 57f51a5f2a1d76cec0dd75148ff9524b3eceacc8931ad469aeb91acdc4a6292b.
", "path": "quantspark-finance-wave-2026-09-09/canadian-awards.md", "sourceIds": [ "w26-ver-s-ca21415-original", "w26-ver-s-ca21415-amended", "w26-ver-s-ca22841", "w26-ver-s-crf-terms2026", "w26-ver-s-ca-disclosure-schema" ], "markdown": "9 September 2026. Bounded public acquisition closed. No account, outreach, paid request, harmful-content access, new case record or canonical mutation. This lane read actual proactive-disclosure record pages, not just indexed snippets. Nine original public responses and their hashes/scopes are retained in canadian-award-captures.json. Executed agreements and payment evidence were not acquired.\n\n## The new evidence\n\nThe Canadian disclosures supply the recipient/project connection that the earlier announcement left indirect. They name **QuantSpark Foundation** for TCAP agreement **21415**, and **Online Harms Foundation** for Phase 2 agreement **22841**. Their fields explicitly identify these as recipient legal names. The current official data schema says that field should come from the agreement, and that agreement value means the **total stated in the agreement, in CAD**, not an amendment increment. These are government-published disclosure assertions about agreements, not the signed instruments themselves.\n\n| Public disclosure | Named recipient and amount state | Dates and project scope |\n| --- | --- | --- |\n| Original TCAP record (see source references) reference 214-2019-2020-Q4-0048; agreement 21415 | QuantSpark Foundation; **CAD 1,155,622.44 agreement value**; contribution, international nongovernment recipient | Displayed agreement period26February 2020–30 September 2021. Purpose: knowledge development and rapid identification/removal of terrorist content, disrupting its online distribution and limiting radicalization. |\n| Amended TCAP record (see source references) reference 088-2021-2022-Q3-00035; same agreement 21415 | QuantSpark Foundation; **CAD 1,256,688.09 revised total**. Additional Information explicitly says the amendment increases funding from1,155,622.44 to1,256,688.09. | Amendment date **22 November 2021**; displayed agreement period25February 2020–31 March 2022. Same project purpose. Amendment panel (see source references) displays one Current entry with that amendment date and total. |\n| TCAP Phase 2 record (see source references) reference 088-2022-2023-Q4-00199; agreement 22841 | Online Harms Foundation; **CAD 1,945,904 agreement value**; contribution, international nongovernment recipient | Displayed agreement period **1 April 2023–31 March 2026**. Note says signed in Q4 with start in Q1 of 2023–24; exact signature date is not given. Two workstreams: archiving/classifying terrorist content and constructing a content-moderation workflow tool for technology platforms. |\n\nAll three identify **Public Safety Canada** and **(CRF) Community Resilience Fund**. Full substantive record fields were read, including expected outcomes. Those expected outcomes concern identifying emerging issues and research/programming gaps, applying relevant initiatives, best practices and partner preparedness; they are not measurable project-specific milestone schedules or proof of outcomes delivered.\n\nThe amendment's **derived increase is CAD 101,065.65**. Adding the two displayed 21415 totals would double-count the pre-amendment commitment. The start dates differ by one day—26 February in the original entry versus25February in the amended entry—and the end date changes. Preserve those source values; no signed original/amendment was acquired to resolve the one-day discrepancy. The panel is an amendment disclosure, not evidence of a third payment or award.\n\n## How this changes the earlier announcement\n\nThe retained 20 September 2022 Public Safety release (see source references) announces up toCAD 1.9million over three years to TAT for Phase 2 and retrospectively says approximatelyCAD 1million was provided in June 2019 to create TCAP. Its original was acquired and fully read in the preceding private-verification lane, not reacquired here; SHA256 4f830f539c2839ae245a4c04e29f2b8a41f072f58e2feee0a6b27faea4cfb619. Prior scope: ../private-verification-wave-2026-09-09/tat.md and its capture manifest.\n\nThe newly acquired records add precise totals, an explicit amendment, legal-name fields, agreement identifiers and later periods. The 2022 announcement's1.9million formulation is not exactly the1,945,904 later disclosed total. It may be announcement rounding or later agreement setting, but no acquired record explains the difference. Likewise, the 2019 statement and the 2020 agreement entry are not automatically separate funding: same project language is a strong reconciliation lead, but does not prove the timing or allocation of the reported 2019 provision. Do not add announcement figures to agreement totals or recast agreement value as paid cash.\n\n**Identity boundary.** The accepted Companies House packet identifies company 11656320 as Online Harms Foundation, previously QuantSpark Foundation until11May 2021. This supports reading these labels as the historical names of the same company, distinct from QuantSpark Limited company 10304694. That legal-name history is retained prior primary research, not a fresh registry inspection in this lane. The November 2021 Canadian amendment continues to display the former Foundation name after the reported UK name change; this may be a legacy agreement label. It is not evidence of a second foundation, an undisclosed recipient or the Limited company itself receiving the Canadian award. No UK company number appears in the inspected Canadian records, so the registry-number crosswalk remains separately sourced.\n\n## Public funding controls — actual current terms, not invented historical schedules\n\nThe CRF Terms and Conditions (see source references) page modified **25 May 2026**, were read completely, sections1–15.6. They provide department-approved project budgets (§8), Steering Committee proposal/cost decisions (§12), and conditional, auditable contributions tied to milestones (§13). Contributions require financial/nonfinancial reporting and final expenditure accounting; frequency depends on risk (§15.2). Progress payments require departmental acceptance of interim financial/project reports, with risk/cash-flow-based advances possible (§15.4).\n\nRecipient copyright remains; a royalty-free, permanent, nonexclusive departmental licence **may** be given (§10), not automatic government ownership. Onward redistribution requires prior departmental approval, while recipients independently choose beneficiaries with minimal guidance and do not act as government agents; selection records may be reviewed (§15.6). This is approval of redistribution plus recipient selection discretion, not unconstrained pass-through or automatic control of every subcontract. No actual redistribution, licence or content-classification instruction is proved.\n\nThese currently published terms show a concrete programme architecture. They do **not** establish which version or bespoke schedules governed 21415/22841, when a particular milestone was accepted, whether money was advanced/reimbursed, or whether a donor exercised any selection right. The precise executed agreements, incorporated terms, approved budgets, reporting schedules, amendment and acceptance/payment records remain the decisive missing documents.\n\n## Source ancestry and finite frontier\n\nThe original announcement led to a narrowly targeted search of official grants disclosures. Indexed results surfaced TCAP identifiers but their very large paginated URLs had changed contents when opened; those stale snippets were discovery only. The live filtered public search returned the exact record hrefs subsequently read and retained. A web-reader failure on the TCAP search, a403 on the dataset landing and cache miss on the Phase 2 detail were resolved by one ordinary unauthenticated GET of each public page; there was no proxy, account, authentication workaround or repeated failed route. Other acquisition GETs succeeded on first attempt.\n\nThe official dataset landing (see source references) links the public schema (see source references) Only the agreement-value, start/end-date, amendment-date/number and recipient-legal-name definitions were read. The schema distinguishes projected end dates from completed performance. Its current guidance helps interpret displayed records but does not retroactively certify historical data quality. The million-record bulk CSV was not downloaded. No broad recipient roster was harvested.\n\nTen targeted search queries covered TCAP/TAT/OHF in official grants records, agreement 21415/22841, the dataset, CRF reporting and programme terms. A 2019–20 departmental-results-report hit and general programme-plan snippets were discovery only; they were not acquired or added as evidence. The old and current award record bodies and amendment panel were read in full. The four-record extract is a derivative for local navigation; no original text is altered. Full metadata and SHA256 appear in the adjacent manifest.\n\n**Forest implication:** the useful path is now Public Safety Canada→identified contribution→named foundation→specified TCAP workstreams. This does not prove Canada→QuantSpark consultancy payment, classification command, individual removal or Australian deployment. Those would require their own allocation/implementation records. The most discriminating next record is the executed 21415 amendment or 22841 agreement with its approved budget and payment/milestone annex, then a corresponding acceptance/payment record. Public Safety Canada's CRF grants/contributions administrators and the recipient hold them. No request or follow-up branch was initiated.\n\n## Original custody\n\n| Retained original | SHA256 |\n| --- | --- |\n| canada-grants-search.html | 90d5434e9d8b815c60fe38710fa8f08f01b62596c474b63ffca4aeaa4c850554 |\n| canada-grants-dataset.html | 1c1d0b427330b7dee079325124c2c828b147f2d48317c58ceb80a8705f5fcece |\n| canada-phase2-record.html | bd99ed33d85637435c4436f8f9e1ccd8466f16ff8c2850f476b67650ddd65aab |\n| canada-quantspark-search.html | 3003297c743b611f5d3f2598d7640c19830e355947b972201bcdd1f4892dd4a4 |\n| canada-phase1-original-record.html | 96ff7704f383ffd60d6e63d464369cc2cb0211397f2caa030bd700fbf02361ae |\n| canada-phase1-amended-record.html | af134718afb4fab14bf577303bcc47d27dd73d50be594e54e21467ce64ee4a4d |\n| canada-phase1-amendments.html | 26ed0dfcd910e3fc7ec7683cf91d3d5438e7869600eedfb3a33ce9431d609f29 |\n| canada-crf-terms-current.html | 67410f69545551534785721c81ff75ac07d248b0ddb3a5e6e5cbbccaac50a0c0 |\n| canada-grants-schema.json | a9b230e6c1750dc5f555ee12464568330ae0882f606e772a64b4fcd933bf919a |\n\nFour-record normalized extract (derivative) SHA256: 57f51a5f2a1d76cec0dd75148ff9524b3eceacc8931ad469aeb91acdc4a6292b.\n", "file": "research/quantspark-finance-wave-2026-09-09-canadian-awards.md" }, { "id": "quantspark-finance-wave-2026-09-09-corporate-control", "title": "QuantSpark: filed control and the reciprocal OHF accounts", "description": "Acquisition closed 9 September 2026. This is the bounded corporate lane selected after the forest review “Privately governed verification can travel farther than its correction.” No account, outreach, paid acquisition, database access or canonical case edit. Findings omit personal contact, residential, birth and family information.", "html": "Acquisition closed 9 September 2026. This is the bounded corporate lane selected after the forest review “Privately governed verification can travel farther than its correction.” No account, outreach, paid acquisition, database access or canonical case edit. Findings omit personal contact, residential, birth and family information.
\nAdam Hadley is QuantSpark Limited's sole currently listed director and its registered person with significant control. The official register lists one officer, no resignations, appointed 1 August 2016. The active PSC entry, notified the same date, states ownership of 75% or more of shares, 75% or more of voting rights, the right to appoint or remove directors, and significant influence or control. These are filed rights, not a precise share percentage or proof that nobody else owns shares. Companies House warns that it does not check the accuracy of filed information. Sources: [officers](http[local research file] active-role/appointment entry; [PSC](http[local research file] sole active entry and nature of control.
\n[Company overview](http[local research file] identifies company 10304694, an active private company incorporated 1 August 2016. It was formerly Congent Strategy and Analytics Limited until 10 July 2017. The latest accounts are for 31 July 2025. The accounts themselves specify a private company limited by shares (PDF5/printed3). This is a different legal entity from Online Harms Foundation, company 11656320, which is limited by guarantee without share capital.
\nThe prior retained OHF PSC record identifies QuantSpark Limited as its controller with 75%+ voting rights and director appointment/removal rights. That yields a documented corporate-control chain Hadley → QuantSpark → OHF, with different rights at each step. TAT identifies OHF as its implementer. The chain does not by itself show Hadley or QuantSpark choosing a particular TCAP entry, setting the verification taxonomy, or receiving instructions from a public donor. Reused source: [OHF PSC](http[local research file] retained ../private-verification-wave-2026-09-09/tat-psc.html; current packet reread its nature-of-control passage. The earlier tat.md holds identity and original custody.
[Latest original QuantSpark accounts](http[local research file] year ended 31 July 2025, were approved by director Mr A I Hadley on 9 July 2026 and filed 15 July 2026. The filing is unaudited and uses the small-companies regime. All 11 pages were visually read. The director elected not to file the profit-and-loss account (PDF3/printed1); the public statements therefore do not disclose annual turnover, annual profit, customer concentration or a project margin. Amounts are sterling, rounded to the nearest pound (PDF5/printed3).
\nThe decisive reciprocal evidence is note 10, PDF10/printed8, identifying OHF as connected through common control and a common director:
\n| QuantSpark's disclosed relation to OHF | \nFY/as at 31 July 2025 | \n2024 comparative | \nEvidence state | \n
|---|---|---|---|
| Consultancy and other services invoiced, net of VAT | \n£913,095 | \n£2,160,943 | \nInvoiced service amount; not a bank receipt | \n
| Amount owed in respect of those services at year end | \n£1,374,157 | \n£1,234,895 | \nOutstanding receivable; not current-year revenue | \n
| Amount owed for recharged expenses at year end | \n£180,769 | \n£52,816 | \nSeparate outstanding receivable | \n
| Sum of the two disclosed outstanding components | \n£1,554,926 | \n£1,287,711 | \nOur arithmetic; not an additional transaction | \n
The retained OHF accounts, PDF7/printed5 note7, report £913,015 invoiced by QuantSpark during 2025 and £1,554,927 owed at year end. Its 2024 comparatives are £2,160,943 invoiced and £1,287,711 owed. The 2025 statements therefore differ by £80 on the invoice amount and £1 on the sum of outstanding components; the 2024 figures reconcile. Preserve each filing's figure and description. The QS invoice figure expressly says net of VAT; the OHF sentence does not specify a VAT basis. The acquired records do not explain the small differences, and they do not warrant an allegation of misstatement or a manufactured reconciliation.
\nThis is useful clarification of OHF's creditor dependence: the supplier describes substantial unpaid services and expenses, rather than leaving the entire balance unclassified. It does not prove that the entire balance is a cash loan. OHF's separate going-concern note says it relies on loans from QuantSpark and records the director's assurance that amounts would not be recalled in the following 12 months. The public reciprocal note does not supply the loan agreements, interest/security terms, repayment schedule, whether any amount is subordinated, or which balance components the assurance covers. Source reused: ../private-verification-wave-2026-09-09/tat-accounts-2025.pdf, SHA256 134815173e7eec20b1be9171b23da097d0db5b3eb9d5ffa7fe71e5b0c8c38ab9; PDF7 visually reread now, PDF4 going-concern read in the preceding accepted lane.
QuantSpark reports positive net assets of £3,961,806 and cash of £922,739 at 31 July 2025 (PDF3/printed1). Those company-wide balance-sheet figures do not establish the source of funds lent to OHF, available unrestricted cash for TCAP, or the profitability of the OHF relationship. The filing's note 10 also includes an unrelated connected-company loan; that does not explain the OHF amounts and no branch into that company was opened.
\nQuantSpark's [20 April 2026 article](http[local research file] describes its AI/software consultancy business and identifies Adam Hadley as CEO and founder. Its author capsule expressly says he also runs Tech Against Terrorism. This is a direct company-owned statement of shared leadership, corroborating a professional connection beyond the register; it is not an OHF management contract or a specification of his classification-approval powers.
\nThe article markets services to private-equity firms and their portfolio companies. That establishes its stated customer focus, not ownership of QuantSpark by a private-equity fund, funding of TAT by those customers, or a transfer from an unnamed asset manager. No company-owned contract-level explanation of the OHF loan/service arrangements was located in this bounded pass. A later exact-domain query repeated the leadership capsule; no wider biography or client-network expansion followed.
\nThe next discriminating records are the QuantSpark–OHF consultancy agreement and related-party approval record, the loan/credit agreement and any standstill confirmation, and the intercompany reconciliation/payment ledger with project codes. Custodians are the two companies and their finance advisers. Joined to actual donor agreements, those would establish rates, earmarked projects, invoice settlement, repayment conditions and any reserved approval rights. They would also resolve the small reciprocal differences. A controlling person's registration or matching year-end balance cannot supply those facts.
\nThe strongest defensible finding is concentrated corporate control combined with a disclosed supplier-creditor relationship. A commercially useful support structure, creditor dependence, and project-specific donor conditions can coexist. Which decisions each participant actually controlled still requires the operative agreements or decision records. No public-sector award is routed into these invoices or into an individual moderation decision in this packet. UK contracts and Canadian awards are separate concurrent lanes.
\nSix originals were retained by ordinary unauthenticated GET; qs-captures.json records exact URLs, timestamps, hashes and reading scopes. The official accounts web-reader click returned an Internal Error; one ordinary GET of the filing's direct PDF succeeded. No other retry or access workaround occurred. The scanned 11-page PDF has no usable text layer. qs-accounts2025.txt contains only empty page headings and is not a readable transcription. Local qs-accounts-p1.png through p11.png were all visually inspected. No earlier accounts or confirmation statement was acquired: the latest filing supplies 2024 comparatives and no material dated control change required another filing.
The company overview, sole PSC entry, officer role/date and latest-account filing row were read. The complete substantive company article and author capsule were read from retained HTML; unrelated client-performance claims are not findings. The initial forest-tail output was truncated; the complete selected latest entry was then reread. Web search output is discovery only except the explicitly read company source. Original custody includes public-register fields that are deliberately excluded from this narrative.
\n| Retained original | \nSHA256 | \n
|---|---|
| qs-company.html | \n9090a78477705be4adb0b3b075d995a8a515a83bacf1b4621942825490de3ffb | \n
| qs-psc.html | \n8b323d5c4006dbc9b859e67475eecf4013cb522edbf1e75886ea6bc21f2a62a9 | \n
| qs-officers.html | \na83a9f8c0ac1d422abb49bbf190eb0103b369013a11a83d2af24794e571b936d | \n
| qs-filings.html | \n5f87516df1a2c5bf78f4a611a38ff3dfca8c68d104076c6c5fbb2bc4da25953b | \n
| qs-company-statement.html | \n36cf4792c960b5b3fd5450805ed34694e83d4524b640920dfa6a307198ca6edb | \n
| qs-accounts2025.pdf | \n196702f52a7dfd545a60f016ecf0f48ee0bc68f568b77b18544d7f934d24bfab | \n
Root lane closed9September2026. Read public official notice and guidance; no tender response, account, recipient contact or underlying content access. This establishes a prospective purchase and a documented recommendation, not the accounting destination of TAT's government support.
\n[Sell2Wales notice SEP653567](http[local research file] published2September2026, identifies Home Office and TAT in a UK5 notice for a pilot covering discovery, analysis and bulk referral of terrorist content to technology companies. OCID ocds-h6vhtk-06efa9; supplier PPON PQQZ-1663-NWMN. The contract is explicitly pending, with estimated signature14September2026 and term14September2026–15March2027. It displays GBP192,000 excluding VAT and GBP240,000 including VAT. These displayed amounts are unreconciled here, not payment evidence. Its direct-award justification invokes prototype/development services. The conflict-of-interest document is marked unpublished; no conclusions about its contents follow.
\nThis government-hosted record has greater evidentiary value for stage than an aggregator's Awarded label. It does not establish an executed agreement, actual delivery, the legal recipient's company number, specific target selection rights, OHF–QuantSpark subcontract terms or a project allocation. Its future dates cannot explain FY2025 invoices.
\nCustody: complete substantive notice read via primary web reader turn1201view0 lines24-252, including pending state and future estimated dates. Original ordinary GET returned403; no original HTML/PDF bytes or hash retained. The saved uk-home-pilot-reader-scope.md is a derivative reading record, not a facsimile. No download form or unpublished attachment was pursued.
\nThe [Home Office guidance dated15February2024](http[local research file] explicitly recommends TCAP and the Knowledge Sharing Platform where applicable. It encourages proportionate customer-identification checks, contractual enforcement, reporting/appeals and service-specific restrictions across infrastructure types. It also says the guidance creates no legal obligation and targets unlawful terrorism-related material. Its discussion of separate statutory/police-notice duties is not a demonstration that any particular TAT alert triggers those duties. Recommendations cover providers regardless of location, while existing laws still matter.
\nThis establishes a public endorsement and intended implementation route. It does not establish a compulsory TCAP subscription, a universal filtering requirement, actual supplier deployment or government direction of particular TCAP classifications. Advice to consider measures and legal compulsion must remain separate in the map.
\nCustody: original HTML117756bytes SHA256 bdcfb8fd4f85590d86a0fc7f73f9df8c1b81d5f665a0f45606d1e28240576455; complete substantive body read across primary web reader and local extraction, including introduction, how-to-use qualifications, Sections1–2, next steps and definitions. Initial combined output truncated; missing introductory scope was reread separately. Statutory discussion is attributed to this dated guidance, not a fresh complete current-law opinion.
\n[The D3 procurement index](http[local research file] reports an Ofcom purchase of twelve-month TCAP access, with a GBP34,000 contract value, a3March2026 signing date and23March2026–23March2027 period. It identifies FindTender032552-2026. That original URL returned a nonretryable web-reader error and was closed; a scoped search found no alternate official Ofcom record. The actual official notice, agreement, permitted access scope and payment are unacquired. Do not convert this secondary listing into a proved government database-access right or a contradiction of the older TCAP FAQ. No database access was attempted.
\nQuantSpark's public Digital Marketplace listing also appeared in search. That is a supplier offer, not evidence of a buyer's call-off or funding TAT; its detailed contract was not acquired/read in this lane. Broad policy reports and consultation responses appeared but were not opened into new branches.
\nThe useful outstanding records are the Home Office's eventual signed pilot agreement and performance/selection clauses, a verified Ofcom instrument defining purchased access, and an actual project-to-supplier allocation or subcontract term. A public procurement label cannot bridge these on its own. Canadian contribution disclosures and the two companies' accounts are separate lanes of this bounded run. All root acquisition has stopped for the combined forest once those lanes close.
", "path": "quantspark-finance-wave-2026-09-09/uk-procurement.md", "sourceIds": [ "w26-ver-s-uk-voluntary2024", "w26-ver-s-uk-pilot2026" ], "markdown": "Root lane closed9September2026. Read public official notice and guidance; no tender response, account, recipient contact or underlying content access. This establishes a prospective purchase and a documented recommendation, not the accounting destination of TAT's government support.\n\n## A named prospective Home Office purchase\n\nSell2Wales notice SEP653567 (see source references) published2September2026, identifies Home Office and TAT in a UK5 notice for a pilot covering discovery, analysis and bulk referral of terrorist content to technology companies. OCID ocds-h6vhtk-06efa9; supplier PPON PQQZ-1663-NWMN. The contract is explicitly pending, with estimated signature14September2026 and term14September2026–15March2027. It displays GBP192,000 excluding VAT and GBP240,000 including VAT. These displayed amounts are unreconciled here, not payment evidence. Its direct-award justification invokes prototype/development services. The conflict-of-interest document is marked unpublished; no conclusions about its contents follow.\n\nThis government-hosted record has greater evidentiary value for stage than an aggregator's Awarded label. It does not establish an executed agreement, actual delivery, the legal recipient's company number, specific target selection rights, OHF–QuantSpark subcontract terms or a project allocation. Its future dates cannot explain FY2025 invoices.\n\nCustody: complete substantive notice read via primary web reader turn1201view0 lines24-252, including pending state and future estimated dates. Original ordinary GET returned403; no original HTML/PDF bytes or hash retained. The saved uk-home-pilot-reader-scope.md is a derivative reading record, not a facsimile. No download form or unpublished attachment was pursued.\n\n## Existing policy endorsement has a different force\n\nThe Home Office guidance dated15February2024 (see source references) explicitly recommends TCAP and the Knowledge Sharing Platform where applicable. It encourages proportionate customer-identification checks, contractual enforcement, reporting/appeals and service-specific restrictions across infrastructure types. It also says the guidance creates no legal obligation and targets unlawful terrorism-related material. Its discussion of separate statutory/police-notice duties is not a demonstration that any particular TAT alert triggers those duties. Recommendations cover providers regardless of location, while existing laws still matter.\n\nThis establishes a public endorsement and intended implementation route. It does not establish a compulsory TCAP subscription, a universal filtering requirement, actual supplier deployment or government direction of particular TCAP classifications. Advice to consider measures and legal compulsion must remain separate in the map.\n\nCustody: original HTML117756bytes SHA256 bdcfb8fd4f85590d86a0fc7f73f9df8c1b81d5f665a0f45606d1e28240576455; complete substantive body read across primary web reader and local extraction, including introduction, how-to-use qualifications, Sections1–2, next steps and definitions. Initial combined output truncated; missing introductory scope was reread separately. Statutory discussion is attributed to this dated guidance, not a fresh complete current-law opinion.\n\n## A finite unconfirmed purchase lead\n\nThe D3 procurement index (see source references) reports an Ofcom purchase of twelve-month TCAP access, with a GBP34,000 contract value, a3March2026 signing date and23March2026–23March2027 period. It identifies FindTender032552-2026. That original URL returned a nonretryable web-reader error and was closed; a scoped search found no alternate official Ofcom record. The actual official notice, agreement, permitted access scope and payment are unacquired. Do not convert this secondary listing into a proved government database-access right or a contradiction of the older TCAP FAQ. No database access was attempted.\n\nQuantSpark's public Digital Marketplace listing also appeared in search. That is a supplier offer, not evidence of a buyer's call-off or funding TAT; its detailed contract was not acquired/read in this lane. Broad policy reports and consultation responses appeared but were not opened into new branches.\n\n## What remains revealing\n\nThe useful outstanding records are the Home Office's eventual signed pilot agreement and performance/selection clauses, a verified Ofcom instrument defining purchased access, and an actual project-to-supplier allocation or subcontract term. A public procurement label cannot bridge these on its own. Canadian contribution disclosures and the two companies' accounts are separate lanes of this bounded run. All root acquisition has stopped for the combined forest once those lanes close.\n", "file": "research/quantspark-finance-wave-2026-09-09-uk-procurement.md" }, { "id": "red-lines-wave-2026-09-09-athens-support-layer", "title": "athens support layer", "description": "Athens 2025 support layer — bounded follow-through Observed 9 September 2026. No canonical case edits.", "html": "Athens 2025 support layer — bounded follow-through\nObserved 9 September 2026. No canonical case edits.
\nThe support layer now has two concrete operational links: Cooperative AI Foundation (CAIF) reports hosting the speakers' reception and contributing to the private red-lines discussion; deliberAIde publishes a named TFS testimonial about its platform helping with the event and report. Fathom is resolved to the AI-governance nonprofit at fathom.org, FEIN 93-4840479, whose own current site describes a TFS partnership on verified redlines. No inspected source assigns a 2025 Athens payment amount or grants editorial approval rights.
\nFathom identity and relationship. The official seventh-edition participant page names Andrew Freedman as Fathom's Chief Strategic Officer (reader lines 149–152). The current fathom.org team page names the same person as co-founder and CEO (lines 115–120); its footer supplies FEIN 93-4840479. Preserve those different role dates. Fathom's current partnership entry for TFS describes “Advancing systems of verified redlines to make AI safer” (lines 96–112). This is an explicit organizational collaboration, with no visible commencement date, contract, allocation or evidence that it funded the September 2025 Call. The Athens website's general history separately says Fathom has financially supported the Roundtable (lines 3–7); that statement covers the series and does not specify the 2025 edition. Sources: [participants](http[local research file] [Fathom team](http[local research file] [Fathom partnership](http[local research file] [Athens history](http[local research file]
\nThe matching ProPublica IRS-derived index labels EIN 93-4840479 “Fathom Ai Inc” and lists one return for the fiscal year ending December 2024, filed 14 November 2025. This index is read, but the underlying return remains unread. Search discovery surfaced a $20,000 TFS/Athens grant claim with conflicting year labels across secondary sites. Do not import that amount as a verified grant or assign it to Athens 2025. The tax period, filing date and event date must be resolved from the original Schedule I. [Index and original-document links](http[local research file]
\nCAIF contribution. Its own event recap says it hosted a reception for the speakers at the House of Lords the previous evening. With the independently established 4 December 2025 event date, 3 December is an inferred reception date. In the private “Defining and Governing Unacceptable AI Risks” dialogue, CAIF reports arguing that red-line monitoring needs to assess interactions among agents as well as individual models. This provides an attributed policy contribution, without evidence that participants accepted it. The article names David Norman and Lewis Hammond. Its displayed date is 9 December 2025; the homepage index says 19 December. Preserve that publication-date discrepancy. [CAIF recap, opening paragraph, private-dialogue passage and byline](http[local research file]
\nCAIF identity and funding/governance. Its footer identifies charity 1201294 and company 13485176. Companies House independently identifies COOPERATIVE AI FOUNDATION, incorporated 30 June 2021, as a company limited by guarantee without share capital. CAIF's undated foundation page describes a $15 million philanthropic commitment from Macroscopic Ventures; this is a commitment to CAIF, without payment dates or an Athens allocation. The same page lists trustee Jesse Clifton as a Macroscopic Ventures grantmaking officer; Companies House records his director appointment effective 6 June 2025. This is a documented funder-associated governance connection, not proof of donor appointment rights or control over the event or TFS. [Company overview](http[local research file] [foundation page, introduction and trustees](http[local research file] [filing history, 18 June 2025 entry](http[local research file]
\nThe public CAIF grant-summaries page was checked for Athens/TFS and returned neither name. Its text says it was last updated 20 April 2026 and explicitly warns that recently approved grants or outputs may be missing. It therefore does not establish that no event grant existed. The company filing index lists 33-page accounts to 31 December 2025, filed 24 June 2026, but their document body was not retrieved in this pass. Those accounts remain the exact next potentially discriminating public record. [Grant summary, introductory scope](http[local research file] [accounts filing](http[local research file]
\ndeliberAIde is resolved at brand/product level to deliberaide.com. Its homepage publishes a testimonial attributed to Niki Iliadis, Director, Global AI Governance, TFS, explicitly concerning Athens 2025. It reports that the platform helped the team concentrate on stakeholder engagement and deliver the event report faster. This supports reported platform use in the event workflow, not a cash contribution, donated license, exclusive processing role, or authority to decide the final report. The provider's current general description leaves review, approval and publication with humans; that is a product claim, not an inspected Athens contract. The exact legal operator remains unresolved because the linked Impressum was unreadable and the browser substitute was refused. [Homepage, testimonial lines 27–35 and human-control description lines 259–268](http[local research file]
\nCandidate directional relations for root: Fathom → TFS, publicly described verified-redlines partnership, date unknown/current observation; Fathom → Athens series, organizer-reported financial support, date and amount unknown; CAIF → Athens 2025 speakers, hosted reception, reported; CAIF → Athens private red-lines dialogue, advocated multi-agent monitoring, reported contribution with acceptance unknown; Macroscopic Ventures → CAIF, philanthropic commitment, USD 15,000,000, organizational rather than event allocation; Jesse Clifton → CAIF, director from 6 June 2025, with separately attributed funder role; deliberAIde platform → Athens 2025 workflow, use reported in an attributed client testimonial. Do not turn reception venue into parliamentary endorsement, a software workflow into editorial control, or organizational backing into a campaign-restricted grant.
\nThe strongest rival remains ordinary sponsorship, convening and service provision by organizations with aligned interests. It fits all inspected records. The next discriminating records are the 2025 CAIF accounts and event-support agreement, original Fathom Schedule I with tax-period/recipient/purpose, the deliberAIde legal notice after the administrative access issue is resolved, and an event-specific contract or processing/report-approval record. No outreach is authorized or needed for this bounded closure.
\nSource custody and access boundaries
\nSuccessful ordinary GET hashes below identify returned bytes; raw HTML was read in memory and is not separately retained as an original artifact. Other cited records were read through the web tool with the exact scopes above. Search-only $20,000 claims remain discovered and excluded from accepted financial edges.
\n| Record | \nState and scope | \nSHA-256 of GET body | \n
|---|---|---|
| fathom.org/partnerships | \nRead; TFS entry and nonprofit footer | \n55336b68d8168a612fcd86eaa3fae8343fa9671e386db0ea86830c3020070d38 | \n
| ProPublica EIN 934840479 index | \nRead; identity, tax period, filing date and document links; original not read | \nafb402134024741f5be4eadc2431ac62f36bfa4a448ea8d43bf9edaea20fc259 | \n
| Companies House 13485176 overview | \nRead; identity, incorporation, accounts period | \n2f00fda43a7484b3c4aa165d26d6508d4a5ffd69e3e6d571192ef669d4627d42 | \n
| CAIF grant-summaries | \nInspected; introductory scope and full-page name search, not a grant-by-grant audit | \n1e5f7d3db6bf3184fc16ee3633b8701cd2650d1c483934941eb23e33fe0b881e | \n
| deliberAIde Impressum | \nInaccessible substantive text; HTTP 200 returned only title/skip-link shell | \n1baf97e64cd726d26ee97b3828a614b448c3629908619703d26f012997222cff | \n
The ProPublica original XML link (object_id=202533189349302983) returned HTTP 403. The one alternate original PDF link was refused by web__run.open with “not safe to open (non-retryable error)”; no further attempt was made. The CAIF accounts link in the official filing history failed in web__run.click because its redirected document URL could not be opened; the body remains inaccessible. Filing-history descriptions are not treated as having read the accounts.
\nExact deliberAIde administrative refusal, retained at root's request. Tool: mcp__cua_repl.js. Action: cua.createBrowserTab(athensLegalBrowser.browserId, "http[local research file] { sessionName: "🔎 Athens support" }). Returned text:
\n“Browser Use could not complete this action because a browser security check was unavailable. Reason: The admin-enforced policy could not be verified, so access was not granted. Browser use could not verify the admin-enforced policy before accessing http[local research file] This failure may be temporary. The agent may retry after the issue is resolved, but must not bypass browser security controls or use an indirect workaround.”
\nThis is a browser administrative-policy refusal, not an automatic sandbox approval-review rejection. No further deliberAIde access or indirect workaround was attempted. Earlier homepage observations are retained with their original scope.
", "path": "red-lines-wave-2026-09-09/athens-support-layer.md", "sourceIds": [ "w8-follow-s-fathom", "w8-follow-s-fathom-about", "w8-follow-s-athens-participants", "w8-follow-s-athens-history", "w8-follow-s-caif-event", "w8-follow-s-caif-foundation", "w8-follow-s-caif-filings", "w8-follow-s-deliberaide" ], "markdown": "Athens 2025 support layer — bounded follow-through\nObserved 9 September 2026. No canonical case edits.\n\nThe support layer now has two concrete operational links: Cooperative AI Foundation (CAIF) reports hosting the speakers' reception and contributing to the private red-lines discussion; deliberAIde publishes a named TFS testimonial about its platform helping with the event and report. Fathom is resolved to the AI-governance nonprofit at fathom.org, FEIN 93-4840479, whose own current site describes a TFS partnership on verified redlines. No inspected source assigns a 2025 Athens payment amount or grants editorial approval rights.\n\nFathom identity and relationship. The official seventh-edition participant page names Andrew Freedman as Fathom's Chief Strategic Officer (reader lines 149–152). The current fathom.org team page names the same person as co-founder and CEO (lines 115–120); its footer supplies FEIN 93-4840479. Preserve those different role dates. Fathom's current partnership entry for TFS describes “Advancing systems of verified redlines to make AI safer” (lines 96–112). This is an explicit organizational collaboration, with no visible commencement date, contract, allocation or evidence that it funded the September 2025 Call. The Athens website's general history separately says Fathom has financially supported the Roundtable (lines 3–7); that statement covers the series and does not specify the 2025 edition. Sources: participants (see source references) Fathom team (see source references) Fathom partnership (see source references) Athens history (see source references)\n\nThe matching ProPublica IRS-derived index labels EIN 93-4840479 “Fathom Ai Inc” and lists one return for the fiscal year ending December 2024, filed 14 November 2025. This index is read, but the underlying return remains unread. Search discovery surfaced a $20,000 TFS/Athens grant claim with conflicting year labels across secondary sites. Do not import that amount as a verified grant or assign it to Athens 2025. The tax period, filing date and event date must be resolved from the original Schedule I. Index and original-document links (see source references)\n\nCAIF contribution. Its own event recap says it hosted a reception for the speakers at the House of Lords the previous evening. With the independently established 4 December 2025 event date, 3 December is an inferred reception date. In the private “Defining and Governing Unacceptable AI Risks” dialogue, CAIF reports arguing that red-line monitoring needs to assess interactions among agents as well as individual models. This provides an attributed policy contribution, without evidence that participants accepted it. The article names David Norman and Lewis Hammond. Its displayed date is 9 December 2025; the homepage index says 19 December. Preserve that publication-date discrepancy. CAIF recap, opening paragraph, private-dialogue passage and byline (see source references)\n\nCAIF identity and funding/governance. Its footer identifies charity 1201294 and company 13485176. Companies House independently identifies COOPERATIVE AI FOUNDATION, incorporated 30 June 2021, as a company limited by guarantee without share capital. CAIF's undated foundation page describes a $15 million philanthropic commitment from Macroscopic Ventures; this is a commitment to CAIF, without payment dates or an Athens allocation. The same page lists trustee Jesse Clifton as a Macroscopic Ventures grantmaking officer; Companies House records his director appointment effective 6 June 2025. This is a documented funder-associated governance connection, not proof of donor appointment rights or control over the event or TFS. Company overview (see source references) foundation page, introduction and trustees (see source references) filing history, 18 June 2025 entry (see source references)\n\nThe public CAIF grant-summaries page was checked for Athens/TFS and returned neither name. Its text says it was last updated 20 April 2026 and explicitly warns that recently approved grants or outputs may be missing. It therefore does not establish that no event grant existed. The company filing index lists 33-page accounts to 31 December 2025, filed 24 June 2026, but their document body was not retrieved in this pass. Those accounts remain the exact next potentially discriminating public record. Grant summary, introductory scope (see source references) accounts filing (see source references)\n\ndeliberAIde is resolved at brand/product level to deliberaide.com. Its homepage publishes a testimonial attributed to Niki Iliadis, Director, Global AI Governance, TFS, explicitly concerning Athens 2025. It reports that the platform helped the team concentrate on stakeholder engagement and deliver the event report faster. This supports reported platform use in the event workflow, not a cash contribution, donated license, exclusive processing role, or authority to decide the final report. The provider's current general description leaves review, approval and publication with humans; that is a product claim, not an inspected Athens contract. The exact legal operator remains unresolved because the linked Impressum was unreadable and the browser substitute was refused. Homepage, testimonial lines 27–35 and human-control description lines 259–268 (see source references)\n\nCandidate directional relations for root: Fathom → TFS, publicly described verified-redlines partnership, date unknown/current observation; Fathom → Athens series, organizer-reported financial support, date and amount unknown; CAIF → Athens 2025 speakers, hosted reception, reported; CAIF → Athens private red-lines dialogue, advocated multi-agent monitoring, reported contribution with acceptance unknown; Macroscopic Ventures → CAIF, philanthropic commitment, USD 15,000,000, organizational rather than event allocation; Jesse Clifton → CAIF, director from 6 June 2025, with separately attributed funder role; deliberAIde platform → Athens 2025 workflow, use reported in an attributed client testimonial. Do not turn reception venue into parliamentary endorsement, a software workflow into editorial control, or organizational backing into a campaign-restricted grant.\n\nThe strongest rival remains ordinary sponsorship, convening and service provision by organizations with aligned interests. It fits all inspected records. The next discriminating records are the 2025 CAIF accounts and event-support agreement, original Fathom Schedule I with tax-period/recipient/purpose, the deliberAIde legal notice after the administrative access issue is resolved, and an event-specific contract or processing/report-approval record. No outreach is authorized or needed for this bounded closure.\n\nSource custody and access boundaries\n\nSuccessful ordinary GET hashes below identify returned bytes; raw HTML was read in memory and is not separately retained as an original artifact. Other cited records were read through the web tool with the exact scopes above. Search-only $20,000 claims remain discovered and excluded from accepted financial edges.\n\n| Record | State and scope | SHA-256 of GET body |\n| --- | --- | --- |\n| fathom.org/partnerships | Read; TFS entry and nonprofit footer | 55336b68d8168a612fcd86eaa3fae8343fa9671e386db0ea86830c3020070d38 |\n| ProPublica EIN 934840479 index | Read; identity, tax period, filing date and document links; original not read | afb402134024741f5be4eadc2431ac62f36bfa4a448ea8d43bf9edaea20fc259 |\n| Companies House 13485176 overview | Read; identity, incorporation, accounts period | 2f00fda43a7484b3c4aa165d26d6508d4a5ffd69e3e6d571192ef669d4627d42 |\n| CAIF grant-summaries | Inspected; introductory scope and full-page name search, not a grant-by-grant audit | 1e5f7d3db6bf3184fc16ee3633b8701cd2650d1c483934941eb23e33fe0b881e |\n| deliberAIde Impressum | Inaccessible substantive text; HTTP 200 returned only title/skip-link shell | 1baf97e64cd726d26ee97b3828a614b448c3629908619703d26f012997222cff |\n\nThe ProPublica original XML link (object_id=202533189349302983) returned HTTP 403. The one alternate original PDF link was refused by web__run.open with “not safe to open (non-retryable error)”; no further attempt was made. The CAIF accounts link in the official filing history failed in web__run.click because its redirected document URL could not be opened; the body remains inaccessible. Filing-history descriptions are not treated as having read the accounts.\n\nExact deliberAIde administrative refusal, retained at root's request. Tool: mcp__cua_repl.js. Action: cua.createBrowserTab(athensLegalBrowser.browserId, \"http[local research file] { sessionName: \"🔎 Athens support\" }). Returned text:\n\n“Browser Use could not complete this action because a browser security check was unavailable. Reason: The admin-enforced policy could not be verified, so access was not granted. Browser use could not verify the admin-enforced policy before accessing http[local research file] This failure may be temporary. The agent may retry after the issue is resolved, but must not bypass browser security controls or use an indirect workaround.”\n\nThis is a browser administrative-policy refusal, not an automatic sandbox approval-review rejection. No further deliberAIde access or indirect workaround was attempted. Earlier homepage observations are retained with their original scope.\n", "file": "research/red-lines-wave-2026-09-09-athens-support-layer.md" }, { "id": "red-lines-wave-2026-09-09-campaign-formation", "title": "Global Call for AI Red Lines: formation and agenda pathway", "description": "Observed 9 September 2026. Bounded primary-source return; no canonical edits, outreach, payments or account actions. The revealing record is the campaign's own acknowledgements: it attributes drafting, text feedback, management and announcement brokerage to named people. These establish work performed, not final approval rights, donor control or government a", "html": "Observed 9 September 2026. Bounded primary-source return; no canonical edits, outreach, payments or account actions. The revealing record is the campaign's own acknowledgements: it attributes drafting, text feedback, management and announcement brokerage to named people. These establish work performed, not final approval rights, donor control or government adoption.
\nSource: [campaign acknowledgements](http[local research file] undated; complete substantive page read, reader lines 2–24. Retrospective organizer attribution, access_state read. The February and July references have no year; interpreting them as 2025 follows the dated launch, rather than an explicit year here. Short extract: “drafted the first versions of the call in February”.
Candidate directional edges from this source:
\n| Actor | \nSupported action → object | \nLocator | \n
|---|---|---|
| Charbel-Raphael Segerie | \nResearched and drafted → initial call; coordinated stakeholders until July TFS partnership | \nInitiation | \n
| Florent Berthet; Manuel Bimich | \nDiscussed wording → statement | \nStatement text | \n
| ControlAI; FLI; SaferAI; Marc Rotenberg | \nProvided feedback → statement | \nStatement beta testing | \n
| Arthur Grimonpont | \nDrafted → initial FAQ; pitched → Maria Ressa | \nFAQ; UN Announcement | \n
| Niki Iliadis; Su Cizem; Tereza Zoumpalova | \nSignificantly revised → FAQ | \nFAQ | \n
| Niki Iliadis; Tereza Zoumpalova | \nManaged → project within TFS | \nManagement | \n
| Florent Berthet; Manuel Bimich | \nIdentified → key signatories | \nSignatory collection | \n
| Niki Iliadis; Anoush Tatevossian | \nConducted policy/CSO outreach → signatory collection | \nSignatory collection | \n
| Jared Perlo | \nSuggested → Segerie's TFS collaboration | \nAdvisors | \n
| Anoush Rima Tatevossian | \nAdvised → UN engagement | \nAdvisors | \n
These are attributed contribution edges, not contracts or delegated approval powers.
\nSource: [Ten AI Governance Priorities: Survey of 44 Civil Society Organizations](http[local research file] The Future Society, 2 April 2025. Access_state read; substantive methodology, results and authorship sections read, especially reader lines 79–109 and 187–190.
Its methodology reports that a Paris AI Action Summit side event on 11 February 2025 generated 15 priorities from 61 participants. Renaissance Numérique and TFS consolidated those into 10 voting options, then surveyed 44 anonymous, self-identified organizational representatives on 15–25 February. Red lines ranked first at 7.25/10 under a weighted ranking method. Authors: Caroline Jeanmaire, Jessica Galissaire, Tereza Zoumpalova and Niki Iliadis. The authors explicitly limit representativeness.
\nCandidate edges: Renaissance Numérique + TFS → consolidated voting options; respondent group → ranked the offered red-lines option highest; TFS publication → supplied a survey subsequently linked by the campaign FAQ. This establishes agenda curation and a cited justification. It does not show an unprompted population-wide consensus, identify every respondent, or establish that the survey caused the first draft. The actual 15-to-10 mapping and original questionnaire would discriminate what was merged or excluded.
\nSource: [22 September 2025 campaign press release](http[local research file] The current site also links the [Foreview-hosted copy](http[local research file] Access_state read; all three PDF pages read. These are the same release family, not two independent confirmations; original byte equality was not tested.
PDF page 1 identifies CeSIA, TFS and Berkeley CHAI as co-organizers alongside 20 partners, and reports over 200 individuals and 70 organizations. Page 3 says the call was “announced this morning by Maria Ressa” at the high-level meeting marking the UN's 80th anniversary, and schedules presentation at AI Safety Connect at UN Headquarters for 25 September. The latter is a planned event in this dated document; the announcement is an organizer-reported completed action. Neither establishes UN endorsement. The release seeks an enforced international agreement by end-2026; it does not present a negotiated treaty text.
\nCandidate edges: CeSIA/TFS/CHAI → co-organized call; Maria Ressa → announced call on 22 September 2025; campaign → scheduled presentation at AI Safety Connect on 25 September. The government-uptake lane owns independent UN records.
\nSource: [Athens Roundtable agenda](http[local research file] nine-page program for 4 December 2025, London. Access_state read for PDF pages 1–6; remaining pages only encountered as returned context. Its filename labels it V0; no assertion that this was the last issued agenda.
Page 1 names TFS as organizer with OECD and UNESCO and calls the event an official pre-summit event of the 2026 AI Impact Summit. Pages 4–6 explicitly build Dialogue 1 on the Global Call and a pre-event survey; participants would be grouped using survey responses and backgrounds, then produce a roadmap for later diplomatic forums. This is a designed agenda pathway, not just shared attendance.
\nCandidate edge: TFS-organized Athens Dialogue 1 → used call and survey to structure agenda. The [TFS outcome article dated 12 February 2026](http[local research file] reports the dialogue happened and informed its report. Its substantive body was read; the linked report PDF returned a reader error and was not read. Agenda design and organizer-reported execution remain separate evidence states.
\nSource: [India workshop findings](http[local research file] three pages, dated 18 February 2026; all read. Page 1 names TFS and CeSIA as conveners at AI Safety Connect Day during the India AI Impact Summit, and expressly links the session to the call. Pages 2–3 record unresolved disagreements over global harmonization, standards and balancing immediate versus frontier harms. Candidate edge: TFS + CeSIA → convened call-linked summit workshop. This report does not establish that governments adopted the call or agreed a single red-line regime. The March URL folder is not treated as the event date.
\n[A Narrow Path's own changelog](http[local research file] substantive entries read, dates its public V1.0.0 release to 2 October 2024, V1.1.0 to 19 December 2024, and V1.2.0 to 3 April 2025. Its [main page](http[local research file] and [About page](http[local research file] portions read, identify Andrea Miotti, Tolga Bilge, Dave Kasten and James Newport as authors and describe a three-phase strategy, beginning with preventing superintelligence development for 20 years. This is a different document and project; its named authorship and dated release cannot be transferred to the September 2025 call.
\nThe shared network does matter: S1 specifically credits ControlAI feedback, while the current [call and FAQ](http[local research file] read in the statement, organizer, submission and FAQ sections, seek international agreement but leave particular red lines for scientific-diplomatic definition. The FAQ expressly says it need not reflect every signatory's views. Endorsing the call therefore does not establish endorsement of every FAQ example or of A Narrow Path's full program.
\nThe inspected site reviews signature submissions before publication, but that does not identify who selected partner organizations or who could veto wording. No founding agreement, tracked draft, editorial approval record, partner invitation criteria or decision-rights instrument was obtained. Nor was a complete version history of the call or FAQ found: S1 describes early versions, and S3 anchors public launch, but present FAQ wording must not automatically be backdated to that launch.
\nThe strongest rival to a centrally controlled private-policy project remains an openly assembled coalition whose members provided different contributions without accepting a common detailed platform. The next discriminating records are precise: the February draft and revision history; the July CeSIA–TFS partnership record; the final approval record for the September statement and later FAQ; the partner-selection record; and the survey's option-consolidation mapping. Custodians are CeSIA/TFS and their named project managers, with Renaissance Numérique for the survey. Pursue only if publicly released or newly linked; no outreach is authorized.
\nAcquisition stops at this coherent evidence breakpoint. Public-reader source text and PDF scopes were inspected; this Markdown is the only local artifact created by this lane, so no original-file hashes or preservation claims are made. The broad Athens recap PDF also returned a reader error; neither failed PDF is represented as read. Search-result-only material, including the Manifund funding appeal, was passed to the appropriate lane and is not evidence for this return.
", "path": "red-lines-wave-2026-09-09/campaign-formation.md", "sourceIds": [ "w8-red-s-ack", "w8-red-s-launch", "w8-red-s-survey" ], "markdown": "Observed 9 September 2026. Bounded primary-source return; no canonical edits, outreach, payments or account actions. The revealing record is the campaign's own acknowledgements: it attributes drafting, text feedback, management and announcement brokerage to named people. These establish work performed, not final approval rights, donor control or government adoption.\n\n## S1 — Specific formation roles\n\nSource: campaign acknowledgements (see source references) undated; complete substantive page read, reader lines 2–24. Retrospective organizer attribution, access_state `read`. The February and July references have no year; interpreting them as 2025 follows the dated launch, rather than an explicit year here. Short extract: “drafted the first versions of the call in February”.\n\nCandidate directional edges from this source:\n\n| Actor | Supported action → object | Locator |\n| --- | --- | --- |\n| Charbel-Raphael Segerie | Researched and drafted → initial call; coordinated stakeholders until July TFS partnership | Initiation |\n| Florent Berthet; Manuel Bimich | Discussed wording → statement | Statement text |\n| ControlAI; FLI; SaferAI; Marc Rotenberg | Provided feedback → statement | Statement beta testing |\n| Arthur Grimonpont | Drafted → initial FAQ; pitched → Maria Ressa | FAQ; UN Announcement |\n| Niki Iliadis; Su Cizem; Tereza Zoumpalova | Significantly revised → FAQ | FAQ |\n| Niki Iliadis; Tereza Zoumpalova | Managed → project within TFS | Management |\n| Florent Berthet; Manuel Bimich | Identified → key signatories | Signatory collection |\n| Niki Iliadis; Anoush Tatevossian | Conducted policy/CSO outreach → signatory collection | Signatory collection |\n| Jared Perlo | Suggested → Segerie's TFS collaboration | Advisors |\n| Anoush Rima Tatevossian | Advised → UN engagement | Advisors |\n\nThese are attributed contribution edges, not contracts or delegated approval powers.\n\n## S2 — A concrete choice upstream of the campaign\n\nSource: Ten AI Governance Priorities: Survey of 44 Civil Society Organizations (see source references) The Future Society, 2 April 2025. Access_state `read`; substantive methodology, results and authorship sections read, especially reader lines 79–109 and 187–190.\n\nIts methodology reports that a Paris AI Action Summit side event on 11 February 2025 generated 15 priorities from 61 participants. **Renaissance Numérique and TFS consolidated those into 10 voting options**, then surveyed 44 anonymous, self-identified organizational representatives on 15–25 February. Red lines ranked first at 7.25/10 under a weighted ranking method. Authors: Caroline Jeanmaire, Jessica Galissaire, Tereza Zoumpalova and Niki Iliadis. The authors explicitly limit representativeness.\n\nCandidate edges: Renaissance Numérique + TFS → consolidated voting options; respondent group → ranked the offered red-lines option highest; TFS publication → supplied a survey subsequently linked by the campaign FAQ. This establishes agenda curation and a cited justification. It does not show an unprompted population-wide consensus, identify every respondent, or establish that the survey caused the first draft. The actual 15-to-10 mapping and original questionnaire would discriminate what was merged or excluded.\n\n## S3 — Dated launch and initial public framing\n\nSource: 22 September 2025 campaign press release (see source references) The current site also links the Foreview-hosted copy (see source references) Access_state `read`; all three PDF pages read. These are the same release family, not two independent confirmations; original byte equality was not tested.\n\nPDF page 1 identifies CeSIA, TFS and Berkeley CHAI as co-organizers alongside 20 partners, and reports over 200 individuals and 70 organizations. Page 3 says the call was “announced this morning by Maria Ressa” at the high-level meeting marking the UN's 80th anniversary, and schedules presentation at AI Safety Connect at UN Headquarters for 25 September. The latter is a planned event in this dated document; the announcement is an organizer-reported completed action. Neither establishes UN endorsement. The release seeks an enforced international agreement by end-2026; it does not present a negotiated treaty text.\n\nCandidate edges: CeSIA/TFS/CHAI → co-organized call; Maria Ressa → announced call on 22 September 2025; campaign → scheduled presentation at AI Safety Connect on 25 September. The government-uptake lane owns independent UN records.\n\n## S4 — Documented route into summit programming\n\nSource: Athens Roundtable agenda (see source references) nine-page program for 4 December 2025, London. Access_state `read` for PDF pages 1–6; remaining pages only encountered as returned context. Its filename labels it V0; no assertion that this was the last issued agenda.\n\nPage 1 names TFS as organizer with OECD and UNESCO and calls the event an official pre-summit event of the 2026 AI Impact Summit. Pages 4–6 explicitly build Dialogue 1 on the Global Call and a pre-event survey; participants would be grouped using survey responses and backgrounds, then produce a roadmap for later diplomatic forums. This is a designed agenda pathway, not just shared attendance.\n\nCandidate edge: TFS-organized Athens Dialogue 1 → used call and survey to structure agenda. The TFS outcome article dated 12 February 2026 (see source references) reports the dialogue happened and informed its report. Its substantive body was read; the linked report PDF returned a reader error and was not read. Agenda design and organizer-reported execution remain separate evidence states.\n\nSource: India workshop findings (see source references) three pages, dated 18 February 2026; all read. Page 1 names TFS and CeSIA as conveners at AI Safety Connect Day during the India AI Impact Summit, and expressly links the session to the call. Pages 2–3 record unresolved disagreements over global harmonization, standards and balancing immediate versus frontier harms. Candidate edge: TFS + CeSIA → convened call-linked summit workshop. This report does not establish that governments adopted the call or agreed a single red-line regime. The March URL folder is not treated as the event date.\n\n## S5 — Distinct from A Narrow Path\n\nA Narrow Path's own changelog (see source references) substantive entries read, dates its public V1.0.0 release to **2 October 2024**, V1.1.0 to 19 December 2024, and V1.2.0 to 3 April 2025. Its main page (see source references) and About page (see source references) portions read, identify Andrea Miotti, Tolga Bilge, Dave Kasten and James Newport as authors and describe a three-phase strategy, beginning with preventing superintelligence development for 20 years. This is a different document and project; its named authorship and dated release cannot be transferred to the September 2025 call.\n\nThe shared network does matter: S1 specifically credits ControlAI feedback, while the current call and FAQ (see source references) read in the statement, organizer, submission and FAQ sections, seek international agreement but leave particular red lines for scientific-diplomatic definition. The FAQ expressly says it need not reflect every signatory's views. Endorsing the call therefore does not establish endorsement of every FAQ example or of A Narrow Path's full program.\n\n## Remaining decision frontier\n\nThe inspected site reviews signature submissions before publication, but that does not identify who selected partner organizations or who could veto wording. No founding agreement, tracked draft, editorial approval record, partner invitation criteria or decision-rights instrument was obtained. Nor was a complete version history of the call or FAQ found: S1 describes early versions, and S3 anchors public launch, but present FAQ wording must not automatically be backdated to that launch.\n\nThe strongest rival to a centrally controlled private-policy project remains an openly assembled coalition whose members provided different contributions without accepting a common detailed platform. The next discriminating records are precise: the February draft and revision history; the July CeSIA–TFS partnership record; the final approval record for the September statement and later FAQ; the partner-selection record; and the survey's option-consolidation mapping. Custodians are CeSIA/TFS and their named project managers, with Renaissance Numérique for the survey. Pursue only if publicly released or newly linked; no outreach is authorized.\n\nAcquisition stops at this coherent evidence breakpoint. Public-reader source text and PDF scopes were inspected; this Markdown is the only local artifact created by this lane, so no original-file hashes or preservation claims are made. The broad Athens recap PDF also returned a reader error; neither failed PDF is represented as read. Search-result-only material, including the Manifund funding appeal, was passed to the appropriate lane and is not evidence for this return.\n", "file": "research/red-lines-wave-2026-09-09-campaign-formation.md" }, { "id": "red-lines-wave-2026-09-09-followthrough-finance", "title": "CeSIA follow-through finance: December 2025 Longview recommendation", "description": "Observed 2026-09-09. One bounded follow-up pass. Only this file written; no canonical edits, outreach, submissions, account access or closed-route retries.", "html": "Observed 2026-09-09. One bounded follow-up pass. Only this file written; no canonical edits, outreach, submissions, account access or closed-route retries.
\nResult: unresolved. No inspected later public disclosure establishes that CeSIA's reported December 9, 2025 recommendation of $2 million over two years became an executed grant or a received payment. The legal grantor, payment schedule and restrictions remain unknown. This is a bounded search result, not evidence that payment never happened.
\nThe earlier [CeSIA Manifund diplomatic-sprint proposal](http[local research file] final funding-history answer, reports the recommendation and expressly distinguished it from money received. The page's creation date is December 4, 2025, earlier than the recommendation date cited in its text. The funding answer's update date was not visible. It cannot establish the receipt position on September 9, 2026. Its separate $0 platform display also does not establish total off-platform funding. This pass did not replace that dated-state qualification with a current unpaid claim.
\n| Source, date and exact locator | \nObservation relevant to receipt | \n
|---|---|
| [CeSIA year-end newsletter](http[local research file] December 30, 2025; introductory organizational review and “La vie du CeSIA / Appel aux dons pour soutenir notre action” | \nDescribes the call and other activities, thanks individual donors and solicits support for 2026. No matching institutional amount, legal grantor, executed agreement or receipt is disclosed in these passages. A general appeal does not demonstrate that the Longview recommendation was unfulfilled. | \n
| [Garde-à-vous @Grok](http[local research file] January 8, 2026; “La vie du CeSIA” | \nReports selection for a three-year European Commission evaluation project with consortium partners. This is evidence of another activity, not receipt of the Longview-recommended funding. No payment value or matching grant identified. | \n
| [Le Bon, la Brute et le Truand](http[local research file] March 5, 2026; French “La vie du CeSIA” section read from public RSS | \nReports return from New Delhi and a Red Lines workshop at UNESCO for the IASEAI annual meeting. This supports reported implementation of work contemplated in the proposal; it does not identify its financing. | \n
| [L'enfer est pavé de bonnes inventions](http[local research file] March 12, 2026; French “La vie du CeSIA” section read from public RSS | \nReports a Sciences Po governance roundtable and a previous-week London workshop involving lab staff and safety researchers. The inspected section names no matching grant or payment. | \n
| [The Sweet Escape](http[local research file] August 27, 2026; “Inside CeSIA” | \nReports educational and policy activity; no institutional receipt update. Other claims in the newsletter were not fact-checked or adopted by this finance inquiry. | \n
| [CeSIA About](http[local research file] current undated page; “They trust us” | \nContinues to describe institutional/philanthropic support in aggregate. It does not identify the $2 million recommendation, subsequent receipt, legal grantor or campaign restriction. The currentness of the page is not a date for its underlying grants. | \n
These are primary organizer self-reports. The absence of a payment notice in them cannot resolve a private ledger entry. Reported staffing, events and output likewise cannot identify which funding source paid for them.
\nThe ordinary public [RSS feed](http[local research file] returned HTTP200 and 1,001,547 bytes. SHA-256 of bytes inspected in memory: 0507e8729d65882a42087c392d6eaabcdd9e68b7ff9ab9fde4f55a3da8bd9fc7. Original RSS bytes were not retained. It contained 20 entries dated November 26, 2025–August 27, 2026, including 16 entries dated 2026 (one French/English pair is not two independent reports).
The article bodies supplied in the feed were programmatically screened for Longview, $2M / two-million numeric forms and same-line CeSIA/funding language; there were no matches under that screen. This is keyword-screening scope, not a claim of exhaustive manual reading. The organizational sections itemized above were read; the year-end, January 8 and August 27 pages were additionally opened through the web reader. A section-extraction pass also returned February organizational announcements, but some bilingual output was truncated and no full-archive reading is claimed. The archive web page itself was stale relative to the RSS feed, showing June as its newest visible entry; the RSS supplied the fresher date boundary.
\nTargeted English/French searches combined CeSIA with Longview, funding, grants, financing, annual reports and the specific amount, including organizer, newsletter and funder domains. They surfaced the existing proposal, older funding acknowledgments and unrelated current projects, but no newly inspected executed-grant or payment document. This is not a comprehensive search of every possible publication or spelling.
\nLongview's [Organizational Structure](http[local research file] opening paragraph and the two entity sections, describes Longview Philanthropy as an umbrella name for separate independent organizations: Longview Philanthropy USA Inc., U.S. EIN 93-2664730; and Longview Inc. Ltd, England and Wales company 14444004. It states that each is responsible for its own activities. Current webpage; no publication date shown. Substantive organizational description read in full; no board/personnel relations added.
\nNeither this page nor CeSIA's recommendation claim selects one of those entities as the $2 million legal grantor. It also remains possible that a donor advised by Longview would grant directly or through another intermediary. That is an unresolved alternative, not an asserted payment route. Do not create an edge from either legal entity to CeSIA on the basis of the shared Longview name.
\nThe next decisive record is the grant award/agreement or remittance advice explicitly referencing the December 9, 2025 recommendation, paired with CeSIA's corresponding donor-receipt ledger entry. It needs the counterparty's legal name, signature/award date, paid amount and date, currency, period covered, restrictions and any instalment conditions. A grant promise or unsigned schedule would still not prove receipt; an accounting receivable would need separation from cash received.
\nLikely custodians are CeSIA's operations/finance function and Longview's relevant grants administrator. If an intermediary handled the gift, its remittance record should identify that route. No assumption is made that a particular named form of document exists. No contact was made. A future recipient financial statement or public grant announcement could satisfy the same question if it identifies the recommendation unambiguously; generic total-income growth cannot.
\nNo new executed-grant, received-payment or campaign-restriction edge is recommended. Preserve the earlier recipient-reported recommendation and its undated non-receipt statement. The new activity self-reports can be considered in the formation/follow-through lane independently of finance. Do not reopen the closed Longview Drive-report or scanned TFS990 acquisition routes based on this note.
", "path": "red-lines-wave-2026-09-09/followthrough-finance.md", "sourceIds": [ "w8-follow-s-longview-legal" ], "markdown": "Observed 2026-09-09. One bounded follow-up pass. Only this file written; no canonical edits, outreach, submissions, account access or closed-route retries.\n\n**Result: unresolved.** No inspected later public disclosure establishes that CeSIA's reported December 9, 2025 recommendation of $2 million over two years became an executed grant or a received payment. The legal grantor, payment schedule and restrictions remain unknown. This is a bounded search result, not evidence that payment never happened.\n\n## Starting claim preserved\n\nThe earlier CeSIA Manifund diplomatic-sprint proposal (see source references) final funding-history answer, reports the recommendation and expressly distinguished it from money received. The page's creation date is December 4, 2025, earlier than the recommendation date cited in its text. The funding answer's update date was not visible. It cannot establish the receipt position on September 9, 2026. Its separate $0 platform display also does not establish total off-platform funding. This pass did not replace that dated-state qualification with a current unpaid claim.\n\n## Later sources actually inspected\n\n| Source, date and exact locator | Observation relevant to receipt |\n| --- | --- |\n| CeSIA year-end newsletter (see source references) December 30, 2025; introductory organizational review and “La vie du CeSIA / Appel aux dons pour soutenir notre action” | Describes the call and other activities, thanks individual donors and solicits support for 2026. No matching institutional amount, legal grantor, executed agreement or receipt is disclosed in these passages. A general appeal does not demonstrate that the Longview recommendation was unfulfilled. |\n| Garde-à-vous @Grok (see source references) January 8, 2026; “La vie du CeSIA” | Reports selection for a three-year European Commission evaluation project with consortium partners. This is evidence of another activity, not receipt of the Longview-recommended funding. No payment value or matching grant identified. |\n| Le Bon, la Brute et le Truand (see source references) March 5, 2026; French “La vie du CeSIA” section read from public RSS | Reports return from New Delhi and a Red Lines workshop at UNESCO for the IASEAI annual meeting. This supports reported implementation of work contemplated in the proposal; it does not identify its financing. |\n| L'enfer est pavé de bonnes inventions (see source references) March 12, 2026; French “La vie du CeSIA” section read from public RSS | Reports a Sciences Po governance roundtable and a previous-week London workshop involving lab staff and safety researchers. The inspected section names no matching grant or payment. |\n| The Sweet Escape (see source references) August 27, 2026; “Inside CeSIA” | Reports educational and policy activity; no institutional receipt update. Other claims in the newsletter were not fact-checked or adopted by this finance inquiry. |\n| CeSIA About (see source references) current undated page; “They trust us” | Continues to describe institutional/philanthropic support in aggregate. It does not identify the $2 million recommendation, subsequent receipt, legal grantor or campaign restriction. The currentness of the page is not a date for its underlying grants. |\n\nThese are primary organizer self-reports. The absence of a payment notice in them cannot resolve a private ledger entry. Reported staffing, events and output likewise cannot identify which funding source paid for them.\n\n## What the newsletter pass covered\n\nThe ordinary public RSS feed (see source references) returned HTTP200 and 1,001,547 bytes. SHA-256 of bytes inspected in memory: `0507e8729d65882a42087c392d6eaabcdd9e68b7ff9ab9fde4f55a3da8bd9fc7`. Original RSS bytes were not retained. It contained 20 entries dated November 26, 2025–August 27, 2026, including 16 entries dated 2026 (one French/English pair is not two independent reports).\n\nThe article bodies supplied in the feed were programmatically screened for Longview, $2M / two-million numeric forms and same-line CeSIA/funding language; there were no matches under that screen. This is keyword-screening scope, not a claim of exhaustive manual reading. The organizational sections itemized above were read; the year-end, January 8 and August 27 pages were additionally opened through the web reader. A section-extraction pass also returned February organizational announcements, but some bilingual output was truncated and no full-archive reading is claimed. The archive web page itself was stale relative to the RSS feed, showing June as its newest visible entry; the RSS supplied the fresher date boundary.\n\nTargeted English/French searches combined CeSIA with Longview, funding, grants, financing, annual reports and the specific amount, including organizer, newsletter and funder domains. They surfaced the existing proposal, older funding acknowledgments and unrelated current projects, but no newly inspected executed-grant or payment document. This is not a comprehensive search of every possible publication or spelling.\n\n## Legal grantor: a necessary distinction, now directly documented\n\nLongview's Organizational Structure (see source references) opening paragraph and the two entity sections, describes Longview Philanthropy as an umbrella name for separate independent organizations: Longview Philanthropy USA Inc., U.S. EIN 93-2664730; and Longview Inc. Ltd, England and Wales company 14444004. It states that each is responsible for its own activities. Current webpage; no publication date shown. Substantive organizational description read in full; no board/personnel relations added.\n\nNeither this page nor CeSIA's recommendation claim selects one of those entities as the $2 million legal grantor. It also remains possible that a donor advised by Longview would grant directly or through another intermediary. That is an unresolved alternative, not an asserted payment route. Do not create an edge from either legal entity to CeSIA on the basis of the shared Longview name.\n\n## Narrowed next record and integration decision\n\nThe next decisive record is the **grant award/agreement or remittance advice explicitly referencing the December 9, 2025 recommendation**, paired with CeSIA's corresponding donor-receipt ledger entry. It needs the counterparty's legal name, signature/award date, paid amount and date, currency, period covered, restrictions and any instalment conditions. A grant promise or unsigned schedule would still not prove receipt; an accounting receivable would need separation from cash received.\n\nLikely custodians are CeSIA's operations/finance function and Longview's relevant grants administrator. If an intermediary handled the gift, its remittance record should identify that route. No assumption is made that a particular named form of document exists. No contact was made. A future recipient financial statement or public grant announcement could satisfy the same question if it identifies the recommendation unambiguously; generic total-income growth cannot.\n\nNo new executed-grant, received-payment or campaign-restriction edge is recommended. Preserve the earlier recipient-reported recommendation and its undated non-receipt statement. The new activity self-reports can be considered in the formation/follow-through lane independently of finance. Do not reopen the closed Longview Drive-report or scanned TFS990 acquisition routes based on this note.\n", "file": "research/red-lines-wave-2026-09-09-followthrough-finance.md" }, { "id": "red-lines-wave-2026-09-09-followthrough-france-g7", "title": "France/G7 follow-through: CeSIA's proposed diplomatic sprint versus public outcomes", "description": "Observed 2026-09-09. Bounded acquisition note for parent review; no canonical case changes. Source identifiers below are local candidate labels, not existing-case IDs. Reuse canonical CeSIA, France/G7, Charbel-Raphaël Ségerie and other established identities during integration.", "html": "Observed 2026-09-09. Bounded acquisition note for parent review; no canonical case changes. Source identifiers below are local candidate labels, not existing-case IDs. Reuse canonical CeSIA, France/G7, Charbel-Raphaël Ségerie and other established identities during integration.
\nThere is a documented connection through an event programme and co-organizer relationship, plus a partial functional overlap with the G7 programme. The inspected outcomes do not establish CeSIA's proposed G7 commitment to prepare a coordinated slowdown, and no public transmission or drafting record attributes the relevant official language to CeSIA. That is a bounded result about these records, not proof of no private influence or a verdict on the campaign's whole success.
\nThe strongest new government-side record is the Digital and Technology Ministers' declaration from their 29 May 2026 Paris meeting. Its risk-assessment discussion commitment is more specific than generic AI-safety rhetoric. The strongest new access-side record is a 10 March 2026 Sciences Po/CeSIA programme naming Alexandre Mirlesse, a French foreign-ministry AI/tech diplomat, as a moderator. Neither record closes the connection between campaign proposal and negotiated wording.
\nS1 — CeSIA's original public fundraising proposal. The Manifund page, created 4 December 2025, identifies France as its preferred champion state. Its June target is a G7 statement preparing multilateral capacity to slow AI development together following extreme warning events or crossed red lines; it expressly distinguishes preparation from immediate action. Its separate technical track seeks common risk thresholds. It also targets a September 2026 UNGA resolution and claims prior high-level French briefings without naming recipients, dates or documents. These are project aims and self-reported experience. The page's fundraising status does not establish the campaign's total receipts or execution. Locator: “Why CeSIA,” “Track 1,” “Track 2,” June and September calendar items, and “Policy Influence.” [Original proposal](http[local research file]
\nS2 — Actual ministerial declaration. The five-page French original says ministers met on 29 May. PDF page 2 commits to French-presidency discussions involving the international AI measurement/evaluation/science network and OECD to improve mutual understanding and comparability of AI risk-assessment frameworks, alongside innovation and adoption. This overlaps functionally with harmonization, but does not adopt a common threshold or a slowdown trigger. The complete extracted text has no matches for ralent, seuil, ligne rouge, lignes rouges, CeSIA, red line, moratoire or suspend; the conclusion also rests on reading, not keyword absence alone. Page 1 expressly continues the Hiroshima process dating to 2023 and its later reporting work. Pages 3–5 name OSI, OECD and Inria contributions to other specific outputs, but do not credit CeSIA. Those credits demonstrate that attributable contributions can appear in this instrument; omission does not exclude uncredited input. [Ministerial original](http[local research file]
S3 — Presidential account and announced continuation. On 17 June, Macron's official closing transcript describes a prospective platform among democracies for AI-security cooperation and common standards, with a September meeting planned to finalize work. He frames frontier-model regulation partly around authoritarian access and cyber threats. This is the French President's account of discussions and announced follow-up; it is not the text of a collectively adopted slowdown instrument. His list of nine adopted leaders' statements includes the separate minors call. Locator: AI discussion paragraphs beginning “Lastly” through the September meeting, and opening adopted-statements list. [Official transcript, marked check against delivery](http[local research file]
\nS4 — Earlier government agenda. An Élysée page published 5 February already identifies safe AI, AI adoption, environmental effects and minors' protection as the digital track's four priorities. It names the responsible ministers and the planned 29 May meeting. This predates the March event. It cannot show that CeSIA had no earlier input, but it prevents attributing the priority's origin to that later event. Locator: four digital axes and meeting calendar. [Official priorities](http[local research file]
\nS5 — Host institutional programme. Sciences Po identifies its School of Management and Impact, Les Mardis de l'IA and CeSIA as co-organizers of the 10 March event. Mirlesse is scheduled to moderate Alexei Grinbaum's fireside discussion and appears in the closing patronage/organizing acknowledgment. The programme identifies him as an AI/tech diplomat at France's foreign ministry. It also schedules Nicolas Miailhe, Lord Tim Clement-Jones, Indian deputy mission chief Eenam Gambhir and Swiss foreign-ministry official Miriam Minder. Charbel Ségerie introduces Yoshua Bengio. These are named programme roles, not independent attendance proof, civil-service instructions, or evidence that a G7 negotiator received a red-lines draft. Locator: organizers, agenda and closing acknowledgment. [Host programme](http[local research file]
\nS6 — Organizer's retrospective. CeSIA's 25-page conference report recounts the event and says the French presidency awarded it a G7 label (PDF page 22). Page 10 connects CeSIA's red-lines campaign to its French-presidency ambitions. The report describes France as responding to laboratory leaders' calls, but bundles Davos and the later Delhi summit beside an early-February French announcement. That narrative cannot establish that a later Delhi appeal caused the earlier priority. Earlier appeals or earlier advocacy remain possible. The report also overstates the UN scientific panel's mandate as translating consensus into binding commitments; do not inherit that characterization. A label award letter, its criteria, and any government receipt of the report were not acquired. Label recognition would still require a separate basis before being treated as endorsement of the proposed slowdown. Locator: PDF pages 10 and 22; event agenda pages 23–24. [Organizer's original report](http[local research file]
\nThe report is linked from CeSIA's 30 April retrospective by Pauline Charazac and Charbel-Raphaël Ségerie. This date is a publication marker for the retrospective, not necessarily an embedded PDF version date. [Retrospective page](http[local research file]
\nThese are separate edges; do not join them into a causal path without the missing transmission record.
\n| Candidate ID | \nTyped relationship | \nEvidence and stage | \nBoundary | \n
|---|---|---|---|
| fg7-e1 | \nCeSIA → French G7 presidency: targets as champion-state opportunity | \nS1; reported intention / proposed | \nNo government acceptance established | \n
| fg7-e2 | \nCeSIA → March10 Sciences Po event: co-organizer | \nS5; source record of programme; retrospective S6 reports event held | \nOrganization is not state endorsement | \n
| fg7-e3 | \nAlexandre Mirlesse → March10 event: scheduled moderator and named in patronage acknowledgment | \nS5; source record / scheduled role; S6 retrospective is organizer account | \nNo inferred G7 drafting mandate or agency-wide approval | \n
| fg7-e4 | \nFrench G7 presidency → March10 event: awarded G7 label | \nS6 PDF22; attributed organizer report | \nAward instrument, criteria and exact endorsement scope missing | \n
| fg7-e5 | \nG7 digital ministers → risk-framework comparability discussions: public commitment | \nS2 PDF2; source record / announced future work | \nNot a completed workshop, common threshold, or coordinated slowdown | \n
| fg7-e6 | \nFrench President → democracies AI-security platform and September meeting: announces follow-up | \nS3 AI passage; source record of presidential statement / announced | \nNot a separately inspected collective treaty or CeSIA-attributed outcome | \n
Do not create CeSIA → G7 language “drafted,” “secured,” “caused,” or “adopted” edges from these materials. A possible semantic-overlap observation belongs in an analytical note rather than as an influence fact. Mirlesse's exact programme title is contemporaneous; do not silently substitute an independently assumed senior envoy role.
\nThe official documents support an independent institutional pathway: earlier Hiroshima/GPAI/OECD work, France's already published agenda, and leaders' stated security/adoption concerns. CeSIA may have reinforced that agenda, sought access after it was established, contributed through an uncredited channel, or failed to secure its specific threshold/slowdown ask. Current evidence discriminates against treating the June outcome as fulfillment of that ask; it does not choose among all these influence alternatives.
\nThe most useful missing object is a dated CeSIA submission and its recipient acknowledgment or a negotiator's draft/change record identifying whether risk-comparability or slowdown language was proposed, accepted, softened or rejected. Likely custodians are CeSIA's policy team and the relevant French foreign-ministry/digital-track or presidential G7 negotiating teams. Event attendance, a G7 label, and matching vocabulary cannot substitute for that object. The label award letter is a separate, narrower record about recognition scope.
\nAs of 9 September, September follow-up and the proposal's September UNGA target are still within their stated target month. Absence from this bounded pass is not a finding that either later event or resolution has failed. No public correspondence, meeting minutes or red-lines draft receipt was recovered in the inspected French/G7 sources.
\nAll reads were public originals or publisher-owned accounts. No private APIs, access bypass, outreach or payment. Source bodies were inspected through web readers or ordinary public GET. Only this note is saved under this assignment; the hashes below identify response bytes inspected in memory, not locally retained original files.
\n| Source | \nOrigin and extent read | \nSHA-256 of inspected HTTP200 bytes where acquired | \n
|---|---|---|
| S1 | \nOriginal Manifund proposal; substantive proposal, calendar, experience and success-level sections, current metadata | \nWeb reader; no byte hash | \n
| S2 | \nFrench government-hosted PDF; all five pages extracted and read across two outputs; whole-text keyword search | \nd60f5a67035807d2a9ba1c53e0542538585c975037ec966c82445af68c835f43 | \n
| S3 | \nOfficial English presidential transcript; AI section, statement list and relevant procedural framing; duplicate body not a second source | \n178b5a8f6c51df2610593d88bbe33cb0d3f3ae184123d1389910c1bfc79f318b | \n
| S4 | \nOfficial priorities page; complete substantive main body | \nec35904be329b85c96a0654a582844df9825e16e87b63d2145bc199e206d6e7d | \n
| S5 | \nHost's event page; complete substantive event description, programme and acknowledgments; navigation excluded | \n5cf4bfa296cd4865370bfff7381abdbdd85e4346dc39d9f812c4c5e90390fddc | \n
| S6 | \nCeSIA PDF; scoped read pages3–4,10,19–24, with pages10/22 rechecked directly; not a complete close reading of every page | \n3cd31074aef822a3062f00104c3ca2e58d5a24310ccd917cf9128bd891b91895 | \n
The [Élysée 16 June digital-priorities summary](http[local research file] was also read and supplied the original ministerial PDF route. Its publication date must not replace the PDF's 29 May meeting date. The separate minors annex was only initially scoped; this note makes no fresh detailed findings about its provisions.
\nBounded searches combined CeSIA/G7 with elysee.fr, diplomatie.gouv.fr and economie.gouv.fr and searched the target slowdown concept. They did not return an attributable government drafting/briefing receipt. This is search-scope evidence, not an exhaustive registry audit. The web reader failed on the ministerial PDF; one ordinary GET substitute succeeded. A transcript link-number click failed, and no claim relies on that unresolved link. Closed UN Geneva403, Security Council PDF202 and UNESCO SharePoint401 routes were not retried.
", "path": "red-lines-wave-2026-09-09/followthrough-france-g7.md", "sourceIds": [ "w8-follow-s-sciencespo", "w8-follow-s-cesia-report", "w8-follow-s-g7-ministerial", "w8-follow-s-elysee-priorities", "w8-follow-s-elysee-june" ], "markdown": "Observed 2026-09-09. Bounded acquisition note for parent review; no canonical case changes. Source identifiers below are local candidate labels, not existing-case IDs. Reuse canonical CeSIA, France/G7, Charbel-Raphaël Ségerie and other established identities during integration.\n\n## Finding\n\nThere is a documented connection through an event programme and co-organizer relationship, plus a partial functional overlap with the G7 programme. The inspected outcomes do not establish CeSIA's proposed G7 commitment to prepare a coordinated slowdown, and no public transmission or drafting record attributes the relevant official language to CeSIA. That is a bounded result about these records, not proof of no private influence or a verdict on the campaign's whole success.\n\nThe strongest new government-side record is the Digital and Technology Ministers' declaration from their **29 May 2026 Paris meeting**. Its risk-assessment discussion commitment is more specific than generic AI-safety rhetoric. The strongest new access-side record is a **10 March 2026 Sciences Po/CeSIA programme naming Alexandre Mirlesse, a French foreign-ministry AI/tech diplomat**, as a moderator. Neither record closes the connection between campaign proposal and negotiated wording.\n\n## Proposed objective and observed outcome\n\n**S1 — CeSIA's original public fundraising proposal.** The Manifund page, created 4 December 2025, identifies France as its preferred champion state. Its June target is a G7 statement preparing multilateral capacity to slow AI development together following extreme warning events or crossed red lines; it expressly distinguishes preparation from immediate action. Its separate technical track seeks common risk thresholds. It also targets a September 2026 UNGA resolution and claims prior high-level French briefings without naming recipients, dates or documents. These are project aims and self-reported experience. The page's fundraising status does not establish the campaign's total receipts or execution. Locator: “Why CeSIA,” “Track 1,” “Track 2,” June and September calendar items, and “Policy Influence.” Original proposal (see source references)\n\n**S2 — Actual ministerial declaration.** The five-page French original says ministers met on 29 May. PDF page 2 commits to French-presidency discussions involving the international AI measurement/evaluation/science network and OECD to improve mutual understanding and comparability of AI risk-assessment frameworks, alongside innovation and adoption. This overlaps functionally with harmonization, but does not adopt a common threshold or a slowdown trigger. The complete extracted text has no matches for `ralent`, `seuil`, `ligne rouge`, `lignes rouges`, `CeSIA`, `red line`, `moratoire` or `suspend`; the conclusion also rests on reading, not keyword absence alone. Page 1 expressly continues the Hiroshima process dating to 2023 and its later reporting work. Pages 3–5 name OSI, OECD and Inria contributions to other specific outputs, but do not credit CeSIA. Those credits demonstrate that attributable contributions can appear in this instrument; omission does not exclude uncredited input. Ministerial original (see source references)\n\n**S3 — Presidential account and announced continuation.** On 17 June, Macron's official closing transcript describes a prospective platform among democracies for AI-security cooperation and common standards, with a September meeting planned to finalize work. He frames frontier-model regulation partly around authoritarian access and cyber threats. This is the French President's account of discussions and announced follow-up; it is not the text of a collectively adopted slowdown instrument. His list of nine adopted leaders' statements includes the separate minors call. Locator: AI discussion paragraphs beginning “Lastly” through the September meeting, and opening adopted-statements list. Official transcript, marked check against delivery (see source references)\n\n**S4 — Earlier government agenda.** An Élysée page published 5 February already identifies safe AI, AI adoption, environmental effects and minors' protection as the digital track's four priorities. It names the responsible ministers and the planned 29 May meeting. This predates the March event. It cannot show that CeSIA had no earlier input, but it prevents attributing the priority's origin to that later event. Locator: four digital axes and meeting calendar. Official priorities (see source references)\n\n## The named access connection\n\n**S5 — Host institutional programme.** Sciences Po identifies its School of Management and Impact, Les Mardis de l'IA and CeSIA as co-organizers of the 10 March event. Mirlesse is scheduled to moderate Alexei Grinbaum's fireside discussion and appears in the closing patronage/organizing acknowledgment. The programme identifies him as an AI/tech diplomat at France's foreign ministry. It also schedules Nicolas Miailhe, Lord Tim Clement-Jones, Indian deputy mission chief Eenam Gambhir and Swiss foreign-ministry official Miriam Minder. Charbel Ségerie introduces Yoshua Bengio. These are named programme roles, not independent attendance proof, civil-service instructions, or evidence that a G7 negotiator received a red-lines draft. Locator: organizers, agenda and closing acknowledgment. Host programme (see source references)\n\n**S6 — Organizer's retrospective.** CeSIA's 25-page conference report recounts the event and says the French presidency awarded it a G7 label (PDF page 22). Page 10 connects CeSIA's red-lines campaign to its French-presidency ambitions. The report describes France as responding to laboratory leaders' calls, but bundles Davos and the later Delhi summit beside an early-February French announcement. That narrative cannot establish that a later Delhi appeal caused the earlier priority. Earlier appeals or earlier advocacy remain possible. The report also overstates the UN scientific panel's mandate as translating consensus into binding commitments; do not inherit that characterization. A label award letter, its criteria, and any government receipt of the report were not acquired. Label recognition would still require a separate basis before being treated as endorsement of the proposed slowdown. Locator: PDF pages 10 and 22; event agenda pages 23–24. Organizer's original report (see source references)\n\nThe report is linked from CeSIA's 30 April retrospective by Pauline Charazac and Charbel-Raphaël Ségerie. This date is a publication marker for the retrospective, not necessarily an embedded PDF version date. Retrospective page (see source references)\n\n## Candidate relationships for integration\n\nThese are separate edges; do not join them into a causal path without the missing transmission record.\n\n| Candidate ID | Typed relationship | Evidence and stage | Boundary |\n| --- | --- | --- | --- |\n| fg7-e1 | CeSIA → French G7 presidency: targets as champion-state opportunity | S1; reported intention / proposed | No government acceptance established |\n| fg7-e2 | CeSIA → March10 Sciences Po event: co-organizer | S5; source record of programme; retrospective S6 reports event held | Organization is not state endorsement |\n| fg7-e3 | Alexandre Mirlesse → March10 event: scheduled moderator and named in patronage acknowledgment | S5; source record / scheduled role; S6 retrospective is organizer account | No inferred G7 drafting mandate or agency-wide approval |\n| fg7-e4 | French G7 presidency → March10 event: awarded G7 label | S6 PDF22; attributed organizer report | Award instrument, criteria and exact endorsement scope missing |\n| fg7-e5 | G7 digital ministers → risk-framework comparability discussions: public commitment | S2 PDF2; source record / announced future work | Not a completed workshop, common threshold, or coordinated slowdown |\n| fg7-e6 | French President → democracies AI-security platform and September meeting: announces follow-up | S3 AI passage; source record of presidential statement / announced | Not a separately inspected collective treaty or CeSIA-attributed outcome |\n\nDo not create CeSIA → G7 language “drafted,” “secured,” “caused,” or “adopted” edges from these materials. A possible semantic-overlap observation belongs in an analytical note rather than as an influence fact. Mirlesse's exact programme title is contemporaneous; do not silently substitute an independently assumed senior envoy role.\n\n## Rival explanations and the decisive remaining record\n\nThe official documents support an independent institutional pathway: earlier Hiroshima/GPAI/OECD work, France's already published agenda, and leaders' stated security/adoption concerns. CeSIA may have reinforced that agenda, sought access after it was established, contributed through an uncredited channel, or failed to secure its specific threshold/slowdown ask. Current evidence discriminates against treating the June outcome as fulfillment of that ask; it does not choose among all these influence alternatives.\n\nThe most useful missing object is a dated CeSIA submission and its recipient acknowledgment or a negotiator's draft/change record identifying whether risk-comparability or slowdown language was proposed, accepted, softened or rejected. Likely custodians are CeSIA's policy team and the relevant French foreign-ministry/digital-track or presidential G7 negotiating teams. Event attendance, a G7 label, and matching vocabulary cannot substitute for that object. The label award letter is a separate, narrower record about recognition scope.\n\nAs of 9 September, September follow-up and the proposal's September UNGA target are still within their stated target month. Absence from this bounded pass is not a finding that either later event or resolution has failed. No public correspondence, meeting minutes or red-lines draft receipt was recovered in the inspected French/G7 sources.\n\n## Acquisition and reading scope\n\nAll reads were public originals or publisher-owned accounts. No private APIs, access bypass, outreach or payment. Source bodies were inspected through web readers or ordinary public GET. Only this note is saved under this assignment; the hashes below identify response bytes inspected in memory, **not locally retained original files**.\n\n| Source | Origin and extent read | SHA-256 of inspected HTTP200 bytes where acquired |\n| --- | --- | --- |\n| S1 | Original Manifund proposal; substantive proposal, calendar, experience and success-level sections, current metadata | Web reader; no byte hash |\n| S2 | French government-hosted PDF; all five pages extracted and read across two outputs; whole-text keyword search | `d60f5a67035807d2a9ba1c53e0542538585c975037ec966c82445af68c835f43` |\n| S3 | Official English presidential transcript; AI section, statement list and relevant procedural framing; duplicate body not a second source | `178b5a8f6c51df2610593d88bbe33cb0d3f3ae184123d1389910c1bfc79f318b` |\n| S4 | Official priorities page; complete substantive main body | `ec35904be329b85c96a0654a582844df9825e16e87b63d2145bc199e206d6e7d` |\n| S5 | Host's event page; complete substantive event description, programme and acknowledgments; navigation excluded | `5cf4bfa296cd4865370bfff7381abdbdd85e4346dc39d9f812c4c5e90390fddc` |\n| S6 | CeSIA PDF; scoped read pages3–4,10,19–24, with pages10/22 rechecked directly; not a complete close reading of every page | `3cd31074aef822a3062f00104c3ca2e58d5a24310ccd917cf9128bd891b91895` |\n\nThe Élysée 16 June digital-priorities summary (see source references) was also read and supplied the original ministerial PDF route. Its publication date must not replace the PDF's 29 May meeting date. The separate minors annex was only initially scoped; this note makes no fresh detailed findings about its provisions.\n\nBounded searches combined CeSIA/G7 with elysee.fr, diplomatie.gouv.fr and economie.gouv.fr and searched the target slowdown concept. They did not return an attributable government drafting/briefing receipt. This is search-scope evidence, not an exhaustive registry audit. The web reader failed on the ministerial PDF; one ordinary GET substitute succeeded. A transcript link-number click failed, and no claim relies on that unresolved link. Closed UN Geneva403, Security Council PDF202 and UNESCO SharePoint401 routes were not retried.\n", "file": "research/red-lines-wave-2026-09-09-followthrough-france-g7.md" }, { "id": "red-lines-wave-2026-09-09-organizer-funding", "title": "Organizer resources behind the Global Call for AI Red Lines", "description": "Observed 2026-09-09. Bounded organizer-funding slice; primary public records and organizer self-reports. This file contains proposed relations only; no canonical case changes. Web-reader text was inspected unless an access note says otherwise. No source originals are alongside this file.", "html": "Observed 2026-09-09. Bounded organizer-funding slice; primary public records and organizer self-reports. This file contains proposed relations only; no canonical case changes. Web-reader text was inspected unless an access note says otherwise. No source originals are retained alongside this file.
\nThe strongest financial connection is a shared intermediary: Effektiv Spenden reports supporting both CeSIA and The Future Society after Longview Philanthropy recommendations. These are organizational funding records. A separate CeSIA solicitation explicitly budgets Red Lines diplomatic follow-through. No inspected record establishes a paid, campaign-restricted grant, the launch's total cost, or donor authority over its wording.
\nCeSIA. Its [legal notice](http[local research file] section 1, identifies the publisher as Centre pour la Sécurité de l'Intelligence Artificielle, an association loi 1901, RNA W751277787. This is its own legal-identity statement, not an independently acquired RNA registry certificate. It does not establish which entity received every earlier grant.
\nThe Future Society. Its [overview](http[local research file] “Who We Are” and “Transparency,” says it became a U.S.-registered 501(c)(3) in 2016. SFF identifies the receiving charity as The Future Society, Inc. The overview describes philanthropic contributions plus donations and service contracts involving governments, intergovernmental and private organizations. Its statement that contributors receive no substantial return benefit uses tax terminology; it does not disclose grant approval, publication-clearance, veto or ownership clauses. No such contractual rights were found in the inspected sources.
\nThe [2024–2025 Impact Report](http[local research file] cover dated April 2026, identifies Red Lines as co-led with CHAI and CeSIA (PDF6). Financial summaries are PDF7–8: 2024 revenue $3,406,697, expenses $1,927,009 and closing net assets $3,630,507; 2025 revenue $1,237,109, expenses $2,854,677 and net assets $2,081,163. The 2024 summary is attributed to audited statements; 2025 is explicitly unaudited. The report explains that a two-year grant received in 2024 supported program spending in 2025, without naming that donor, the grant amount, or a Red Lines allocation. These organizational totals must not become the campaign budget. Reading scope: nine-page text, especially PDF2,6–8; report graphics were not independently visually inspected.
\n| Primary record and inspected locator | \nAmount / period / state | \nPrecise relation and limit | \n
|---|---|---|
| Effektiv Spenden, [H1/2024 fund update](http[local research file] opening paragraphs and “CeSIA”; [fund history](http[local research file] | \n€195,000 collected in Germany and Switzerland during H1/2024; fund history says forwarded September 2024 | \nEffektiv Spenden reports forwarding the fund's entire H1 allocation to CeSIA following Longview's recommendation. The update describes research, teaching, public engagement and roundtables. Ségerie's attributed acknowledgment says support enabled hiring. No Red Lines restriction or project-by-project spending appears. | \n
| Effektiv Spenden, [H2/2024 fund update](http[local research file] allocation introduction and “The Future Society”; fund history, H2/2024 | \n€114,000: €60,000 from the AI fund plus €54,000 from the future-preservation fund; collections H2/2024, forwarded March 2025 | \nEffektiv Spenden reports supporting TFS following Longview's recommendation. Its rationale concerns AI governance and planned U.S. staffing, not a named Red Lines grant. Do not treat the two fund contributions as two additional €114,000 awards. | \n
| SFF [recommendation table](http[local research file] header and SFF-2024 CeSIA row | \n$48,000, 2024 recommendation; source Jaan Tallinn; receiving charity EffiSciences | \nSFF recommended general support of CeSIA through a named receiving charity. This identifies an intermediary; it does not itself disclose a fiscal-sponsorship agreement or confirm disbursement. | \n
| Same SFF table, TFS rows | \n$97,000 in 2024; $336,000 in 2025; Jaan Tallinn named source, TFS Inc. receiving charity | \nSeparate general-support recommendations, not campaign awards or independently confirmed payments. | \n
| Same SFF table, SFF-2023-H2 TFS row | \n$365,000 sourced FLI and $11,000 sourced Jaan Tallinn, respectively | \nRecommended general support. This predates the call and documents a financial relationship with a later partner, without establishing campaign financing or instructions. | \n
SFF amounts above are dollars as displayed; no exchange conversion or sum across currencies was made. The table describes its records as recommendations. CHAI-related general-support entries also exist in that database, but no campaign-specific CHAI award emerged, so that broader funding lane was not expanded.
\nCeSIA's [“From Nobel Signatures to Binding Red Lines: The 2026 Diplomatic Sprint”](http[local research file] is marked created December 4, 2025. The current page shows “Not funded” and $0 raised on Manifund. It requests $150,000 for Q1–Q2 2026 and estimates roughly $400,000 for twelve months: staff, diplomatic travel, workshops, event organization and professional UN lobbying consultation. The team section lists APCO among advisors/partners; this is not an executed contract or payment record.
\nThe final funding-history answer reports Longview recommended $2 million over two years on December 9, 2025 and explicitly says this was only a recommendation, with no money received yet. That passage necessarily postdates page creation; no update date was visible. Its receipt statement must not become a verified September 2026 bank-balance claim. Nor does placement in this proposal establish that the entire $2 million was restricted to Red Lines. Scope read: complete page, especially project summary, funding-use section and final funding-history answer; embedded budget images not read.
\nCharbel-Raphaël Ségerie's [December 20, 2024 announcement](http[local research file] “Funding,” reports grants from Effektiv Spenden, AISTOF, Longview, an unnamed Swedish foundation and SFF. It distinguishes general operations from targeted initiatives without assigning each donor to a category. It specifically attributes some ML4Good bootcamp funding to Erasmus+ and textbook support mainly to Ryan Kidd through Manifund. This gives concrete examples of other projects; it does not justify attributing those funds to Red Lines. Read relevant funding and adjacent activity passages; no amounts supplied there.
\nCeSIA's current [about page](http[local research file] “They trust us,” identifies the European Commission, ENS, UK AISI, philanthropic organizations and individuals as funders and says it does not accept funding from private AI companies. This is organizational self-report without amounts, periods or campaign allocation. It is not evidence that every supporter is independent of the AI industry, nor a substitute for grant terms.
\nLongview's Emerging Challenges Fund page.
\n| Subject → precise verb → object | \nEvidence and boundary | \n
|---|---|
| Longview → recommended funding to → CeSIA; Effektiv Spenden → reported forwarding €195,000 to → CeSIA | \nH1/2024 donor update plus September 2024 transfer statement; organization support, purpose unallocated. | \n
| Longview → recommended funding to → TFS; Effektiv Spenden → reported forwarding €114,000 to → TFS | \nH2/2024 donor update plus March 2025 transfer statement; organization support. | \n
| SFF → recommended general support for → named recipients | \nPreserve the separate rows above; receiving-charity designation does not establish control. | \n
| CeSIA → solicited funding for → Red Lines follow-through | \nSolicitation and reported institutional recommendation only; preserve the dated distinctions above. | \n
No candidate donor-to-policy-instruction, donor-to-red-line-text approval, APCO-payment, campaign-expenditure-total, or CHAI campaign-award relation is supported at this breakpoint.
\nTFS's organizer-hosted [2024 public Form 990](http[local research file] returned HTTP200, 3,854,546 bytes, 43 pages. SHA-256 of the original bytes read in memory: 0bb2940d980c822f7eeb280a3dd0ba50472dc02709c5d58043df4d3c1990d4be. Those bytes were not retained. Web reader had no text, and sampled pages produced no extractable text. The screenshot bridge did not expose images to this session; an in-memory render handoff failed on a warning/JSON parsing issue. The filing's substantive content was not inspected and is not used for an EIN, contributor or contractual claim. Stop this acquisition branch rather than treating a file title as content.
Longview links a [public 2025 ECF report](http[local research file] The web reader exposed only a loading page; one ordinary public download substitute returned HTML rather than PDF. Its contents remain unread. No authentication, donor-only report request or other access workaround was attempted.
\nThe decisive next records are the Red Lines cost-center ledger and launch/vendor invoices held by TFS/CeSIA finance; CeSIA's December 2025 Longview recommendation letter and resulting executed grant agreement/disbursement schedule held by CeSIA and Longview's grant team; the EffiSciences–CeSIA receiving/sponsorship agreement and transfer ledger; and restricted-grant schedules for TFS's unnamed two-year 2024 grant. These would distinguish funded capacity from campaign expenditure and reveal any approval, reporting, publication or repayment clauses. Their existence and terms have not been established here. No outreach made.
\nStopping condition reached: public records establish shared funding intermediaries, reported organizational transfers, general-support recommendations and an explicit campaign follow-through funding request. They do not establish campaign-restricted paid money or donor control of the call.
", "path": "red-lines-wave-2026-09-09/organizer-funding.md", "sourceIds": [ "w8-red-s-sprint", "w8-red-s-eff-2024h1", "w8-red-s-eff-2024h2", "w8-red-s-eff-history", "w8-red-s-sff" ], "markdown": "Observed 2026-09-09. Bounded organizer-funding slice; primary public records and organizer self-reports. This file contains proposed relations only; no canonical case changes. Web-reader text was inspected unless an access note says otherwise. No source originals are retained alongside this file.\n\nThe strongest financial connection is a shared intermediary: Effektiv Spenden reports supporting both CeSIA and The Future Society after Longview Philanthropy recommendations. These are organizational funding records. A separate CeSIA solicitation explicitly budgets Red Lines diplomatic follow-through. No inspected record establishes a paid, campaign-restricted grant, the launch's total cost, or donor authority over its wording.\n\n## Legal identity and accounting scope\n\n**CeSIA.** Its legal notice (see source references) section 1, identifies the publisher as Centre pour la Sécurité de l'Intelligence Artificielle, an association loi 1901, RNA W751277787. This is its own legal-identity statement, not an independently acquired RNA registry certificate. It does not establish which entity received every earlier grant.\n\n**The Future Society.** Its overview (see source references) “Who We Are” and “Transparency,” says it became a U.S.-registered 501(c)(3) in 2016. SFF identifies the receiving charity as The Future Society, Inc. The overview describes philanthropic contributions plus donations and service contracts involving governments, intergovernmental and private organizations. Its statement that contributors receive no substantial return benefit uses tax terminology; it does not disclose grant approval, publication-clearance, veto or ownership clauses. No such contractual rights were found in the inspected sources.\n\nThe 2024–2025 Impact Report (see source references) cover dated April 2026, identifies Red Lines as co-led with CHAI and CeSIA (PDF6). Financial summaries are PDF7–8: 2024 revenue $3,406,697, expenses $1,927,009 and closing net assets $3,630,507; 2025 revenue $1,237,109, expenses $2,854,677 and net assets $2,081,163. The 2024 summary is attributed to audited statements; 2025 is explicitly unaudited. The report explains that a two-year grant received in 2024 supported program spending in 2025, without naming that donor, the grant amount, or a Red Lines allocation. These organizational totals must not become the campaign budget. Reading scope: nine-page text, especially PDF2,6–8; report graphics were not independently visually inspected.\n\n## Reported transfers and grant recommendations\n\n| Primary record and inspected locator | Amount / period / state | Precise relation and limit |\n| --- | --- | --- |\n| Effektiv Spenden, H1/2024 fund update (see source references) opening paragraphs and “CeSIA”; fund history (see source references) | €195,000 collected in Germany and Switzerland during H1/2024; fund history says forwarded September 2024 | Effektiv Spenden reports forwarding the fund's entire H1 allocation to CeSIA following Longview's recommendation. The update describes research, teaching, public engagement and roundtables. Ségerie's attributed acknowledgment says support enabled hiring. No Red Lines restriction or project-by-project spending appears. |\n| Effektiv Spenden, H2/2024 fund update (see source references) allocation introduction and “The Future Society”; fund history, H2/2024 | €114,000: €60,000 from the AI fund plus €54,000 from the future-preservation fund; collections H2/2024, forwarded March 2025 | Effektiv Spenden reports supporting TFS following Longview's recommendation. Its rationale concerns AI governance and planned U.S. staffing, not a named Red Lines grant. Do not treat the two fund contributions as two additional €114,000 awards. |\n| SFF recommendation table (see source references) header and SFF-2024 CeSIA row | $48,000, 2024 recommendation; source Jaan Tallinn; receiving charity EffiSciences | SFF recommended general support of CeSIA through a named receiving charity. This identifies an intermediary; it does not itself disclose a fiscal-sponsorship agreement or confirm disbursement. |\n| Same SFF table, TFS rows | $97,000 in 2024; $336,000 in 2025; Jaan Tallinn named source, TFS Inc. receiving charity | Separate general-support recommendations, not campaign awards or independently confirmed payments. |\n| Same SFF table, SFF-2023-H2 TFS row | $365,000 sourced FLI and $11,000 sourced Jaan Tallinn, respectively | Recommended general support. This predates the call and documents a financial relationship with a later partner, without establishing campaign financing or instructions. |\n\nSFF amounts above are dollars as displayed; no exchange conversion or sum across currencies was made. The table describes its records as recommendations. CHAI-related general-support entries also exist in that database, but no campaign-specific CHAI award emerged, so that broader funding lane was not expanded.\n\n## Campaign-specific solicitation: strong purpose evidence, no payment proof\n\nCeSIA's “From Nobel Signatures to Binding Red Lines: The 2026 Diplomatic Sprint” (see source references) is marked created December 4, 2025. The current page shows “Not funded” and $0 raised on Manifund. It requests $150,000 for Q1–Q2 2026 and estimates roughly $400,000 for twelve months: staff, diplomatic travel, workshops, event organization and professional UN lobbying consultation. The team section lists APCO among advisors/partners; this is not an executed contract or payment record.\n\nThe final funding-history answer reports Longview recommended $2 million over two years on December 9, 2025 and explicitly says this was only a recommendation, with no money received yet. That passage necessarily postdates page creation; no update date was visible. Its receipt statement must not become a verified September 2026 bank-balance claim. Nor does placement in this proposal establish that the entire $2 million was restricted to Red Lines. Scope read: complete page, especially project summary, funding-use section and final funding-history answer; embedded budget images not read.\n\n## Corroborating disclosures, kept separate from campaign allocation\n\nCharbel-Raphaël Ségerie's December 20, 2024 announcement (see source references) “Funding,” reports grants from Effektiv Spenden, AISTOF, Longview, an unnamed Swedish foundation and SFF. It distinguishes general operations from targeted initiatives without assigning each donor to a category. It specifically attributes some ML4Good bootcamp funding to Erasmus+ and textbook support mainly to Ryan Kidd through Manifund. This gives concrete examples of other projects; it does not justify attributing those funds to Red Lines. Read relevant funding and adjacent activity passages; no amounts supplied there.\n\nCeSIA's current about page (see source references) “They trust us,” identifies the European Commission, ENS, UK AISI, philanthropic organizations and individuals as funders and says it does not accept funding from private AI companies. This is organizational self-report without amounts, periods or campaign allocation. It is not evidence that every supporter is independent of the AI industry, nor a substitute for grant terms.\n\nLongview's Emerging Challenges Fund page.\n\n## Candidate relations for integration\n\n| Subject → precise verb → object | Evidence and boundary |\n| --- | --- |\n| Longview → recommended funding to → CeSIA; Effektiv Spenden → reported forwarding €195,000 to → CeSIA | H1/2024 donor update plus September 2024 transfer statement; organization support, purpose unallocated. |\n| Longview → recommended funding to → TFS; Effektiv Spenden → reported forwarding €114,000 to → TFS | H2/2024 donor update plus March 2025 transfer statement; organization support. |\n| SFF → recommended general support for → named recipients | Preserve the separate rows above; receiving-charity designation does not establish control. |\n| CeSIA → solicited funding for → Red Lines follow-through | Solicitation and reported institutional recommendation only; preserve the dated distinctions above. |\n\nNo candidate donor-to-policy-instruction, donor-to-red-line-text approval, APCO-payment, campaign-expenditure-total, or CHAI campaign-award relation is supported at this breakpoint.\n\n## Access gaps and next records\n\nTFS's organizer-hosted 2024 public Form 990 (see source references) returned HTTP200, 3,854,546 bytes, 43 pages. SHA-256 of the original bytes read in memory: `0bb2940d980c822f7eeb280a3dd0ba50472dc02709c5d58043df4d3c1990d4be`. Those bytes were not retained. Web reader had no text, and sampled pages produced no extractable text. The screenshot bridge did not expose images to this session; an in-memory render handoff failed on a warning/JSON parsing issue. The filing's substantive content was not inspected and is not used for an EIN, contributor or contractual claim. Stop this acquisition branch rather than treating a file title as content.\n\nLongview links a public 2025 ECF report (see source references) The web reader exposed only a loading page; one ordinary public download substitute returned HTML rather than PDF. Its contents remain unread. No authentication, donor-only report request or other access workaround was attempted.\n\nThe decisive next records are the Red Lines cost-center ledger and launch/vendor invoices held by TFS/CeSIA finance; CeSIA's December 2025 Longview recommendation letter and resulting executed grant agreement/disbursement schedule held by CeSIA and Longview's grant team; the EffiSciences–CeSIA receiving/sponsorship agreement and transfer ledger; and restricted-grant schedules for TFS's unnamed two-year 2024 grant. These would distinguish funded capacity from campaign expenditure and reveal any approval, reporting, publication or repayment clauses. Their existence and terms have not been established here. No outreach made.\n\nStopping condition reached: public records establish shared funding intermediaries, reported organizational transfers, general-support recommendations and an explicit campaign follow-through funding request. They do not establish campaign-restricted paid money or donor control of the call.\n", "file": "research/red-lines-wave-2026-09-09-organizer-funding.md" }, { "id": "red-lines-wave-2026-09-09-policy-uptake", "title": "Global Call for AI Red Lines: policy access and uptake", "description": "Observed 2026-09-09. Bounded public-source inquiry; candidate findings only. No canonical-case edits, outreach, publication, or paid access. Scope: identifiable carriers, institutional venues, and the difference between receiving a demand and adopting it. Six evidence clusters follow. The result is documented access plus a continuing civil-society workstream", "html": "Observed 2026-09-09. Bounded public-source inquiry; candidate findings only. No canonical-case edits, outreach, publication, or paid access. Scope: identifiable carriers, institutional venues, and the difference between receiving a demand and adopting it. Six evidence clusters follow. The result is documented access plus a continuing civil-society workstream, with no binding adoption of the campaign established by the records inspected here.
\nU1 — UN-hosted prepared statement, 22 September 2025. [Speech for UN 80th Anniversary General Assembly](http[local research file] Maria A. Ressa. Read all four pages; decisive locator PDF page 3, “Three Solutions,” first solution. Ressa explicitly names the Global Call, says more than 200 people will launch it later that day, and asks governments to define international boundaries for unacceptable AI risks. This is a submitted/prepared speech on the UN's statement service, not a General Assembly resolution or proof every word was delivered. It supports a named public carrier and recipient venue. It does not establish an agreed UN position, a campaign drafting contract, or who obtained her speaking invitation.
\nAcquisition: web PDF parser failed; ordinary public GET succeeded (HTTP 200), parsed in memory with pypdf. Original-byte SHA-256: add56fc6bb85bf06b39a110e9aa8f3b49064859125287af9378df636a941e471. Bytes were not retained because this lane was restricted to this note.
U2 — UN-hosted automatic transcript, Security Council meeting 10005, 24 September 2025. [Transcript](http[local research file] Scoped read of Bengio's intervention, especially the paragraph beginning “Earlier this week,” and the United States intervention. Bengio refers to the newly supported international red lines and calls for government visibility and control. The US speaker rejects international bodies asserting centralized AI governance. These are evidence of presentation and disagreement, not a Council decision or proof the US statement targeted this campaign specifically.
\nThe service explicitly states its automatic transcripts are not official records or official UN documents. Its speaker labelling contains errors; use the original meeting catalogue to identify Bengio, not an automated label alone. [Official catalogue S/PV.10005](http[local research file] identifies the meeting and Bengio. The English official-record PDF was indexed but not fully acquired: web parser error, ordinary GET HTTP 202 with empty body. The automatic transcript was the bounded substitute. Hash of obtained transcript HTML: 9b6731f529a5f7ab53d673bffccac05409ac2e9db1a4eca7a0db79f51bb1fa34.
U3a — [International AI Red Lines subgroup](http[local research file] current page, inspected 9 September 2026. Full substantive page read. Locators: “Subgroup Activities,” “Subgroup Members,” and final disclaimer. The Future Society and Ada Lovelace Institute are co-leads; CeSIA is a member. The page describes a six-month research process and three workshops: New Delhi 18 February, Paris/UNESCO headquarters 26 February, and an Ada-organized virtual roundtable 16 March 2026. It distinguishes future explainer/discussion-paper outputs from a linked two-page brief. Activities are independent CSO/academic work and do not necessarily express UNESCO, member-state or partner views or commit UNESCO. This is an organizational seat and a work programme, not delegated rulemaking authority. HTML hash: 5f2d5268da959589a8352a9ec105dd614f3314acd443445db6dd39f11073df5c.
**U3b — [Multi-Stakeholder Coordination Group](http[local research file] Scoped read of International AI Red Lines subgroup entries. UNESCO names Cami Rincon of Ada and Tereza Zoumpalova of TFS. The latter's entry identifies her as a Global Call organizer and leader of TFS's red-lines workstream. This connects an identifiable campaign organizer to the network's coordination structure. It does not specify personal veto, budget, or member-state approval rights. HTML hash: 9a142f15220a0ab3a8b3c4dd17ae3bfcf04697f3f573844e3f700b3c1a45630d.
**U3c — [Umbrella network page](http[local research file] Read purpose, chronology, activities and funding paragraphs. The network launched in June 2025, before the September campaign; members shape its activities. UNESCO reports initial Patrick J. McGovern Foundation funding through its 2025 inclusive ethical-governance project, and an earlier UKRI-funded pilot. This concerns the umbrella network. It establishes neither a Global Call grant nor campaign-specific donor control. HTML hash: 532614bd3086d8a9d30339eadb688bf283a979577c435b67e95b2ce42da9b6f8.
Output-access frontier: the subgroup's visible Download PDF link resolves to [UNESCO's SharePoint share](http[local research file] Public link GET returned viewer HTML; the exact file path exposed by that redirect (CSO_Redlines_compressed.pdf) returned HTTP 401. No authentication or bypass attempted; full brief uninspected. Its title/description are website evidence only. Missing decisive record: readable brief plus version/approval record and any transmission to a member-state delegation.
U4 — [Hansard, 29 January 2026](http[local research file] “Superintelligent AI.” Scoped read: Hunt opening at 3.03 pm, Clement-Jones at 3.42 pm, and opening ministerial response at 3.50 pm. Hunt thanks ControlAI for support. Clement-Jones says he joined the Global Call, promotes it, then separately describes ControlAI's UK proposals. He declares DLA Piper consultancy and APPG AI co-chairing. The minister, Baroness Lloyd of Effra, describes context-based regulation and AISI work without accepting a campaign treaty or moratorium in the inspected passage.
\nThe earned mechanism is acknowledged NGO support and a signatory's explicit parliamentary advocacy. Support is not evidence of a payment, authorship of Hunt's speech, or government commissioning. Clement-Jones's interpretation of the call must not silently replace the campaign's own text. The minister's response is not a comprehensive proof of permanent rejection. Web reader content inspected; native GET returned403, so no original-HTML hash or retained bytes claimed. Needed: the actual briefing sent to Hunt/Clement-Jones, correspondence on requested wording, and any departmental response to the specific campaign proposal.
\nU5a — Organizer's account: [TFS, Key Takeaways from the India AI Impact Summit](http[local research file] February 2026 event; publication date not independently established. Read “Defining and Governing Unacceptable AI Risks” and “General Reflections.” TFS says it and CeSIA co-organized a session within AI Safety Connect Day, explicitly building on the campaign. It reports interventions from Denmark's tech ambassador Anne Marie Engtoft Meldgaard, Brazil foreign-ministry AI adviser Guilherme Fitzgibbon Alves Pereira, Japan AISI's Akiko Murakami, Singapore IMDA's Wan Sie Lee, and MEP Brando Benifei. This is named organizer-reported access, not independently confirmed government endorsement. Its own reflections call for moving beyond voluntary commitments. Web reader inspected; ordinary GET403; no captured-original hash claimed.
\n**U5b — Government outcome: [PIB/MeitY, New Delhi Declaration](http[local research file] Posted 21 February 2026; live page includes an endorsement list dated 5 March, so it is not an immutable launch-day snapshot. Full declaration annex read, especially “Secure and Trusted AI” and “Conclusion.” It supports industry-led voluntary measures and a voluntary, non-binding Trusted AI Commons; the conclusion likewise emphasizes voluntary/non-binding instruments. The declaration does not name this campaign or enact its requested internationally enforceable prohibitions. No drafting or transmission record links the side event to this text. HTML hash: cf42f235597ee26c30934bd44d5c0e7b5f25dba4737595a7cec02fa09c98cab7.
Needed to move from attendance to uptake: the workshop paper actually transmitted to the Safe and Trusted AI working group, recipient acknowledgment, tracked draft/declaration edits and decision record. Independent government priorities and pre-existing summit commitments remain adequate alternative explanations for shared language.
\nU6a — [UN Global Dialogue website](http[local research file] and [FAQ](http[local research file] Indexed official-source text was inspected; direct web and ordinary GET access returned403. The homepage advertises a Co-Chairs' Summary for Geneva 6–7 July 2026. The FAQ says the Dialogue is not a negotiating forum and ends with a co-chair summary. The summary itself was not acquired or read**, so no claim is made about whether it incorporates particular red-lines recommendations.
\n**U6b — [UN automatic closing-session transcript](http[local research file] Scoped review: final co-chair remarks (parsed lines1038–1053) and red-lines search matches. The co-chairs describe diverse views and a summary still to be drafted. The shared-red-lines passage at parsed lines915-918 is a stakeholder intervention, not a negotiated decision. This ASR source has the same non-official-record limitation as U2. It supplies useful context, not an authoritative final outcome. A separate [UN Web TV programme, 1 July 2026](http[local research file] names Bengio and Ressa as scientific-panel co-chairs and describes independent scientific assessments; it does not give them treaty-making authority. Programme metadata inspected; video not watched.
\nPriority missing record: actual final Co-Chairs' Summary and annexed submissions, especially any TFS/CeSIA/CHAI contribution, followed by the attribution/version trail. Root should treat this as a defined public-document access gap, not evidence of concealment or adoption.
\nIDs below are provisional within this note; match existing entity IDs before integration. Evidence status describes the source claim, not a finding of control. No graph edge should jump directly from campaign involvement to a state adopting rules.
\n| Candidate ID | \nDirection and relation | \nDate / action stage | \nSource and limitation | \n
|---|---|---|---|
| rl-policy-01 | \nMaria Ressa → Global Call: publicly advocated | \n2025-09-22 / prepared statement | \nU1; named campaign | \n
| rl-policy-02 | \nMaria Ressa → UNGA80 anniversary session: presented policy demand | \n2025-09-22 / submitted speech | \nU1; not a resolution | \n
| rl-policy-03 | \nYoshua Bengio → UNSC10005: advocated international AI red lines | \n2025-09-24 / speech reported by ASR | \nU2; official-record corroboration outstanding | \n
| rl-policy-04 | \nThe Future Society → UNESCO-network red-lines subgroup: co-lead | \nobserved2026-09-09 / current listed role | \nU3a; independent CSO work | \n
| rl-policy-05 | \nAda Lovelace Institute → same subgroup: co-lead | \nobserved2026-09-09 / current listed role | \nU3a | \n
| rl-policy-06 | \nCeSIA → same subgroup: member | \nobserved2026-09-09 / current listed role | \nU3a; no control right inferred | \n
| rl-policy-07 | \nTereza Zoumpalova → Global Call: organizer | \nhistorical, observed2026-09-09 | \nU3b; UNESCO-hosted biography | \n
| rl-policy-08 | \nTereza Zoumpalova → UNESCO network coordination group: named subgroup representative | \nobserved2026-09-09 / listed participation | \nU3b; not UNESCO officer status | \n
| rl-policy-09 | \nControlAI → Lord Hunt: acknowledged support | \n2026-01-29 / support reported by recipient | \nU4; modality/amount/briefing contents unknown | \n
| rl-policy-10 | \nLord Clement-Jones → Global Call: self-reported signatory and parliamentary advocate | \n2026-01-29 / delivered speech | \nU4 | \n
| rl-policy-11 | \nTFS and CeSIA → India AI Safety Connect workshop: co-organized | \n2026-02 / organizer-reported event | \nU5a; no state endorsement inferred | \n
| rl-policy-12 | \nNamed officials in U5a → workshop: interventions reported | \n2026-02 / organizer account | \nU5a; preserve each person's contemporaneous role | \n
No adoption edge proposed for India, UNESCO, UK Government, UNGA, UNSC or Geneva based on this pass. No campaign-funding edge proposed from the UNESCO umbrella funding paragraph. A current institutional co-chair appointment is a separate governance relation, not evidence the appointment was caused by campaign participation.
\nThe September2025 campaign cannot have caused the February2025 Paris summit, June2025 UNESCO umbrella-network launch, or earlier Seoul2024 commitments. Its organizers can have contributed to earlier processes in other capacities; that requires its own dated evidence. Hosting on OECD/UNESCO sites is not by itself institutional endorsement. Attendance and parliamentary speech are useful mechanisms of access, but do not establish agreed policy, government sponsorship, donor instruction, or exclusive access.
\nClosed routes: UN Security Council official PDF202/empty → UN ASR substitute; UN Dialogue homepage/FAQ403 → indexed text and closing ASR only; UNESCO brochure public viewer → exact file401, parked; TFS/Hansard native403 → already-inspected web reader. No continued retries, authenticated sessions, or evasion. An initial Python terminal encoding error occurred after relevant India extraction; it did not invalidate bytes or the web-read annex.
\nThe strongest next acquisition is the actual briefing/transmission document at one conversion point: UNESCO subgroup to a government delegation, TFS/CeSIA workshop to India's working group, or ControlAI to Hunt. These can distinguish policy influence from ordinary participation without enlarging the membership roster.
", "path": "red-lines-wave-2026-09-09/policy-uptake.md", "sourceIds": [ "w8-red-s-unesco-subgroup", "w8-red-s-unesco-coordination", "w8-red-s-unesco-network", "w8-red-s-ressa", "w8-red-s-bengio", "w8-red-s-india", "w8-red-s-india-declaration" ], "markdown": "Observed 2026-09-09. Bounded public-source inquiry; candidate findings only. No canonical-case edits, outreach, publication, or paid access. Scope: identifiable carriers, institutional venues, and the difference between receiving a demand and adopting it. Six evidence clusters follow. The result is documented access plus a continuing civil-society workstream, with no binding adoption of the campaign established by the records inspected here.\n\n## 1. Maria Ressa carried the named campaign into a UN General Assembly speech\n\n**U1 — UN-hosted prepared statement, 22 September 2025.** Speech for UN 80th Anniversary General Assembly (see source references) Maria A. Ressa. Read all four pages; decisive locator PDF page 3, “Three Solutions,” first solution. Ressa explicitly names the Global Call, says more than 200 people will launch it later that day, and asks governments to define international boundaries for unacceptable AI risks. This is a submitted/prepared speech on the UN's statement service, not a General Assembly resolution or proof every word was delivered. It supports a named public carrier and recipient venue. It does not establish an agreed UN position, a campaign drafting contract, or who obtained her speaking invitation.\n\nAcquisition: web PDF parser failed; ordinary public GET succeeded (HTTP 200), parsed in memory with pypdf. Original-byte SHA-256: `add56fc6bb85bf06b39a110e9aa8f3b49064859125287af9378df636a941e471`. Bytes were not retained because this lane was restricted to this note.\n\n## 2. Yoshua Bengio presented the red-lines initiative at the Security Council; the same meeting contained opposition\n\n**U2 — UN-hosted automatic transcript, Security Council meeting 10005, 24 September 2025.** Transcript (see source references) Scoped read of Bengio's intervention, especially the paragraph beginning “Earlier this week,” and the United States intervention. Bengio refers to the newly supported international red lines and calls for government visibility and control. The US speaker rejects international bodies asserting centralized AI governance. These are evidence of presentation and disagreement, not a Council decision or proof the US statement targeted this campaign specifically.\n\nThe service explicitly states its automatic transcripts are **not official records or official UN documents**. Its speaker labelling contains errors; use the original meeting catalogue to identify Bengio, not an automated label alone. Official catalogue S/PV.10005 (see source references) identifies the meeting and Bengio. The English official-record PDF was indexed but not fully acquired: web parser error, ordinary GET HTTP 202 with empty body. The automatic transcript was the bounded substitute. Hash of obtained transcript HTML: `9b6731f529a5f7ab53d673bffccac05409ac2e9db1a4eca7a0db79f51bb1fa34`.\n\n## 3. UNESCO hosts a continuing workstream with campaign organizers, under an express independence disclaimer\n\n**U3a — International AI Red Lines subgroup (see source references) current page, inspected 9 September 2026.** Full substantive page read. Locators: “Subgroup Activities,” “Subgroup Members,” and final disclaimer. The Future Society and Ada Lovelace Institute are co-leads; CeSIA is a member. The page describes a six-month research process and three workshops: New Delhi 18 February, Paris/UNESCO headquarters 26 February, and an Ada-organized virtual roundtable 16 March 2026. It distinguishes future explainer/discussion-paper outputs from a linked two-page brief. Activities are independent CSO/academic work and do not necessarily express UNESCO, member-state or partner views or commit UNESCO. This is an organizational seat and a work programme, not delegated rulemaking authority. HTML hash: `5f2d5268da959589a8352a9ec105dd614f3314acd443445db6dd39f11073df5c`.\n\n**U3b — Multi-Stakeholder Coordination Group (see source references) Scoped read of International AI Red Lines subgroup entries. UNESCO names Cami Rincon of Ada and Tereza Zoumpalova of TFS. The latter's entry identifies her as a Global Call organizer and leader of TFS's red-lines workstream. This connects an identifiable campaign organizer to the network's coordination structure. It does not specify personal veto, budget, or member-state approval rights. HTML hash: `9a142f15220a0ab3a8b3c4dd17ae3bfcf04697f3f573844e3f700b3c1a45630d`.\n\n**U3c — Umbrella network page (see source references) Read purpose, chronology, activities and funding paragraphs. The network launched in June 2025, before the September campaign; members shape its activities. UNESCO reports initial Patrick J. McGovern Foundation funding through its 2025 inclusive ethical-governance project, and an earlier UKRI-funded pilot. This concerns the umbrella network. It establishes neither a Global Call grant nor campaign-specific donor control. HTML hash: `532614bd3086d8a9d30339eadb688bf283a979577c435b67e95b2ce42da9b6f8`.\n\n**Output-access frontier:** the subgroup's visible Download PDF link resolves to UNESCO's SharePoint share (see source references) Public link GET returned viewer HTML; the exact file path exposed by that redirect (`CSO_Redlines_compressed.pdf`) returned HTTP 401. No authentication or bypass attempted; full brief uninspected. Its title/description are website evidence only. Missing decisive record: readable brief plus version/approval record and any transmission to a member-state delegation.\n\n## 4. The campaign reached a UK parliamentary demand; ControlAI support is separately acknowledged\n\n**U4 — Hansard, 29 January 2026 (see source references) “Superintelligent AI.”** Scoped read: Hunt opening at 3.03 pm, Clement-Jones at 3.42 pm, and opening ministerial response at 3.50 pm. Hunt thanks ControlAI for support. Clement-Jones says he joined the Global Call, promotes it, then separately describes ControlAI's UK proposals. He declares DLA Piper consultancy and APPG AI co-chairing. The minister, Baroness Lloyd of Effra, describes context-based regulation and AISI work without accepting a campaign treaty or moratorium in the inspected passage.\n\nThe earned mechanism is acknowledged NGO support and a signatory's explicit parliamentary advocacy. Support is not evidence of a payment, authorship of Hunt's speech, or government commissioning. Clement-Jones's interpretation of the call must not silently replace the campaign's own text. The minister's response is not a comprehensive proof of permanent rejection. Web reader content inspected; native GET returned403, so no original-HTML hash or retained bytes claimed. Needed: the actual briefing sent to Hunt/Clement-Jones, correspondence on requested wording, and any departmental response to the specific campaign proposal.\n\n## 5. India provides a matched access/outcome comparison\n\n**U5a — Organizer's account: TFS, Key Takeaways from the India AI Impact Summit (see source references) February 2026 event; publication date not independently established.** Read “Defining and Governing Unacceptable AI Risks” and “General Reflections.” TFS says it and CeSIA co-organized a session within AI Safety Connect Day, explicitly building on the campaign. It reports interventions from Denmark's tech ambassador Anne Marie Engtoft Meldgaard, Brazil foreign-ministry AI adviser Guilherme Fitzgibbon Alves Pereira, Japan AISI's Akiko Murakami, Singapore IMDA's Wan Sie Lee, and MEP Brando Benifei. This is named organizer-reported access, not independently confirmed government endorsement. Its own reflections call for moving beyond voluntary commitments. Web reader inspected; ordinary GET403; no captured-original hash claimed.\n\n**U5b — Government outcome: PIB/MeitY, New Delhi Declaration (see source references) Posted 21 February 2026; live page includes an endorsement list dated 5 March, so it is not an immutable launch-day snapshot. Full declaration annex read, especially “Secure and Trusted AI” and “Conclusion.” It supports industry-led voluntary measures and a voluntary, non-binding Trusted AI Commons; the conclusion likewise emphasizes voluntary/non-binding instruments. The declaration does not name this campaign or enact its requested internationally enforceable prohibitions. No drafting or transmission record links the side event to this text. HTML hash: `cf42f235597ee26c30934bd44d5c0e7b5f25dba4737595a7cec02fa09c98cab7`.\n\nNeeded to move from attendance to uptake: the workshop paper actually transmitted to the Safe and Trusted AI working group, recipient acknowledgment, tracked draft/declaration edits and decision record. Independent government priorities and pre-existing summit commitments remain adequate alternative explanations for shared language.\n\n## 6. Geneva: do not promote speeches or a summary into a negotiated agreement\n\n**U6a — UN Global Dialogue website (see source references) and FAQ (see source references) Indexed official-source text was inspected; direct web and ordinary GET access returned403. The homepage advertises a Co-Chairs' Summary for Geneva 6–7 July 2026. The FAQ says the Dialogue is not a negotiating forum and ends with a co-chair summary. **The summary itself was not acquired or read**, so no claim is made about whether it incorporates particular red-lines recommendations.\n\n**U6b — UN automatic closing-session transcript (see source references) Scoped review: final co-chair remarks (parsed lines1038–1053) and red-lines search matches. The co-chairs describe diverse views and a summary still to be drafted. The shared-red-lines passage at parsed lines915-918 is a stakeholder intervention, not a negotiated decision. This ASR source has the same non-official-record limitation as U2. It supplies useful context, not an authoritative final outcome. A separate UN Web TV programme, 1 July 2026 (see source references) names Bengio and Ressa as scientific-panel co-chairs and describes independent scientific assessments; it does not give them treaty-making authority. Programme metadata inspected; video not watched.\n\nPriority missing record: actual final Co-Chairs' Summary and annexed submissions, especially any TFS/CeSIA/CHAI contribution, followed by the attribution/version trail. Root should treat this as a defined public-document access gap, not evidence of concealment or adoption.\n\n## Candidate typed relations (not yet canonical)\n\nIDs below are provisional within this note; match existing entity IDs before integration. Evidence status describes the source claim, not a finding of control. No graph edge should jump directly from campaign involvement to a state adopting rules.\n\n| Candidate ID | Direction and relation | Date / action stage | Source and limitation |\n| --- | --- | --- | --- |\n| rl-policy-01 | Maria Ressa → Global Call: publicly advocated | 2025-09-22 / prepared statement | U1; named campaign |\n| rl-policy-02 | Maria Ressa → UNGA80 anniversary session: presented policy demand | 2025-09-22 / submitted speech | U1; not a resolution |\n| rl-policy-03 | Yoshua Bengio → UNSC10005: advocated international AI red lines | 2025-09-24 / speech reported by ASR | U2; official-record corroboration outstanding |\n| rl-policy-04 | The Future Society → UNESCO-network red-lines subgroup: co-lead | observed2026-09-09 / current listed role | U3a; independent CSO work |\n| rl-policy-05 | Ada Lovelace Institute → same subgroup: co-lead | observed2026-09-09 / current listed role | U3a |\n| rl-policy-06 | CeSIA → same subgroup: member | observed2026-09-09 / current listed role | U3a; no control right inferred |\n| rl-policy-07 | Tereza Zoumpalova → Global Call: organizer | historical, observed2026-09-09 | U3b; UNESCO-hosted biography |\n| rl-policy-08 | Tereza Zoumpalova → UNESCO network coordination group: named subgroup representative | observed2026-09-09 / listed participation | U3b; not UNESCO officer status |\n| rl-policy-09 | ControlAI → Lord Hunt: acknowledged support | 2026-01-29 / support reported by recipient | U4; modality/amount/briefing contents unknown |\n| rl-policy-10 | Lord Clement-Jones → Global Call: self-reported signatory and parliamentary advocate | 2026-01-29 / delivered speech | U4 |\n| rl-policy-11 | TFS and CeSIA → India AI Safety Connect workshop: co-organized | 2026-02 / organizer-reported event | U5a; no state endorsement inferred |\n| rl-policy-12 | Named officials in U5a → workshop: interventions reported | 2026-02 / organizer account | U5a; preserve each person's contemporaneous role |\n\nNo adoption edge proposed for India, UNESCO, UK Government, UNGA, UNSC or Geneva based on this pass. No campaign-funding edge proposed from the UNESCO umbrella funding paragraph. A current institutional co-chair appointment is a separate governance relation, not evidence the appointment was caused by campaign participation.\n\n## Causal checks and stopping boundary\n\nThe September2025 campaign cannot have caused the February2025 Paris summit, June2025 UNESCO umbrella-network launch, or earlier Seoul2024 commitments. Its organizers can have contributed to earlier processes in other capacities; that requires its own dated evidence. Hosting on OECD/UNESCO sites is not by itself institutional endorsement. Attendance and parliamentary speech are useful mechanisms of access, but do not establish agreed policy, government sponsorship, donor instruction, or exclusive access.\n\nClosed routes: UN Security Council official PDF202/empty → UN ASR substitute; UN Dialogue homepage/FAQ403 → indexed text and closing ASR only; UNESCO brochure public viewer → exact file401, parked; TFS/Hansard native403 → already-inspected web reader. No continued retries, authenticated sessions, or evasion. An initial Python terminal encoding error occurred after relevant India extraction; it did not invalidate bytes or the web-read annex.\n\nThe strongest next acquisition is the actual briefing/transmission document at one conversion point: UNESCO subgroup to a government delegation, TFS/CeSIA workshop to India's working group, or ControlAI to Hunt. These can distinguish policy influence from ordinary participation without enlarging the membership roster.\n", "file": "research/red-lines-wave-2026-09-09-policy-uptake.md" }, { "id": "red-lines-wave-2026-09-09-root-findings", "title": "AI Red Lines: the operational structure", "description": "Research breakpoint observed 9 September 2026. This is a source-bounded synthesis of the completed first Red Lines wave. The user subsequently authorised unattended continuation; the follow-through lanes remain separate until reviewed. Canonical integration is case-r12; case-r11 and the earlier UK reader stay preserved.", "html": "Research breakpoint observed 9 September 2026. This is a source-bounded synthesis of the completed first Red Lines wave. The user subsequently authorised unattended continuation; the follow-through lanes remain separate until reviewed. Canonical integration is case-r12; case-r11 and the earlier UK reader stay preserved.
\nThe useful change is that several lines now describe work performed. The campaign's own [acknowledgements](http[local research file] identify Charbel-Raphael Segerie as initial drafter, Niki Iliadis and Tereza Zoumpalova as TFS project managers, and Anoush Rima Tatevossian as an adviser on UN engagement. Arthur Grimonpont approached Maria Ressa and coordinated the press conference. Apollo Strategic Communications is credited with preparing/moderating that conference. Miotti, Adam Shimi, ControlAI and FLI supplied feedback in different capacities. These are retrospective organizer credits; they establish a reported division of labour, not final approval rights or donor instructions. February/July references lack an explicit year.
\nThis matters to the Hubinger inquiry. The campaign [lists him as an individual signatory](http[local research file] while the acknowledgements identify Shimi as a contributor. The roles differ. Neither record establishes Hubinger as a campaign manager or an author of A Narrow Path. Kidron also appears on the signatory roster; roster inclusion alone is a weaker bridge than the policy presentation below.
\nEffektiv Spenden's H1/2024 disclosure.
\nThe [SFF table](http[local research file] adds a second channel. It records a 2024 USD48,000 CeSIA recommendation, naming Tallinn as source and EffiSciences as receiving charity; TFS recommendations are USD97,000 in2024 and USD336,000 in2025. Its2023H2 TFS row names USD365,000 from FLI and USD11,000 from Tallinn. These are expressly recommendations for general support. Neither disbursement nor campaign restriction follows from a row in that table.
\nCeSIA's [diplomatic-sprint solicitation](http[local research file] is more specific about intended action. It seeks USD150,000 for early2026 work, estimates roughly USD400,000 for a year, and proposes combining lab-threshold harmonization with a champion-state/UN-resolution strategy. It anticipates that Anthropic and DeepMind would have competitive reasons to support common thresholds. That is the organizer's stated incentive model, not proof of a deal with either company. Its fallback includes seeking externally imposed thresholds if labs decline harmonization.
\nThe same page reports a USD2million Longview recommendation dated9December2025, expressly unreceived when that passage was written. The update date is unknown. Manifund's displayed USD0 only describes that platform. We cannot turn either statement into a current total of CeSIA funding, an executed award, or a Red Lines-restricted grant. APCO is listed among proposed advisers/partners; a paid lobbying contract has not been acquired.
\nIn its [25September2025 article](http[local research file] ControlAI explicitly acknowledges that the general call leaves the particular red lines unspecified. It then promotes its preferred ASI prohibition and A Narrow Path. This is observable advocacy using a coalition's common objective to advance a particular implementation. It does not show all signatories consented to that implementation, nor that partners hid their differences. CeSIA's threshold-harmonization proposal supplies an independently articulated route within the same campaign environment.
\nThat suggests a useful working interpretation: a coalition assembles endorsements around a broad demand, while its participating organizations develop and promote more specific policy instruments. Coordination on the campaign is documented through attributed work. A single command centre, common final regime and donor editorial control remain separate hypotheses requiring different evidence.
\nThe [UNESCO-hosted subgroup page](http[local research file] lists TFS and Ada Lovelace Institute as co-leads and CeSIA as a member. Tereza is named in its coordination structure. This supplies continuity beyond the launch. The page expressly says the independent CSO work does not commit UNESCO or member states. The umbrella network began inJune2025; its reported McGovern support and earlier UKRI-funded pilot must not be relabelled September campaign funding.
\nThe [Athens recap announcement](http[local research file] ties the4December2025 London event to the Global Call. Root retained the65-page original PDF and visually read pages2 and23. Page2 credits TFS organization, OECD/UNESCO partnership and support from Fathom, Arnold & Porter, Cooperative AI Foundation and deliberAIde. Support form, amount, precise contracting entities and programme rights are undisclosed by those credits.
\nPage23 reports 5Rights executive director Leanda Barrington-Leach advocating the UK's Age Appropriate Design Code as a model for AI safety by design. This is a specific policy model being carried into the AI-governance forum. The same page separately attributes Greece's kids-wallet age-verification example to Ambassador George Pagoulatos. Do not attribute his example to Leanda, infer that the Global Call demands it, or say age verification was absent from the event. Video links were extracted but the video was not watched; the organizer's recap remains the source of the account.
\n[Hansard](http[local research file] records Clement-Jones advocating the call and Hunt thanking ControlAI. A UN-hosted prepared Ressa speech names the campaign. Bengio's Security Council intervention is available in a clearly non-official automatic transcript. These locate recipients and carriers. None is a resolution adopting the campaign. The government's New Delhi declaration separately emphasises voluntary instruments; an event held alongside the summit is not itself a negotiated outcome.
\nThe independent review also distinguishes two Delhi events on18February2026: the UNESCO-network Global South roundtable and the TFS/CeSIA AI Safety Connect session. Date and city do not make them one workshop.
\nThe most valuable missing material sits at conversion points: the recommendation becoming an executed grant; organizer money becoming campaign expenditure; collaborative drafting becoming final approval; a public statement becoming a government submission; a submission becoming an accepted amendment. Recommendation letters, remittance schedules, restricted-fund ledgers, PR contracts, version history and briefing-transmission records have identifiable likely custodians. We have not established that every proposed agreement was executed or that every sought record exists in final form.
\nThe continued inquiry is testing later CeSIA financing disclosures, France/G7 outcomes against the organizer's stated plan, and the Athens support layer. It does not reopen the closed UN/UNESCO access routes, Longview Drive acquisition or TFS scanned990 branch. No outreach, paid acquisition or public release has occurred. The broad geopolitical account remains the user's attributed prior, not a verified conclusion of this case.
\nSource custody: campaign-formation.md, organizer-funding.md and policy-uptake.md retain exact reading scopes; independent-review.md records the challenge. originals-manifest.json describes this wave's additional captures. The Athens original is recorded separately in the canonical source with its SHA256; only two pages were visually inspected. A failed native release-PDF retrieval does not invalidate the earlier public-reader text, and no original bytes are claimed for that release.
\nFathom is now resolved to the AI-governance nonprofit at fathom.org. Its current partnership page identifies TFS collaboration on verified redlines; the event history separately acknowledges financial support for the Athens series without allocating it to2025. CAIF's own account adds reception hosting and a contribution advocating multi-agent monitoring. Its foundation page reports a USD15million Macroscopic commitment and lists a funder-associated trustee; this does not establish donor appointment rights or an event grant. deliberAIde's provider-hosted TFS testimonial supports product use in the event/report workflow. The operator's legal identity remains unresolved after a browser administrative-policy refusal. See athens-support-layer.md for exact sources and access scope.
\nThe France/G7 comparison distinguishes an institutional event programme from a completed briefing. The May29 official declaration commits to discussion of risk-framework comparability and also promotes AI openness and adoption; it does not adopt the specific coordinated-slowdown trigger sought in the CeSIA proposal. September targets remain within their target month at observation. Root securely captured the official five-page PDF and read page2; the policy worker read the complete text. See followthrough-france-g7.md and followthrough-review.md.
\nThe next promising instruments are the Fathom-sponsored California verification bill, the newer CAIF/ARIA/industry multi-agent-safety fund, and the G7 minors declaration. These remain acquisition leads. Striving is active through the hourly task heartbeat and current-work.md.
", "path": "red-lines-wave-2026-09-09/root-findings.md", "sourceIds": [ "w8-red-s-signatories", "w8-red-s-controlai", "w8-red-s-athens", "w8-red-s-athens-page" ], "markdown": "Research breakpoint observed 9 September 2026. This is a source-bounded synthesis of the completed first Red Lines wave. The user subsequently authorised unattended continuation; the follow-through lanes remain separate until reviewed. Canonical integration is case-r12; case-r11 and the earlier UK reader stay preserved.\n\nThe useful change is that several lines now describe work performed. The campaign's own acknowledgements (see source references) identify Charbel-Raphael Segerie as initial drafter, Niki Iliadis and Tereza Zoumpalova as TFS project managers, and Anoush Rima Tatevossian as an adviser on UN engagement. Arthur Grimonpont approached Maria Ressa and coordinated the press conference. Apollo Strategic Communications is credited with preparing/moderating that conference. Miotti, Adam Shimi, ControlAI and FLI supplied feedback in different capacities. These are retrospective organizer credits; they establish a reported division of labour, not final approval rights or donor instructions. February/July references lack an explicit year.\n\nThis matters to the Hubinger inquiry. The campaign lists him as an individual signatory (see source references) while the acknowledgements identify Shimi as a contributor. The roles differ. Neither record establishes Hubinger as a campaign manager or an author of A Narrow Path. Kidron also appears on the signatory roster; roster inclusion alone is a weaker bridge than the policy presentation below.\n\n## Resources and the strategy they were intended to support\n\nEffektiv Spenden's H1/2024 disclosure.\n\nThe SFF table (see source references) adds a second channel. It records a 2024 USD48,000 CeSIA recommendation, naming Tallinn as source and EffiSciences as receiving charity; TFS recommendations are USD97,000 in2024 and USD336,000 in2025. Its2023H2 TFS row names USD365,000 from FLI and USD11,000 from Tallinn. These are expressly recommendations for general support. Neither disbursement nor campaign restriction follows from a row in that table.\n\nCeSIA's diplomatic-sprint solicitation (see source references) is more specific about intended action. It seeks USD150,000 for early2026 work, estimates roughly USD400,000 for a year, and proposes combining lab-threshold harmonization with a champion-state/UN-resolution strategy. It anticipates that Anthropic and DeepMind would have competitive reasons to support common thresholds. That is the organizer's stated incentive model, not proof of a deal with either company. Its fallback includes seeking externally imposed thresholds if labs decline harmonization.\n\nThe same page reports a USD2million Longview recommendation dated9December2025, expressly unreceived when that passage was written. The update date is unknown. Manifund's displayed USD0 only describes that platform. We cannot turn either statement into a current total of CeSIA funding, an executed award, or a Red Lines-restricted grant. APCO is listed among proposed advisers/partners; a paid lobbying contract has not been acquired.\n\n## A broad call leaves room for different implementations\n\nIn its 25September2025 article (see source references) ControlAI explicitly acknowledges that the general call leaves the particular red lines unspecified. It then promotes its preferred ASI prohibition and A Narrow Path. This is observable advocacy using a coalition's common objective to advance a particular implementation. It does not show all signatories consented to that implementation, nor that partners hid their differences. CeSIA's threshold-harmonization proposal supplies an independently articulated route within the same campaign environment.\n\nThat suggests a useful working interpretation: a coalition assembles endorsements around a broad demand, while its participating organizations develop and promote more specific policy instruments. Coordination on the campaign is documented through attributed work. A single command centre, common final regime and donor editorial control remain separate hypotheses requiring different evidence.\n\n## Institutional channels and a substantive age-policy bridge\n\nThe UNESCO-hosted subgroup page (see source references) lists TFS and Ada Lovelace Institute as co-leads and CeSIA as a member. Tereza is named in its coordination structure. This supplies continuity beyond the launch. The page expressly says the independent CSO work does not commit UNESCO or member states. The umbrella network began inJune2025; its reported McGovern support and earlier UKRI-funded pilot must not be relabelled September campaign funding.\n\nThe Athens recap announcement (see source references) ties the4December2025 London event to the Global Call. Root retained the65-page original PDF and visually read pages2 and23. Page2 credits TFS organization, OECD/UNESCO partnership and support from Fathom, Arnold & Porter, Cooperative AI Foundation and deliberAIde. Support form, amount, precise contracting entities and programme rights are undisclosed by those credits.\n\nPage23 reports 5Rights executive director Leanda Barrington-Leach advocating the UK's Age Appropriate Design Code as a model for AI safety by design. This is a specific policy model being carried into the AI-governance forum. The same page separately attributes Greece's kids-wallet age-verification example to Ambassador George Pagoulatos. Do not attribute his example to Leanda, infer that the Global Call demands it, or say age verification was absent from the event. Video links were extracted but the video was not watched; the organizer's recap remains the source of the account.\n\nHansard (see source references) records Clement-Jones advocating the call and Hunt thanking ControlAI. A UN-hosted prepared Ressa speech names the campaign. Bengio's Security Council intervention is available in a clearly non-official automatic transcript. These locate recipients and carriers. None is a resolution adopting the campaign. The government's New Delhi declaration separately emphasises voluntary instruments; an event held alongside the summit is not itself a negotiated outcome.\n\nThe independent review also distinguishes two Delhi events on18February2026: the UNESCO-network Global South roundtable and the TFS/CeSIA AI Safety Connect session. Date and city do not make them one workshop.\n\n## The next records that could change the interpretation\n\nThe most valuable missing material sits at conversion points: the recommendation becoming an executed grant; organizer money becoming campaign expenditure; collaborative drafting becoming final approval; a public statement becoming a government submission; a submission becoming an accepted amendment. Recommendation letters, remittance schedules, restricted-fund ledgers, PR contracts, version history and briefing-transmission records have identifiable likely custodians. We have not established that every proposed agreement was executed or that every sought record exists in final form.\n\nThe continued inquiry is testing later CeSIA financing disclosures, France/G7 outcomes against the organizer's stated plan, and the Athens support layer. It does not reopen the closed UN/UNESCO access routes, Longview Drive acquisition or TFS scanned990 branch. No outreach, paid acquisition or public release has occurred. The broad geopolitical account remains the user's attributed prior, not a verified conclusion of this case.\n\nSource custody: campaign-formation.md, organizer-funding.md and policy-uptake.md retain exact reading scopes; independent-review.md records the challenge. originals-manifest.json describes this wave's additional captures. The Athens original is recorded separately in the canonical source with its SHA256; only two pages were visually inspected. A failed native release-PDF retrieval does not invalidate the earlier public-reader text, and no original bytes are claimed for that release.\n\n## Reviewed follow-through integrated into case-r13\n\nFathom is now resolved to the AI-governance nonprofit at fathom.org. Its current partnership page identifies TFS collaboration on verified redlines; the event history separately acknowledges financial support for the Athens series without allocating it to2025. CAIF's own account adds reception hosting and a contribution advocating multi-agent monitoring. Its foundation page reports a USD15million Macroscopic commitment and lists a funder-associated trustee; this does not establish donor appointment rights or an event grant. deliberAIde's provider-hosted TFS testimonial supports product use in the event/report workflow. The operator's legal identity remains unresolved after a browser administrative-policy refusal. See athens-support-layer.md for exact sources and access scope.\n\nThe France/G7 comparison distinguishes an institutional event programme from a completed briefing. The May29 official declaration commits to discussion of risk-framework comparability and also promotes AI openness and adoption; it does not adopt the specific coordinated-slowdown trigger sought in the CeSIA proposal. September targets remain within their target month at observation. Root securely captured the official five-page PDF and read page2; the policy worker read the complete text. See followthrough-france-g7.md and followthrough-review.md.\n\nThe next promising instruments are the Fathom-sponsored California verification bill, the newer CAIF/ARIA/industry multi-agent-safety fund, and the G7 minors declaration. These remain acquisition leads. Striving is active through the hourly task heartbeat and current-work.md.\n", "file": "research/red-lines-wave-2026-09-09-root-findings.md" }, { "id": "red-threads-wave-2026-09-08-australia-findings", "title": "Australia trial: objections, revisions and review authority", "description": "The missing public board sequence is now acquired. It documents responses, some changed procedures and limits. It also identifies a consequential boundary: Corby's declared exclusion covered testing and euCONSENT submission, while July minutes record his intention to review technical and vendor report sections. That is a question about the scope and operatio", "html": "Nova · 8 September 2026 · Bounded wave w1-australia
\nThe missing public board sequence is now acquired. It documents responses, some changed procedures and retained limits. It also identifies a consequential boundary: Corby's declared exclusion covered testing and euCONSENT submission, while July minutes record his intention to review technical and vendor report sections. That is a question about the scope and operation of the exclusion, not proof of its breach.
\nMarch meeting. The published ten-page minutes record objections concerning children, sample diversity, safeguarding and circumvention, with actions to circulate the safeguarding policy and arrange expert follow-up. The document likely corresponds to DIGI's March14 reference, but its cover says March13 and its body says Friday13February. Preserve those conflicting dates pending the meeting invitation or corrected record. [Original minutes, pp5–10](http[local research file]
\nApril29 response. The next minutes report follow-up correspondence, a completed mid-project impartiality report, an eSafety boundary MOU, and introduced vendor-selection criteria. Scout Insights clarifies that recruitment uses its registered panel, not employees/their children. Whether this corrected a misunderstanding or changed a plan remains unresolved. A request to record parent-child consent conflicts was kept outside scope. [Original minutes, PDF pp3–4 and9](http[local research file]
\nJune17 challenge continued. DIGI questioned the privacy comparison between independent providers and platform systems; participants challenged how practical testing would be described. The team reported all testing complete, while explaining that some technologies were assessed through documents/interviews. Minutes record a commitment to revisit wording. These are concerns and responses, not a unanimous endorsement. [Original minutes, pp5–7](http[local research file]
\nJuly8 review access. SAB received A/B/K; vendors received extracts concerning their own systems. Detailed C–J sections remained unavailable to the board during legal review. Corby committed to review B–J for consistency. Vendor replies were sought by July18. This establishes differentiated review access and a recorded intention, not actual edits or a recusal breach. [Original minutes, pp3 and6–8](http[local research file] Its control header also has stale dates/meeting numbering, so use the explicit July8 body date with the discrepancy retained.
\nFinal methodology account. Part B reports changes from stakeholder feedback, including usability clarification and more realistic circumvention scenarios; it preserves exclusions. It identifies minuted ethics meetings and eSafety comments under the MOU. The stakeholder-response table derives from prior minutes, so it is not independent corroboration of them. [Part B, B.14/B.17/B.18, PDF pp52–55,66–68,74](http[local research file]
\nThe evaluation proposal is dated February6,2025. Its pp51–55 already describe usability and circumvention work with limits; this does not settle whether the planned coverage was adequate. Cover version0.3 conflicts with a history extending to version2.0. [Original proposal](http[local research file]
\nThe human-subject protocol's filename says V1.3, but its cover, approvals and revision history have inconsistent years. Its history describes collection-notice revisions. We have the documents and reported responses, not a clean approved-version comparison; no change is attributed specifically to DIGI without the intervening record. [Original DOCX](http[local research file]
\n| Evidence stage | \nCurrent result | \n
|---|---|
| Objections raised | \nOriginal March, June and July minutes; retained DIGI letter. | \n
| Responses recorded | \nApril decisions/clarifications, follow-up email identified, final response table. | \n
| Safeguards said to operate | \nMOU signed, selection criteria introduced, procedural/recruitment accounts. | \n
| Actual individual recusal operation | \nAccess/assignment/signoff and conflict-decision records not obtained. | \n
| Approved before/after protocol | \nPublic versions acquired, but inconsistent controls and missing intervening versions prevent reconstruction. | \n
| Independent outcome evidence | \nNo independent recusal audit acquired; raw testing data not reanalysed in this lane. | \n
DIGI's April16 complaint remains an interested party's contemporaneous representation. Its members faced regulation and some submitted trial technologies. AVPA representation and ACCS's commercial certification interests remain on the other side of the analysis. Neither industry's stake decides whose criticism is correct. [Retained letter, PDF pp10–13](http[local research file] [retained conflict-register source](http[local research file] The latter was read from the September7 capture, not refreshed in this wave.
\nAn original departmental briefing now directly confirms contract CN4111198 at A$3,830,277, awarded November8,2024. It describes seven tenders and the department's selection rationale. Locator: FOI26-145 Document12, SB25-000302 version02 dated September24,2025, PDF pp87–88 / printed9–10. This upgrades the old e15 award edge; it is the same procurement, not another transfer. [Official briefing](http[local research file]
\nThe newly identified FOI26-176 search index exposes this row under a GST-inclusive value heading:
\n| Contract | \nSupplier as indexed | \nStart | \nEnd | \nIndexed value | \n
|---|---|---|---|---|
| CN4111198 | \nAVID CERTIFICATION SERVICES LTD AGE CHECK CERTIFICATION SERVICES P/ | \n08-Nov-24 | \n30-Jun-25 | \n3,830,277.00 | \n
[Indexed official register](http[local research file] The full PDF failed retrieval, so this is still index-level tax/row evidence. The inspected briefing does not specify GST. No contract-payment total or consortium allocation is inferred; the broader A$6.5m programme envelope is a different quantity.
\nThe official publication index timed out in the web parser; one ordinary public GET succeeded and exposed the linked originals. Captured files are listed below. No private collaboration system was accessed.
\nFOI26-154 returned a web error and its public GET timed out at35seconds. The independent FOI25-288 register copy returned a web error and timed out at15seconds. FOI26-176 exceeded the web tool's14.25MB limit; one ordinary GET timed out at15seconds. Each route stopped. The original [AusTender entry](http[local research file] returned403 and was not retried. Its link came from TenderTracker; the aggregator's inconsistent status and shifted dates were excluded.
\nThe obtained public record supports a useful correction: a generic “safeguards were merely declared and objections unanswered” account is too strong. There were recorded responses. It also does not establish adequate enforcement, a representative achieved sample, or that reviewers respected the operative boundary. Separate regulator input, vendor review, advisory challenge and final authorial control instead of collapsing them into one committee.
\nThe decisive next instruments are already identifiable. The secretariat/ACCS and Amanda Third hold the referenced follow-up correspondence; project leads hold approved revision history and task assignments. The ethics/impartiality function should hold the applicable exclusion decision and review signoffs. eSafety and project legal hold the signed MOU and draft comments. Scout Insights holds the recruitment specification and achieved-sample record. The department and AVID/ACCS hold CN4111198, variations and invoices. These are specific holder gaps, not evidence of refusal or concealed wrongdoing. No outreach occurred.
\n| File in this wave folder | \nExtent | \n
|---|---|
| australia-publications.html | \nOriginal index inspected for links | \n
| australia-march-minutes.pdf (+.txt) | \nAll10 pages read | \n
| australia-april-minutes.pdf (+.txt) | \nSelected pp3–6,9 read | \n
| australia-june-minutes.pdf (+.txt) | \nSelected pp3–7 passages read | \n
| australia-july-minutes.pdf (+.txt) | \nAll8 pages read | \n
| australia-evaluation-proposal.pdf (+.txt) | \n88pages captured; selected control/method sections read | \n
| australia-human-protocol-v1-3.docx (+.txt) | \nControl, recruitment and collection-notice introduction read | \n
| australia-final-part-b.pdf (+.txt) | \n138pages captured; selected governance, response, liaison and plan sections read | \n
Exact SHA-256 values, source locators, ancestry, failed routes and integration hints are in australia-candidate.json. Raw capture is not a claim of full substantive reading. The packet contains four new nodes, seven relationships, one proposed evidence upgrade to e15, and four parked holder-record leads. Alias retained source records by URL; do not double-count the same source, contract or derivative response table.
\nNo canonical case, explorer, automation or other campaign file was changed.
", "path": "red-threads-wave-2026-09-08/australia-findings.md", "sourceIds": [ "src-a61d78086a00", "src-81bf267f91d0", "w1-australia-src-publications", "w1-australia-src-march-minutes", "w1-australia-src-april-minutes", "w1-australia-src-june-minutes", "w1-australia-src-july-minutes", "w1-australia-src-evaluation", "w1-australia-src-human-protocol", "w1-australia-src-final-responses", "w1-australia-src-procurement-brief", "w1-australia-src-gst-register", "w1-australia-src-foi26154-failed", "w1-australia-src-foi25288-failed", "w1-australia-src-austender-failed" ], "markdown": "Nova · 8 September 2026 · Bounded wave w1-australia\n\nThe missing public board sequence is now acquired. It documents responses, some changed procedures and retained limits. It also identifies a consequential boundary: Corby's declared exclusion covered testing and euCONSENT submission, while July minutes record his intention to review technical and vendor report sections. That is a question about the scope and operation of the exclusion, not proof of its breach.\n\n## The newly recovered decision trail\n\n**March meeting.** The published ten-page minutes record objections concerning children, sample diversity, safeguarding and circumvention, with actions to circulate the safeguarding policy and arrange expert follow-up. The document likely corresponds to DIGI's March14 reference, but its cover says March13 and its body says Friday13February. Preserve those conflicting dates pending the meeting invitation or corrected record. Original minutes, pp5–10 (see source references)\n\n**April29 response.** The next minutes report follow-up correspondence, a completed mid-project impartiality report, an eSafety boundary MOU, and introduced vendor-selection criteria. Scout Insights clarifies that recruitment uses its registered panel, not employees/their children. Whether this corrected a misunderstanding or changed a plan remains unresolved. A request to record parent-child consent conflicts was kept outside scope. Original minutes, PDF pp3–4 and9 (see source references)\n\n**June17 challenge continued.** DIGI questioned the privacy comparison between independent providers and platform systems; participants challenged how practical testing would be described. The team reported all testing complete, while explaining that some technologies were assessed through documents/interviews. Minutes record a commitment to revisit wording. These are concerns and responses, not a unanimous endorsement. Original minutes, pp5–7 (see source references)\n\n**July8 review access.** SAB received A/B/K; vendors received extracts concerning their own systems. Detailed C–J sections remained unavailable to the board during legal review. Corby committed to review B–J for consistency. Vendor replies were sought by July18. This establishes differentiated review access and a recorded intention, not actual edits or a recusal breach. Original minutes, pp3 and6–8 (see source references) Its control header also has stale dates/meeting numbering, so use the explicit July8 body date with the discrepancy retained.\n\n**Final methodology account.** Part B reports changes from stakeholder feedback, including usability clarification and more realistic circumvention scenarios; it preserves exclusions. It identifies minuted ethics meetings and eSafety comments under the MOU. The stakeholder-response table derives from prior minutes, so it is not independent corroboration of them. Part B, B.14/B.17/B.18, PDF pp52–55,66–68,74 (see source references)\n\n## What can and cannot be compared before/after\n\nThe evaluation proposal is dated February6,2025. Its pp51–55 already describe usability and circumvention work with limits; this does not settle whether the planned coverage was adequate. Cover version0.3 conflicts with a history extending to version2.0. Original proposal (see source references)\n\nThe human-subject protocol's filename says V1.3, but its cover, approvals and revision history have inconsistent years. Its history describes collection-notice revisions. We have the documents and reported responses, not a clean approved-version comparison; no change is attributed specifically to DIGI without the intervening record. Original DOCX (see source references)\n\n| Evidence stage | Current result |\n| --- | --- |\n| Objections raised | Original March, June and July minutes; retained DIGI letter. |\n| Responses recorded | April decisions/clarifications, follow-up email identified, final response table. |\n| Safeguards said to operate | MOU signed, selection criteria introduced, procedural/recruitment accounts. |\n| Actual individual recusal operation | Access/assignment/signoff and conflict-decision records not obtained. |\n| Approved before/after protocol | Public versions acquired, but inconsistent controls and missing intervening versions prevent reconstruction. |\n| Independent outcome evidence | No independent recusal audit acquired; raw testing data not reanalysed in this lane. |\n\nDIGI's April16 complaint remains an interested party's contemporaneous representation. Its members faced regulation and some submitted trial technologies. AVPA representation and ACCS's commercial certification interests remain on the other side of the analysis. Neither industry's stake decides whose criticism is correct. Retained letter, PDF pp10–13 (see source references) retained conflict-register source (see source references) The latter was read from the September7 capture, not refreshed in this wave.\n\n## Procurement: evidence upgraded, tax/perimeter gap retained\n\nAn original departmental briefing now directly confirms contract CN4111198 at **A$3,830,277**, awarded November8,2024. It describes seven tenders and the department's selection rationale. Locator: FOI26-145 Document12, SB25-000302 version02 dated September24,2025, PDF pp87–88 / printed9–10. This upgrades the old e15 award edge; it is the same procurement, not another transfer. Official briefing (see source references)\n\nThe newly identified FOI26-176 search index exposes this row under a GST-inclusive value heading:\n\n| Contract | Supplier as indexed | Start | End | Indexed value |\n| --- | --- | --- | --- | --- |\n| CN4111198 | AVID CERTIFICATION SERVICES LTD AGE CHECK CERTIFICATION SERVICES P/ | 08-Nov-24 | 30-Jun-25 | 3,830,277.00 |\n\nIndexed official register (see source references) The full PDF failed retrieval, so this is still index-level tax/row evidence. The inspected briefing does not specify GST. No contract-payment total or consortium allocation is inferred; the broader A$6.5m programme envelope is a different quantity.\n\n## Acquisition, custody and stopping points\n\nThe official publication index timed out in the web parser; one ordinary public GET succeeded and exposed the linked originals. Captured files are listed below. No private collaboration system was accessed.\n\nFOI26-154 returned a web error and its public GET timed out at35seconds. The independent FOI25-288 register copy returned a web error and timed out at15seconds. FOI26-176 exceeded the web tool's14.25MB limit; one ordinary GET timed out at15seconds. Each route stopped. The original AusTender entry (see source references) returned403 and was not retried. Its link came from TenderTracker; the aggregator's inconsistent status and shifted dates were excluded.\n\nThe obtained public record supports a useful correction: a generic “safeguards were merely declared and objections unanswered” account is too strong. There were recorded responses. It also does not establish adequate enforcement, a representative achieved sample, or that reviewers respected the operative boundary. Separate regulator input, vendor review, advisory challenge and final authorial control instead of collapsing them into one committee.\n\nThe decisive next instruments are already identifiable. The secretariat/ACCS and Amanda Third hold the referenced follow-up correspondence; project leads hold approved revision history and task assignments. The ethics/impartiality function should hold the applicable exclusion decision and review signoffs. eSafety and project legal hold the signed MOU and draft comments. Scout Insights holds the recruitment specification and achieved-sample record. The department and AVID/ACCS hold CN4111198, variations and invoices. These are specific holder gaps, not evidence of refusal or concealed wrongdoing. No outreach occurred.\n\n## Files acquired\n\n| File in this wave folder | Extent |\n| --- | --- |\n| australia-publications.html | Original index inspected for links |\n| australia-march-minutes.pdf (+.txt) | All10 pages read |\n| australia-april-minutes.pdf (+.txt) | Selected pp3–6,9 read |\n| australia-june-minutes.pdf (+.txt) | Selected pp3–7 passages read |\n| australia-july-minutes.pdf (+.txt) | All8 pages read |\n| australia-evaluation-proposal.pdf (+.txt) | 88pages captured; selected control/method sections read |\n| australia-human-protocol-v1-3.docx (+.txt) | Control, recruitment and collection-notice introduction read |\n| australia-final-part-b.pdf (+.txt) | 138pages captured; selected governance, response, liaison and plan sections read |\n\nExact SHA-256 values, source locators, ancestry, failed routes and integration hints are in australia-candidate.json. Raw capture is not a claim of full substantive reading. The packet contains four new nodes, seven relationships, one proposed evidence upgrade to e15, and four parked holder-record leads. Alias retained source records by URL; do not double-count the same source, contract or derivative response table.\n\nNo canonical case, explorer, automation or other campaign file was changed.\n", "file": "research/red-threads-wave-2026-09-08-australia-findings.md" }, { "id": "red-threads-wave-2026-09-08-baseline-evidence-scope", "title": "Baseline evidence custody", "description": "The reviewed atlas and every original node/relationship field were in case-r1.json and atlas-baseline-original.json. This import does not reopen or newly inspect the 117 cited URLs. Sources initially use discovered because their individual historical reading scopes are not all reconciled here. Existing evidence_class and review_scope fields, titles, qualifie", "html": "The reviewed atlas and every original node/relationship field were retained in case-r1.json and atlas-baseline-original.json. This import does not reopen or newly inspect the 117 cited URLs. Sources initially use discovered because their individual historical reading scopes are not all reconciled here. Existing evidence_class and review_scope fields, titles, qualifiers and source objects remain beside the converted records. Later source-specific inspections must record the actual scope and location without claiming full-corpus freshness. Historical is temporal; each proposal record has an explicitly staged verb. No effective interval was invented from an observation date.
", "path": "red-threads-wave-2026-09-08/baseline-evidence-scope.md", "sourceIds": [ "src-477c83c16019", "src-b7bb716eacc0", "src-c401141d51b7", "src-b9795844a727", "src-24fdb5685974", "src-44c03bd0a530", "src-82c15a7cd2b7", "src-547a5b3ea786", "src-c5c76c4cf21e", "src-bad5f974d0c4", "src-d17247535809", "src-ef55ca30efaa", "src-a82329c94eac", "src-aaae21f08c81", "src-f762980e9c03", "src-ef8d4d859fc2", "src-73e997a6d1ef", "src-71475aa32b34", "src-95a63bfefeca", "src-d7fd807078d3", "src-0d707744efd9", "src-b8a63a1347f2", "src-94263f452256", "src-d451e2202e83", "src-033fd1e2f487", "src-4711920c1730", "src-cf6c0895df42", "src-29f8a10f712c", "src-8016e12c6894", "src-0764ac1a884e", "src-1a8f761e951d", "src-e07be722d815", "src-2ed7c939cd31", "src-1d6bd7e4b8fa", "src-6a9425acbb0c", "src-23397b6fa896", "src-a96ee8404df0", "src-1b802ead5bff", "src-e7bf79d5bdc8", "src-42725da209c6", "src-8638b6bbbdfc", "src-f40ef2b0252e", "src-e996011534a6", "src-604eca7e0531", "src-eb7a62848703", "src-22304018d7f5", "src-f4fef5547775", "src-b2d0264369e7", "src-5747555ff9d7", "src-9982667e53da", "src-b6e607a7fd65", "src-1bdb2fcb24d8", "src-8efa5883d584", "src-a69f9d4e2a03", "src-210ecbb4dfbb", "src-11e9bb8515f1", "src-7f2794f6a070", "src-4fe7b92dcece", "src-5e4e930cec53", "src-dc138c91500c", "src-1bc441d0528b", "src-dbb8d89ffb52", "src-6dbd58c5e8f8", "src-8e230246a046", "src-c7e428b699d3", "src-3db31f719414", "src-0f5ce69a6784", "src-8806cf368b7f", "src-9d9e4fbe80b2", "src-90f14526fca6", "src-478c52027034", "src-81ff54e1c713", "src-1e184fdc4ff9", "src-1f9134b8bbcc", "src-33a85b0bc3bd", "src-421fa29c369c", "src-c2da4caaaa8e", "src-953ce6b745ed", "src-2af30dede1cb", "src-50b299128d80", "src-5c7f3011d794", "src-1f84b8a8e5b4", "src-9537facf4ef2", "src-ac43e1b525d3", "src-4cc57af3ad93", "src-b4a2b9e3261c", "src-ed3d2ac6a354", "src-6df46b543f84", "src-7e9b476e322d", "src-685569dc8514", "src-27738cfd6708", "src-808a0af999fb", "src-cda852240051", "src-306689bfa358", "src-9e1956e1c0db", "src-3bc8f45060eb", "src-718684d59ccd", "src-767c5b45b782", "src-b1a700934b28", "src-f3950ad6f80e", "src-14c9e7f34568", "src-2615708ffd63" ], "markdown": "The reviewed atlas and every original node/relationship field were retained in case-r1.json and atlas-baseline-original.json. This import does not reopen or newly inspect the 117 cited URLs. Sources initially use discovered because their individual historical reading scopes are not all reconciled here. Existing evidence_class and review_scope fields, titles, qualifiers and source objects remain beside the converted records. Later source-specific inspections must record the actual scope and location without claiming full-corpus freshness. Historical is temporal; each proposal record has an explicitly staged verb. No effective interval was invented from an observation date.\n", "file": "research/red-threads-wave-2026-09-08-baseline-evidence-scope.md" }, { "id": "red-threads-wave-2026-09-08-controlai-findings", "title": "ControlAI acquisition findings — 8 September 2026", "description": "This bounded pass recovered a new original US tax return and read the UK company's filed decision rules. It did not identify the US related-party creditor or resolve the UK combined accrual/deferred-income balance. No outreach, paid access or canonical-map edits were undertaken. Facts below distinguish a party's official filing from independent verification ", "html": "This bounded pass recovered a new original US tax return and read the UK company's filed decision rules. It did not identify the US related-party creditor or resolve the UK combined accrual/deferred-income balance. No outreach, paid access or canonical-map edits were undertaken. Facts below distinguish a party's official filing from independent verification of its substance.
\nThe IRS official [2026 XML index](http[local research file] selects EIN 33-3361748, CONTROL AI US INC, tax period 202512, return 990EZ, object202611749349201056, batch2026_TEOS_XML_06A. The original member 202611749349201056_public.xml was recovered from the [official IRS ZIP](http[local research file] using ordinary public byte-range requests. This is a new custodian route; earlier failed third-party original-return endpoints were not repeated.
The return covers 4 February–31 December 2025, uses accrual accounting, marks itself initial, and was signed by Andrea Miotti on 23 June 2026. It reports $136,570 contributions, $377 bank-fee refunds, total revenue $136,947, expenses $93,682, assets $69,178, liabilities $25,913 and net assets $43,265. These totals were previously available through a secondary listing; they now have original-return support.
\nThe consequential new detail is Schedule O, Form 990-EZ Part II line 26. Exact filed excerpt: DESCRIPTION: DUE TO RELATED ORGANIZATION. BEG. OF YEAR AMOUNT: 0. END OF YEAR AMOUNT: 20,955. The same line separately reports $4,958 accounts payable and accrued expenses. The two liabilities sum to $25,913. This is a balance owed as of 31 December 2025, not a disclosed cash payment, annual grant or loan commitment.
The creditor is unnamed. The released XML does not say whether this is a shared-cost reimbursement, loan/advance or other related-party accrual. No UK-to-US funding edge is earned. An actual relationship of obligation is now evidenced by the filing, while its identity and conditions remain unresolved. The return also records $3,008 currency gains/losses and $3,180 research/production costs, but these do not identify a country, donor or creditor.
\nMiotti is the filed President, Treasurer and Secretary, the officer signatory and named books custodian. The officer schedule reports 6 hours/week, zero compensation, benefits and allowances. This is a 2025-period declaration and does not erase later officers or establish that he performed every operational approval. Gelman Rosenberg & Freedman is the paid preparer firm and Elizabeth W Heller the named preparer; that does not make this an audited return.
\nSchedule B contains RESTRICTED placeholders for contributor number, name, address and amount. It yields neither donor names nor a defensible donor count. The controlled-entity answer is negative but is a narrower question than the related-organisation liability; those statements are not contradictory on their face. Political-campaign answers cannot be recast as proof of no lobbying.
Preserved original: w1-controlai-us-2025.xml,9,572 bytes, SHA2565ef05de70e7b197cd5272c885477f73edfae997a5e9a2b3cd4dfe1470c3f649e. w1-controlai-us-2025-selection-receipt.json preserves the exact CSV row, object/member, URL, byte ranges, CRC and retrieval time. XML content consists of the main 990-EZ and SchedulesB/O; no binary attachments are declared. Exact finance locators are IRS990EZ balance groups and IRS990Schedule O SupplementalInformationDetail; officer locators are ReturnHeader/BusinessOfficerGrp and IRS990EZ/OfficerDirectorTrusteeEmplGrp.
The next discriminating record is the creditor ledger entry and underlying balance confirmation/invoice/agreement, then its clearing entry. Control AI US is the custodian; the filed books contact is Miotti, and the preparer may hold workpapers. These are private records. A later public return could reveal a remaining balance without naming the creditor; that is a reopening trigger, not a promise of public donor identification.
\nThe [Companies House articles](http[local research file] were filed 20 March 2024; their cover has a 15March 2024 stamp. This pass does not establish a separate effective-adoption date. Original image-only PDF retained as w1-controlai-uk-articles-2024.pdf (710,846 bytes; SHA256b81271a477b7fe2d4c56c0726c304153733f85489b3859eaf3067bac7bb98977). Selected relevant pages were rendered and visually read after text extraction returned blank. One-based PDF page numbers below avoid the document's later repeated printed numbering.
The articles call Secure Future Research Ltd the Foundation and define a Trustee as its company director (Schedule 1, PDF p 20). These are drafting terms for the existing company, not evidence of a second foundation or hidden board. Article 19.1 makes the directors the only legal members, and Article 17.2 allows the board to appoint directors (PDF pp 12–13). Article 7 assigns management and powers to the directors; Article 8 allows delegation, including further delegation if permitted, and lets the directors alter or terminate it. Article 8.4 requires committee expenditure/liabilities to be board-authorized or within a board-approved budget (PDF p 8).
\nDecisions are collective under Article 11; Article 13 provides minimum quorum two and, where the chair is eligible, a casting vote; Article 14 provides majority decisions without a meeting (PDF pp 9–10). This is the formal decision framework, not proof of who was chair or how a particular approval occurred. Conflict rules include disclosure, potential-conflict participation subject to specified exceptions, and exclusion in the benefit cases covered by Article 15.5 (PDF pp 10–12); no conflict breach is alleged here.
\nThe purposes/powers allow accepting or disclaiming gifts, supporting other organisations and advocacy within the stated charitable limits (Articles 1–2, PDF pp 3–5). This permits relevant operating forms but proves no particular transfer. The charitable drafting does not independently establish registered-charity status.
\nArticle 27 requires minutes of officer appointments, all company/board resolutions including decisions without meetings, and proceedings of board/company/committee meetings. Article 28 provides for accounting and company records (PDF p 17). This makes resolutions, delegation terms and budget authority a specific expected record family with Secure Future Research as custodian; actual existence/compliance and public availability remain unproved. No donor veto or external appointment power was identified in the inspected governance provisions. Grant agreements could nevertheless create contractual constraints and must be checked separately.
\nThe [current filing history](http[local research file] still lists 21August 2026 accounts to 31December 2025 as the latest financial filing. The retained [original XHTML](http[local research file] is unaudited, filleted micro-entity reporting under FRS105, with income/expenditure not publicly delivered under section 444(5A). It records £7,596,691 accruals and deferred income combined. No new breakdown or fuller replacement was found. The line is not all established donor advances, spendable cash or yearly expenditure.
\nThe needed records are the year-end combined-line reconciliation, deferred-income/restricted-grant schedules and underlying grant agreements, with the full approved accounts and board decisions. Custodians are the company and its finance/board/accountant. Reopen on fuller approved accounts, a named grant disclosure or agreement; repeated identical public filing requests will not create the omitted schedule.
\nBoth [ControlAI US R007795 v 3](http[local research file] and [Amir Benjamin Balde R007864 v 2](http[local research file] returned original JSON 200 and were saved as w1-controlai-german-us-entry.json and w1-controlai-german-balde-entry.json. Their substantive versions remain 16March 2026. ControlAI lists Miotti and lobbying operated through a third party; Balde declares ControlAI US as client for AI-risk political education and contacts with MPs, parliamentary staff and officials. This refreshes an already retained commissioned-lobbying relation.
Balde's zero financial figures refer to calendar 2025; they do not show whether 2026 lobbying was paid. ControlAI's register finance declarations still say no completed fiscal year. The newly acquired US return provides a completed 2025 tax period and was signed later, inJune. These are different dated filings; this pass does not infer a false statement or legal reporting breach. No new donor names,2026 compensation or contract terms were obtained.
\nThe US-original-return gap is closed by a precise new IRS route. The related-party counterparty question replaces a generic suspicion of cross-border funding with a traceable balance; autonomy with cost sharing, internal advance financing and externally conditioned funding remain distinguishable rivals. UK constitutional authority is now located, but exercised approval and any external covenant remain to be acquired. German official refresh is complete and does not answer 2026 payment. Earlier CauseIQ/ProPublica and US lobbying portal failures were not reopened. The Companies House web-tool fetch failed on an expired cached signed URL; direct ordinary PowerShell succeeded. Image-only PDF extraction was replaced once with visual inspection. No failed endpoint was hammered.
\nCandidate graph records are in controlai-candidate.json, with all new IDs prefixedw1-controlai. Keep the related-organisation endpoint visibly unresolved and do not collapse it into Secure Future Research. No historical-job-to-government-funding, CASIR-draft-to-award, preparer-to-auditor, payable-to-paid, or restricted-placeholder-to-donor-count inference is supported.
The public record now contains one specific creator-disclosed, grant-supported output and a clearer distinction between three operating forms: grants with a stated independence condition, a jointly acknowledged advocacy collaboration, and vendor production under prior approval. The signed instruments, payment entries and Sabine's original offer remain holder records. No outreach, signup, paid access or publication was undertaken.
\nFrame links its named fellow Mateus De Souza to [The Dark Philosophy Driving AI Development](http[local research file] published 8 May 2026 by Future Tense. The original public YouTube description explicitly acknowledges production grant funding from BlueDot Impact, Frame Fellowship and Future of Life Institute. It states that the funders had no editorial control and the opinions are the creator's own.
\nThis is a new firsthand creator record for one actual published output, beyond Frame's own success story that Mateus secured FLI support. The two are separate speakers, though both have interests in the programme. The creator's description establishes its public acknowledgment and independence claim; it does not independently verify payment, amount, recipient legal entity, a contractual clause, the absence of informal feedback, or which FLI programme supplied the grant. Do not call the three support edges three independent confirmations: they all come from one description. The description as captured today may have changed after publication; no historical description version was recovered.
\nRaw public HTML is creator-mateus.html; the precise extracted description and publication metadata are creator-mateus-metadata.json. creator-capture-receipts.json preserves URL, capture time, hash and bytes for this and other captures. The useful next records are the grant letter(s), named recipient, milestones/renewal terms, payment entry and any revisions for this specific video. Holders are the creator and the relevant FLI/Frame/BlueDot administrators. A public donor award entry could corroborate amount and legal identity, but no new broad funding hunt was launched.
The [Frame homepage](http[local research file] supplies its public FAQ dynamically through the linked [frontend asset index-CUMxxB1e.js](http[local research file] Reading that delivered public content recovered a consequential answer omitted by the static-page text reader: Frame names Long Term Future Fund and AI Safety Tactical Opportunities Fund as its lead funders. It says funders do not select fellows or influence their content, conclusions or framing.
\nThis is an organisation's disclosure, with no amount, award date, agreement, precise grant-making legal entity or donor-side corroboration. It improves the upstream lead while also preserving the explicit contrary evidence to a donor-editorial-control hypothesis. The live FAQ provides no fiscal-sponsor or legal contracting answer; Mox/Manifund contextual ties cannot fill that gap.
\nThe FAQ says selection evaluates storytelling, niche ownership, clarity and commitment to communication; published work is an important signal. Frame may place applicants in a different track. It describes the stipend as a private nonprofit grant for educational participation and warns that exceptional tailored offers may omit it. Programme deliverables focus on AI; the rest of a creator's channel need not. These are current programme terms, not an executed placement contract. The independent-track promise and amplifier-track institutional brief remain distinct. September 7 is the announced new cohort start, not evidence that its work has already been delivered.
\nRaw delivered script: creator-frame-public-ui-script.txt. The small creator-frame-faq-extract.json contains 23 question/answer pairs and exact source offsets. It was produced by reading public string literals, without executing code or accessing a private endpoint. Same website/source family; the script is not independent evidence. Holders for the next grant/selection/placement instruments are Frame operations and the named fund administrators. No fresh donor tree was pursued.
Neither source names the paying legal entity, fee, contractual script veto, acceptance process or actual payment. A partnership label alone does not make it paid. The original agreement and accepted brief should be sought from Rational Animations and ControlAI's partnerships/finance holders. Keep the brand ControlAI separate from the unresolved UK/US contracting selection. Captures are creator-rational.txt, creator-rational-metadata.json, and creator-controlai-rational.html.
The inspected SciShow video of 31 October 2025 likewise names ControlAI and its lawmaker tool in its description. The Hey AI video linked by Frame does not name CAIS in the inspected description, but Frame's claim of a CAIS contract does not establish this particular video was the contract output. Neither an absent description reference nor absent paid-promotion metadata proves a disclosure violation. Those narrower checks are preserved as boundary findings, not accusations.
\nPeople First's current general terms.
\nSections 1–2 place deliverables and price in the SOW, permit Company/Advertiser review and revisions, and require approvals before publication. Sections 4–6 require disclosure compliance and make payment conditional on completed compliant work and evidence of publication. Section 6 names Dots as the platform payout provider, with multiple payout methods; this proves no particular campaign disbursement. Sections 9–10 retain creator ownership while licensing organic reuse and, subject to SOW limits, paid amplification. These are current general permissions and conditions, not evidence that a right was exercised against a named creator. They are also not the terms of the separate FLI Digital Media Accelerator.
\nThe actual decision record is a particular accepted SOW, its dated incorporated terms, accepted brief, revision/approval log, invoice/settlement and live post. Holders are People First, the selected creator and actual advertiser-side contract holder. Dots may know payout, not editorial approval. There is no evidentiary link from this named platform provider to the unidentified bank in the separate games-payment workstream.
\nThe original [YouTube publication](http[local research file] dated 5 September 2026, was recovered as public HTML after the web tool could not fetch it. Its description supplies no offer sender or correspondence. It links her Patreon/Substack transcripts. The public Patreon sitemap identifies [post 168573904](http[local research file] but the web fetch failed and one direct public fallback returned 403. That route was stopped. Exact-title and sender searches yielded commentary/mirrors, not original offer correspondence; none was promoted to evidence.
\nHer own rejected offer remains unnamed. In the supplied transcript, her account of another creator's screenshot mentioning CAIS is a separate statement. A causal arrow from any named NGO to her offer remains unsupported. The next record is her original email thread with dates, sender domain, attachments, proposed wording/references, fee and disclosure terms; Sabine and the sender are its holders. Reopen on release or supplied correspondence, without further identity guessing.
\nCandidate nodes, typed relations, source ancestry, current-versus-publication-date limits and holder records are in creator-candidate.json. The bounded route is complete: public outputs and closest public terms acquired; original offer, executed contracts, payment proof and exercised approvals remain expressly unresolved.
Bounded public acquisition, Nova. Read retained tegmark-lightcone-followup.md, FLI workstream notes, source-ledger entries S010/S011/S033 and campaign-plan section7. No outreach, spend or publication. This packet adds an original 2025 audit and identifies a precise Swiss governing-document frontier. Candidate records use prefix w2-fli.
The reserve-to-operations transfer continues into 2025. The newly recovered [full FLI 2025 audit](http[local research file] issued 16 June 2026, reports $45,197,838 of grants received from Lightcone Foundation (Note6, printed p12 / PDF14). Total contributions were $45,234,659 (printed p4 / PDF6). Their difference is $36,821, reconciling arithmetically with the website’s approximately $37,000 fundraising figure. Lightcone supplied 99.9186% of contributions (own calculation). The audit is FLI standalone, not a consolidated affiliate/endowment roll-forward. It also records contractual cost sharing: $959,308 due from FLF and $430,343 due to FLF at year-end; these are reciprocal balances, not two new donations. The audit’s $46,191,762 expense and $5,794,889 marketing/communications totals substantiate the website’s rounded figures. This remains a 2025 accounting state.
\nThe [FLI finance page](http[local research file] read fresh, separates limited fundraising from the endowment-funded difference. It asserts that Buterin’s contribution was unconditional and he lacks a decision role. The arithmetic now explains the apparent discrepancy between small new fundraising and large audited contributions; it does not independently establish original gift conditions or rights. The [German register](http[local research file] updated 30 June 2026, reports Lightcone support in a €38.460–38.470 million band, describing it as unrestricted. Do not add this differently denominated disclosure to the audited dollar receipt or assume an exact conversion without its method.
\nThe entity identified is Future of Life Institute Europe, Swiss foundation CHE-450.633.132, commercial register CH-660.1.343.022-6, Geneva. [Moneyhouse’s reproduction of the founding Swiss Gazette notice](http[local research file] publication HR01-1005455911, 21 April 2022, records a deed dated 11 April 2022 and registration on 14 April 2022. It names Jaan Tallinn as chair, Anthony Nicholas Aguirre, Mihaela Chita-Tegmark, Anna Béatrice Ramsay Fihn and Max Erik Tegmark as board members, each with joint signature by two. ECHO SA is auditor. Its “six management persons” display includes that corporate auditor; it is not evidence of six human trustees.
\nThe reproduced notice mentions a founder’s reserved purpose-modification right under Article86a. It does not name the founder in that passage or reveal appointment/removal, distribution-vote or dissolution clauses. A joint-signature rule is not itself a grant-approval rule. The full deed and current statutes were not recovered; treat these as transfer-era facts from a registry reproduction, not certified-current governance.
\nAn [official Gazette notice mirrored by LIXT](http[local research file] HR02-1005510667, 1 July 2022, confirms the same legal name/UID and identifies the Federal Department of Home Affairs as supervisor. This one-page original-format notice was captured. Supervision is not evidence of government funding or editorial direction.
\nThe [original FLI 2022 Form990](http[local research file] was captured and re-read. ScheduleN, PDF49, reports $180,297,079 of cash/financial assets transferred to Future of Life Institute Europe on 30 December 2022. It says Tegmark, Aguirre, Chita-Tegmark and Tallinn were directors of all three transferees before and after the transfers. This establishes transfer-era overlap, not the power to appoint successors. ScheduleO, PDF51, reproduces the US Institute’s amended purpose supporting Future of Life Foundation, EIN88-0926067. Lightcone Foundation is separately identified by EIN92-0636259. No link to Lightcone Infrastructure/LessWrong is inferred.
\nThe US filing uses a Geneva correspondence address different from the Swiss registered office. Exact name, jurisdiction and overlapping named governors support matching the Swiss entity; a recipient-side receipt/accounting statement would directly close that reconciliation. The register date should not be replaced by the Gazette publication date.
\nNo inspected public source supplies FLI Europe’s standalone 2022 closing balance, subsequent investment returns, external distributions, interaffiliate transfers and closing reserves through 2025. Accordingly do not subtract the Lightcone-to-FLI receipt from Europe’s $180.297 million or treat it as that money returning. Those are different legal entities. The US audit names “Future of Life Europe” among related organisations but gives no Europe balance or received grant in its Note6 table; that is not evidence of zero European activity.
\nTwo records now define the narrow frontier. First, the 11 April 2022 deed and current statutes for CHE-450.633.132, including amendments, identify the founder and appointment, distribution, purpose-change and dissolution powers. [Geneva’s official guidance](http[local research file] says its free registry search can expose statute PDFs for relevant post-November2007 filings. It does not establish that this foundation’s complete instrument is online. The specific search route is http[local research file] . A certified copy or physical dossier is a separate paid route; no order was made.
\nSecond, seek FLI Europe’s signed 2022–25 financial statements with investment and related-party schedules, plus the 30 December2022 transfer acceptance/resolution. The foundation board/finance function holds these; ECHO SA is the publicly identified transfer-era auditor. The federal supervisor is another potential holder, with access eligibility to be assessed before any official request. A voluntary holder request is available without inventing an entitlement. These distinguish an independently administered reserve from centrally appointable affiliates or legally restricted assets. None of those models requires the original donor to retain command.
\nThe official legacy SHAB links for the two named notices returned web cache misses. The July notice’s public mirror was a successful substitute; the founding notice remains a third-party registry reproduction. A direct Moneyhouse capture returned403 and was not saved; web-rendered contents were inspected. An exploratory Zefix URL with an unverified numeric identifier returned InternalError and was not treated as an entity record.
\nGeneva’s JavaScript registry returned no text to the web extractor. The browser then explicitly denied access because its admin-enforced security policy could not be verified. That route was stopped without workaround or bypass; no page inside the denied registry was inspected. This is a browser security-policy failure, not an automatic approval-review rejection and not proof that the deed is unavailable. Existing lawful sources and the independently recovered audit carry the deliverable.
\nCaptured files and hashes are recorded in the candidate source metadata. The 2022 return capture succeeded; its first console print encountered Unicode encoding trouble, resolved by reading the saved PDF with ASCII-safe output. PDF extraction was targeted to financial and governance pages; no new visual table inspection is claimed. Documents issued by the same organisation and their registry copies are one underlying origin, not independent confirmation. Broader current affiliate boards, gift instruments and consolidated reserves remain unverified.
", "path": "red-threads-wave-2026-09-08/fli-findings.md", "sourceIds": [ "src-5a885dee5911", "src-e4e7930e8216", "w2-fli-s-audit25", "w2-fli-s-founding", "w2-fli-s-supervision", "w2-fli-s-bundestag", "w2-fli-s-geneva-route", "w2-fli-s-geneva-dossier" ], "markdown": "Bounded public acquisition, Nova. Read retained `tegmark-lightcone-followup.md`, FLI workstream notes, source-ledger entries S010/S011/S033 and campaign-plan section7. No outreach, spend or publication. This packet adds an original 2025 audit and identifies a precise Swiss governing-document frontier. Candidate records use prefix `w2-fli`.\n\n## What this pass resolves\n\n**The reserve-to-operations transfer continues into 2025.** The newly recovered full FLI 2025 audit (see source references) issued 16 June 2026, reports $45,197,838 of grants received from Lightcone Foundation (Note6, printed p12 / PDF14). Total contributions were $45,234,659 (printed p4 / PDF6). Their difference is **$36,821**, reconciling arithmetically with the website’s approximately $37,000 fundraising figure. Lightcone supplied 99.9186% of contributions (own calculation). The audit is FLI standalone, not a consolidated affiliate/endowment roll-forward. It also records contractual cost sharing: $959,308 due from FLF and $430,343 due to FLF at year-end; these are reciprocal balances, not two new donations. The audit’s $46,191,762 expense and $5,794,889 marketing/communications totals substantiate the website’s rounded figures. This remains a 2025 accounting state.\n\nThe FLI finance page (see source references) read fresh, separates limited fundraising from the endowment-funded difference. It asserts that Buterin’s contribution was unconditional and he lacks a decision role. The arithmetic now explains the apparent discrepancy between small new fundraising and large audited contributions; it does not independently establish original gift conditions or rights. The German register (see source references) updated 30 June 2026, reports Lightcone support in a €38.460–38.470 million band, describing it as unrestricted. Do not add this differently denominated disclosure to the audited dollar receipt or assume an exact conversion without its method.\n\n## Exact European entity and transfer-era governance\n\nThe entity identified is **Future of Life Institute Europe**, Swiss foundation **CHE-450.633.132**, commercial register **CH-660.1.343.022-6**, Geneva. Moneyhouse’s reproduction of the founding Swiss Gazette notice (see source references) publication **HR01-1005455911, 21 April 2022**, records a deed dated **11 April 2022** and registration on **14 April 2022**. It names Jaan Tallinn as chair, Anthony Nicholas Aguirre, Mihaela Chita-Tegmark, Anna Béatrice Ramsay Fihn and Max Erik Tegmark as board members, each with joint signature by two. ECHO SA is auditor. Its “six management persons” display includes that corporate auditor; it is not evidence of six human trustees.\n\nThe reproduced notice mentions a founder’s reserved purpose-modification right under Article86a. It does **not** name the founder in that passage or reveal appointment/removal, distribution-vote or dissolution clauses. A joint-signature rule is not itself a grant-approval rule. The full deed and current statutes were not recovered; treat these as transfer-era facts from a registry reproduction, not certified-current governance.\n\nAn official Gazette notice mirrored by LIXT (see source references) **HR02-1005510667, 1 July 2022**, confirms the same legal name/UID and identifies the Federal Department of Home Affairs as supervisor. This one-page original-format notice was captured. Supervision is not evidence of government funding or editorial direction.\n\nThe original FLI 2022 Form990 (see source references) was captured and re-read. ScheduleN, PDF49, reports **$180,297,079** of cash/financial assets transferred to Future of Life Institute Europe on **30 December 2022**. It says Tegmark, Aguirre, Chita-Tegmark and Tallinn were directors of all three transferees before and after the transfers. This establishes transfer-era overlap, not the power to appoint successors. ScheduleO, PDF51, reproduces the US Institute’s amended purpose supporting **Future of Life Foundation, EIN88-0926067**. Lightcone Foundation is separately identified by **EIN92-0636259**. No link to Lightcone Infrastructure/LessWrong is inferred.\n\nThe US filing uses a Geneva correspondence address different from the Swiss registered office. Exact name, jurisdiction and overlapping named governors support matching the Swiss entity; a recipient-side receipt/accounting statement would directly close that reconciliation. The register date should not be replaced by the Gazette publication date.\n\n## Roll-forward and authority still missing\n\nNo inspected public source supplies FLI Europe’s standalone 2022 closing balance, subsequent investment returns, external distributions, interaffiliate transfers and closing reserves through 2025. Accordingly **do not subtract the Lightcone-to-FLI receipt from Europe’s $180.297 million** or treat it as that money returning. Those are different legal entities. The US audit names “Future of Life Europe” among related organisations but gives no Europe balance or received grant in its Note6 table; that is not evidence of zero European activity.\n\nTwo records now define the narrow frontier. First, **the 11 April 2022 deed and current statutes for CHE-450.633.132**, including amendments, identify the founder and appointment, distribution, purpose-change and dissolution powers. Geneva’s official guidance (see source references) says its free registry search can expose statute PDFs for relevant post-November2007 filings. It does not establish that this foundation’s complete instrument is online. The specific search route is http[local research file] . A certified copy or physical dossier is a separate paid route; no order was made.\n\nSecond, seek **FLI Europe’s signed 2022–25 financial statements with investment and related-party schedules**, plus the 30 December2022 transfer acceptance/resolution. The foundation board/finance function holds these; ECHO SA is the publicly identified transfer-era auditor. The federal supervisor is another potential holder, with access eligibility to be assessed before any official request. A voluntary holder request is available without inventing an entitlement. These distinguish an independently administered reserve from centrally appointable affiliates or legally restricted assets. None of those models requires the original donor to retain command.\n\n## Access, custody and stop boundary\n\nThe official legacy SHAB links for the two named notices returned web cache misses. The July notice’s public mirror was a successful substitute; the founding notice remains a third-party registry reproduction. A direct Moneyhouse capture returned403 and was not saved; web-rendered contents were inspected. An exploratory Zefix URL with an unverified numeric identifier returned InternalError and was not treated as an entity record.\n\nGeneva’s JavaScript registry returned no text to the web extractor. The browser then explicitly denied access because its admin-enforced security policy could not be verified. That route was stopped without workaround or bypass; no page inside the denied registry was inspected. This is a browser security-policy failure, not an automatic approval-review rejection and not proof that the deed is unavailable. Existing lawful sources and the independently recovered audit carry the deliverable.\n\nCaptured files and hashes are recorded in the candidate source metadata. The 2022 return capture succeeded; its first console print encountered Unicode encoding trouble, resolved by reading the saved PDF with ASCII-safe output. PDF extraction was targeted to financial and governance pages; no new visual table inspection is claimed. Documents issued by the same organisation and their registry copies are one underlying origin, not independent confirmation. Broader current affiliate boards, gift instruments and consolidated reserves remain unverified.\n", "file": "research/red-threads-wave-2026-09-08-fli-findings.md" }, { "id": "red-threads-wave-2026-09-08-icmec-findings", "title": "ICMEC acquisition findings — 8 September 2026", "description": "Public-record wave by Nova. Findings concern the records and versions identified below; no outreach, paid access, publication or implementation testing. Raw PDFs were inspected in memory, not saved. Candidate graph imports are in icmec-candidate.json.", "html": "Public-record wave by Nova. Findings concern the records and versions identified below; no outreach, paid access, publication or implementation testing. Raw PDFs were inspected in memory, not saved. Candidate graph imports are in icmec-candidate.json.
The public financial record changes the picture from a presumed well-resourced intermediary to an organisation with substantial financing dependence. It does not identify an age-policy funder. Legislative records reveal negotiated adaptation, including identifiable alternatives and an author’s response, rather than unchanged adoption of an NGO template. A separate historical record does establish payment-company funding of ICMEC policy roundtables.
\nICMEC’s [2024 consolidated audit](http[local research file] issued 29 January 2026, covers US ICMEC and its Singapore affiliate, not ICMEC Switzerland. It reports a $2,293,789 net-asset deficit at 31 December 2024 and $817,886 of programme-restricted assets. Its going-concern discussion identifies dependence on unrestricted fundraising; management reports signed 2026 grants/contracts. Note 2 records board-member and Swiss-affiliate borrowing. Note 9 records a US-government stop-work order dated 27 January 2025. These are historical accounts, not a September 2026 balance sheet. Locators: printed pp9,12–16; going concern p15; one-based PDF pages10,13–17. The audit says staff allocations use biweekly timesheets; underlying programme allocation records should therefore exist.
\nThe [2024 Form 990](http[local research file] names board members Franz Humer, Sally Paul and Rick Li as interested persons in operating-fund loans, with $1,117,000 aggregate outstanding in Schedule L, Part II; explanatory Part V follows (PDF pages42–43). “Interested person” is not automatically the legal lender: the audit identifies ICMEC Switzerland for some loans. Row extraction did not justify assigning every amount to a person. Schedule C, Part II-A (PDF p28) reports $996 direct lobbying and zero grassroots lobbying in FY2024. This predates AB1043 and does not disclose its budget. The programme reporting inspected does not isolate age-policy expenditure.
\nICMEC’s [supporters page](http[local research file] lists Meta among $25,000-plus general supporters as of 1 May 2023; a separate Financial Coalition section names payment/financial firms without amounts. Neither establishes a 2025 legislative earmark.
\nA more concrete historical route appears in [ICMEC’s APFC roundtable account](http[local research file] a PayPal grant enabled country roundtables; Microsoft and law firms supplied in-kind resources; resulting working groups included Australia, whose institutional home was the Australian Payments Network’s Fraud in Banking Forum. This establishes recipient-reported resource mobilisation and institutional follow-through. The undated page refers to activity since 2013; no grant amount, age-policy earmark or current financing arrangement is disclosed.
\n| Record | \nNew decision or boundary established | \n
|---|---|
| [ICMEC model legislation](http[local research file] pp2–4 | \nProposes OS/device age signals, parental consent for under-16 downloads, and adult-content blocking. Title calls it an Act of 2025; the upload-directory date is not independently verified publication timing. | \n
| [Technical whitepaper, 7 February 2025](http[local research file] p5 | \nProposes OS-provider review/approval of website API integrations and agreed use cases. This is a proposed admission function, not proof that a law or deployed API grants those powers. | \n
| [AB1043 APCP analysis, 22 April 2025](http[local research file] March28 version, pp1–4,16 | \nLists ICMEC and Children Now as co-sponsors; the then-current design includes six age bands and parental download consent. | \n
| [Senate Judiciary analysis, 15 July 2025](http[local research file] May23 version, pp16–17,19 | \nRecords stakeholder technical/scoping concerns and the author’s agreement to narrow the application definition by removing online services/products/features. It identifies speech concerns about download-consent restrictions. Only Children Now is listed as sponsor; this roster difference alone does not establish ICMEC’s withdrawal or motive. | \n
| [AB1043 chaptered text](http[local research file] Chapter675, 13 October 2025 | \nEnacted age-signal framework, operative January2027, differs from the model: four bands and account-entered age; the inspected text lacks the earlier parental-download consent and broad website scheme. It includes data minimisation and non-discrimination. The whitepaper’s per-website API approval mechanism is not established by this text. | \n
| [AB1856 APCP analysis, 21 April 2026](http[local research file] April6 version, pp5–6 | \nIdentifies an MPA memo seeking a family-account carveout and an option to block minors. The author instead chooses a generally applicable clarification of conflicting age evidence. Committee discussion criticises a general minor-exclusion option. Proposed browser/website duties are explicitly described. | \n
| [AB1856 enrolled text](http[local research file] and [official status](http[local research file] checked 8 September 2026 | \nEnrolled September1, presented to Governor September4 at4pm; no signature shown. Current text contains an open-source OS exclusion and account-setup limitation. It omits the April browser/website scheme. Treat this as pending legislation, not operative law. | \n
The April2026 committee account is a credible counterpart to the previously failed Wicks-page capture. It supplies named requests, deliberation and an alternative actually chosen, although the MPA memorandum itself remains unacquired. The current record supports selective template diffusion and local negotiation; it does not establish donor control of drafting.
\nBoth financial PDFs returned ordinary HTTP200 and were read using pypdf. Audit: 429,674 bytes; SHA256 277651b6bccf9ca39e565f8e515fed219a95a56cf5f3996ecfce0ceb61b6b2fc. Return: 557,097 bytes; SHA256 553fb43c3deed524f2288fa96661f756d83fa9925535a7a7de890328553ee991. Web PDF screenshot routes produced no usable financial image (cache misses on attempted audit/990 pages); no visual row alignment is claimed. The two documents provide different financial disclosures but remain records supplied by the same organisation, not independent donor confirmation.
The official AB1043 version selector was read through ordinary HTTP200. It exposes introduced February20; amendments March28, April24, May23, July23, September5; enrolled September16; chaptered October13, all2025. Direct web opens using &version=20250AB104399INT and &version=20250AB104395AMD failed with InternalError; these failures do not mean historical texts are absent. Version-specific committee PDFs and the chaptered statute were the successful substitute. Exact amendment-by-amendment attribution of every deletion remains unfinished. No repeat of the unchanged Wicks raw-capture failure was attempted.
ICMEC model and whitepaper are one organisational source family; copies elsewhere are not corroboration. Committee analyses are official accounts, but attributed stakeholder arguments are not committee findings. This bounded pass inspected public finance, organisational programme/supporter pages and official legislative routes; it was not an exhaustive donor-record search.
\nThe best next financial holder record is ICMEC’s 2024–26 age-assurance project ledger/cost-centre allocation, linked grant schedules and approval record. Ask ICMEC voluntarily for who paid for model/whitepaper production and legislative engagement, amounts, dates, restrictions and approval rights. The audit indicates time-allocation records exist; it does not prove a distinct age-policy cost centre exists. Unrestricted funding, earmarked drafting support and commercial deliverables imply materially different relationships.
\nFor design lineage, seek the MPA memorandum already described by the April2026 committee and the author/committee’s amendment instructions or stakeholder submissions. Public committee holdings are the first route; otherwise request from the holders without assuming a particular legal access entitlement. The discriminating issue is who proposed the surviving and deleted provisions, not who appeared on a general supporter list. Actual OS API admission terms would distinguish a mandatory neutral signal service from discretionary commercial gatekeeping; the whitepaper alone cannot settle that implementation question.
", "path": "red-threads-wave-2026-09-08/icmec-findings.md", "sourceIds": [ "s2-icmec-donors", "s2-ab1043-analysis", "s2-age-paper", "w1-icmec-s-audit", "w1-icmec-s-990", "w1-icmec-s-apfc", "w1-icmec-s-model", "w1-icmec-s-sjud25", "w1-icmec-s-ab1043", "w1-icmec-s-apcp26", "w1-icmec-s-ab1856", "w1-icmec-s-ab1856-status" ], "markdown": "Public-record wave by Nova. Findings concern the records and versions identified below; no outreach, paid access, publication or implementation testing. Raw PDFs were inspected in memory, not saved. Candidate graph imports are in `icmec-candidate.json`.\n\n## What moved\n\nThe public financial record changes the picture from a presumed well-resourced intermediary to an organisation with substantial financing dependence. It does not identify an age-policy funder. Legislative records reveal negotiated adaptation, including identifiable alternatives and an author’s response, rather than unchanged adoption of an NGO template. A separate historical record does establish payment-company funding of ICMEC policy roundtables.\n\n## Financial perimeter and funding\n\nICMEC’s 2024 consolidated audit (see source references) issued 29 January 2026, covers US ICMEC and its Singapore affiliate, not ICMEC Switzerland. It reports a $2,293,789 net-asset deficit at 31 December 2024 and $817,886 of programme-restricted assets. Its going-concern discussion identifies dependence on unrestricted fundraising; management reports signed 2026 grants/contracts. Note 2 records board-member and Swiss-affiliate borrowing. Note 9 records a US-government stop-work order dated 27 January 2025. These are historical accounts, not a September 2026 balance sheet. Locators: printed pp9,12–16; going concern p15; one-based PDF pages10,13–17. The audit says staff allocations use biweekly timesheets; underlying programme allocation records should therefore exist.\n\nThe 2024 Form 990 (see source references) names board members Franz Humer, Sally Paul and Rick Li as interested persons in operating-fund loans, with $1,117,000 aggregate outstanding in Schedule L, Part II; explanatory Part V follows (PDF pages42–43). “Interested person” is not automatically the legal lender: the audit identifies ICMEC Switzerland for some loans. Row extraction did not justify assigning every amount to a person. Schedule C, Part II-A (PDF p28) reports $996 direct lobbying and zero grassroots lobbying in FY2024. This predates AB1043 and does not disclose its budget. The programme reporting inspected does not isolate age-policy expenditure.\n\nICMEC’s supporters page (see source references) lists Meta among $25,000-plus general supporters **as of 1 May 2023**; a separate Financial Coalition section names payment/financial firms without amounts. Neither establishes a 2025 legislative earmark.\n\nA more concrete historical route appears in ICMEC’s APFC roundtable account (see source references) a PayPal grant enabled country roundtables; Microsoft and law firms supplied in-kind resources; resulting working groups included Australia, whose institutional home was the Australian Payments Network’s Fraud in Banking Forum. This establishes recipient-reported resource mobilisation and institutional follow-through. The undated page refers to activity since 2013; no grant amount, age-policy earmark or current financing arrangement is disclosed.\n\n## Drafts, decisions and operative design\n\n| Record | New decision or boundary established |\n| --- | --- |\n| ICMEC model legislation (see source references) pp2–4 | Proposes OS/device age signals, parental consent for under-16 downloads, and adult-content blocking. Title calls it an Act of 2025; the upload-directory date is not independently verified publication timing. |\n| Technical whitepaper, 7 February 2025 (see source references) p5 | Proposes OS-provider review/approval of website API integrations and agreed use cases. This is a proposed admission function, not proof that a law or deployed API grants those powers. |\n| AB1043 APCP analysis, 22 April 2025 (see source references) March28 version, pp1–4,16 | Lists ICMEC and Children Now as co-sponsors; the then-current design includes six age bands and parental download consent. |\n| Senate Judiciary analysis, 15 July 2025 (see source references) May23 version, pp16–17,19 | Records stakeholder technical/scoping concerns and the author’s agreement to narrow the application definition by removing online services/products/features. It identifies speech concerns about download-consent restrictions. Only Children Now is listed as sponsor; this roster difference alone does not establish ICMEC’s withdrawal or motive. |\n| AB1043 chaptered text (see source references) Chapter675, 13 October 2025 | Enacted age-signal framework, operative January2027, differs from the model: four bands and account-entered age; the inspected text lacks the earlier parental-download consent and broad website scheme. It includes data minimisation and non-discrimination. The whitepaper’s per-website API approval mechanism is not established by this text. |\n| AB1856 APCP analysis, 21 April 2026 (see source references) April6 version, pp5–6 | Identifies an MPA memo seeking a family-account carveout and an option to block minors. The author instead chooses a generally applicable clarification of conflicting age evidence. Committee discussion criticises a general minor-exclusion option. Proposed browser/website duties are explicitly described. |\n| AB1856 enrolled text (see source references) and official status (see source references) checked 8 September 2026 | Enrolled September1, presented to Governor September4 at4pm; no signature shown. Current text contains an open-source OS exclusion and account-setup limitation. It omits the April browser/website scheme. Treat this as pending legislation, not operative law. |\n\nThe April2026 committee account is a credible counterpart to the previously failed Wicks-page capture. It supplies named requests, deliberation and an alternative actually chosen, although the MPA memorandum itself remains unacquired. The current record supports selective template diffusion and local negotiation; it does not establish donor control of drafting.\n\n## Access and ancestry\n\nBoth financial PDFs returned ordinary HTTP200 and were read using pypdf. Audit: 429,674 bytes; SHA256 `277651b6bccf9ca39e565f8e515fed219a95a56cf5f3996ecfce0ceb61b6b2fc`. Return: 557,097 bytes; SHA256 `553fb43c3deed524f2288fa96661f756d83fa9925535a7a7de890328553ee991`. Web PDF screenshot routes produced no usable financial image (cache misses on attempted audit/990 pages); no visual row alignment is claimed. The two documents provide different financial disclosures but remain records supplied by the same organisation, not independent donor confirmation.\n\nThe official AB1043 version selector was read through ordinary HTTP200. It exposes introduced February20; amendments March28, April24, May23, July23, September5; enrolled September16; chaptered October13, all2025. Direct web opens using `&version=20250AB104399INT` and `&version=20250AB104395AMD` failed with InternalError; these failures do not mean historical texts are absent. Version-specific committee PDFs and the chaptered statute were the successful substitute. Exact amendment-by-amendment attribution of every deletion remains unfinished. No repeat of the unchanged Wicks raw-capture failure was attempted.\n\nICMEC model and whitepaper are one organisational source family; copies elsewhere are not corroboration. Committee analyses are official accounts, but attributed stakeholder arguments are not committee findings. This bounded pass inspected public finance, organisational programme/supporter pages and official legislative routes; it was not an exhaustive donor-record search.\n\n## Consequential missing joint\n\nThe best next financial holder record is ICMEC’s 2024–26 age-assurance project ledger/cost-centre allocation, linked grant schedules and approval record. Ask ICMEC voluntarily for who paid for model/whitepaper production and legislative engagement, amounts, dates, restrictions and approval rights. The audit indicates time-allocation records exist; it does not prove a distinct age-policy cost centre exists. Unrestricted funding, earmarked drafting support and commercial deliverables imply materially different relationships.\n\nFor design lineage, seek the MPA memorandum already described by the April2026 committee and the author/committee’s amendment instructions or stakeholder submissions. Public committee holdings are the first route; otherwise request from the holders without assuming a particular legal access entitlement. The discriminating issue is who proposed the surviving and deleted provisions, not who appeared on a general supporter list. Actual OS API admission terms would distinguish a mandatory neutral signal service from discretionary commercial gatekeeping; the whitepaper alone cannot settle that implementation question.\n", "file": "research/red-threads-wave-2026-09-08-icmec-findings.md" }, { "id": "red-threads-wave-2026-09-08-payments-findings", "title": "Payment permissions: bounded acquisition return", "description": "The strongest new addition is Steam's own attributed notice chain. Valve supplied statements to the journalist who published them, including a rejected alternative and the precise Mastercard rule reportedly invoked. Separately, original itch records expose an earlier platform payout decision and a contemporaneous UK geographic-compliance decision. The bank b", "html": "Nova · 8 September 2026 · Candidate wave w1-payments
\nThe strongest new addition is Steam's own attributed notice chain. Valve supplied statements to the journalist who published them, including a rejected alternative and the precise Mastercard rule reportedly invoked. Separately, original itch records expose an earlier platform payout decision and a contemporaneous UK geographic-compliance decision. The bank behind itch's July 2025 restriction remains unidentified.
\nThis wave adds source-backed candidates to the retained 175-node/236-edge edition. It does not alter the canonical case. Current revised webpages were read; raw correspondence, immutable historical policy payloads and account-specific agreements were not obtained.
\nSteam: the recipient published Valve's response. GamingOnLinux's July 17 update records Valve saying processors had flagged games, and that it retired affected titles to protect payment methods used elsewhere in the store. Valve also described developer notification and future app credits. The reporter explicitly says Valve press replied to the publication; raw message headers and attachments are absent. Locator: July 17 update, article lines 8–11. [Recipient's July article](http[local research file]
\nThe August 1 update is more discriminating. Valve says Mastercard did not communicate directly with Valve; processors relayed communications involving Mastercard and acquiring banks. Valve says those processors rejected its legal-content argument and cited Rule 5.12.7. This is a company account supplied to its recipient journalist, not recovered upstream instructions. Locator: Update 23:36 BST, lines 8–10; last update 22:38 UTC. [Recipient's August article](http[local research file]
\nThis permits a Steam-specific attributed chain without importing Stripe's itch role. The processors remain unnamed. The existing graph node “Payment-letter addressees” is not their identity. Mastercard's denial and the June 3, 2025 rulebook remain separate retained records. The apparent disagreement could concern who communicated, what was requested, or downstream interpretation of a broad rule; it cannot be resolved by treating either public statement as the missing instruction.
\nItch: internal eligibility and exit machinery predate July 2025. An official admin reply dated April 26, 2023 says selected developers had lost eligibility for collected payouts and links the platform's notice. It distinguishes selected exclusions from a universal change and says raising a seller's revenue share will not reverse the decision. Locator: pinned reply opening and revenue-share discussion; original HTML timestamp title 2023-04-26 17:07:52, timezone unspecified. [Original admin reply](http[local research file]
\nThe linked notice currently explains content/risk/review-resource factors and requires affected sellers to leave collected payouts. Final-balance timing and a migration window govern the transition; existing purchases remain distinct from accepting new payments. Locator: opening paragraphs and final-payout FAQ. [Original notice template](http[local research file]
\nThe template is undated current text, linked by a dated 2023 post with later edits. We did not recover immutable 2023 wording. Its value is a documented decision function and switching constraints; it does not establish why July 2025 restrictions occurred.
\nItch: UK page blocking was a separate compliance response. Its July 27, 2025 official update describes geographic restrictions attributed to OSA compliance, tag classification with staff override, a dashboard appeal, and preserved downloads for existing purchases. It also announces exploration of unnamed digital-ID providers. Locator: lines 30–36 and 47; topic created 22:08 UTC and now marked five edits. [Original OSA update](http[local research file]
\nA dated July 29 compiler entry reproduces material paragraphs and links that original. It supports contemporaneous circulation, not independent corroboration or a deployed provider relationship. [Compiler entry](http[local research file] The candidate therefore keeps geographic blocking, payment eligibility and prospective ID procurement separate.
\nThe edited Stripe account has a contemporaneous reproduction. GamingOnLinux's August 1 article, published 15:53 UTC and updated 15:55 UTC, reproduces itch's account of Stripe's bank restriction and sexual-content boundary. It narrows the risk that this central explanation appeared only in a later edit of itch's July 31 post. Locator: lines 9–13. It depends on that post and is not another Stripe confirmation. [Contemporaneous reproduction](http[local research file]
\nAn ordinary public request returned the Internet Archive's CDX index for Stripe's restricted-businesses page. Selected exact records:
\n| Capture (UTC) | \nCDX digest | \n
|---|---|
| 2025-07-08 05:08:09 | \nHNQ3PWU6LEKUJDDDBLCEIJO2L6UNZLSI | \n
| 2025-07-16 03:57:34 | \nY6IHCIODMO7GZPK2DIX2NFX4HCTLKHGN | \n
| 2025-07-21 11:38:01 | \nSPBM6EM327CWVJQ5YGRLJQ45SEMGZY7E | \n
| 2025-07-24 06:25:46 | \nRK6F3PRVKVHI4UA7M5SABV2U7XH7HRJ5 | \n
| 2025-07-30 08:02:27 | \n7BWWDYBGRVVVIR6JKXTRUG4C4AGDX237 | \n
| 2025-07-30 13:58:37 | \nMBJBZ4JBCCYJOOFIXUXHLQVPWYQAFQHP | \n
| 2025-07-31 17:25:09 | \nYOGAFSZ53AHNPYSFVMMGLUL5Q72EV3YG | \n
[Exact CDX query](http[local research file] Different digests do not establish changed policy text; dynamic HTML is an alternative.
\nPayload replay using id_ URLs failed in the web tool with a nonretryable URL-safety/parser error. One standard-replay substitute also failed: [identified July 21 capture](http[local research file] This route stopped. A separate itch archive-index request returned HTTP 429 and was not retried. Current [Stripe general terms](http[local research file] were inspected only as background after the /legal/ssa/us route failed; they were not treated as itch's July 2025 contract.
\nThe itch joint is the original July 2025 conversation/communication, whatever its medium; the bank instruction it references; the merchant/acquirer arrangement; and the accepted terms and implementation exchange. Public language about speaking with Stripe does not establish an email. Relevant custodians are itch payments/operations and Stripe's account/risk team. The bank becomes a named lead only when account-specific routing identifies it.
\nThe Steam joint is Valve's processor notice and response, the corresponding Mastercard/acquirer message and the actual counterparty identities. Valve payments/legal and its processor counterparts should hold operational communications; Valve press and GamingOnLinux hold the identified company replies. The published replies establish an attributed account, not the contents of unrecovered upstream records.
\nNo original bank instruction or account-specific historical agreement was located in the bounded official, archive-index and released-correspondence/exhibit search. This is not an exhaustive court-docket search and does not establish nonexistence or a holder refusal. No outreach occurred. Reopen on a released notice, a functioning authorized archive route, routing disclosure or substantive holder response.
\nA compiler's Discord transcription offered a potentially useful Stripe case comment, but supplied no original message link or screenshot. Its detailed entry dates the message July 25 at 20:13 -0700, the linked timeline dates it July 25 at 03:13 UTC, and the entry creation metadata is July 25 at 06:20:34. These do not reconcile. It is excluded from event chronology and mechanism claims. [Transcription](http[local research file]
\nThe evidence now supports a specific bargaining step in Steam's account and documented platform decision machinery at itch. It remains compatible with outside instruction, downstream underwriting interpretation and platform precaution operating together. Geographic compliance creates an additional, separately evidenced explanation for some restrictions. None of these records establishes a common command structure, a named bank behind itch, a Stripe–Steam processing relationship or a deployed itch identity-provider contract.
\npayments-candidate.json contains five new role/event/program nodes, seven typed edges, fourteen source records (including explicit retained-origin alias and inaccessible/excluded records), four leads and provenance/dependency notes. Its external node IDs refer to the retained explorer. Alias the retained itch reindexing source by exact URL during integration. Source counts are bibliographic records, not independent confirmations. No existing relationship is upgraded by adding a derivative report. No effective start date is inferred from September 8 inspection.
\nOnly this note and the candidate packet were written. They contain citations, bounded paraphrases, access outcomes and index metadata; no full-text local captures are claimed.
", "path": "red-threads-wave-2026-09-08/payments-findings.md", "sourceIds": [ "src-094404b86089", "w1-payments-src-itch-admin-20230426", "w1-payments-src-itch-payout-notice", "w1-payments-src-valve-reply-20250717", "w1-payments-src-gol-valve-20250717", "w1-payments-src-valve-reply-20250801", "w1-payments-src-gol-valve-20250801", "w1-payments-src-gol-itch-20250801", "w1-payments-src-itch-osa-20250727", "w1-payments-src-exo-osa-20250729", "w1-payments-src-stripe-policy-cdx", "w1-payments-src-stripe-policy-20250721", "w1-payments-src-exo-discord-excluded", "w1-payments-src-stripe-ssa-current" ], "markdown": "Nova · 8 September 2026 · Candidate wave w1-payments\n\nThe strongest new addition is Steam's own attributed notice chain. Valve supplied statements to the journalist who published them, including a rejected alternative and the precise Mastercard rule reportedly invoked. Separately, original itch records expose an earlier platform payout decision and a contemporaneous UK geographic-compliance decision. The bank behind itch's July 2025 restriction remains unidentified.\n\nThis wave adds source-backed candidates to the retained 175-node/236-edge edition. It does not alter the canonical case. Current revised webpages were read; raw correspondence, immutable historical policy payloads and account-specific agreements were not obtained.\n\n## What is new\n\n**Steam: the recipient published Valve's response.** GamingOnLinux's July 17 update records Valve saying processors had flagged games, and that it retired affected titles to protect payment methods used elsewhere in the store. Valve also described developer notification and future app credits. The reporter explicitly says Valve press replied to the publication; raw message headers and attachments are absent. Locator: July 17 update, article lines 8–11. Recipient's July article (see source references)\n\nThe August 1 update is more discriminating. Valve says Mastercard did not communicate directly with Valve; processors relayed communications involving Mastercard and acquiring banks. Valve says those processors rejected its legal-content argument and cited Rule 5.12.7. This is a company account supplied to its recipient journalist, not recovered upstream instructions. Locator: Update 23:36 BST, lines 8–10; last update 22:38 UTC. Recipient's August article (see source references)\n\nThis permits a Steam-specific attributed chain without importing Stripe's itch role. The processors remain unnamed. The existing graph node “Payment-letter addressees” is not their identity. Mastercard's denial and the June 3, 2025 rulebook remain separate retained records. The apparent disagreement could concern who communicated, what was requested, or downstream interpretation of a broad rule; it cannot be resolved by treating either public statement as the missing instruction.\n\n**Itch: internal eligibility and exit machinery predate July 2025.** An official admin reply dated April 26, 2023 says selected developers had lost eligibility for collected payouts and links the platform's notice. It distinguishes selected exclusions from a universal change and says raising a seller's revenue share will not reverse the decision. Locator: pinned reply opening and revenue-share discussion; original HTML timestamp title 2023-04-26 17:07:52, timezone unspecified. Original admin reply (see source references)\n\nThe linked notice currently explains content/risk/review-resource factors and requires affected sellers to leave collected payouts. Final-balance timing and a migration window govern the transition; existing purchases remain distinct from accepting new payments. Locator: opening paragraphs and final-payout FAQ. Original notice template (see source references)\n\nThe template is undated current text, linked by a dated 2023 post with later edits. We did not recover immutable 2023 wording. Its value is a documented decision function and switching constraints; it does not establish why July 2025 restrictions occurred.\n\n**Itch: UK page blocking was a separate compliance response.** Its July 27, 2025 official update describes geographic restrictions attributed to OSA compliance, tag classification with staff override, a dashboard appeal, and preserved downloads for existing purchases. It also announces exploration of unnamed digital-ID providers. Locator: lines 30–36 and 47; topic created 22:08 UTC and now marked five edits. Original OSA update (see source references)\n\nA dated July 29 compiler entry reproduces material paragraphs and links that original. It supports contemporaneous circulation, not independent corroboration or a deployed provider relationship. Compiler entry (see source references) The candidate therefore keeps geographic blocking, payment eligibility and prospective ID procurement separate.\n\n**The edited Stripe account has a contemporaneous reproduction.** GamingOnLinux's August 1 article, published 15:53 UTC and updated 15:55 UTC, reproduces itch's account of Stripe's bank restriction and sexual-content boundary. It narrows the risk that this central explanation appeared only in a later edit of itch's July 31 post. Locator: lines 9–13. It depends on that post and is not another Stripe confirmation. Contemporaneous reproduction (see source references)\n\n## Historical policy: identified capture, unavailable payload\n\nAn ordinary public request returned the Internet Archive's CDX index for Stripe's restricted-businesses page. Selected exact records:\n\n| Capture (UTC) | CDX digest |\n| --- | --- |\n| 2025-07-08 05:08:09 | HNQ3PWU6LEKUJDDDBLCEIJO2L6UNZLSI |\n| 2025-07-16 03:57:34 | Y6IHCIODMO7GZPK2DIX2NFX4HCTLKHGN |\n| 2025-07-21 11:38:01 | SPBM6EM327CWVJQ5YGRLJQ45SEMGZY7E |\n| 2025-07-24 06:25:46 | RK6F3PRVKVHI4UA7M5SABV2U7XH7HRJ5 |\n| 2025-07-30 08:02:27 | 7BWWDYBGRVVVIR6JKXTRUG4C4AGDX237 |\n| 2025-07-30 13:58:37 | MBJBZ4JBCCYJOOFIXUXHLQVPWYQAFQHP |\n| 2025-07-31 17:25:09 | YOGAFSZ53AHNPYSFVMMGLUL5Q72EV3YG |\n\nExact CDX query (see source references) Different digests do not establish changed policy text; dynamic HTML is an alternative.\n\nPayload replay using id_ URLs failed in the web tool with a nonretryable URL-safety/parser error. One standard-replay substitute also failed: identified July 21 capture (see source references) This route stopped. A separate itch archive-index request returned HTTP 429 and was not retried. Current Stripe general terms (see source references) were inspected only as background after the /legal/ssa/us route failed; they were not treated as itch's July 2025 contract.\n\n## The records still held elsewhere\n\nThe itch joint is the original July 2025 conversation/communication, whatever its medium; the bank instruction it references; the merchant/acquirer arrangement; and the accepted terms and implementation exchange. Public language about speaking with Stripe does not establish an email. Relevant custodians are itch payments/operations and Stripe's account/risk team. The bank becomes a named lead only when account-specific routing identifies it.\n\nThe Steam joint is Valve's processor notice and response, the corresponding Mastercard/acquirer message and the actual counterparty identities. Valve payments/legal and its processor counterparts should hold operational communications; Valve press and GamingOnLinux hold the identified company replies. The published replies establish an attributed account, not the contents of unrecovered upstream records.\n\nNo original bank instruction or account-specific historical agreement was located in the bounded official, archive-index and released-correspondence/exhibit search. This is not an exhaustive court-docket search and does not establish nonexistence or a holder refusal. No outreach occurred. Reopen on a released notice, a functioning authorized archive route, routing disclosure or substantive holder response.\n\n## Exclusions and comparison\n\nA compiler's Discord transcription offered a potentially useful Stripe case comment, but supplied no original message link or screenshot. Its detailed entry dates the message July 25 at 20:13 -0700, the linked timeline dates it July 25 at 03:13 UTC, and the entry creation metadata is July 25 at 06:20:34. These do not reconcile. It is excluded from event chronology and mechanism claims. Transcription (see source references)\n\nThe evidence now supports a specific bargaining step in Steam's account and documented platform decision machinery at itch. It remains compatible with outside instruction, downstream underwriting interpretation and platform precaution operating together. Geographic compliance creates an additional, separately evidenced explanation for some restrictions. None of these records establishes a common command structure, a named bank behind itch, a Stripe–Steam processing relationship or a deployed itch identity-provider contract.\n\n## Integration handoff\n\npayments-candidate.json contains five new role/event/program nodes, seven typed edges, fourteen source records (including explicit retained-origin alias and inaccessible/excluded records), four leads and provenance/dependency notes. Its external node IDs refer to the retained explorer. Alias the retained itch reindexing source by exact URL during integration. Source counts are bibliographic records, not independent confirmations. No existing relationship is upgraded by adding a derivative report. No effective start date is inferred from September 8 inspection.\n\nOnly this note and the candidate packet were written. They contain citations, bounded paraphrases, access outcomes and index metadata; no full-text local captures are claimed.\n", "file": "research/red-threads-wave-2026-09-08-payments-findings.md" }, { "id": "red-threads-wave-2026-09-08b-5rights-findings", "title": "5Rights–Capitol Connection: the 2022 operation now has a documentary middle", "description": "Research packet for root integration; captured September 8, 2026 in America/Chicago (some receipt timestamps are September 9 UTC). This is a bounded historical acquisition, not a statement about the Act's present enforceability. Canonical case and reader edition were not changed.", "html": "Research packet for root integration; captured September 8, 2026 in America/Chicago (some receipt timestamps are September 9 UTC). This is a bounded historical acquisition, not a statement about the Act's present enforceability. Canonical case and reader edition were not changed.
\nThe retained $50,000 contract is now connected to actual filed advocacy materials, dated amendment correspondence and published legislative arguments. The strongest addition is not another funding association. It is an identifiable policy dispute: preserve protection through age 17 while conceding changes elsewhere. The correspondence records 5Rights' chair advocating that position; California's enacted 2022 text retains it. Causation of individual amendments and donor approval remain unresolved.
\nThe original [Exhibit AB](http[local research file] PDF pp22–23, covers May 1–September 30, 2022. Nick Martlew signs as 5Rights executive director on May 24; Doug Subers signs for Capitol Connection on May 17. Its scope includes coalition calls, lobbying, letter development/review and stakeholder-amendment negotiation. It assigns resulting work, including drafts, to 5Rights. The prior-written-consent provision concerns disclosure of non-public information, with legal and already-public exceptions; it is not a separately stated requirement that every legislative position receive editorial approval. Some expenses require advance client approval. Monthly invoices are contractually specified but are not attached.
\nPDF p21 contains five $10,000 entries under a “Date Received” column. Each gives a range, not a single bank date: June 2–29; July 18–August 12; August 2–12; August 27–September 21; September 22–October 4, 2022. These are registrant-reported receipts, not inspected bank transfers. The five entries total $50,000; their overlap should be preserved rather than silently regularized into five monthly payment dates. Work-product ownership, release permission and actual approval decisions must remain separate.
\nThis refresh enriches existing e157, not a new money edge. The original source duplicates src-55256b9868ae. No upstream earmark was established; the 2023 Oak award cannot fund this chain by chronological assumption.
\n[Kidron to Anthony Portantino, August 3, 2022](http[local research file] p1, thanks him for a meeting and says she worked with her team and the bill's authors on amendments responding to his issue and stakeholder needs. She seeks support for retaining under-18 coverage and alignment with the UK code. This is a named participant's contemporary account, subsequently filed by CapConn, rather than independent verification of everything described.
\n[Kidron to Henry Stern, August 3, 2022](http[local research file] p1, is more discriminating: it says industry participants sought an under-13 or under-16 boundary, asks Stern to preserve under-18 coverage, and reports amendments addressing Portantino's and other stakeholders' feedback. It also thanks Stern for willingness to become a co-author. The two letters bear House of Lords stationery. That establishes the institutional presentation used, not a UK-government funding or instruction chain. The letters do not identify which industry party sought which age cutoff.
\nThe files distinguish a substantive preference from generic sponsorship. They support an active bargaining role for Kidron; they do not show her directing legislators or prove her preferences caused the final provisions.
\n[Filed coalition letter of June 10, 2022](http[local research file] pp1–2, addresses Senator Umberg and Senate Judiciary and bears 5Rights plus numerous coalition logos. The original [Assembly concurrence analysis, August 30 selection](http[local research file] analysisId 356279, PDF p4, reproduces distinctive passages from this letter under Arguments in Support. Four normalized matching blocks contain 59, 47, 40 and 34 words; the comparison receipt is 5rights-text-comparison.json. These are repeated advocacy passages, not independent confirmations of the underlying health statistics.
The FARA activity log, Exhibit AB p15, records a June 21 email/fax submission of a coalition support letter to Senate Judiciary. Those records form a strong document-family match; the public record does not include the email attachment hash or message headers proving the filed June 10 copy was the exact submitted version. Nor does filing the letter establish CapConn authored it. Legislative quotation establishes uptake as an argument, not committee endorsement or a legislator's changed vote.
\nThere is a second, narrower example: the log p14 records submission of a Tim Kendall support letter to Senate Judiciary on June 13. The original June 24-indexed [Judiciary analysis](http[local research file] analysisId 352148, for the June 28 hearing, quotes Kendall's support on p19. An exact original Kendall letter was not identified among the thirteen filed informational attachments.
\nThe visually checked log provides better next-record selectors than a generic request for correspondence. On June 9–10, drafts of a committee background sheet go to Shannon McKinley, Wicks' chief of staff (p14). On August 5, proposed author amendments go to Aaron Skaggs in Portantino's office, and a UK/California comparison chart goes to Eric Dang in the Senate President pro Tempore policy unit (p18). On August 10, a meeting/email with Christine Aurre in Governor Newsom's office includes a summary of amendments; August 15 and 23 entries describe updates responding to opposition concerns (pp18–19). June and July entries record repeated stakeholder negotiations, but a full actor-to-clause attribution was not completed in this pass.
\nA [filed comparison chart](http[local research file] p1, explicitly presents the bill as proposed for amendment in Senate Appropriations and advocates cross-jurisdiction consistency to ease compliance and establish a global standard. It is undated. It plausibly belongs to the August 5 chart transaction; exact attachment identity is unproved. Its assertions about legal equivalence should not be imported as this investigation's own legal conclusion.
\nThe official [August 23 Senate analysis](http[local research file] analysisId 355438, p1, says the August 22 amendments refined definitions, exempted products/services, changed timelines and expanded cure rights. The Assembly concurrence analysis 356279, pp1–2, itemizes changes including assessments supplied on Attorney General request instead of routine public-agency submission; assessment confidentiality; a qualified 90-day cure opportunity; and replacement of a taskforce with a working group. Its p4 also publishes opponents' request for guidance and opportunity to correct mistakes. These changes are documentary evidence of a changing proposal, not proof which participant won each concession.
\nThe [chaptered 2022 Act](http[local research file] chapter320, approved September15, retains under-18 at Civil Code1798.99.30(b)(1). Section1798.99.31(a)(5) permits businesses to apply child protections to everyone instead of estimating age. Thus the historical text does not mandate identity-document submission by every user. This is a consequential qualification to an undifferentiated universal-ID interpretation. No contemporary litigation status was researched here.
\nThe official [FARA registrant index](http[local research file] returned 18 rows for7427: ExhibitAB, registration, ExhibitC, two short forms and thirteen informational-material PDFs14–26. No supplemental statement appears in those returned rows. This is bounded index coverage, not proof none exists elsewhere. The web-reader variant of a different FARA index URL failed; an ordinary public HTML route succeeded. No access-control bypass, credentials or outreach were used.
\nAll thirteen material PDFs were acquired. Full argumentative text was inspected for19,20,21,26; material18 was inspected at first and final pages;17 at first/final pages. Other materials were screened at first-page/title level only. They include background material, a May6 5Rights support letter, a Frances Haugen letter, a July27 5Rights letter, poll tables, and Assembly/Senate floor alerts. Their existence does not make every assertion verified. The original contract and log were read in targeted text and visually at pp14–15,17–19,21,23; other log pages received keyword screening. Legislative analyses were inspected at the precise cited pages, not represented as exhaustively legally reviewed.
\n5rights-acquisition-receipts.json preserves original URL, selected public form parameters, byte length, response status, local path and SHA-256. California analyses were returned by the site's normal public POST form; use the index labels and analysisIds to reproduce the selection. 5rights-originals-manifest.json hashes the retained packet files. All documents are source captures; OCR text is subordinate to page images. Original FARA materials and activity descriptions share CapConn filing ancestry. Legislative publication independently establishes what its analysis printed, while quoted coalition content retains coalition ancestry.
No invoice, line-item account, internal approval message, final tracked-change draft or exact August5 amendment attachment was found in the inspected documents. These absences do not mean no such record exists.
\nThe highest-value next record is the August5,2022 CapConn email to Aaron Skaggs with proposed author amendments, including attachment versions and its associated exchanges from July18 through August11. Likely holders are CapConn's engagement archive, 5Rights' policy archive and the California Senate custodian for Portantino/Appropriations records. It can distinguish who originated an amendment, who merely relayed it, and which draft survived. A substantive public reopening needs an attached correspondence release, legislative bill file or newly indexed exhibit; repeating general searches adds little.
\nNext are the June9–10 background-sheet exchanges with McKinley and the June21 coalition-letter transmission/approval chain, held by CapConn, 5Rights and the relevant legislative office/committee record custodians. These can distinguish commissioned drafting from transmission of already-approved coalition material, and routine coordination from exercise of client release permission.
\nFor funding allocation, 5Rights' finance custodian and CapConn's finance custodian hold the specified monthly invoices, remittance records and ledger coding. These could identify a restricted funding allocation or a general operating pool. Nick Martlew and Doug Subers are evidenced contract signatories in2022, not asserted current record custodians. No proposed request has been sent.
", "path": "red-threads-wave-2026-09-08b/5rights-findings.md", "sourceIds": [ "src-55256b9868ae", "w3-5r-s-index", "w3-5r-s-material-14", "w3-5r-s-material-15", "w3-5r-s-material-16", "w3-5r-s-material-17", "w3-5r-s-material-18", "w3-5r-s-material-19", "w3-5r-s-material-20", "w3-5r-s-material-21", "w3-5r-s-material-22", "w3-5r-s-material-23", "w3-5r-s-material-24", "w3-5r-s-material-25", "w3-5r-s-material-26", "w3-5r-s-analysis-352148", "w3-5r-s-analysis-355438", "w3-5r-s-analysis-356279", "w3-5r-s-chapter" ], "markdown": "Research packet for root integration; captured September 8, 2026 in America/Chicago (some receipt timestamps are September 9 UTC). This is a bounded historical acquisition, not a statement about the Act's present enforceability. Canonical case and reader edition were not changed.\n\n## Finding\n\nThe retained $50,000 contract is now connected to actual filed advocacy materials, dated amendment correspondence and published legislative arguments. The strongest addition is not another funding association. It is an identifiable policy dispute: preserve protection through age 17 while conceding changes elsewhere. The correspondence records 5Rights' chair advocating that position; California's enacted 2022 text retains it. Causation of individual amendments and donor approval remain unresolved.\n\n### 1. Original contract: what the client actually owned, and what the money record actually says\n\nThe original Exhibit AB (see source references) PDF pp22–23, covers May 1–September 30, 2022. Nick Martlew signs as 5Rights executive director on May 24; Doug Subers signs for Capitol Connection on May 17. Its scope includes coalition calls, lobbying, letter development/review and stakeholder-amendment negotiation. It assigns resulting work, including drafts, to 5Rights. The prior-written-consent provision concerns disclosure of non-public information, with legal and already-public exceptions; it is not a separately stated requirement that every legislative position receive editorial approval. Some expenses require advance client approval. Monthly invoices are contractually specified but are not attached.\n\nPDF p21 contains five $10,000 entries under a “Date Received” column. Each gives a range, not a single bank date: June 2–29; July 18–August 12; August 2–12; August 27–September 21; September 22–October 4, 2022. These are registrant-reported receipts, not inspected bank transfers. The five entries total $50,000; their overlap should be preserved rather than silently regularized into five monthly payment dates. Work-product ownership, release permission and actual approval decisions must remain separate.\n\nThis refresh enriches existing **e157**, not a new money edge. The original source duplicates **src-55256b9868ae**. No upstream earmark was established; the 2023 Oak award cannot fund this chain by chronological assumption.\n\n### 2. A real client-side decision, recorded in two contemporary letters\n\nKidron to Anthony Portantino, August 3, 2022 (see source references) p1, thanks him for a meeting and says she worked with her team and the bill's authors on amendments responding to his issue and stakeholder needs. She seeks support for retaining under-18 coverage and alignment with the UK code. This is a named participant's contemporary account, subsequently filed by CapConn, rather than independent verification of everything described.\n\nKidron to Henry Stern, August 3, 2022 (see source references) p1, is more discriminating: it says industry participants sought an under-13 or under-16 boundary, asks Stern to preserve under-18 coverage, and reports amendments addressing Portantino's and other stakeholders' feedback. It also thanks Stern for willingness to become a co-author. The two letters bear House of Lords stationery. That establishes the institutional presentation used, not a UK-government funding or instruction chain. The letters do not identify which industry party sought which age cutoff.\n\nThe files distinguish a substantive preference from generic sponsorship. They support an active bargaining role for Kidron; they do not show her directing legislators or prove her preferences caused the final provisions.\n\n### 3. Work reached the legislative record\n\nFiled coalition letter of June 10, 2022 (see source references) pp1–2, addresses Senator Umberg and Senate Judiciary and bears 5Rights plus numerous coalition logos. The original Assembly concurrence analysis, August 30 selection (see source references) analysisId **356279**, PDF p4, reproduces distinctive passages from this letter under Arguments in Support. Four normalized matching blocks contain 59, 47, 40 and 34 words; the comparison receipt is `5rights-text-comparison.json`. These are repeated advocacy passages, not independent confirmations of the underlying health statistics.\n\nThe FARA activity log, Exhibit AB p15, records a June 21 email/fax submission of a coalition support letter to Senate Judiciary. Those records form a strong document-family match; the public record does not include the email attachment hash or message headers proving the filed June 10 copy was the exact submitted version. Nor does filing the letter establish CapConn authored it. Legislative quotation establishes uptake as an argument, not committee endorsement or a legislator's changed vote.\n\nThere is a second, narrower example: the log p14 records submission of a Tim Kendall support letter to Senate Judiciary on June 13. The original June 24-indexed Judiciary analysis (see source references) analysisId **352148**, for the June 28 hearing, quotes Kendall's support on p19. An exact original Kendall letter was not identified among the thirteen filed informational attachments.\n\n### 4. The amendment corridor is now dated and has specific holders\n\nThe visually checked log provides better next-record selectors than a generic request for correspondence. On June 9–10, drafts of a committee background sheet go to Shannon McKinley, Wicks' chief of staff (p14). On August 5, proposed author amendments go to **Aaron Skaggs in Portantino's office**, and a UK/California comparison chart goes to **Eric Dang in the Senate President pro Tempore policy unit** (p18). On August 10, a meeting/email with **Christine Aurre in Governor Newsom's office** includes a summary of amendments; August 15 and 23 entries describe updates responding to opposition concerns (pp18–19). June and July entries record repeated stakeholder negotiations, but a full actor-to-clause attribution was not completed in this pass.\n\nA filed comparison chart (see source references) p1, explicitly presents the bill as proposed for amendment in Senate Appropriations and advocates cross-jurisdiction consistency to ease compliance and establish a global standard. It is undated. It plausibly belongs to the August 5 chart transaction; exact attachment identity is unproved. Its assertions about legal equivalence should not be imported as this investigation's own legal conclusion.\n\n### 5. California's own records show both survival and narrowing\n\nThe official August 23 Senate analysis (see source references) analysisId **355438**, p1, says the August 22 amendments refined definitions, exempted products/services, changed timelines and expanded cure rights. The Assembly concurrence analysis **356279**, pp1–2, itemizes changes including assessments supplied on Attorney General request instead of routine public-agency submission; assessment confidentiality; a qualified 90-day cure opportunity; and replacement of a taskforce with a working group. Its p4 also publishes opponents' request for guidance and opportunity to correct mistakes. These changes are documentary evidence of a changing proposal, not proof which participant won each concession.\n\nThe chaptered 2022 Act (see source references) chapter320, approved September15, retains under-18 at Civil Code1798.99.30(b)(1). Section1798.99.31(a)(5) permits businesses to apply child protections to everyone instead of estimating age. Thus the historical text does not mandate identity-document submission by every user. This is a consequential qualification to an undifferentiated universal-ID interpretation. No contemporary litigation status was researched here.\n\n## Acquisition, ancestry and meaningful negatives\n\nThe official FARA registrant index (see source references) returned 18 rows for7427: ExhibitAB, registration, ExhibitC, two short forms and thirteen informational-material PDFs14–26. No supplemental statement appears in those returned rows. This is bounded index coverage, not proof none exists elsewhere. The web-reader variant of a different FARA index URL failed; an ordinary public HTML route succeeded. No access-control bypass, credentials or outreach were used.\n\nAll thirteen material PDFs were acquired. Full argumentative text was inspected for19,20,21,26; material18 was inspected at first and final pages;17 at first/final pages. Other materials were screened at first-page/title level only. They include background material, a May6 5Rights support letter, a Frances Haugen letter, a July27 5Rights letter, poll tables, and Assembly/Senate floor alerts. Their existence does not make every assertion verified. The original contract and log were read in targeted text and visually at pp14–15,17–19,21,23; other log pages received keyword screening. Legislative analyses were inspected at the precise cited pages, not represented as exhaustively legally reviewed.\n\n`5rights-acquisition-receipts.json` preserves original URL, selected public form parameters, byte length, response status, local path and SHA-256. California analyses were returned by the site's normal public POST form; use the index labels and analysisIds to reproduce the selection. `5rights-originals-manifest.json` hashes the retained packet files. All documents are source captures; OCR text is subordinate to page images. Original FARA materials and activity descriptions share CapConn filing ancestry. Legislative publication independently establishes what its analysis printed, while quoted coalition content retains coalition ancestry.\n\nNo invoice, line-item account, internal approval message, final tracked-change draft or exact August5 amendment attachment was found in the inspected documents. These absences do not mean no such record exists.\n\n## What the next holder record would change\n\nThe highest-value next record is the **August5,2022 CapConn email to Aaron Skaggs with proposed author amendments**, including attachment versions and its associated exchanges from July18 through August11. Likely holders are CapConn's engagement archive, 5Rights' policy archive and the California Senate custodian for Portantino/Appropriations records. It can distinguish who originated an amendment, who merely relayed it, and which draft survived. A substantive public reopening needs an attached correspondence release, legislative bill file or newly indexed exhibit; repeating general searches adds little.\n\nNext are the June9–10 background-sheet exchanges with McKinley and the June21 coalition-letter transmission/approval chain, held by CapConn, 5Rights and the relevant legislative office/committee record custodians. These can distinguish commissioned drafting from transmission of already-approved coalition material, and routine coordination from exercise of client release permission.\n\nFor funding allocation, 5Rights' finance custodian and CapConn's finance custodian hold the specified monthly invoices, remittance records and ledger coding. These could identify a restricted funding allocation or a general operating pool. Nick Martlew and Doug Subers are evidenced contract signatories in2022, not asserted current record custodians. No proposed request has been sent.\n", "file": "research/red-threads-wave-2026-09-08b-5rights-findings.md" }, { "id": "red-threads-wave-2026-09-08b-icmec-findings", "title": "ICMEC counterparts: money, participation and an observed drafting step", "description": "Read 8 September 2026 local time. Nine original documents/pages were captured, hashed and read at the locators in icmec-candidate.json. The international 36MB report exceeded the web reader limit; one ordinary public download succeeded. The Australian report also succeeded through ordinary public download. icmec-capture-manifest.json preserves the bytes and ", "html": "Read 8 September 2026 local time. Nine original documents/pages were captured, hashed and read at the locators in icmec-candidate.json. The international 36MB report exceeded the web reader limit; one ordinary public download succeeded. The Australian report also succeeded through ordinary public download. icmec-capture-manifest.json preserves the bytes and URLs.
\nWestpac's 2024 programme report records a $25 million six-year commitment beginning in 2021 for ICMEC data and knowledge sharing. Its Australian-project section names Lighthouse, a data application for banks and payment platforms, and Westpac participation in the SaferAI launch. The report uses both commitment and provided language; this is not a verified payment schedule. Its outcome footnotes explicitly say recipient information was not independently verified. The bank website's three-year grant description and this six-year programme description need the original agreement and amendments, not arithmetic invention. AUD is inferred from the Australian reporting context; the line itself prints only a dollar sign.
\nThe September 2020 Westpac response plan already promised $25 million for cross-border and cross-industry data sharing after AUSTRAC's proceedings. This amount is a precursor, not another funding flow to add. The recovered plan is bank-issued, not a court order mandating a grant. Westpac's April2026 update reports FY25 data projects supported by its funding and separately describes continued SaferAI convening; it supplies no coalition-specific allocation or editorial approval term.
\nICMEC Australia's own current account says it began as an affiliate in late2021 and now operates independently, with its own governance. It also describes foundational funding ending and a search for ongoing sponsors and project support. That creates a specific renewal/dependence question. It does not establish that the Australian and international organisations have common current command. The transition instrument remains missing.
\nSaferAI's terms, drafted September2025 and circulated January2026, allocate membership and partnership determination and project guidance to ICMEC Australia in collaboration with members. They also permit a partner category that participates without normal public website listing or project attribution. This proves the public roster is not designed as a complete attendance record; it does not prove that a particular unlisted institution participated. Attribution consent, output clearance, attendance and the category register are the decisive next records. Meeting confidentiality alone is not evidence of improper conduct.
\nThe Australian 2024-25 impact report describes operational Lighthouse searching and analytic functions for suspected child-exploitation data, with financial-sector and law-enforcement uses. It lists Jodie Arthur both as a non-executive director and a Westpac financial-crime executive. This is a resource-plus-governance connection. It does not establish donor nomination rights, her vote on any policy, or a bank action taken from an alert. Provider licences, provenance, access agreements, correction rules and de-identified alert-to-action records could distinguish those steps without seeking sensitive underlying data.
\nColm Gannon's 4May2026 article advocates device-centred age signals, arguing for reduced repeated disclosure, while acknowledging friction and circumvention. It reports a meeting with Mongolian parliamentarians and officials. This is actual public advocacy; third-party technical implementation claims in the article were not adopted as independent facts. No Westpac age-policy earmark or US drafting instruction was found.
\nVirginia's JCOTS annual report records Cunningham's November6,2024 device-based age presentation and the commission's questions. Its December3 proceedings explicitly say Senator Craig's device-age bill concept resulted from that presentation and that she submitted a DLS draft request. This is an unusually clear institutional counterpart for agenda uptake. It is not proof of enacted legislation or authorship of every later provision. The original drafting request and resulting versions are highly targeted next evidence.
\nThe international ICMEC2024 report lists digital age assurance among its programmes and identifies general2024 donors. That refreshes the earlier2023 general acknowledgement snapshot, but it does not disclose amounts or any age-policy earmark. Named general support must not be attributed to the Australian operation or to a specific bill without allocation records.
\nTargeted UK follow-up recovered official schedule pages for Sobel's September8 motion, not its outcome. The Hansard and bills open routes failed; a separate official schedule search did not supply the missing outcome. Preserve scheduled stage. Do not rerun identical failed routes or call introduction established.
", "path": "red-threads-wave-2026-09-08b/icmec-findings.md", "sourceIds": [ "w3-icmec-s-us-impact2024", "w3-icmec-s-au-impact2025", "w3-icmec-s-westpac2024", "w3-icmec-s-westpac2025", "w3-icmec-s-saferai-tor", "w3-icmec-s-virginia2024", "w3-icmec-s-au-about", "w3-icmec-s-au-age", "w3-icmec-s-westpac-response" ], "markdown": "Read 8 September 2026 local time. Nine original documents/pages were captured, hashed and read at the locators in icmec-candidate.json. The international 36MB report exceeded the web reader limit; one ordinary public download succeeded. The Australian report also succeeded through ordinary public download. icmec-capture-manifest.json preserves the bytes and URLs.\n\nWestpac's 2024 programme report records a $25 million six-year commitment beginning in 2021 for ICMEC data and knowledge sharing. Its Australian-project section names Lighthouse, a data application for banks and payment platforms, and Westpac participation in the SaferAI launch. The report uses both commitment and provided language; this is not a verified payment schedule. Its outcome footnotes explicitly say recipient information was not independently verified. The bank website's three-year grant description and this six-year programme description need the original agreement and amendments, not arithmetic invention. AUD is inferred from the Australian reporting context; the line itself prints only a dollar sign.\n\nThe September 2020 Westpac response plan already promised $25 million for cross-border and cross-industry data sharing after AUSTRAC's proceedings. This amount is a precursor, not another funding flow to add. The recovered plan is bank-issued, not a court order mandating a grant. Westpac's April2026 update reports FY25 data projects supported by its funding and separately describes continued SaferAI convening; it supplies no coalition-specific allocation or editorial approval term.\n\nICMEC Australia's own current account says it began as an affiliate in late2021 and now operates independently, with its own governance. It also describes foundational funding ending and a search for ongoing sponsors and project support. That creates a specific renewal/dependence question. It does not establish that the Australian and international organisations have common current command. The transition instrument remains missing.\n\nSaferAI's terms, drafted September2025 and circulated January2026, allocate membership and partnership determination and project guidance to ICMEC Australia in collaboration with members. They also permit a partner category that participates without normal public website listing or project attribution. This proves the public roster is not designed as a complete attendance record; it does not prove that a particular unlisted institution participated. Attribution consent, output clearance, attendance and the category register are the decisive next records. Meeting confidentiality alone is not evidence of improper conduct.\n\nThe Australian 2024-25 impact report describes operational Lighthouse searching and analytic functions for suspected child-exploitation data, with financial-sector and law-enforcement uses. It lists Jodie Arthur both as a non-executive director and a Westpac financial-crime executive. This is a resource-plus-governance connection. It does not establish donor nomination rights, her vote on any policy, or a bank action taken from an alert. Provider licences, provenance, access agreements, correction rules and de-identified alert-to-action records could distinguish those steps without seeking sensitive underlying data.\n\nColm Gannon's 4May2026 article advocates device-centred age signals, arguing for reduced repeated disclosure, while acknowledging friction and circumvention. It reports a meeting with Mongolian parliamentarians and officials. This is actual public advocacy; third-party technical implementation claims in the article were not adopted as independent facts. No Westpac age-policy earmark or US drafting instruction was found.\n\nVirginia's JCOTS annual report records Cunningham's November6,2024 device-based age presentation and the commission's questions. Its December3 proceedings explicitly say Senator Craig's device-age bill concept resulted from that presentation and that she submitted a DLS draft request. This is an unusually clear institutional counterpart for agenda uptake. It is not proof of enacted legislation or authorship of every later provision. The original drafting request and resulting versions are highly targeted next evidence.\n\nThe international ICMEC2024 report lists digital age assurance among its programmes and identifies general2024 donors. That refreshes the earlier2023 general acknowledgement snapshot, but it does not disclose amounts or any age-policy earmark. Named general support must not be attributed to the Australian operation or to a specific bill without allocation records.\n\nTargeted UK follow-up recovered official schedule pages for Sobel's September8 motion, not its outcome. The Hansard and bills open routes failed; a separate official schedule search did not supply the missing outcome. Preserve scheduled stage. Do not rerun identical failed routes or call introduction established.\n", "file": "research/red-threads-wave-2026-09-08b-icmec-findings.md" }, { "id": "red-threads-wave-2026-09-08b-oak-weprotect-findings", "title": "Oak → WeProtect: where resources become decisions", "description": "The new evidence identifies institutional budget rights, an amended funding route, a commissioned research product and a named Oak contribution to that product. It does not identify Alan Parker as the decisionmaker behind the two existing grants, or establish a donor veto over WeProtect publications. The useful geometry is two partly connected chains: Oak → ", "html": "Nova · 8 September 2026 · Bounded public-acquisition return against case-r6
\nThe new evidence identifies institutional budget rights, an amended funding route, a commissioned research product and a named Oak contribution to that product. It does not identify Alan Parker as the decisionmaker behind the two existing grants, or establish a donor veto over WeProtect publications. The useful geometry is two partly connected chains: Oak → Safe Futures Hub → commissioned review, with named advisory input; and external donors → Dutch WeProtect foundation → budget-constrained UK operating company. Their exact grant instruments still matter.
\nOak's [three-page grant-process guide](http[local research file] pp2–3, describes programme-officer screening and collaborative application development, review involving finance and programme leadership, and a recommendation to unspecified senior foundation members. Both parties sign a grant letter before funds are wired. The guide is undated; its URL contains 2019/11 and the current grant-making page still links it.
\nThe [current grant-making page](http[local research file] says renewals are separate awards, core support is unrestricted, and applicants are told who will decide. Its general 20%-of-organisational-budget / 50%-of-project-budget guidelines permit exceptions discussed with staff. These are policy statements, not proof that either named grant complied with a ratio or retained a specific termination right. The [FAQ](http[local research file] also describes recipient-led strategy and advance discussion of funding exits. This supplies a serious alternative explanation: selection of aligned, independently directed recipients followed by ordinary financial oversight.
\nMissing: the signed 2023 5Rights and 2025 WeProtect letters, concept notes, approval record and applicable renewal/reporting clauses. No new personal direction edge for Alan Parker is proposed. Existing e159–e161 remain intact.
\nThe [director-signed 2023 accounts](http[local research file] pp10,15,17,21, separate Stichting The WePROTECT Global Alliance (Dutch registration 76585379) from its controlled UK company. They disclose a 2021 agreement requiring a submitted annual budget, prior written permission for spending outside it, and a discretionary deduction of underspend from later grants. The €1,520,482 UK subgrant is an internal 2023 expenditure, eliminated in consolidation.
\nPage17 also records a revealing amendment: Together for Girls' original $400,000 Safe Futures commitment became $100,000 when the remaining $300,000 moved to “Swiss Philanthropy”; the replacement $430,000 agreement added $130,000, for March2024–February2027. Do not add $400,000 and $430,000. Exact Swiss legal identity, hosted fund and upstream donor remain unestablished.
\nThe source is byte-identical to the retained September7 PDF: newly inspected content, not independent confirmation. Its final page contains an auditor-report heading without a substantive opinion; this return does not call it independently audited. [December2023 minutes](http[local research file] separately record Policy Board approval of the 2024 budget.
\nThe [2024 evidence review](http[local research file] goes further. Page105 expressly identifies Safe Futures Hub as commissioner, Oak as supporter, three lead authors, and a named group of advisers/key informants—including Duch—credited with shaping the review. WeProtect's Iain Drennan appears in leadership and Shailey Hingorani in the core team. Page104 describes over 750 comments, suggestions and edits being reviewed. Pages17–18 distinguish systematically screened studies from additional literature; stakeholder feedback contributed particularly to the latter.
\nThis earns a contribution relationship to a specific output. It does not identify Duch's individual comments, an enforced editorial condition, a veto, or final manuscript approver. A broad expert consultation with recipient-selected content remains compatible with the record. The [publication page](http[local research file] dates publication to 22 October2024 and its update to 15 September2025. No attribution of NetID adoption or a specific statute is made.
\nThe later [2025 guidance framework](http[local research file] pp3,9–11,79–80, is a useful discrimination check. It acknowledges Oak support, credits Arti Mohan as author, and describes leadership/core and external draft review. The three advisory members specifically thanked for direct contribution are Manjeer Mukherjee, Sabine Rakotomalala and Shanaaz Mathews. Duch is not among those three. Generic advisory membership must not be projected into every publication. The guidance calls its tools adaptable rather than prescriptive; this is not a binding policy instrument.
\nWeProtect's [2023 annual review](http[local research file] pp5–7, identifies the Oak-supported Safe Futures launch and describes commissioning research, including the Economist Impact European survey. It reports using that survey in direct engagement with more than 60 national policymakers. The accounts' narrative links it to targeted government engagement and ECLAG-letter work. These are statements from the same institution, not two independent confirmations of influence or adoption.
\nThe [original European survey](http[local research file] pp3,7–8,20, identifies Economist Impact's role and February–March2023 fieldwork. It limits its population to retrospective accounts from 2,000 eighteen-year-olds across four countries. The questionnaire derived from the 2021 study; the original questionnaire/IRB record was not acquired. It expressly distinguishes expert advice from endorsement. The commissioning SOW, payment/funding code, draft changes and final signoff remain holder records. No Oak earmark to this survey is established.
\nThe packet contains 13 proposed nodes, 21 relationships, 13 source records and five parked holder-record leads. All new IDs use w3-oak-. The capture manifest stores exact URLs, response states, raw filenames, extracted-text filenames, sizes and SHA256 values. Reading extent is source-specific: the 128-page review was read for methods and production/review credits; the 83-page guidance for methodology, scope and named reviewers. Their scientific evidence bases were not audited.
\nThe compressed review URL returned HTTP403 once. One official alternate filename yielded the report; the inaccessible and accessible files are not asserted byte-identical. An erroneous Oak team URL failed once; the indexed official team page supplied the legitimate substitute. Targeted grant-letter and Swiss Safe Futures searches did not yield the missing instruments. Unrelated Oak–Swiss Philanthropy Foundation collaborations do not establish the donor behind this particular agreement.
\nKeep the existing umbrella WeProtect node. Add Dutch and UK legal actors separately; do not turn every member into a subsidiary. The new specific Policy Board is a component of the old wp-boards grouping, not an alias for both boards. Do not automatically merge “Swiss Philanthropy” with a named foundation, hosted fund or Oak. Neither the internal subgrant nor the amendment produces additional external money by being mapped twice. The already retained 2024 comparative-account mismatch remains unresolved; this lane does not invent a financial trend.
\nThe next discriminating records are concrete: named award letters; the 2021 Dutch-to-UK agreement and actual budget exception decisions; the Together for Girls amendment and replacement Swiss fund identifier; the review's tracked drafts and signoff; and the European survey SOW and cost code. No outreach, paid access, publication, automation or canonical-case change was performed.
", "path": "red-threads-wave-2026-09-08b/oak-weprotect-findings.md", "sourceIds": [ "w3-oak-grant-process", "w3-oak-grant-making", "w3-oak-faq", "w3-oak-wp-accounts-2023", "w3-oak-wp-minutes-dec2023", "w3-oak-wp-review2023", "w3-oak-european-survey", "w3-oak-safe-futures-about", "w3-oak-safe-futures-review", "w3-oak-safe-futures-guidance", "w3-oak-safe-futures-review-substitute", "w3-oak-team", "w3-oak-safe-futures-publication" ], "markdown": "Nova · 8 September 2026 · Bounded public-acquisition return against case-r6\n\nThe new evidence identifies institutional budget rights, an amended funding route, a commissioned research product and a named Oak contribution to that product. It does not identify Alan Parker as the decisionmaker behind the two existing grants, or establish a donor veto over WeProtect publications. The useful geometry is two partly connected chains: Oak → Safe Futures Hub → commissioned review, with named advisory input; and external donors → Dutch WeProtect foundation → budget-constrained UK operating company. Their exact grant instruments still matter.\n\n## Oak's public process identifies a gate, not a named grant approver\n\nOak's three-page grant-process guide (see source references) pp2–3, describes programme-officer screening and collaborative application development, review involving finance and programme leadership, and a recommendation to unspecified senior foundation members. Both parties sign a grant letter before funds are wired. The guide is undated; its URL contains 2019/11 and the current grant-making page still links it.\n\nThe current grant-making page (see source references) says renewals are separate awards, core support is unrestricted, and applicants are told who will decide. Its general 20%-of-organisational-budget / 50%-of-project-budget guidelines permit exceptions discussed with staff. These are policy statements, not proof that either named grant complied with a ratio or retained a specific termination right. The FAQ (see source references) also describes recipient-led strategy and advance discussion of funding exits. This supplies a serious alternative explanation: selection of aligned, independently directed recipients followed by ordinary financial oversight.\n\nMissing: the signed 2023 5Rights and 2025 WeProtect letters, concept notes, approval record and applicable renewal/reporting clauses. No new personal direction edge for Alan Parker is proposed. Existing e159–e161 remain intact.\n\n## The operating constraint is inside WeProtect\n\nThe director-signed 2023 accounts (see source references) pp10,15,17,21, separate Stichting The WePROTECT Global Alliance (Dutch registration 76585379) from its controlled UK company. They disclose a 2021 agreement requiring a submitted annual budget, prior written permission for spending outside it, and a discretionary deduction of underspend from later grants. The €1,520,482 UK subgrant is an internal 2023 expenditure, eliminated in consolidation.\n\nPage17 also records a revealing amendment: Together for Girls' original $400,000 Safe Futures commitment became $100,000 when the remaining $300,000 moved to “Swiss Philanthropy”; the replacement $430,000 agreement added $130,000, for March2024–February2027. Do not add $400,000 and $430,000. Exact Swiss legal identity, hosted fund and upstream donor remain unestablished.\n\nThe source is byte-identical to the retained September7 PDF: newly inspected content, not independent confirmation. Its final page contains an auditor-report heading without a substantive opinion; this return does not call it independently audited. December2023 minutes (see source references) separately record Policy Board approval of the 2024 budget.\n\n## Oak support reaches an actual commissioned review\n\nThe Safe Futures team page.\n\nThe 2024 evidence review (see source references) goes further. Page105 expressly identifies Safe Futures Hub as commissioner, Oak as supporter, three lead authors, and a named group of advisers/key informants—including Duch—credited with shaping the review. WeProtect's Iain Drennan appears in leadership and Shailey Hingorani in the core team. Page104 describes over 750 comments, suggestions and edits being reviewed. Pages17–18 distinguish systematically screened studies from additional literature; stakeholder feedback contributed particularly to the latter.\n\nThis earns a contribution relationship to a specific output. It does not identify Duch's individual comments, an enforced editorial condition, a veto, or final manuscript approver. A broad expert consultation with recipient-selected content remains compatible with the record. The publication page (see source references) dates publication to 22 October2024 and its update to 15 September2025. No attribution of NetID adoption or a specific statute is made.\n\nThe later 2025 guidance framework (see source references) pp3,9–11,79–80, is a useful discrimination check. It acknowledges Oak support, credits Arti Mohan as author, and describes leadership/core and external draft review. The three advisory members specifically thanked for direct contribution are Manjeer Mukherjee, Sabine Rakotomalala and Shanaaz Mathews. Duch is not among those three. Generic advisory membership must not be projected into every publication. The guidance calls its tools adaptable rather than prescriptive; this is not a binding policy instrument.\n\n## The European research-to-advocacy route is also explicit\n\nWeProtect's 2023 annual review (see source references) pp5–7, identifies the Oak-supported Safe Futures launch and describes commissioning research, including the Economist Impact European survey. It reports using that survey in direct engagement with more than 60 national policymakers. The accounts' narrative links it to targeted government engagement and ECLAG-letter work. These are statements from the same institution, not two independent confirmations of influence or adoption.\n\nThe original European survey (see source references) pp3,7–8,20, identifies Economist Impact's role and February–March2023 fieldwork. It limits its population to retrospective accounts from 2,000 eighteen-year-olds across four countries. The questionnaire derived from the 2021 study; the original questionnaire/IRB record was not acquired. It expressly distinguishes expert advice from endorsement. The commissioning SOW, payment/funding code, draft changes and final signoff remain holder records. No Oak earmark to this survey is established.\n\n## Custody, boundaries and merge instructions\n\nThe packet contains 13 proposed nodes, 21 relationships, 13 source records and five parked holder-record leads. All new IDs use w3-oak-. The capture manifest stores exact URLs, response states, raw filenames, extracted-text filenames, sizes and SHA256 values. Reading extent is source-specific: the 128-page review was read for methods and production/review credits; the 83-page guidance for methodology, scope and named reviewers. Their scientific evidence bases were not audited.\n\nThe compressed review URL returned HTTP403 once. One official alternate filename yielded the report; the inaccessible and accessible files are not asserted byte-identical. An erroneous Oak team URL failed once; the indexed official team page supplied the legitimate substitute. Targeted grant-letter and Swiss Safe Futures searches did not yield the missing instruments. Unrelated Oak–Swiss Philanthropy Foundation collaborations do not establish the donor behind this particular agreement.\n\nKeep the existing umbrella WeProtect node. Add Dutch and UK legal actors separately; do not turn every member into a subsidiary. The new specific Policy Board is a component of the old wp-boards grouping, not an alias for both boards. Do not automatically merge “Swiss Philanthropy” with a named foundation, hosted fund or Oak. Neither the internal subgrant nor the amendment produces additional external money by being mapped twice. The already retained 2024 comparative-account mismatch remains unresolved; this lane does not invent a financial trend.\n\nThe next discriminating records are concrete: named award letters; the 2021 Dutch-to-UK agreement and actual budget exception decisions; the Together for Girls amendment and replacement Swiss fund identifier; the review's tracked drafts and signoff; and the European survey SOW and cost code. No outreach, paid access, publication, automation or canonical-case change was performed.\n", "file": "research/red-threads-wave-2026-09-08b-oak-weprotect-findings.md" }, { "id": "red-threads-wave-2026-09-08b-opposing-findings", "title": "Opposing intervention: a filed vendor route, then a contract frontier", "description": "Ready for independent review. This pass follows one operation: Leading the Future (LTF) → Think Big → Lantern Production Consultants LLC → March 2026 opposition to Alexander Bores in the NY-12 federal primary. It establishes a named vendor and itemized services beyond the existing reported $5,000 creator offer. It does not establish ultimate subcontractors, ", "html": "Ready for independent review. This pass follows one operation: Leading the Future (LTF) → Think Big → Lantern Production Consultants LLC → March 2026 opposition to Alexander Bores in the NY-12 federal primary. It establishes a named vendor and itemized services beyond the existing reported $5,000 creator offer. It does not establish ultimate subcontractors, an original donor's instruction, or the effect of the spending. Public records retrieved September 8, 2026. No messages, paid acquisition or publication.
\nLTF's Q1 Schedule B, page 13, reports a $5 million contribution to Think Big on February 24, 2026, transaction SB23.I31388. This is a component of the retained e147 $40 million contributions-to-other-committees aggregate. It is not a new external donation and is not a transfer to Build American AI. The source is the exact original already represented by src-6460d3b558f5; alias it instead of manufacturing corroboration. [LTF original](http[local research file]
The controlling March return is Think Big's May 20 amended April monthly report, covering March 1–31. Its seven Bores-opposition Schedule E rows total $787,404.29. Each names Lantern as payee and checks opposition, House NY-12, primary 2026. The filed dissemination and disbursement-or-obligation dates are different fields; their order is preserved below. These are reported independent expenditures, not independently verified bank settlements or receipts from a creator. [Amended return, pages 11–14](http[local research file]
\n| Transaction | \nPDF page | \nFiled purpose | \nUSD | \nDissemination | \nDisbursement or obligation | \n
|---|---|---|---|---|---|
SE.4218 | \n11 | \nDIGITAL MEDIA PRODUCTION | \n35,000.00 | \n2026-03-01 | \n2026-03-02 | \n
SE.4219 | \n11 | \nDIGITAL MEDIA PLACEMENT | \n171,136.11 | \n2026-03-01 | \n2026-03-02 | \n
SE.4220 | \n12 | \nPRINT ADVERTISING | \n111,136.11 | \n2026-03-01 | \n2026-03-02 | \n
SE.4221 | \n13 | \nTEXTING | \n50,000.00 | \n2026-03-12 | \n2026-03-11 | \n
SE.4222 | \n13 | \nDIGITAL ADVERTISING / PRODUCTION | \n256,691.67 | \n2026-03-15 | \n2026-03-16 | \n
SE.4223 | \n14 | \nPRINT ADVERTISING | \n83,799.29 | \n2026-03-14 | \n2026-03-16 | \n
SE.4224 | \n14 | \nCABLE ADVERTISING / PRODUCTION | \n79,641.11 | \n2026-03-30 | \n2026-03-27 | \n
The same Schedule E includes $277,827.78 for digital placement supporting Jesse Jackson Jr (SE.4217, p12). That is excluded from the Bores total. Together they reconcile to March independent expenditures of $1,065,232.07. These are the ordinary return's rows, not rapid notices added onto an ordinary return.
Schedule B pages 7–9 separately report $215,037 of operating disbursements to Lantern: polling $121,500; research $13,500; management consulting $15,012; strategy consulting $20,025; political strategy consulting $25,000; media production $20,000. The candidate fields are blank; several rows carry an “Other 2025” election label despite March 2026 disbursement dates. The record does not allocate all this work to Bores. A separate $30,976 direct disbursement to Manhattan Strategies makes total operating spending $246,013; it does not establish that firm as Lantern's subcontractor. Operating plus independent expenditures reconcile to the return's $1,311,245.07 total disbursements. These partitions are not additional flows to add onto that total.
\nFunding is also not an earmark. Think Big's 2025 year-end receipt pages disclose $500,000 from Ronald Conway on October 21 and $5 million from LTF on December 15. Its first reported Lantern expenditure on December 10 predates that December LTF receipt. Accordingly, even the record's short payment route cannot justify assigning every early ad to that particular transfer or to one LTF donor. This pass uses the February contribution followed by the March operation and makes no exact-dollar tracing claim. [Year-end return, pp2,6–8](http[local research file]
\nThe original April 20 return left the $13,500 Lantern purpose blank and used “political consulting” for a $25,000 item. The FEC's April 26 request asked for the missing purpose and a clearer description. The May 20 amendment supplies research and political strategy consulting respectively. The March expense amounts and selected Schedule E rows are unchanged. The public index shows this amendment; the original must not be presented as the current uncorrected disclosure. The request concerns those two reporting descriptions, not a Commission finding about vendor ownership or unlawful concealment. [Original](http[local research file] [FEC request](http[local research file] [filing index](http[local research file]
\nCity & State's January 26 report records Josh Vlasto, working with both LTF and Think Big, criticizing Bores' state AI approach and advocating national solutions. It also records the campaign's dispute over the ads' ICE/Palantir claims. This is an attributable account of the policy conflict, not evidence that those attack allegations were true. The original linked December YouTube page yielded only the title Alex Bores is Wrong for AI & Wrong for Congress. No video or transcript was reviewed and no particular March row was matched to that creative. No claim here establishes which March ad wording aired. [City & State](http[local research file] [linked original page, metadata only](http[local research file]
\nThe strongest new result therefore concerns purchased electoral capacity and service types. It does not close the retained Build American AI → SM4 → creator-prospect offer-only route. Public First Action remains a different entity with a different funding instrument; no inference here moves Anthropic's restricted gift into election spending.
\nThe May 5 CLC/Gonsalves-Brown complaint alleges that Think Big used Lantern to obscure ultimate providers. The inspected passage derives its expenditure case from FEC records; it is not an independent payment witness. Its ownership, exclusivity, lack of arm's-length dealing and unlawful-concealment conclusions remain complainant allegations and inferences, not findings adopted by this pass. Its all-period totals were not recomputed or integrated. Read pp7–11 and 15–18; the other case, exhibits and complete current enforcement docket were not reviewed. [Complaint](http[local research file]
\nA general contractor may coordinate media, polling and consultants under an ordinary commercial agreement. Alternatively, a vendor may be integrated with a campaign and function mainly as a payment intermediary. The contract, actual personnel overlap and subvendor instructions discriminate those explanations. The FEC's historical Advisory Opinion 1983-25 permitted less detailed reporting for the specific separately operated, arms-length media arrangement described there, with attention to other clients and independent principals. This does not decide Lantern's legal status; it explains why vendor concentration alone is insufficient. [Official opinion, pp2–3](http[local research file]
\nStop here: the specific next record is Think Big/Lantern's executed March agreement and invoice/placement bundle keyed to SE.4218–SE.4224 (excluding SE.4217). It should identify the commissioned creative, media placement orders, actual producers and approval rights, with a subvendor ledger to reconcile net fees against pass-through costs. Think Big and Lantern are the identified holders; a publicly released station order tied to one transaction could reopen a smaller part of the route. No holder request was sent. A shell-company accusation, complete creator fee or policy-effect claim cannot replace those records.
opposing-candidate.json contains five proposed edges, four new nodes, one holder lead, source scopes, local capture hashes and the seven-row extraction. Candidate source w3-opp-s-ltf-q1 aliases retained src-6460d3b558f5; contribution edge w3-opp-e-ltf-thinkbig is a subset of retained e147. Existing ltf, buildai, sm4, pitchcreator and publicfirst remain distinct and are not recreated. No existing Think Big, Lantern or Bores node was found by label/ID scan of r6.
Text and rendered-page files derive from the corresponding PDFs and are not extra independent sources. The complaint's cited FEC originals, its appendix-based claims, and summaries of that complaint share ancestry. The June17 notice was read for initial vendor discovery but was not added to the March operation. The committee indices were used to locate amendments, not as independent corroboration of transaction amounts. The source packet's original and amended March forms are versions of the same return. No full-cycle spend total, no complete source coverage and no private contract access is claimed.
", "path": "red-threads-wave-2026-09-08b/opposing-findings.md", "sourceIds": [], "markdown": "**Ready for independent review.** This pass follows one operation: Leading the Future (LTF) → Think Big → Lantern Production Consultants LLC → March 2026 opposition to Alexander Bores in the NY-12 federal primary. It establishes a named vendor and itemized services beyond the existing reported $5,000 creator offer. It does not establish ultimate subcontractors, an original donor's instruction, or the effect of the spending. Public records retrieved September 8, 2026. No messages, paid acquisition or publication.\n\n## What the money demonstrably reaches\n\nLTF's Q1 Schedule B, page 13, reports a **$5 million contribution to Think Big on February 24, 2026**, transaction `SB23.I31388`. This is a component of the retained `e147` $40 million contributions-to-other-committees aggregate. It is not a new external donation and is not a transfer to Build American AI. The source is the exact original already represented by `src-6460d3b558f5`; alias it instead of manufacturing corroboration. LTF original (see source references)\n\nThe controlling March return is Think Big's **May 20 amended April monthly report**, covering March 1–31. Its seven Bores-opposition Schedule E rows total **$787,404.29**. Each names Lantern as payee and checks opposition, House NY-12, primary 2026. The filed dissemination and disbursement-or-obligation dates are different fields; their order is preserved below. These are reported independent expenditures, not independently verified bank settlements or receipts from a creator. Amended return, pages 11–14 (see source references)\n\n| Transaction | PDF page | Filed purpose | USD | Dissemination | Disbursement or obligation |\n| --- | --- | --- | --- | --- | --- |\n| `SE.4218` | 11 | DIGITAL MEDIA PRODUCTION | 35,000.00 | 2026-03-01 | 2026-03-02 |\n| `SE.4219` | 11 | DIGITAL MEDIA PLACEMENT | 171,136.11 | 2026-03-01 | 2026-03-02 |\n| `SE.4220` | 12 | PRINT ADVERTISING | 111,136.11 | 2026-03-01 | 2026-03-02 |\n| `SE.4221` | 13 | TEXTING | 50,000.00 | 2026-03-12 | 2026-03-11 |\n| `SE.4222` | 13 | DIGITAL ADVERTISING / PRODUCTION | 256,691.67 | 2026-03-15 | 2026-03-16 |\n| `SE.4223` | 14 | PRINT ADVERTISING | 83,799.29 | 2026-03-14 | 2026-03-16 |\n| `SE.4224` | 14 | CABLE ADVERTISING / PRODUCTION | 79,641.11 | 2026-03-30 | 2026-03-27 |\n\nThe same Schedule E includes $277,827.78 for digital placement **supporting Jesse Jackson Jr** (`SE.4217`, p12). That is excluded from the Bores total. Together they reconcile to March independent expenditures of $1,065,232.07. These are the ordinary return's rows, not rapid notices added onto an ordinary return.\n\nSchedule B pages 7–9 separately report **$215,037 of operating disbursements to Lantern**: polling $121,500; research $13,500; management consulting $15,012; strategy consulting $20,025; political strategy consulting $25,000; media production $20,000. The candidate fields are blank; several rows carry an “Other 2025” election label despite March 2026 disbursement dates. The record does not allocate all this work to Bores. A separate $30,976 direct disbursement to Manhattan Strategies makes total operating spending $246,013; it does not establish that firm as Lantern's subcontractor. Operating plus independent expenditures reconcile to the return's $1,311,245.07 total disbursements. These partitions are not additional flows to add onto that total.\n\nFunding is also not an earmark. Think Big's 2025 year-end receipt pages disclose $500,000 from Ronald Conway on October 21 and $5 million from LTF on December 15. Its first reported Lantern expenditure on December 10 predates that December LTF receipt. Accordingly, even the record's short payment route cannot justify assigning every early ad to that particular transfer or to one LTF donor. This pass uses the February contribution followed by the March operation and makes no exact-dollar tracing claim. Year-end return, pp2,6–8 (see source references)\n\n## The amendment is consequential\n\nThe original April 20 return left the $13,500 Lantern purpose blank and used “political consulting” for a $25,000 item. The FEC's April 26 request asked for the missing purpose and a clearer description. The May 20 amendment supplies research and political strategy consulting respectively. The March expense amounts and selected Schedule E rows are unchanged. The public index shows this amendment; the original must not be presented as the current uncorrected disclosure. The request concerns those two reporting descriptions, not a Commission finding about vendor ownership or unlawful concealment. Original (see source references) FEC request (see source references) filing index (see source references)\n\n## What the operation was trying to do\n\nCity & State's January 26 report records Josh Vlasto, working with both LTF and Think Big, criticizing Bores' state AI approach and advocating national solutions. It also records the campaign's dispute over the ads' ICE/Palantir claims. This is an attributable account of the policy conflict, not evidence that those attack allegations were true. The original linked December YouTube page yielded only the title **Alex Bores is Wrong for AI & Wrong for Congress**. No video or transcript was reviewed and no particular March row was matched to that creative. No claim here establishes which March ad wording aired. City & State (see source references) linked original page, metadata only (see source references)\n\nThe strongest new result therefore concerns purchased electoral capacity and service types. It does not close the retained Build American AI → SM4 → creator-prospect offer-only route. Public First Action remains a different entity with a different funding instrument; no inference here moves Anthropic's restricted gift into election spending.\n\n## The holder frontier, with rival explanations\n\nThe May 5 CLC/Gonsalves-Brown complaint alleges that Think Big used Lantern to obscure ultimate providers. The inspected passage derives its expenditure case from FEC records; it is not an independent payment witness. Its ownership, exclusivity, lack of arm's-length dealing and unlawful-concealment conclusions remain **complainant allegations and inferences**, not findings adopted by this pass. Its all-period totals were not recomputed or integrated. Read pp7–11 and 15–18; the other case, exhibits and complete current enforcement docket were not reviewed. Complaint (see source references)\n\nA general contractor may coordinate media, polling and consultants under an ordinary commercial agreement. Alternatively, a vendor may be integrated with a campaign and function mainly as a payment intermediary. The contract, actual personnel overlap and subvendor instructions discriminate those explanations. The FEC's historical Advisory Opinion 1983-25 permitted less detailed reporting for the specific separately operated, arms-length media arrangement described there, with attention to other clients and independent principals. This does not decide Lantern's legal status; it explains why vendor concentration alone is insufficient. Official opinion, pp2–3 (see source references)\n\n**Stop here:** the specific next record is Think Big/Lantern's executed March agreement and invoice/placement bundle keyed to `SE.4218`–`SE.4224` (excluding `SE.4217`). It should identify the commissioned creative, media placement orders, actual producers and approval rights, with a subvendor ledger to reconcile net fees against pass-through costs. Think Big and Lantern are the identified holders; a publicly released station order tied to one transaction could reopen a smaller part of the route. No holder request was sent. A shell-company accusation, complete creator fee or policy-effect claim cannot replace those records.\n\n## Custody and integration\n\n`opposing-candidate.json` contains five proposed edges, four new nodes, one holder lead, source scopes, local capture hashes and the seven-row extraction. Candidate source `w3-opp-s-ltf-q1` aliases retained `src-6460d3b558f5`; contribution edge `w3-opp-e-ltf-thinkbig` is a subset of retained `e147`. Existing `ltf`, `buildai`, `sm4`, `pitchcreator` and `publicfirst` remain distinct and are not recreated. No existing Think Big, Lantern or Bores node was found by label/ID scan of r6.\n\nText and rendered-page files derive from the corresponding PDFs and are not extra independent sources. The complaint's cited FEC originals, its appendix-based claims, and summaries of that complaint share ancestry. The June17 notice was read for initial vendor discovery but was not added to the March operation. The committee indices were used to locate amendments, not as independent corroboration of transaction amounts. The source packet's original and amended March forms are versions of the same return. No full-cycle spend total, no complete source coverage and no private contract access is claimed.\n", "file": "research/red-threads-wave-2026-09-08b-opposing-findings.md" }, { "id": "roost-adoption-wave-2026-09-09-adopters", "title": "Named adopters: Osprey at Discord and Bluesky", "description": "9 September 2026. Bounded primary-source adoption inquiry closed after six discovery queries and four original HTML acquisitions. Read adopters-captures.json for custody and exact scopes. No new organization branch, harmful dataset, account, model inference, code execution, deployment inspection or case edit.", "html": "9 September 2026. Bounded primary-source adoption inquiry closed after six discovery queries and four original HTML acquisitions. Read adopters-captures.json for custody and exact scopes. No new organization branch, harmful dataset, account, model inference, code execution, deployment inspection or case edit.
There are substantive, named reported production deployments, beyond partner logos. Discord provides its own dated technical account and describes a reciprocal development path. Bluesky's account is a ROOST-published case study quoting participating staff and identifying the platform's action machinery. Neither is an independent operational/outcome audit. Both locate important choices with the deploying platform rather than establishing a common policy imposed by the tool provider.
\nDiscord Engineering's [19 February 2026 technical article](http[local research file] credited to Jared Miller and Ayu, reports that as of December 2025, Discord used open-source Osprey for approximately 400 million actions/day, across 204 action types and 2,288 rules, with 99 custom user-defined functions. These are operator-reported processing quantities, not counts of abusive users or successful takedowns.
\nIts System Components section says Discord distributes rules through ETCD and routes outputs to configurable sinks. Handling Sensitive Data describes Discord's own access controls, justification requirements and audit trails. Customizing Osprey for the Public expressly says Discord's full rules, functions and sinks were not supplied; internal configurations were placed behind plugin integrations. This documents a configurable engine and retained local configuration. The shared source does not expose every internal detection rule or imply that ROOST receives Discord's user data. Exact deployed commit, rule bundle and measured error rates remain unacquired. Source: adopters-discord-engineering.html; complete substantive body read, figures not inspected and examples not executed.
Clint Smith's [16 March 2026 account](http[local research file] says Osprey began as Discord's production engine, was donated and improved through ROOST, and the community version was reintegrated into Discord production. His author biography identifies him as Discord's chief legal officer; the article also identifies his ROOST board-chair role. This dual position matters to provenance: it is participant testimony from someone involved in both bodies. It supports a reported reciprocal resource flow, not independently proved performance. The article refers to the public v1 release but does not pin the exact Discord production revision. Its broad ecosystem user count and unrelated initiative references do not establish common policies, joined databases or additional named deployments. Source: adopters-discord2026.html, What Osprey Does and Results & Impact.
ROOST's [4 February 2026 case study](http[local research file] by Andrew Chang, reports Osprey in Bluesky production, quoting its trust-and-safety engineering staff and department head by role. It reports over 45 million events and 100,000 enforcement actions daily. Those figures are attributed operating claims; actions are not necessarily unique accounts, bans or proven violations.
\nThe described path is atproto event stream → Osprey with Bluesky-specific functions → Kafka output sink → effects receiver on Bluesky servers → action in Ozone. The account says analysts migrated and developed rules; its named example labels suspected t-shirt spammers. This identifies the platform's rule-writing and action integration, not ROOST approval of each decision. Setup still involved engineering, despite reduced dependence on engineers for subsequent rule changes.
\nIts cost savings are estimates, and the publication date is not an exact go-live date. The case links v1.0 information without proving Bluesky's deployed tag. This remains supplier-published participant evidence, not an independently observed endpoint. Source: adopters-bluesky-case.html, full substantive text read; diagrams and code images not visually verified.
The [21 July 2025 official release](http[local research file] described Bluesky as planning adoption, with a quotation from Aaron Rodericks, then named as its trust-and-safety head. It also described public availability as forthcoming and Osprey as a Discord donation. The later accounts support progression from intention to reported production; July alone does not.
\nThat release credits philanthropic funding and pro bono Perkins Coie legal services with the acquisition of Cove technology and tool launch. It does not allocate those resources to Bluesky or establish a recipient support contract. No paid support terms, service-level guarantees, adoption grant, update obligation or agreement requiring a particular policy was acquired for either selected platform. The public accounts describe collaboration and community assistance; contractual rights remain a separate question. Source: adopters-launch2025.html, Bringing Professional Tools, Early Adopters and What's Next.
The supportable paths are Discord's reported production use / contribution / reintegration, and a ROOST-published account of Bluesky production use, with platform-configured rules and a local action integration. Where the case already holds the organizations and Osprey, reuse them. Do not create an enforcement-authority edge from a logo, model-provider affiliation or working-group attendance.
\nNo acquired adopter account establishes use of the CCF TVEC prompt pack, mandatory ROOST policy, a Canadian-funded deployment, Lantern signal access or a specific law-enforcement referral. Osprey adoption is not evidence of adopting every ROOST-supported model or sample policy. The most useful next record would pair a named deployment's exact configuration/policy revision with its approval process and a documented action or correction. That question remains for the next forest assessment; this lane has not pursued it.
\n| Original | \nSHA256 | \n
|---|---|
| Discord technical account | \n0f2a56a451732d6b4757001b422adbc4fa9bc20d4377af27b78cedbd1420507b | \n
| Discord reciprocal-use account | \n56344330354ef4d4cdbd720aef5824e93e9f06a02b2efdbd5bcc81650a7a7d64 | \n
| ROOST Bluesky case study | \n973be08b940baafe8e865583f9ed2cd716e0fb4839d7bc61f505c4372fec1bd6 | \n
| July 2025 launch release | \n203ad1a26a54aaedca7dc58e30f0d0af0ac7ebd4ed3156e2093cfdd67a2f70bf | \n
All four ordinary unauthenticated GETs returned substantive 200 HTML and were hash-checked. Full substantive article bodies were read. Related articles, linked models, external vendor offerings and screenshots were not treated as inspected sources. Raw HTML and derivative text remain separate; Discord's scoped rich-text body extracts aid reproduction of the read.
\nThe six queries covered Bluesky's own site/docs, Discord's Osprey/ROOST accounts, Matrix's Osprey mention, and the named Bluesky case study. Matrix annual-report search excerpts described prospective 2026 use; they were not expanded into a third case or used to contradict a later current-use claim. No Bluesky-owned deployment report surfaced in these scoped queries; that is a retrieval limitation, not proof none exists. Public social reposts and irrelevant similarly named results were not treated as independent corroboration. No failed routes were retried. Acquisition stops here.
", "path": "roost-adoption-wave-2026-09-09/adopters.md", "sourceIds": [ "w27-roost-s-discord-engineering", "w27-roost-s-discord-roost2026", "w27-roost-s-bluesky-case", "w27-roost-s-osprey-launch2025" ], "markdown": "9 September 2026. Bounded primary-source adoption inquiry closed after six discovery queries and four original HTML acquisitions. Read `adopters-captures.json` for custody and exact scopes. No new organization branch, harmful dataset, account, model inference, code execution, deployment inspection or case edit.\n\n## Finding\n\nThere are substantive, named **reported production deployments**, beyond partner logos. Discord provides its own dated technical account and describes a reciprocal development path. Bluesky's account is a ROOST-published case study quoting participating staff and identifying the platform's action machinery. Neither is an independent operational/outcome audit. Both locate important choices with the deploying platform rather than establishing a common policy imposed by the tool provider.\n\n## Discord: shared engine, retained internal rules\n\nDiscord Engineering's 19 February 2026 technical article (see source references) credited to Jared Miller and Ayu, reports that **as of December 2025**, Discord used open-source Osprey for approximately 400 million actions/day, across 204 action types and 2,288 rules, with 99 custom user-defined functions. These are operator-reported processing quantities, not counts of abusive users or successful takedowns.\n\nIts System Components section says Discord distributes rules through ETCD and routes outputs to configurable sinks. Handling Sensitive Data describes Discord's own access controls, justification requirements and audit trails. Customizing Osprey for the Public expressly says Discord's full rules, functions and sinks were not supplied; internal configurations were placed behind plugin integrations. This documents a configurable engine and retained local configuration. The shared source does not expose every internal detection rule or imply that ROOST receives Discord's user data. Exact deployed commit, rule bundle and measured error rates remain unacquired. Source: `adopters-discord-engineering.html`; complete substantive body read, figures not inspected and examples not executed.\n\nClint Smith's 16 March 2026 account (see source references) says Osprey began as Discord's production engine, was donated and improved through ROOST, and the community version was reintegrated into Discord production. His author biography identifies him as Discord's chief legal officer; the article also identifies his ROOST board-chair role. This dual position matters to provenance: it is participant testimony from someone involved in both bodies. It supports a reported reciprocal resource flow, not independently proved performance. The article refers to the public v1 release but does not pin the exact Discord production revision. Its broad ecosystem user count and unrelated initiative references do not establish common policies, joined databases or additional named deployments. Source: `adopters-discord2026.html`, What Osprey Does and Results & Impact.\n\n## Bluesky: implementation reaches a platform-owned action receiver\n\nROOST's 4 February 2026 case study (see source references) by Andrew Chang, reports Osprey in Bluesky production, quoting its trust-and-safety engineering staff and department head by role. It reports over 45 million events and 100,000 enforcement actions daily. Those figures are attributed operating claims; actions are not necessarily unique accounts, bans or proven violations.\n\nThe described path is **atproto event stream → Osprey with Bluesky-specific functions → Kafka output sink → effects receiver on Bluesky servers → action in Ozone**. The account says analysts migrated and developed rules; its named example labels suspected t-shirt spammers. This identifies the platform's rule-writing and action integration, not ROOST approval of each decision. Setup still involved engineering, despite reduced dependence on engineers for subsequent rule changes.\n\nIts cost savings are estimates, and the publication date is not an exact go-live date. The case links v1.0 information without proving Bluesky's deployed tag. This remains supplier-published participant evidence, not an independently observed endpoint. Source: `adopters-bluesky-case.html`, full substantive text read; diagrams and code images not visually verified.\n\n## Earlier state and support conditions\n\nThe 21 July 2025 official release (see source references) described Bluesky as **planning** adoption, with a quotation from Aaron Rodericks, then named as its trust-and-safety head. It also described public availability as forthcoming and Osprey as a Discord donation. The later accounts support progression from intention to reported production; July alone does not.\n\nThat release credits philanthropic funding and pro bono Perkins Coie legal services with the acquisition of Cove technology and tool launch. It does not allocate those resources to Bluesky or establish a recipient support contract. No paid support terms, service-level guarantees, adoption grant, update obligation or agreement requiring a particular policy was acquired for either selected platform. The public accounts describe collaboration and community assistance; contractual rights remain a separate question. Source: `adopters-launch2025.html`, Bringing Professional Tools, Early Adopters and What's Next.\n\n## What enters the map\n\nThe supportable paths are Discord's **reported production use / contribution / reintegration**, and a **ROOST-published account of Bluesky production use**, with **platform-configured rules and a local action integration**. Where the case already holds the organizations and Osprey, reuse them. Do not create an enforcement-authority edge from a logo, model-provider affiliation or working-group attendance.\n\nNo acquired adopter account establishes use of the CCF TVEC prompt pack, mandatory ROOST policy, a Canadian-funded deployment, Lantern signal access or a specific law-enforcement referral. Osprey adoption is not evidence of adopting every ROOST-supported model or sample policy. The most useful next record would pair a named deployment's exact configuration/policy revision with its approval process and a documented action or correction. That question remains for the next forest assessment; this lane has not pursued it.\n\n## Source custody and finite frontier\n\n| Original | SHA256 |\n| --- | --- |\n| Discord technical account | `0f2a56a451732d6b4757001b422adbc4fa9bc20d4377af27b78cedbd1420507b` |\n| Discord reciprocal-use account | `56344330354ef4d4cdbd720aef5824e93e9f06a02b2efdbd5bcc81650a7a7d64` |\n| ROOST Bluesky case study | `973be08b940baafe8e865583f9ed2cd716e0fb4839d7bc61f505c4372fec1bd6` |\n| July 2025 launch release | `203ad1a26a54aaedca7dc58e30f0d0af0ac7ebd4ed3156e2093cfdd67a2f70bf` |\n\nAll four ordinary unauthenticated GETs returned substantive 200 HTML and were hash-checked. Full substantive article bodies were read. Related articles, linked models, external vendor offerings and screenshots were not treated as inspected sources. Raw HTML and derivative text remain separate; Discord's scoped rich-text body extracts aid reproduction of the read.\n\nThe six queries covered Bluesky's own site/docs, Discord's Osprey/ROOST accounts, Matrix's Osprey mention, and the named Bluesky case study. Matrix annual-report search excerpts described prospective 2026 use; they were not expanded into a third case or used to contradict a later current-use claim. No Bluesky-owned deployment report surfaced in these scoped queries; that is a retrieval limitation, not proof none exists. Public social reposts and irrelevant similarly named results were not treated as independent corroboration. No failed routes were retried. Acquisition stops here.\n", "file": "research/roost-adoption-wave-2026-09-09-adopters.md" }, { "id": "roost-adoption-wave-2026-09-09-implementation", "title": "From model guidance to local rules: the Phoebe boundary", "description": "9 September2026. Bounded acquisition closed after one RMC guide and its single named Phoebe implementation. The public record establishes instructional architecture and static mechanisms for drafting rules. It does not establish a named deployment using CCF's policy, a production moderation decision, or a Canadian-funded delivery. No software installation, m", "html": "9 September2026. Bounded acquisition closed after one RMC guide and its single named Phoebe implementation. The public record establishes instructional architecture and static mechanisms for drafting rules. It does not establish a named deployment using CCF's policy, a production moderation decision, or a Canadian-funded delivery. No software installation, model inference, endpoint call, harmful-content dataset, account or submission occurred.
\nThe [Using Open Safety Models guide](http[local research file] separates detection, investigation, review and enforcement. It describes models classifying against an existing policy, supporting investigation/review and improving policies. Its suggested two-model arrangement routes uncertain cases to a stronger model that recommends an action for human review; review data should then inform policy revision. This is a recommended design, not a compulsory architecture or measured deployment. Locators: retained implementation-rmc-guide.md lines7–39.
The guide places model, hosting and policy choices with the adopting team. It advises choosing by performance, latency, cost and licence, testing against a suitable dataset, choosing relevant metrics, and distinguishing model errors from inconsistent policy. It allows open/closed hybrids. Its policy section says ROOST does not write policies and points to outside-contributed examples. That institutional description is consistent with hosting a CCF contribution; it does not prove that every ROOST tool uses that contribution. Locators: lines45–95. This source is pinned to Model Community revision 22b3bd48562616f9379eb6fa63abe131346a6de6; the retained earlier README supplies the exact guide link. No exact guide-publication date is inferred from the repository revision.
The guide's investigation example links [Phoebe](http[local research file] describing an agent that analyzes threats and suggests Osprey rules. I followed this one implementation only. Public HEAD returned dfb2bf206fc6cf757b5f10fb4c318ce9a3cce062, committer date7February2026. All findings below concern that pinned source, not every current or privately modified Phoebe deployment.
The [README](http[local research file] lines3–20, describes Osprey rule writing, analytics and Ozone moderation. It claims writing, validation and deployment in its feature summary but expressly marks moderation as unimplemented. Its sample model configuration is Claude Sonnet4.5 via Anthropic, with alternate API backends configurable. It is therefore not evidence of gpt-oss-safeguard, the CCF pack or any particular open model operating on Bluesky. Mentioning the ATProtocol network in the README and prompt is not an adopter account.
\nThe concrete code draws a narrower boundary than the feature summary. [Osprey.save_rule](http[local research file] writes an .sml file under the local ./data/ruleset tree after checking its path/extension. If creating a non-index file and a parent index already exists, it appends a Require entry, optionally conditioned by the supplied require_if. It can overwrite an existing local rule. This creates or changes a candidate rule and its local inclusion, not a recorded remote deployment. Lines165–208.
[validate_rules](http[local research file] runs osprey-cli push-rules ../ruleset --dry-run; the code expressly says the flag validates without pushing. The tool wrapper returns success/error from that validator. Validation thus cannot be represented as production activation or content-policy accuracy testing. No remote push/commit, reviewer approval or active-engine acknowledgement is established by these functions. An external deployment arrangement could supply a later step, but none was inspected. Osprey source lines210–233 and tool definitions lines136–149.
The operator configures the model/API, Osprey endpoint and source/ruleset repositories through settings; the inspected defaults include a particular author's ATProtocol ruleset repository, while README examples use an operator's repository. I did not acquire or inspect that separate ruleset. The Osprey adapter requests labels, features, known actions and UDF documentation from the configured service. These determine the available vocabulary/effects that the agent sees. No universal CCF taxonomy or hardwired CCF pack is established. [Config](http[local research file] lines26–57; Osprey adapter lines59–73; tool definitions lines1–59.
\nThe [source prompt](http[local research file] directs the model to understand target behavior, use available effects, write rules and validate them. It includes examples of account labels and reports for moderation review, and a requirement to validate every new rule. Those are source instructions/examples, not proof an effect ran. The save function itself neither validates nor awaits a human approval; validation is a separate tool plus a prompt instruction. This is a scoped observation about these functions, not an audit of every possible deployment's permission system. Prompt lines31–38,172–209; tool definitions lines103–149.
\nThe MIT licence permits modification and redistribution subject to retaining its notices and disclaimer. It provides software freedom, not access to an Osprey installation, a platform's content or separately licensed models. The source files can consequently be changed by a downstream operator, while the actual chosen version/configuration remains an evidentiary question. Entire21-line LICENSE read. No named actor's production permission or exercise of a right is inferred.
\nBoth direct Ozone methods, [apply_label and remove_label](http[local research file] raise NotImplementedError; the client constructor is also a TODO. The README explicitly marks these as unfinished. Prospective allowed-label/moderator settings and comments in config do not establish enforcement of that permission scheme. The absence of a working direct removal method is not proof that all Osprey deployments lack reversal; they are distinct mechanisms.
The initialization routine clones configured repositories only if their local directories are absent, otherwise logs that they already exist and does not fetch them in that routine. This does not establish any comprehensive update policy. No mandatory corrected-policy notification, recipient synchronization acknowledgement, review reopening or restored-item trace was acquired. Local rule updates and the guide's recommendation to iterate policies are weaker claims than those actual duties/outcomes. Osprey initialization lines32–57 and source-fetch methods75–115.
\nThe relevant chain stops here: operator-selected context and instructions → model-generated candidate rule → local write/inclusion → separate dry-run validation. Activation in the live rules engine, the runtime effect handler, any moderation decision and any later correction require their own records. The generic README term deploy cannot supply the missing steps.
\nThis implementation narrows the adoption claim rather than closing it. ROOST's guide promotes an example that can turn analysis into editable local rules. The selected example's own source exposes unfinished direct moderation and a non-pushing validator. It supports a route for assistance and transferable rules, not a proved uniform restriction system or a government-enforced CCF policy.
\nThe useful next record is a named operator's exact model/prompt/ruleset revision plus the approved activation mechanism and effect-handler configuration, joined to one action and reconsideration outcome. A later Phoebe release or a separate deploying system might resolve the missing step; this pass did not follow either. Root's Pathways and adopter accounts remain independent evidence lanes. No automatic Phoebe/Osprey/Coop→CCF/Elevate/contribution26007 relation is warranted.
\nimplementation-captures.json records10 successful original/API responses plus one reader-failure observation. Original bytes and hashes are retained. Complete reads: guide105lines; Phoebe README133; prompt254; Osprey adapter233; Ozone adapter18; Osprey tool definitions149; config60; MIT licence21. Commit metadata supplies the source pin; repository-tree metadata was used only to choose these files. The remainder of the repository, agent loop, executor/sandbox and external runtime effect handlers were not audited. README claims about sandbox security/performance were not independently tested or adopted as findings.
The pinned guide web reader failed with a cache miss; one ordinary GET succeeded. Subsequent sources used normal public GET. No installation or repository execution was performed, and installation commands in the source remained unevaluated text. No additional implementation link, dataset, model weights or CCF policy test was opened. The prior case, current-work, atlas and ZIP were not changed.
", "path": "roost-adoption-wave-2026-09-09/implementation.md", "sourceIds": [ "w27-roost-s-rmc-guide", "w27-roost-s-phoebe-pin", "w27-roost-s-phoebe-readme", "w27-roost-s-phoebe-prompt", "w27-roost-s-phoebe-osprey", "w27-roost-s-phoebe-ozone", "w27-roost-s-phoebe-tools", "w27-roost-s-phoebe-config", "w27-roost-s-phoebe-mit" ], "markdown": "9 September2026. Bounded acquisition closed after one RMC guide and its single named Phoebe implementation. The public record establishes instructional architecture and static mechanisms for drafting rules. It does not establish a named deployment using CCF's policy, a production moderation decision, or a Canadian-funded delivery. No software installation, model inference, endpoint call, harmful-content dataset, account or submission occurred.\n\n## What the RMC guide actually recommends\n\nThe Using Open Safety Models guide (see source references) separates detection, investigation, review and enforcement. It describes models classifying against an existing policy, supporting investigation/review and improving policies. Its suggested two-model arrangement routes uncertain cases to a stronger model that recommends an action for human review; review data should then inform policy revision. This is a recommended design, not a compulsory architecture or measured deployment. Locators: retained `implementation-rmc-guide.md` lines7–39.\n\nThe guide places model, hosting and policy choices with the adopting team. It advises choosing by performance, latency, cost and licence, testing against a suitable dataset, choosing relevant metrics, and distinguishing model errors from inconsistent policy. It allows open/closed hybrids. Its policy section says ROOST does not write policies and points to outside-contributed examples. That institutional description is consistent with hosting a CCF contribution; it does not prove that every ROOST tool uses that contribution. Locators: lines45–95. This source is pinned to Model Community revision `22b3bd48562616f9379eb6fa63abe131346a6de6`; the retained earlier README supplies the exact guide link. No exact guide-publication date is inferred from the repository revision.\n\n## One linked implementation, inspected rather than presumed deployed\n\nThe guide's investigation example links Phoebe (see source references) describing an agent that analyzes threats and suggests Osprey rules. I followed this one implementation only. Public HEAD returned `dfb2bf206fc6cf757b5f10fb4c318ce9a3cce062`, committer date7February2026. All findings below concern that pinned source, not every current or privately modified Phoebe deployment.\n\nThe README (see source references) lines3–20, describes Osprey rule writing, analytics and Ozone moderation. It claims writing, validation and deployment in its feature summary but expressly marks moderation as unimplemented. Its sample model configuration is Claude Sonnet4.5 via Anthropic, with alternate API backends configurable. It is therefore not evidence of gpt-oss-safeguard, the CCF pack or any particular open model operating on Bluesky. Mentioning the ATProtocol network in the README and prompt is not an adopter account.\n\nThe concrete code draws a narrower boundary than the feature summary. Osprey.save_rule (see source references) writes an `.sml` file under the local `./data/ruleset` tree after checking its path/extension. If creating a non-index file and a parent index already exists, it appends a Require entry, optionally conditioned by the supplied `require_if`. It can overwrite an existing local rule. This creates or changes a candidate rule and its local inclusion, not a recorded remote deployment. Lines165–208.\n\nvalidate_rules (see source references) runs `osprey-cli push-rules ../ruleset --dry-run`; the code expressly says the flag validates without pushing. The tool wrapper returns success/error from that validator. Validation thus cannot be represented as production activation or content-policy accuracy testing. No remote push/commit, reviewer approval or active-engine acknowledgement is established by these functions. An external deployment arrangement could supply a later step, but none was inspected. Osprey source lines210–233 and tool definitions lines136–149.\n\n## Who can change what\n\nThe operator configures the model/API, Osprey endpoint and source/ruleset repositories through settings; the inspected defaults include a particular author's ATProtocol ruleset repository, while README examples use an operator's repository. I did not acquire or inspect that separate ruleset. The Osprey adapter requests labels, features, known actions and UDF documentation from the configured service. These determine the available vocabulary/effects that the agent sees. No universal CCF taxonomy or hardwired CCF pack is established. Config (see source references) lines26–57; Osprey adapter lines59–73; tool definitions lines1–59.\n\nThe source prompt (see source references) directs the model to understand target behavior, use available effects, write rules and validate them. It includes examples of account labels and reports for moderation review, and a requirement to validate every new rule. Those are source instructions/examples, not proof an effect ran. The save function itself neither validates nor awaits a human approval; validation is a separate tool plus a prompt instruction. This is a scoped observation about these functions, not an audit of every possible deployment's permission system. Prompt lines31–38,172–209; tool definitions lines103–149.\n\nThe MIT licence permits modification and redistribution subject to retaining its notices and disclaimer. It provides software freedom, not access to an Osprey installation, a platform's content or separately licensed models. The source files can consequently be changed by a downstream operator, while the actual chosen version/configuration remains an evidentiary question. Entire21-line LICENSE read. No named actor's production permission or exercise of a right is inferred.\n\n## Action and correction are not finished interfaces here\n\nBoth direct Ozone methods, apply_label and remove_label (see source references) raise `NotImplementedError`; the client constructor is also a TODO. The README explicitly marks these as unfinished. Prospective allowed-label/moderator settings and comments in config do not establish enforcement of that permission scheme. The absence of a working direct removal method is not proof that all Osprey deployments lack reversal; they are distinct mechanisms.\n\nThe initialization routine clones configured repositories only if their local directories are absent, otherwise logs that they already exist and does not fetch them in that routine. This does not establish any comprehensive update policy. No mandatory corrected-policy notification, recipient synchronization acknowledgement, review reopening or restored-item trace was acquired. Local rule updates and the guide's recommendation to iterate policies are weaker claims than those actual duties/outcomes. Osprey initialization lines32–57 and source-fetch methods75–115.\n\nThe relevant chain stops here: operator-selected context and instructions → model-generated candidate rule → local write/inclusion → separate dry-run validation. Activation in the live rules engine, the runtime effect handler, any moderation decision and any later correction require their own records. The generic README term deploy cannot supply the missing steps.\n\n## Forest implication and next discriminator\n\nThis implementation narrows the adoption claim rather than closing it. ROOST's guide promotes an example that can turn analysis into editable local rules. The selected example's own source exposes unfinished direct moderation and a non-pushing validator. It supports a route for assistance and transferable rules, not a proved uniform restriction system or a government-enforced CCF policy.\n\nThe useful next record is a named operator's exact model/prompt/ruleset revision plus the approved activation mechanism and effect-handler configuration, joined to one action and reconsideration outcome. A later Phoebe release or a separate deploying system might resolve the missing step; this pass did not follow either. Root's Pathways and adopter accounts remain independent evidence lanes. No automatic Phoebe/Osprey/Coop→CCF/Elevate/contribution26007 relation is warranted.\n\n## Custody and limits\n\n`implementation-captures.json` records10 successful original/API responses plus one reader-failure observation. Original bytes and hashes are retained. Complete reads: guide105lines; Phoebe README133; prompt254; Osprey adapter233; Ozone adapter18; Osprey tool definitions149; config60; MIT licence21. Commit metadata supplies the source pin; repository-tree metadata was used only to choose these files. The remainder of the repository, agent loop, executor/sandbox and external runtime effect handlers were not audited. README claims about sandbox security/performance were not independently tested or adopted as findings.\n\nThe pinned guide web reader failed with a cache miss; one ordinary GET succeeded. Subsequent sources used normal public GET. No installation or repository execution was performed, and installation commands in the source remained unevaluated text. No additional implementation link, dataset, model weights or CCF policy test was opened. The prior case, current-work, atlas and ZIP were not changed.\n", "file": "research/roost-adoption-wave-2026-09-09-implementation.md" }, { "id": "roost-adoption-wave-2026-09-09-pathways-support", "title": "Pathways: access conditions, advice and a separate consulting service", "description": "9 September2026. Root acquisition closed. This is the Tech Coalition capacity/support lane of the ROOST adoption run. It establishes a concrete support programme and published obligations, not a named ROOST deployment or a compulsory classification standard.", "html": "9 September2026. Root acquisition closed. This is the Tech Coalition capacity/support lane of the ROOST adoption run. It establishes a concrete support programme and published obligations, not a named ROOST deployment or a compulsory classification standard.
\nROOST's previously retained10February2025 launch letter credits Tech Coalition expertise and quotes Sean Litton supporting wider tool adoption through capacity building. This is collaboration, not proof Pathways supplied ROOST code or licensed a particular model. Source and full reading scope remain in ../roost-programme-wave-2026-09-09/governance-funding.md and governance-gratitude2025.html.
\nThe Coalition's [18July2024 launch](http[local research file] described policy-writing and reporting guidance, technology-licensing opportunities and prospective membership development. Giphy/Rec Room testimonials endorse the programme; they do not demonstrate tool installation. Current eligibility wording should not be silently backdated to this announcement.
\nThe [18December2025 year review](http[local research file] reports67 participating companies that year, five new resources, two fireside chats and an initial Elevate pilot preparing firms for possible membership. These are the operator's activity counts, not independently verified outcomes or a ROOST user count. Root read the complete substantive body; other programme numbers are not new findings here.
\nThe [current Pathways page](http[local research file] offers no-cost guidance and peer support to eligible company employees handling user content/communications. It excludes current member-company staff and vendors without that function; applicants are reviewed. Membership is neither required nor guaranteed. PhotoDNA is an opportunity to apply for a free sublicense through the Coalition, not automatic access, hash-database entitlement or installation. Topics include policy drafting, reporting and generative-AI safeguards. The intake explicitly requires agreement to terms including confidentiality. No intake was submitted, restricted resource hub entered or license application made. Publication/revision date is unverified; this is the9September2026 page.
\nThe [published non-member terms](http[local research file] clauses1–9, make continued use acceptance of discretionary updates, require accurate registration and reasonable efforts to answer information requests within three months, and grant a revocable, nontransferable site licence. They restrict copying to specified internal uses. Mentoring information carries confidentiality obligations and limited internal sharing. Crucially, clause7 leaves whether/how to use the materials to the user's discretion, calls use voluntary and disclaims a legal/community standard. This demonstrates access/IP/confidentiality conditions alongside retained implementation choice, not an exercised direction to remove content. No executed participant agreement or enforcement event was acquired. Root read all nine clauses through the primary reader; no legal enforceability opinion is offered.
\nThe [Tech Coalition Elevate page](http[local research file] offers tailored consulting, risk/maturity assessment, implementation support and mentorship to members, Pathways participants or other companies. Membership is unnecessary. Pricing depends on scope/expertise/support and is set in a tailored proposal; there is no published fixed fee or acquired payment. It can complement free Pathways or membership services. The actual proposal/deliverables/approval rights remain unknown. Its form links a different /terms/ page, which was not acquired: Pathways's detailed clauses must not be treated as the consulting contract.
\nThis is not Christchurch Call Foundation's Elevate. Different hosts and stated programmes establish distinct identities despite the shared name. The Canadian CCF contribution must not finance the Coalition's consulting service by name matching.
\n[Stability AI's announcement](http[local research file] published11February2026 according to retained page metadata, reports Pathways participation in2025 and subsequent Coalition membership. It names guidance/peer learning, not ROOST adoption or Lantern access. Full substantive announcement read; latest modification metadata is2March2026.
\nAn earlier [company email published by the Brennan Center](http[local research file] PDF64–66, is attributed to Ana Guillen and dated21January2025. Page65 instead says Pathways was joined in July2024. The date accounts conflict; exact enrolment is unresolved. This is an attributed company response in a third-party compilation, not independent corroboration of safety performance. Root read those three pages, not all72 or linked safeguards. The other initiatives on those pages were not followed into new branches.
\nAdvice can travel through a selective, confidential support relationship and lead into deeper consulting or membership without transferring final platform decision rights. That structure is more specific than calling all cooperation open or all support coercive. Here it provides an adoption pathway and conditions on resources; no named ROOST configuration, content order, donor veto or actual PhotoDNA sublicense follows. A worked recipient support agreement and its implementation record would distinguish normative influence from exercised authority.
\npathways-captures.json holds seven successful public originals: six HTML pages and one72-page PDF, with exact URLs, UTC capture times and byte hashes. Root read the substantive HTML bodies as scoped there, all nine Pathways clauses and PDF64–66. Derived plain text is labelled separately. The initial fitz import was unavailable; one pypdf substitute succeeded. The Elevate primary reader errored; one ordinary GET succeeded. Closed CCF/GitHub/registry routes were not retried. Four initial discovery queries and one focused eligibility query identified the records; unrelated search results were not evidence. No forms, accounts, outreach, model inference or payment occurred. Acquisition stopped for the combined forest.
", "path": "roost-adoption-wave-2026-09-09/pathways-support.md", "sourceIds": [ "w27-roost-s-pathways-current", "w27-roost-s-pathways-terms", "w27-roost-s-tc-elevate", "w27-roost-s-pathways-launch2024", "w27-roost-s-pathways-year2025", "w27-roost-s-stability-pathways2026", "w27-roost-s-stability-earlier-email" ], "markdown": "9 September2026. Root acquisition closed. This is the Tech Coalition capacity/support lane of the ROOST adoption run. It establishes a concrete support programme and published obligations, not a named ROOST deployment or a compulsory classification standard.\n\n## The connection and chronology\n\nROOST's previously retained10February2025 launch letter credits Tech Coalition expertise and quotes Sean Litton supporting wider tool adoption through capacity building. This is collaboration, not proof Pathways supplied ROOST code or licensed a particular model. Source and full reading scope remain in ../roost-programme-wave-2026-09-09/governance-funding.md and governance-gratitude2025.html.\n\nThe Coalition's 18July2024 launch (see source references) described policy-writing and reporting guidance, technology-licensing opportunities and prospective membership development. Giphy/Rec Room testimonials endorse the programme; they do not demonstrate tool installation. Current eligibility wording should not be silently backdated to this announcement.\n\nThe 18December2025 year review (see source references) reports67 participating companies that year, five new resources, two fireside chats and an initial Elevate pilot preparing firms for possible membership. These are the operator's activity counts, not independently verified outcomes or a ROOST user count. Root read the complete substantive body; other programme numbers are not new findings here.\n\n## Current access and obligation\n\nThe current Pathways page (see source references) offers no-cost guidance and peer support to eligible company employees handling user content/communications. It excludes current member-company staff and vendors without that function; applicants are reviewed. Membership is neither required nor guaranteed. PhotoDNA is an opportunity to apply for a free sublicense through the Coalition, not automatic access, hash-database entitlement or installation. Topics include policy drafting, reporting and generative-AI safeguards. The intake explicitly requires agreement to terms including confidentiality. No intake was submitted, restricted resource hub entered or license application made. Publication/revision date is unverified; this is the9September2026 page.\n\nThe published non-member terms (see source references) clauses1–9, make continued use acceptance of discretionary updates, require accurate registration and reasonable efforts to answer information requests within three months, and grant a revocable, nontransferable site licence. They restrict copying to specified internal uses. Mentoring information carries confidentiality obligations and limited internal sharing. Crucially, clause7 leaves whether/how to use the materials to the user's discretion, calls use voluntary and disclaims a legal/community standard. This demonstrates access/IP/confidentiality conditions alongside retained implementation choice, not an exercised direction to remove content. No executed participant agreement or enforcement event was acquired. Root read all nine clauses through the primary reader; no legal enforceability opinion is offered.\n\n## Paid depth of support, separate identity\n\nThe Tech Coalition Elevate page (see source references) offers tailored consulting, risk/maturity assessment, implementation support and mentorship to members, Pathways participants or other companies. Membership is unnecessary. Pricing depends on scope/expertise/support and is set in a tailored proposal; there is no published fixed fee or acquired payment. It can complement free Pathways or membership services. The actual proposal/deliverables/approval rights remain unknown. Its form links a different /terms/ page, which was not acquired: Pathways's detailed clauses must not be treated as the consulting contract.\n\n**This is not Christchurch Call Foundation's Elevate.** Different hosts and stated programmes establish distinct identities despite the shared name. The Canadian CCF contribution must not finance the Coalition's consulting service by name matching.\n\n## A named participant and a real date discrepancy\n\nStability AI's announcement (see source references) published11February2026 according to retained page metadata, reports Pathways participation in2025 and subsequent Coalition membership. It names guidance/peer learning, not ROOST adoption or Lantern access. Full substantive announcement read; latest modification metadata is2March2026.\n\nAn earlier company email published by the Brennan Center (see source references) PDF64–66, is attributed to Ana Guillen and dated21January2025. Page65 instead says Pathways was joined in July2024. The date accounts conflict; exact enrolment is unresolved. This is an attributed company response in a third-party compilation, not independent corroboration of safety performance. Root read those three pages, not all72 or linked safeguards. The other initiatives on those pages were not followed into new branches.\n\n## What this changes and what it leaves open\n\nAdvice can travel through a selective, confidential support relationship and lead into deeper consulting or membership without transferring final platform decision rights. That structure is more specific than calling all cooperation open or all support coercive. Here it provides an adoption pathway and conditions on resources; no named ROOST configuration, content order, donor veto or actual PhotoDNA sublicense follows. A worked recipient support agreement and its implementation record would distinguish normative influence from exercised authority.\n\n## Custody and frontier\n\npathways-captures.json holds seven successful public originals: six HTML pages and one72-page PDF, with exact URLs, UTC capture times and byte hashes. Root read the substantive HTML bodies as scoped there, all nine Pathways clauses and PDF64–66. Derived plain text is labelled separately. The initial fitz import was unavailable; one pypdf substitute succeeded. The Elevate primary reader errored; one ordinary GET succeeded. Closed CCF/GitHub/registry routes were not retried. Four initial discovery queries and one focused eligibility query identified the records; unrelated search results were not evidence. No forms, accounts, outreach, model inference or payment occurred. Acquisition stopped for the combined forest.\n", "file": "research/roost-adoption-wave-2026-09-09-pathways-support.md" }, { "id": "roost-programme-wave-2026-09-09-ccf-programme", "title": "CCF programme and delivery lane — Elevate", "description": "9 September 2026. Root acquisition closed for this lane; ROOST governance and tools lanes continue. This is the recipient-side programme crosswalk, not a finding of completed software delivery.", "html": "9 September 2026. Root acquisition closed for this lane; ROOST governance and tools lanes continue. This is the recipient-side programme crosswalk, not a finding of completed software delivery.
\n[CCF Canadian-investment announcement](http[local research file] full substantive body/Notes read through primary reader, lines14–43. It names Elevate, the Community Resilience Fund support and ROOST collaboration, with an attributed statement by ROOST president Camille Francois. Its rounded amount matches the separately retained Canadian contribution26007; it is not a second award. It describes a three-year programme spanning interventions, open-source tools, benchmarks and capacity. Earlier leader priorities in2022/23 are described, so this announcement alone cannot show the2025 donor originated the policy. The claimed Canadian role in shaping international technology governance is the recipient's characterization, not a specified contractual right. No payment receipt, subcontract amount, acceptance record or named deployed product appears. Original GET403 retained as error; no retry.
\n[Charting a Collective Response](http[local research file] full substantive indexed-primary text read, direct reader404 and GET403 closed. It describes initial needs assessment/experimentation, later continuing tool development and parallel support for smaller platforms with ROOST. This is programme sequencing and intent, not an implementation report.
\n[Current Elevate programme page](http[local research file] indexed substantive body read, undated snapshot. It ties the programme to Canada's CRF, ROOST and the community, and describes work on AI standards/benchmarks and a toolkit for smaller providers. The page advertises a GitHub link, but the available index supplies no target and its GET403/reader-cache-miss routes are closed. No repository identity or deployed software crosswalk is established by that link label. The history-page indexed Elevate entry says12November2025 launch; other historical claims there were not acquired for use.
\n[ROOST-hosted CCF partnership announcement](http[local research file] original retained and full main body/editor notes read. It identifies itself as a permitted reprint authored by the Christchurch Call team, displayed10February2025; CCF index cards display12February. These publication labels are not proof of two partnerships or executed contract dates. It sets out intentions to help smaller services handle TVEC, redirect users and develop detection/upload-prevention/removal tools. It does not name a particular delivered toolkit, source database, licence or moderation-policy gate. This same-authorship reprint is not independent corroboration of performance. Its existence places announced collaboration before the October2025 contribution period, so the later award must not be described as founding ROOST or initiating the entire partnership.
\nThis establishes a funded programme identity: contribution26007 to CCF is associated with Elevate and an already announced ROOST relationship. It still does not identify which ROOST code or services were financed, nor whether a produced classification was used to restrict an account or content. Coop, Osprey and ROOST model tools appeared in general blog discovery, but none is promoted here to an Elevate deliverable. A programme report/repository attribution, funding schedule or recipient acceptance record would make that next join concrete.
\nThe current claims could describe both greater accessibility of safety software and greater spread of shared classification practices. The source material does not yet choose between those effects. Source rights, defaults and exercised recipient decisions in the parallel tool lane are the discriminator.
\nFour root ordinary GETs: three CCF error responses, one substantive ROOST-hosted original (SHA256fc5ed59a39aa8049dae94186064b0da163a09854eecc47911d52b981d45d348b). ccf-captures.json records exact status and bytes. No forms, contact, accounts, harmful content or live APIs accessed. Root search was bounded to CCF/ROOST/Elevate identity and possible GitHub target; no genuine CCF repository found. Unrelated search results, Catalyst gaming training, third-party newsletter revenue estimates and later report citations were discovery only, not new findings. No new branch before the combined forest.
", "path": "roost-programme-wave-2026-09-09/ccf-programme.md", "sourceIds": [ "w27-roost-s-ccf-investment2025", "w27-roost-s-ccf-elevate", "w27-roost-s-ccf-roost-reprint" ], "markdown": "9 September 2026. Root acquisition closed for this lane; ROOST governance and tools lanes continue. This is the recipient-side programme crosswalk, not a finding of completed software delivery.\n\nCCF Canadian-investment announcement (see source references) full substantive body/Notes read through primary reader, lines14–43. It names Elevate, the Community Resilience Fund support and ROOST collaboration, with an attributed statement by ROOST president Camille Francois. Its rounded amount matches the separately retained Canadian contribution26007; it is not a second award. It describes a three-year programme spanning interventions, open-source tools, benchmarks and capacity. Earlier leader priorities in2022/23 are described, so this announcement alone cannot show the2025 donor originated the policy. The claimed Canadian role in shaping international technology governance is the recipient's characterization, not a specified contractual right. No payment receipt, subcontract amount, acceptance record or named deployed product appears. Original GET403 retained as error; no retry.\n\nCharting a Collective Response (see source references) full substantive indexed-primary text read, direct reader404 and GET403 closed. It describes initial needs assessment/experimentation, later continuing tool development and parallel support for smaller platforms with ROOST. This is programme sequencing and intent, not an implementation report.\n\nCurrent Elevate programme page (see source references) indexed substantive body read, undated snapshot. It ties the programme to Canada's CRF, ROOST and the community, and describes work on AI standards/benchmarks and a toolkit for smaller providers. The page advertises a GitHub link, but the available index supplies no target and its GET403/reader-cache-miss routes are closed. No repository identity or deployed software crosswalk is established by that link label. The history-page indexed Elevate entry says12November2025 launch; other historical claims there were not acquired for use.\n\nROOST-hosted CCF partnership announcement (see source references) original retained and full main body/editor notes read. It identifies itself as a permitted reprint authored by the Christchurch Call team, displayed10February2025; CCF index cards display12February. These publication labels are not proof of two partnerships or executed contract dates. It sets out intentions to help smaller services handle TVEC, redirect users and develop detection/upload-prevention/removal tools. It does not name a particular delivered toolkit, source database, licence or moderation-policy gate. This same-authorship reprint is not independent corroboration of performance. Its existence places announced collaboration before the October2025 contribution period, so the later award must not be described as founding ROOST or initiating the entire partnership.\n\n## Consequence and remaining question\n\nThis establishes a funded programme identity: contribution26007 to CCF is associated with Elevate and an already announced ROOST relationship. It still does not identify which ROOST code or services were financed, nor whether a produced classification was used to restrict an account or content. Coop, Osprey and ROOST model tools appeared in general blog discovery, but none is promoted here to an Elevate deliverable. A programme report/repository attribution, funding schedule or recipient acceptance record would make that next join concrete.\n\nThe current claims could describe both greater accessibility of safety software and greater spread of shared classification practices. The source material does not yet choose between those effects. Source rights, defaults and exercised recipient decisions in the parallel tool lane are the discriminator.\n\nFour root ordinary GETs: three CCF error responses, one substantive ROOST-hosted original (SHA256fc5ed59a39aa8049dae94186064b0da163a09854eecc47911d52b981d45d348b). ccf-captures.json records exact status and bytes. No forms, contact, accounts, harmful content or live APIs accessed. Root search was bounded to CCF/ROOST/Elevate identity and possible GitHub target; no genuine CCF repository found. Unrelated search results, Catalyst gaming training, third-party newsletter revenue estimates and later report citations were discovery only, not new findings. No new branch before the combined forest.\n", "file": "research/roost-programme-wave-2026-09-09-ccf-programme.md" }, { "id": "roost-programme-wave-2026-09-09-governance-funding", "title": "ROOST: governance, operational home and funding", "description": "Bounded public-source acquisition completed 9 September 2026. This is a candidate research packet, not a case mutation. Read alongside root's ccf-programme.md; code/access is a separate lane. The exact Canadian contribution record was already in the accepted Christchurch wave and was not fetched again.", "html": "Bounded public-source acquisition completed 9 September 2026. This is a candidate research packet, not a case mutation. Read alongside root's ccf-programme.md; code/access is a separate lane. The exact Canadian contribution record was already acquired in the accepted Christchurch wave and was not fetched again.
ROOST has a documented route from board appointment to technical roadmap to code-maintainer decisions. It also has a separate employment arrangement, while its published fundraising total combines financial and in-kind resources. The acquired records establish these different kinds of authority; they do not establish that a founding partner, employer intermediary or the Canadian government can direct a particular deployed content decision.
\nROOST's [privacy statement](http[local research file] effective 28 October 2025, identifies Robust Open Online Safety Tools Inc as the entity processing personal data for covered websites/events. Its current site footer reports 501(c)(3) public-charity status and EIN 33-3305434. These are the organization's legal/tax representations, not an independently obtained IRS determination or incorporation instrument. The website notice does not govern every product merely because it uses ROOST code. Source: governance-privacy.html, Privacy Statement and Contact Us; footer independently read on the official web reader.
Columbia SIPA's [19 February 2025 announcement](http[local research file] describes incubation at its Institute of Global Politics, founding president Camille François, and prospective operation as an independent nonprofit retaining close collaboration with Columbia. This proves the institution's account of incubation and intended separation, not a university ownership interest or transfer agreement. Full substantive announcement read through the primary-site web reader, lines 136–156. Ordinary GET returned a retained 403 challenge; that file is not the announcement's original content.
\nA current [Marketing Manager listing linked directly from ROOST's Careers page](http[local research file] says US-based individuals will be employed by Global Impact on behalf of ROOST. The position reports to ROOST's COO. This is a concrete advertised employment/operating relationship, not proof this vacancy was filled, every existing staff member has that employer, or Global Impact owns ROOST's grants. The posting's publication date is not established; it was live on 9 September 2026. No fiscal sponsorship, employer-services agreement or payment record was acquired. Source: governance-marketing-role.html, Compensation and Structure; source-discovery parent governance-careers.html.
The current [Project Roles policy](http[local research file] expressly assigns:
\n| Actor | \nPublished decision/resource role | \n
|---|---|
| ROOST | \nRepository administration; infrastructure, connections, specialists and some engineering capacity. Initially appoints maintainers; may replace them for inactivity, vacancy or misconduct. | \n
| Board | \nAppoints the Technical Design Committee (TDC). | \n
| TDC | \nOverarching technical direction/roadmap; consultation or tie-breaking for maintainer teams. | \n
| Maintainers | \nDay-to-day development, project/release management and merge approval under ROOST branch protections; can appoint additional/replacement maintainers. | \n
| Contributors/community | \nPropose changes; community membership alone grants no explicit permissions. | \n
The policy says employer organizations are not automatically contributors/maintainers. It permits partner GitHub teams for access administration, including prerelease repositories. It identifies open-source forking as a check on repository control. These are published rules, not an audit of their use or proof that downstream platforms must accept an upstream update. Source: governance-roles.html, entire substantive page; undated current snapshot. The actual identity of repository maintainers and runtime behavior belong to the code lane.
The [26 November 2025 governance explanation](http[local research file] describes a light, evolving governance approach: the initial TDC was recruited by ROOST, with future selection arrangements still to be developed. It also says users can offer feedback privately or pseudonymously and distinguishes shared infrastructure from platform-specific signals and decision logic. Do not backdate the more explicit current Roles policy to the earliest partnership. Source: governance-design2025.html, full substantive body, particularly Decision making, TDC and Staying secure sections.
The [current Team page](http[local research file] lists Clint Smith (Discord legal chief) and Eli Sugarman (Hewlett Foundation special-projects director) as co-chairs, François as president/CEO, and other board members including Mark Surman (Mozilla), Vincent Courson (Google), Audrey Tang, Angela Lungati, Jamil Jaffer, Megan Smith and Niki Christoff. Its TDC includes Naren Koneru (Roblox) and ROOST CTO Vinay Rao as co-chairs; other listed people include OpenAI's Sam Toizer, Discord's Hailey Elizabeth, Cash App/Afterpay engineering lead Shu Lei, GitHub's Tom Thorley, Rebecca Portnoff and Tim Pepper. These are named individual roles and listed affiliations, not employer-controlled seats or voting blocs. Source: governance-team.html, Staff/Board/TDC, full body read.
Chronology matters: the [22 December 2025 board announcement](http[local research file] calls this an inaugural board and uses Chair/Vice-Chair for Smith/Sugarman. The February launch already names Smith as board chair. Therefore December is an announcement date, not demonstrated first existence of a board. The exact co-chair transition, board appointment/removal rules, voting thresholds, reserved donor rights and approved budget delegations remain unacquired. Source: governance-board2025.html, full substantive body; compare launch below.
ROOST's [10 February 2025 distributed launch release](http[local research file] says it had raised more than $27 million for its first four operating years. It names Eric Schmidt, Discord, OpenAI, Google, Roblox, Knight Foundation, AI Collaborative, Patrick J. McGovern Foundation and Project Liberty Institute as founding partners. This is an issuer press release distributed by PR Newswire (SOURCE ROOST), not independent reporting. It does not allocate individual contributions, disclose restrictions or establish cash received. Source: governance-launch2025.html, complete release body excluding distributor navigation/marketing; exact currency is expressed as dollars without an acquired currency annotation.
The same-date [president's launch letter](http[local research file] instead uses $28 million in initial funding; the [4 June 2025 first-100-days update](http[local research file] specifies $28.5 million in funding and in-kind contributions. These differing dates/descriptions cannot safely be subtracted to create fresh grants or summed. The June account attributes capacity to founding-partner support, but no cash/in-kind schedule, valuation method, audited reconciliation, individual grant contract or CCF allocation was acquired. Source: governance-100days.html, Building an Organization for Change; governance-gratitude2025.html, opening.
The [current partnerships page](http[local research file] separates founding partners from partners; CCF appears in the latter. Logos are partnership labels, not donor-payment evidence. Its founder label for SCSP does not identify the legal vehicle behind every Schmidt-associated resource. No Carnegie entity is newly introduced by this packet.
\nThe February launch letter credits CCF/Paul Ash for planned joint technical and information solutions supporting the Call. It separately credits Tech Coalition/Sean Litton, Lauren Tharp and Kay Chau with child-safety expertise, alongside a quotation about the Coalition's Pathways capacity-building programme. The author expressly describes ROOST as not a corporate membership organization. These are source-attributed collaboration/expertise relations; they do not establish Lantern data sharing, a Tech Coalition grant, or ROOST membership in the Coalition. Source: governance-gratitude2025.html, CCF/Tech Coalition paragraphs and paragraph preceding philanthropic partners.
Accepted prior record: [Public Safety Canada contribution26007](http[local research file] CAD1,369,064 to The Christchurch Call Foundation, 23 October 2025–31 March 2028, includes a ROOST partnership in three expected workstreams. The retained original is ../christchurch-governance-wave-2026-09-09/transition-canada26007.html, SHA256 df2b85ed3c34f56a465fa769bdba2bee05b829fd67f710ac5d4318e1ab6cf352. It does not turn the whole contribution into a ROOST subaward or identify authority over ROOST's maintainers. Root owns new programme/output findings; this packet does not duplicate or pre-judge them.
For integration, reuse the existing CCF, Columbia, Discord, OpenAI, Google, Roblox, Tech Coalition and other exact identities where they resolve. The meaningful new relationship types are incubation, advertised employer, named board/technical role, appoints, roadmap authority, merge/release authority, fundraising reported and collaboration/expertise reported. Avoid converting the entire roster into control edges. The Canadian edge remains a CCF award with an expected partnership purpose.
\nThe highest-yield missing instruments are the CCF–ROOST project agreement/budget and its acceptance schedule; ROOST's board constitution/budget delegations; donor contribution schedules separating money from donated resources; and the Global Impact service/fiscal arrangement. These would establish who can withhold resources, approve work or redirect it. Ordinary employer administration, charitable resource pooling and maintainers' independent technical judgment remain live alternative explanations to a donor-directed policy chain.
\nTwelve ordinary public GETs: eleven substantive HTML originals and one Columbia 403 challenge. Columbia's primary web reader supplied the substantive substitute; no retry or proxy was used. Companion governance-captures.json records each URL, response status, original SHA256, derivative text, reading scope and exclusions. All substantive originals were read in their stated scope. No repository execution, API account, contact, application submission or money movement occurred.
Finite searches covered ROOST governance/legal identity, its exact EIN, technical committee, and targeted ROOST + Google/McGovern/Global Impact funding combinations. Several combined searches produced irrelevant results or none. They do not establish the absence of financial records. No original IRS determination, constitutional instrument, restricted grant agreement or partner-specific cash receipt was located in this bounded pass. An initial overlong search response was truncated before any results could be inspected; it is not evidence. Acquisition is closed for the combined forest assessment.
", "path": "roost-programme-wave-2026-09-09/governance-funding.md", "sourceIds": [ "w27-roost-s-roost-privacy", "w27-roost-s-columbia-incubation", "w27-roost-s-roost-careers", "w27-roost-s-global-impact-employment", "w27-roost-s-roost-roles", "w27-roost-s-roost-design2025", "w27-roost-s-roost-team", "w27-roost-s-roost-board2025", "w27-roost-s-roost-launch", "w27-roost-s-roost-gratitude", "w27-roost-s-roost-100days" ], "markdown": "Bounded public-source acquisition completed 9 September 2026. This is a candidate research packet, not a case mutation. Read alongside root's `ccf-programme.md`; code/access is a separate lane. The exact Canadian contribution record was already acquired in the accepted Christchurch wave and was not fetched again.\n\n## Finding\n\nROOST has a documented route from board appointment to technical roadmap to code-maintainer decisions. It also has a separate employment arrangement, while its published fundraising total combines financial and in-kind resources. The acquired records establish these different kinds of authority; they do not establish that a founding partner, employer intermediary or the Canadian government can direct a particular deployed content decision.\n\n## Legal and operational home\n\nROOST's privacy statement (see source references) effective 28 October 2025, identifies **Robust Open Online Safety Tools Inc** as the entity processing personal data for covered websites/events. Its current site footer reports 501(c)(3) public-charity status and **EIN 33-3305434**. These are the organization's legal/tax representations, not an independently obtained IRS determination or incorporation instrument. The website notice does not govern every product merely because it uses ROOST code. Source: `governance-privacy.html`, Privacy Statement and Contact Us; footer independently read on the official web reader.\n\nColumbia SIPA's 19 February 2025 announcement (see source references) describes incubation at its Institute of Global Politics, founding president Camille François, and prospective operation as an independent nonprofit retaining close collaboration with Columbia. This proves the institution's account of incubation and intended separation, not a university ownership interest or transfer agreement. Full substantive announcement read through the primary-site web reader, lines 136–156. Ordinary GET returned a retained 403 challenge; that file is **not** the announcement's original content.\n\nA current Marketing Manager listing linked directly from ROOST's Careers page (see source references) says US-based individuals will be employed by **Global Impact on behalf of ROOST**. The position reports to ROOST's COO. This is a concrete advertised employment/operating relationship, not proof this vacancy was filled, every existing staff member has that employer, or Global Impact owns ROOST's grants. The posting's publication date is not established; it was live on 9 September 2026. No fiscal sponsorship, employer-services agreement or payment record was acquired. Source: `governance-marketing-role.html`, Compensation and Structure; source-discovery parent `governance-careers.html`.\n\n## Who can decide what\n\nThe current Project Roles policy (see source references) expressly assigns:\n\n| Actor | Published decision/resource role |\n| --- | --- |\n| ROOST | Repository administration; infrastructure, connections, specialists and some engineering capacity. Initially appoints maintainers; may replace them for inactivity, vacancy or misconduct. |\n| Board | Appoints the Technical Design Committee (TDC). |\n| TDC | Overarching technical direction/roadmap; consultation or tie-breaking for maintainer teams. |\n| Maintainers | Day-to-day development, project/release management and merge approval under ROOST branch protections; can appoint additional/replacement maintainers. |\n| Contributors/community | Propose changes; community membership alone grants no explicit permissions. |\n\nThe policy says employer organizations are not automatically contributors/maintainers. It permits partner GitHub teams for access administration, including prerelease repositories. It identifies open-source forking as a check on repository control. These are published rules, not an audit of their use or proof that downstream platforms must accept an upstream update. **Source:** `governance-roles.html`, entire substantive page; undated current snapshot. The actual identity of repository maintainers and runtime behavior belong to the code lane.\n\nThe 26 November 2025 governance explanation (see source references) describes a light, evolving governance approach: the initial TDC was recruited by ROOST, with future selection arrangements still to be developed. It also says users can offer feedback privately or pseudonymously and distinguishes shared infrastructure from platform-specific signals and decision logic. Do not backdate the more explicit current Roles policy to the earliest partnership. Source: `governance-design2025.html`, full substantive body, particularly Decision making, TDC and Staying secure sections.\n\nThe current Team page (see source references) lists Clint Smith (Discord legal chief) and Eli Sugarman (Hewlett Foundation special-projects director) as co-chairs, François as president/CEO, and other board members including Mark Surman (Mozilla), Vincent Courson (Google), Audrey Tang, Angela Lungati, Jamil Jaffer, Megan Smith and Niki Christoff. Its TDC includes Naren Koneru (Roblox) and ROOST CTO Vinay Rao as co-chairs; other listed people include OpenAI's Sam Toizer, Discord's Hailey Elizabeth, **Cash App/Afterpay engineering lead Shu Lei**, GitHub's Tom Thorley, Rebecca Portnoff and Tim Pepper. These are named individual roles and listed affiliations, not employer-controlled seats or voting blocs. Source: `governance-team.html`, Staff/Board/TDC, full body read.\n\nChronology matters: the 22 December 2025 board announcement (see source references) calls this an inaugural board and uses Chair/Vice-Chair for Smith/Sugarman. The February launch already names Smith as board chair. Therefore December is an announcement date, not demonstrated first existence of a board. The exact co-chair transition, board appointment/removal rules, voting thresholds, reserved donor rights and approved budget delegations remain unacquired. Source: `governance-board2025.html`, full substantive body; compare launch below.\n\n## Financing: the headline is not a cash balance\n\nROOST's 10 February 2025 distributed launch release (see source references) says it had raised **more than $27 million for its first four operating years**. It names Eric Schmidt, Discord, OpenAI, Google, Roblox, Knight Foundation, AI Collaborative, Patrick J. McGovern Foundation and Project Liberty Institute as founding partners. This is an issuer press release distributed by PR Newswire (SOURCE ROOST), not independent reporting. It does not allocate individual contributions, disclose restrictions or establish cash received. Source: `governance-launch2025.html`, complete release body excluding distributor navigation/marketing; exact currency is expressed as dollars without an acquired currency annotation.\n\nThe same-date president's launch letter (see source references) instead uses **$28 million in initial funding**; the 4 June 2025 first-100-days update (see source references) specifies **$28.5 million in funding and in-kind contributions**. These differing dates/descriptions cannot safely be subtracted to create fresh grants or summed. The June account attributes capacity to founding-partner support, but no cash/in-kind schedule, valuation method, audited reconciliation, individual grant contract or CCF allocation was acquired. Source: `governance-100days.html`, Building an Organization for Change; `governance-gratitude2025.html`, opening.\n\nThe current partnerships page (see source references) separates founding partners from partners; CCF appears in the latter. Logos are partnership labels, not donor-payment evidence. Its founder label for SCSP does not identify the legal vehicle behind every Schmidt-associated resource. No Carnegie entity is newly introduced by this packet.\n\n## Connection to the existing case\n\nThe February launch letter credits **CCF/Paul Ash** for planned joint technical and information solutions supporting the Call. It separately credits **Tech Coalition/Sean Litton, Lauren Tharp and Kay Chau** with child-safety expertise, alongside a quotation about the Coalition's Pathways capacity-building programme. The author expressly describes ROOST as not a corporate membership organization. These are source-attributed collaboration/expertise relations; they do not establish Lantern data sharing, a Tech Coalition grant, or ROOST membership in the Coalition. Source: `governance-gratitude2025.html`, CCF/Tech Coalition paragraphs and paragraph preceding philanthropic partners.\n\nAccepted prior record: Public Safety Canada contribution26007 (see source references) CAD1,369,064 to **The Christchurch Call Foundation**, 23 October 2025–31 March 2028, includes a ROOST partnership in three expected workstreams. The retained original is `../christchurch-governance-wave-2026-09-09/transition-canada26007.html`, SHA256 `df2b85ed3c34f56a465fa769bdba2bee05b829fd67f710ac5d4318e1ab6cf352`. It does not turn the whole contribution into a ROOST subaward or identify authority over ROOST's maintainers. Root owns new programme/output findings; this packet does not duplicate or pre-judge them.\n\n## Candidate relations and next decisive record\n\nFor integration, reuse the existing CCF, Columbia, Discord, OpenAI, Google, Roblox, Tech Coalition and other exact identities where they resolve. The meaningful new relationship types are **incubation**, **advertised employer**, **named board/technical role**, **appoints**, **roadmap authority**, **merge/release authority**, **fundraising reported** and **collaboration/expertise reported**. Avoid converting the entire roster into control edges. The Canadian edge remains a CCF award with an expected partnership purpose.\n\nThe highest-yield missing instruments are the CCF–ROOST project agreement/budget and its acceptance schedule; ROOST's board constitution/budget delegations; donor contribution schedules separating money from donated resources; and the Global Impact service/fiscal arrangement. These would establish who can withhold resources, approve work or redirect it. Ordinary employer administration, charitable resource pooling and maintainers' independent technical judgment remain live alternative explanations to a donor-directed policy chain.\n\n## Acquisition frontier\n\nTwelve ordinary public GETs: eleven substantive HTML originals and one Columbia 403 challenge. Columbia's primary web reader supplied the substantive substitute; no retry or proxy was used. Companion `governance-captures.json` records each URL, response status, original SHA256, derivative text, reading scope and exclusions. All substantive originals were read in their stated scope. No repository execution, API account, contact, application submission or money movement occurred.\n\nFinite searches covered ROOST governance/legal identity, its exact EIN, technical committee, and targeted ROOST + Google/McGovern/Global Impact funding combinations. Several combined searches produced irrelevant results or none. They do not establish the absence of financial records. No original IRS determination, constitutional instrument, restricted grant agreement or partner-specific cash receipt was located in this bounded pass. An initial overlong search response was truncated before any results could be inspected; it is not evidence. Acquisition is closed for the combined forest assessment.\n", "file": "research/roost-programme-wave-2026-09-09-governance-funding.md" }, { "id": "roost-programme-wave-2026-09-09-tool-governance", "title": "CCF policy as a public ROOST model input", "description": "9 September2026. Bounded tool-governance acquisition closed. A named CCF contribution is now inspectable: a TVEC policy prompt pack in ROOST's Model Community repository. It is a template and accompanying evaluation/context documents, not proof that Coop, Osprey or a named platform deployed the policy, or that Canada accepted it under contribution26007. No m", "html": "9 September2026. Bounded tool-governance acquisition closed. A named CCF contribution is now inspectable: a TVEC policy prompt pack in ROOST's Model Community repository. It is a template and accompanying evaluation/context documents, not proof that Coop, Osprey or a named platform deployed the policy, or that Canada accepted it under contribution26007. No model, API, underlying harmful-content dataset, account or submission was used.
\nThe [pinned CCF pack](http[local research file] contains four files: a text-classification policy, an annex comparing definitions, an annex supplying actor context and an evaluation report. The repository [README](http[local research file] lines19–26, expressly describes this area as partner-contributed policy prompts organized by submitter for use with policy-following models. Annex2 expressly speaks as the Christchurch Call Foundation. This is stronger attribution than a directory name alone.
\nInspected repository revision: 22b3bd48562616f9379eb6fa63abe131346a6de6, whose public API commit record is dated20August2026. This is the version pin, not the pack's initial-publication date. A path-scoped [27May2026 relocation commit](http[local research file] names Juliet Shen as author, credits Mackenzie as coauthor and describes relocation of the CCF policy from gpt-oss-safeguard/projects/tvec-policy with no content changes. That is repository attribution and a commit-message claim, not proof of each person's employer, drafting approval or independent verification of the rename. A finite five-result query of that stated earlier path returns7May2026 organization commits credited to Mackenzie. It was not an exhaustive history or proof of first publication. The evaluation's17April2026 update date is a document date, not a repository-release date.
The current [roadmap](http[local research file] describes the Model Community as a way to convert traditional policies into model-ready prompts and distribute implementation knowledge. The [11December2025 roadmap announcement](http[local research file] What We're Building, specifically planned2026 access to sample TVEC policies. Neither inspected roadmap nor these four pack files establishes a budget allocation, accepted Canadian milestone or CCF programme exclusivity. Root's ccf-programme.md owns that separate evidence. No generalized Coop/Osprey-to-CCF funding edge is supported.
The template defines terrorism and violent extremism, assigns eight categories to a removal heading, and supplies a separate permissible-content heading. Its credible-threat category adds an instruction to escalate to law enforcement and normally requires at least two listed indicators. The EDSA/newsworthiness category expressly calls for human review. These are desired labels/actions in a prompt template; no inspected execution path sends a report or removes content. Locators: tools-tvec-policy.txt lines5–15,25–43,101–123.
The categories extend beyond official branded material into material support, recruitment, praise/promotion and incitement. Annex2 recommends combining actor, content and behavior signals, warns that designations can be politicized and incomplete, and treats UN/EU/FiveEyes lists as contextual signals rather than an exclusive truth source. CCF supplies its own limited incident/perpetrator list, selected where it engaged its crisis protocol and/or an attack referenced the Christchurch attacker. It expressly says the list is not comprehensive and not every listed person faced a terrorism/violent-extremism charge or investigation. This is CCF selecting context for classification; it does not make each entry a legally certified terrorist. No individual incident/person nodes are warranted. Locators: annex2 lines3–31 and table32–65.
\nAnnex1 compares legal and scholarly approaches and notes absence of a universally agreed definition. Its external legal references remain the author's comparison; this run did not independently verify the current laws, lists or incident entries. The policy text's chosen definition is not itself a statute or proof of mandatory government instructions.
\nThe repository carries a [CC-BY4.0 licence](http[local research file] Sections2–3 permit reproduction and adaptation of licensed material subject to attribution, notices and identifying changes; section2 excludes implied endorsement and preserves distinct noncopyright rights. Section5 supplies an as-is disclaimer. Section6 provides termination for noncompliance and restoration rules; stopping distribution does not itself terminate an otherwise valid licence. These are the published repository terms, not an acquired signed CCF agreement or permission to use external training data, hash databases or separately licensed model weights.
\nThe README invites issue reports, discussions and iterative feedback to model creators, and supports configurable policy-following models. That makes independent adaptation and public feedback possible. It does not establish a required patch/adoption schedule, an affected user's appeal, or mandatory propagation of a corrected prompt into recipients' deployed systems. The precise project-maintainer/TDC rights belong to Linnaeus's governance lane. No case-specific model deployment, policy variant, update acknowledgement, enforcement adapter or restored item was acquired here.
\nThe74-line report is labelled17April2026 and names gpt-oss-safeguard without an exact checkpoint or complete reproducibility package. It reports170 examples:114 true positives,12 false negatives, zero false positives and44 true negatives against evaluator/team judgments. It reports90.5% recall and100% precision for that sample. This does not establish perfect general accuracy or legal ground truth. Recruitment recall is12/17, approximately71%, below its own85% target. The failure discussion explicitly says team experts disagreed with the model's threshold interpretation in four examples and suggests policy refinement; that is a disclosed source of classificatory judgment, not demonstrated political misuse. Locators: tools-tvec-evaluation.md lines3–62.
The report also gives72.2% precision for EDSA classification and explains that confusion concerns two permissible categories, rather than those errors admitting violative content. Preserve this alongside the binary zero-false-positive result. The raw examples, annotation/adjudication protocol, evaluation script, exact model checkpoint and run settings were not identified in the four-file pack; no performance test was run.
\nTwo observable textual tensions deserve reconciliation before interpreting a deployment. TVEC0 references INCITE1, which is undefined within the policy. The EDSA section says its exception does not apply if content falls into any TVEC1–8 category, while the evaluation describes otherwise-TVEC content permitted for an EDSA purpose. The template's examples also include journalism and research. Context precedence and intended exception handling therefore require a version-specific interpretation/test; neither a concrete over-removal incident nor the impossibility of an exception follows from this text alone.
The new bridge is CCF-authored policy/context entering a public collection intended for direct model use. This is a concrete path by which an institution's definitions and judgments can travel, alongside meaningful freedom to adapt them. It is distinct from government financial acceptance, mandatory adoption, a model's actual output and a platform's action. A source contribution can shape defaults without giving its author universal enforcement authority.
\nThe next discriminating records are a named adopter's exact prompt/model/configuration and review/enforcement route; a resolved interpretation of the exception text; and the matched CCF/Canada output schedule and acceptance record. For the claimed evaluation, a benign or suitably redacted reproducibility package would distinguish the documented finite test from general performance. No new branch is opened here.
\ntools-captures.json records14 successful original/API responses with hashes, exact URLs, commit and read scopes, plus two reader/access-history observations. All four CCF files were read completely: policy123lines, annex1 62lines, annex2 65lines, evaluation74lines. The92-line README was read completely. Licence scope is its title, selected definitions and sections2–6 with supplementary terms through line372; not every introductory/footer line. Roadmap and announcement substantive bodies were read; media and links were not automatically treated as read. Repository listings were discovery metadata only. All originals are public unauthenticated response bytes; derived text is separately labelled.
The roadmap reader failed with a cache miss; its sole normal GET substitute succeeded. The exact GitHub policy-directory reader returned429; after a pause, one ordinary public API route supplied the listing and pinned source URLs. No access restriction was bypassed. Previously closed CCF403 routes were not retried. Six targeted queries were used for CCF/Elevate/TVEC/roadmap discovery; irrelevant municipal and similarly named software results were discarded. No code installation, model inference, form, private dataset or outreach occurred. The candidate case, atlas and ZIP remain untouched.
", "path": "roost-programme-wave-2026-09-09/tool-governance.md", "sourceIds": [ "w27-roost-s-roadmap", "w27-roost-s-roadmap2025", "w27-roost-s-rmc-pin", "w27-roost-s-rmc-readme", "w27-roost-s-rmc-ccby", "w27-roost-s-tvec-policy", "w27-roost-s-tvec-annex1", "w27-roost-s-tvec-annex2", "w27-roost-s-tvec-evaluation", "w27-roost-s-tvec-relocation", "w27-roost-s-tvec-prior-path" ], "markdown": "9 September2026. Bounded tool-governance acquisition closed. A named CCF contribution is now inspectable: a TVEC policy prompt pack in ROOST's Model Community repository. It is a template and accompanying evaluation/context documents, not proof that Coop, Osprey or a named platform deployed the policy, or that Canada accepted it under contribution26007. No model, API, underlying harmful-content dataset, account or submission was used.\n\n## What was actually released\n\nThe pinned CCF pack (see source references) contains four files: a text-classification policy, an annex comparing definitions, an annex supplying actor context and an evaluation report. The repository README (see source references) lines19–26, expressly describes this area as partner-contributed policy prompts organized by submitter for use with policy-following models. Annex2 expressly speaks as the Christchurch Call Foundation. This is stronger attribution than a directory name alone.\n\nInspected repository revision: `22b3bd48562616f9379eb6fa63abe131346a6de6`, whose public API commit record is dated20August2026. This is the version pin, not the pack's initial-publication date. A path-scoped 27May2026 relocation commit (see source references) names Juliet Shen as author, credits Mackenzie as coauthor and describes relocation of the CCF policy from `gpt-oss-safeguard/projects/tvec-policy` with no content changes. That is repository attribution and a commit-message claim, not proof of each person's employer, drafting approval or independent verification of the rename. A finite five-result query of that stated earlier path returns7May2026 organization commits credited to Mackenzie. It was not an exhaustive history or proof of first publication. The evaluation's17April2026 update date is a document date, not a repository-release date.\n\nThe current roadmap (see source references) describes the Model Community as a way to convert traditional policies into model-ready prompts and distribute implementation knowledge. The 11December2025 roadmap announcement (see source references) What We're Building, specifically planned2026 access to sample TVEC policies. Neither inspected roadmap nor these four pack files establishes a budget allocation, accepted Canadian milestone or CCF programme exclusivity. Root's `ccf-programme.md` owns that separate evidence. No generalized Coop/Osprey-to-CCF funding edge is supported.\n\n## Where the policy places judgment\n\nThe template defines terrorism and violent extremism, assigns eight categories to a removal heading, and supplies a separate permissible-content heading. Its credible-threat category adds an instruction to escalate to law enforcement and normally requires at least two listed indicators. The EDSA/newsworthiness category expressly calls for human review. These are desired labels/actions in a prompt template; no inspected execution path sends a report or removes content. Locators: `tools-tvec-policy.txt` lines5–15,25–43,101–123.\n\nThe categories extend beyond official branded material into material support, recruitment, praise/promotion and incitement. Annex2 recommends combining actor, content and behavior signals, warns that designations can be politicized and incomplete, and treats UN/EU/FiveEyes lists as contextual signals rather than an exclusive truth source. CCF supplies its own limited incident/perpetrator list, selected where it engaged its crisis protocol and/or an attack referenced the Christchurch attacker. It expressly says the list is not comprehensive and not every listed person faced a terrorism/violent-extremism charge or investigation. This is CCF selecting context for classification; it does not make each entry a legally certified terrorist. No individual incident/person nodes are warranted. Locators: annex2 lines3–31 and table32–65.\n\nAnnex1 compares legal and scholarly approaches and notes absence of a universally agreed definition. Its external legal references remain the author's comparison; this run did not independently verify the current laws, lists or incident entries. The policy text's chosen definition is not itself a statute or proof of mandatory government instructions.\n\n## Adaptation rights and correction limits\n\nThe repository carries a CC-BY4.0 licence (see source references) Sections2–3 permit reproduction and adaptation of licensed material subject to attribution, notices and identifying changes; section2 excludes implied endorsement and preserves distinct noncopyright rights. Section5 supplies an as-is disclaimer. Section6 provides termination for noncompliance and restoration rules; stopping distribution does not itself terminate an otherwise valid licence. These are the published repository terms, not an acquired signed CCF agreement or permission to use external training data, hash databases or separately licensed model weights.\n\nThe README invites issue reports, discussions and iterative feedback to model creators, and supports configurable policy-following models. That makes independent adaptation and public feedback possible. It does not establish a required patch/adoption schedule, an affected user's appeal, or mandatory propagation of a corrected prompt into recipients' deployed systems. The precise project-maintainer/TDC rights belong to Linnaeus's governance lane. No case-specific model deployment, policy variant, update acknowledgement, enforcement adapter or restored item was acquired here.\n\n## What the evaluation supports\n\nThe74-line report is labelled17April2026 and names gpt-oss-safeguard without an exact checkpoint or complete reproducibility package. It reports170 examples:114 true positives,12 false negatives, zero false positives and44 true negatives against evaluator/team judgments. It reports90.5% recall and100% precision for that sample. This does not establish perfect general accuracy or legal ground truth. Recruitment recall is12/17, approximately71%, below its own85% target. The failure discussion explicitly says team experts disagreed with the model's threshold interpretation in four examples and suggests policy refinement; that is a disclosed source of classificatory judgment, not demonstrated political misuse. Locators: `tools-tvec-evaluation.md` lines3–62.\n\nThe report also gives72.2% precision for EDSA classification and explains that confusion concerns two permissible categories, rather than those errors admitting violative content. Preserve this alongside the binary zero-false-positive result. The raw examples, annotation/adjudication protocol, evaluation script, exact model checkpoint and run settings were not identified in the four-file pack; no performance test was run.\n\nTwo observable textual tensions deserve reconciliation before interpreting a deployment. TVEC0 references `INCITE1`, which is undefined within the policy. The EDSA section says its exception does not apply if content falls into any TVEC1–8 category, while the evaluation describes otherwise-TVEC content permitted for an EDSA purpose. The template's examples also include journalism and research. Context precedence and intended exception handling therefore require a version-specific interpretation/test; neither a concrete over-removal incident nor the impossibility of an exception follows from this text alone.\n\n## Forest implication and finite frontier\n\nThe new bridge is CCF-authored policy/context entering a public collection intended for direct model use. This is a concrete path by which an institution's definitions and judgments can travel, alongside meaningful freedom to adapt them. It is distinct from government financial acceptance, mandatory adoption, a model's actual output and a platform's action. A source contribution can shape defaults without giving its author universal enforcement authority.\n\nThe next discriminating records are a named adopter's exact prompt/model/configuration and review/enforcement route; a resolved interpretation of the exception text; and the matched CCF/Canada output schedule and acceptance record. For the claimed evaluation, a benign or suitably redacted reproducibility package would distinguish the documented finite test from general performance. No new branch is opened here.\n\n## Custody and read scope\n\n`tools-captures.json` records14 successful original/API responses with hashes, exact URLs, commit and read scopes, plus two reader/access-history observations. All four CCF files were read completely: policy123lines, annex1 62lines, annex2 65lines, evaluation74lines. The92-line README was read completely. Licence scope is its title, selected definitions and sections2–6 with supplementary terms through line372; not every introductory/footer line. Roadmap and announcement substantive bodies were read; media and links were not automatically treated as read. Repository listings were discovery metadata only. All originals are public unauthenticated response bytes; derived text is separately labelled.\n\nThe roadmap reader failed with a cache miss; its sole normal GET substitute succeeded. The exact GitHub policy-directory reader returned429; after a pause, one ordinary public API route supplied the listing and pinned source URLs. No access restriction was bypassed. Previously closed CCF403 routes were not retried. Six targeted queries were used for CCF/Elevate/TVEC/roadmap discovery; irrelevant municipal and similarly named software results were discarded. No code installation, model inference, form, private dataset or outreach occurred. The candidate case, atlas and ZIP remain untouched.\n", "file": "research/roost-programme-wave-2026-09-09-tool-governance.md" }, { "id": "sf-compliance-wave-2026-09-09-filings", "title": "Briver compliance filings — bounded public-record result", "description": "Research cutoff: 9 September 2026. This lane did not acquire an actual Briver attestation, supporting domain evidence, paragraph 16 disclosure response, satisfaction of judgment, or Briver compliance/enforcement filing. It established the public court route and encountered its explicit human-verification boundary before any case-number search. The result is ", "html": "Research cutoff: 9 September 2026. This lane did not acquire an actual Briver attestation, supporting domain evidence, paragraph 16 disclosure response, satisfaction of judgment, or Briver compliance/enforcement filing. It established the public court route and encountered its explicit human-verification boundary before any case-number search. The result is an access and publication gap, not evidence that Briver failed to comply or that the records do not exist.
\nThe previously read judgment directs paragraph 16 information to the People, and paragraphs 17–18 attestations and supporting evidence to the People through the notice procedure in paragraph 47. It does not in those clauses require every response to be publicly filed with the court. Thus even a complete docket search would not necessarily disclose all compliance material. The City Attorney and Briver/counsel are the apparent primary holders of those responses; the court holds whatever was actually filed. This custody distinction comes from the accepted judgment analysis, PDF pages 4–5 and 10 of the retained instrument, not a newly obtained response.
\n| Public route | \nExact scope and outcome | \n
|---|---|
| First indexed discovery batch | \nQueries: "CGC-24-617237" "compliance"; "Briver" "attestation" "2025"; site.sf.courts.ca.gov case information civil online case query. The combined results exposed the official court access page, the already acquired judgment and a June 13, 2025 Tilak declaration. No actual Briver response was identified. | \n
| Official court access page | \n[Case Information](http[local research file] The web reader returned 403 Forbidden; one ordinary GET substitute succeeded. The complete main Case Calendar & Query content and actual Civil Case Query href were read. This page advertises civil case lookup; it is not the case’s docket. | \n
| Actual Civil Case Query link | \nhttp[local research file]. An ordinary GET returned the court’s gateway, which explicitly verifies a person rather than an automated program. No verification response, CAPTCHA solution or underlying CaseInfo.dll request was submitted. No case-number search or register-of-actions inspection occurred. | \n
| Second indexed discovery batch | \nQueries: site:sf.gov "Briver" "attestation"; site:sfcityattorney.org "Briver" "satisfaction"; "CGC-24-617237" "Briver" "compliance". The combined returned material included the prior judgment and secondary explanatory articles, not a newly acquired Briver response, satisfaction or enforcement record. Secondary summaries were not accepted as implementation evidence. | \n
| Already retained Tilak declaration | \nRoot acquired the [June 13 declaration republished by Ars Technica](http[local research file] This lane performed only a targeted search of root’s local page-text derivatives for Briver, attestation, satisfaction, compliance and comply. The caption names Briver; the “Statement of Compliance” on PDF 47 concerns Companies Act 2006 registration. It is not a Briver judgment attestation. This limited search is not a full visual audit, and sparse/image-only extraction cannot prove absence from the PDF. Root owns substantive provider/disclosure review; it must remain separate from Briver compliance. | \n
The exact gateway statement and successful original-response metadata are retained in filings-captures.json and filings-case-query-gateway.html/.txt. Its anti-automation condition is a site access boundary, not an automatic sandbox approval rejection. No attempt was made to work around it. Closed LegInfo routes were not revisited. The six queries are preserved in filings-query-log.json, a derived search log rather than a complete search-index snapshot.
\nfilings-court-information.html: official court page, HTTP 200, 638,340 bytes, SHA-256 bfe73a633bbe71ef78bc26bfd736eee6183d26d5cede085e78980ec1da5fc3f7. Reading scope is the public main query content and link; navigation scripts are captured, not interpreted as case records.
filings-case-query-gateway.html: actual linked court gateway, HTTP 200, 2,167 bytes, SHA-256 c1cb249fe448f2cd3687036b8bd0a48271f71cd853f0169161ebfefbe4e3adb0. Complete visible text and form action were inspected. The case record remained inaccessible behind that boundary.
Root’s separately retained tilak-declaration-june13-2025.pdf: SHA-256 bbd7cfa2059c35fd203ce7202b1bb6072d7bbfe6412ea061ffdf7aa58d2183a5. No duplicate download occurred. Its primary-instrument content and republisher custody are described in root’s packet; this lane claims only the limited text-search scope above.
The strongest current conclusion is that the judgment created recurrent reporting and discovery rights, but this public pass did not verify their exercise or fulfillment. The missing discriminator remains the actual attestation with domain evidence, paragraph 16 response, payment/satisfaction record, or a court filing enforcing or recording those obligations. A legitimately supplied case file or a newly public city/court publication could reopen that question. No account, fee, contact, records submission, CAPTCHA solution, target-site visit or further research branch was undertaken.
", "path": "sf-compliance-wave-2026-09-09/filings.md", "sourceIds": [ "w25-sf-s-court-gateway" ], "markdown": "Research cutoff: **9 September 2026**. This lane did **not acquire an actual Briver attestation, supporting domain evidence, paragraph 16 disclosure response, satisfaction of judgment, or Briver compliance/enforcement filing**. It established the public court route and encountered its explicit human-verification boundary before any case-number search. The result is an access and publication gap, not evidence that Briver failed to comply or that the records do not exist.\n\nThe previously read judgment directs paragraph 16 information to **the People**, and paragraphs 17–18 attestations and supporting evidence to the People through the notice procedure in paragraph 47. It does not in those clauses require every response to be publicly filed with the court. Thus even a complete docket search would not necessarily disclose all compliance material. The City Attorney and Briver/counsel are the apparent primary holders of those responses; the court holds whatever was actually filed. This custody distinction comes from the [accepted judgment analysis](sf-enforcement-wave-2026-09-09-judgment.md), PDF pages 4–5 and 10 of the retained instrument, not a newly obtained response.\n\n## Routes and limits\n\n| Public route | Exact scope and outcome |\n| --- | --- |\n| First indexed discovery batch | Queries: `\"CGC-24-617237\" \"compliance\"`; `\"Briver\" \"attestation\" \"2025\"`; `site.sf.courts.ca.gov case information civil online case query`. The combined results exposed the official court access page, the already acquired judgment and a June 13, 2025 Tilak declaration. No actual Briver response was identified. |\n| Official court access page | Case Information (see source references) The web reader returned 403 Forbidden; **one ordinary GET substitute succeeded**. The complete main Case Calendar & Query content and actual Civil Case Query href were read. This page advertises civil case lookup; it is not the case’s docket. |\n| Actual Civil Case Query link | `http[local research file]`. An ordinary GET returned the court’s gateway, which explicitly verifies a person rather than an automated program. No verification response, CAPTCHA solution or underlying CaseInfo.dll request was submitted. **No case-number search or register-of-actions inspection occurred.** |\n| Second indexed discovery batch | Queries: `site:sf.gov \"Briver\" \"attestation\"`; `site:sfcityattorney.org \"Briver\" \"satisfaction\"`; `\"CGC-24-617237\" \"Briver\" \"compliance\"`. The combined returned material included the prior judgment and secondary explanatory articles, not a newly acquired Briver response, satisfaction or enforcement record. Secondary summaries were not accepted as implementation evidence. |\n| Already retained Tilak declaration | Root acquired the June 13 declaration republished by Ars Technica (see source references) This lane performed only a targeted search of root’s local page-text derivatives for Briver, attestation, satisfaction, compliance and comply. The caption names Briver; the “Statement of Compliance” on PDF 47 concerns Companies Act 2006 registration. It is not a Briver judgment attestation. This limited search is not a full visual audit, and sparse/image-only extraction cannot prove absence from the PDF. Root owns substantive provider/disclosure review; it must remain separate from Briver compliance. |\n\nThe exact gateway statement and successful original-response metadata are retained in **filings-captures.json** and **filings-case-query-gateway.html/.txt**. Its anti-automation condition is a site access boundary, not an automatic sandbox approval rejection. No attempt was made to work around it. Closed LegInfo routes were not revisited. The six queries are preserved in **filings-query-log.json**, a derived search log rather than a complete search-index snapshot.\n\n## Retained evidence\n\n**filings-court-information.html**: official court page, HTTP 200, 638,340 bytes, SHA-256 `bfe73a633bbe71ef78bc26bfd736eee6183d26d5cede085e78980ec1da5fc3f7`. Reading scope is the public main query content and link; navigation scripts are captured, not interpreted as case records.\n\n**filings-case-query-gateway.html**: actual linked court gateway, HTTP 200, 2,167 bytes, SHA-256 `c1cb249fe448f2cd3687036b8bd0a48271f71cd853f0169161ebfefbe4e3adb0`. Complete visible text and form action were inspected. The case record remained inaccessible behind that boundary.\n\nRoot’s separately retained **tilak-declaration-june13-2025.pdf**: SHA-256 `bbd7cfa2059c35fd203ce7202b1bb6072d7bbfe6412ea061ffdf7aa58d2183a5`. No duplicate download occurred. Its primary-instrument content and republisher custody are described in root’s packet; this lane claims only the limited text-search scope above.\n\n## Breakpoint\n\nThe strongest current conclusion is that the judgment created recurrent reporting and discovery rights, but this public pass did not verify their exercise or fulfillment. The missing discriminator remains the actual attestation with domain evidence, paragraph 16 response, payment/satisfaction record, or a court filing enforcing or recording those obligations. A legitimately supplied case file or a newly public city/court publication could reopen that question. No account, fee, contact, records submission, CAPTCHA solution, target-site visit or further research branch was undertaken.\n", "file": "research/sf-compliance-wave-2026-09-09-filings.md" }, { "id": "sf-compliance-wave-2026-09-09-provider-disclosure", "title": "Provider disclosure and an attributed Australian lead", "description": "9 September 2026. Root provider lane closed. Tilak declaration, filed 13 June 2025(httplocal research file republished by Ars Technica. Original custody and hash: tilak-capture.json. Retained PDF is 61 pages, 6,815,888 bytes, SHA256 bbd7cfa2059c35fd203ce7202b1bb6072d7bbfe6412ea061ffdf7aa58d2183a5.", "html": "9 September 2026. Root provider lane closed. [Tilak declaration, filed 13 June 2025](http[local research file] republished by Ars Technica. Original custody and hash: tilak-capture.json. Retained PDF is 61 pages, 6,815,888 bytes, SHA256 bbd7cfa2059c35fd203ce7202b1bb6072d7bbfe6412ea061ffdf7aa58d2183a5.
\nThe sworn declaration supports a motion for alternative service on Sol Ecom, not a Briver compliance certification. Paragraphs18–23 report subpoenas and responsive contact records from Cloudflare, Namecheap and former registrar Name.com. Paragraph29 reports Google's June6,2025 production. Paragraphs24–30 describe subsequent notices and correspondence. These are counsel's sworn accounts; actual subpoena and production instruments are not attached in this read scope. They establish a reported disclosure route, not provider shutdown instructions.
\nExhibit6 reproduces December2024–January2025 correspondence. A respondent associated by counsel with drawnudes.io claims eSafety requirements and compliance, including alleged January2025 legislation (PDF19,21). The respondent seeks written SF requirements and discusses territorial restrictions (PDF19,22). This is a respondent's account and offer, not independent verification of Australian law, successful compliance, eSafety approval, or regulator-to-regulator coordination. No specific technical restriction or issuing provider is demonstrated by that exchange.
\nThe complete declaration and exhibit do not supply Briver's required report or a named provider's takedown command. The city-announced cooperation and this separate subpoena route must retain separate source ancestry.
\nRoot read the complete six-page declaration through the web reader, including the filed stamp, allegations about service attempts and signature, and complete Exhibit6 (PDF18–23) through retained extraction. Root visually checked PDF1,4,6,21,22. PDF21's visually underlined eSafety reference has no URI annotation on that page in the retained PDF. That is not a finding about external links everywhere in the bundle. All61 pages were extracted for search; corporate-registration/profile/site exhibits outside those twelve pages were not substantively reviewed, and no whole-bundle-read claim is made. Residential addresses, private contact identifiers and unrelated personal profile details are not reproduced here.
\nRoot searched seven queries: "Briver" "registrar"; "Sol Ecom" "registry"; site.sfcityattorney.org "deepfake" "registrars"; site.sf.gov "deepfake" "registrars"; "San Francisco" "registries" "June 2, 2025"; "Chiu" "Briver" "cooperate"; "San Francisco" "registrars" "Namecheap" deepfake. Most results were unrelated, including different Briver entities. They were not identity evidence. The named-company query was a probe, not prior proof of involvement.
The complete main body of Ars Technica's16August2024 article was read. It supplies a lead concerning the complaint's requested provider relief; root did not acquire the linked complaint or promote a report of requested relief into an order. The worker independently found the actual Tilak declaration; root's web reader and one ordinary originalGET succeeded. Search and retained-record read only. No provider takedown record was recovered in this bounded lane; that is not proof no provider acted. The separate worker filings.md records the court-gateway limitation and Briver-record search. No further acquisition branch was opened before the forest review.
", "path": "sf-compliance-wave-2026-09-09/provider-disclosure.md", "sourceIds": [ "w25-sf-s-tilak" ], "markdown": "9 September 2026. Root provider lane closed. Tilak declaration, filed 13 June 2025 (see source references) republished by Ars Technica. Original custody and hash: tilak-capture.json. Retained PDF is 61 pages, 6,815,888 bytes, SHA256 bbd7cfa2059c35fd203ce7202b1bb6072d7bbfe6412ea061ffdf7aa58d2183a5.\n\n## Findings and locators\n\nThe sworn declaration supports a motion for alternative service on Sol Ecom, not a Briver compliance certification. Paragraphs18–23 report subpoenas and responsive contact records from Cloudflare, Namecheap and former registrar Name.com. Paragraph29 reports Google's June6,2025 production. Paragraphs24–30 describe subsequent notices and correspondence. These are counsel's sworn accounts; actual subpoena and production instruments are not attached in this read scope. They establish a reported disclosure route, not provider shutdown instructions.\n\nExhibit6 reproduces December2024–January2025 correspondence. A respondent associated by counsel with drawnudes.io claims eSafety requirements and compliance, including alleged January2025 legislation (PDF19,21). The respondent seeks written SF requirements and discusses territorial restrictions (PDF19,22). This is a respondent's account and offer, not independent verification of Australian law, successful compliance, eSafety approval, or regulator-to-regulator coordination. No specific technical restriction or issuing provider is demonstrated by that exchange.\n\nThe complete declaration and exhibit do not supply Briver's required report or a named provider's takedown command. The city-announced cooperation and this separate subpoena route must retain separate source ancestry.\n\n## Read scope\n\nRoot read the complete six-page declaration through the web reader, including the filed stamp, allegations about service attempts and signature, and complete Exhibit6 (PDF18–23) through retained extraction. Root visually checked PDF1,4,6,21,22. PDF21's visually underlined eSafety reference has no URI annotation on that page in the retained PDF. That is not a finding about external links everywhere in the bundle. All61 pages were extracted for search; corporate-registration/profile/site exhibits outside those twelve pages were not substantively reviewed, and no whole-bundle-read claim is made. Residential addresses, private contact identifiers and unrelated personal profile details are not reproduced here.\n\n## Bounded acquisition log\n\nRoot searched seven queries: `\"Briver\" \"registrar\"`; `\"Sol Ecom\" \"registry\"`; `site.sfcityattorney.org \"deepfake\" \"registrars\"`; `site.sf.gov \"deepfake\" \"registrars\"`; `\"San Francisco\" \"registries\" \"June 2, 2025\"`; `\"Chiu\" \"Briver\" \"cooperate\"`; `\"San Francisco\" \"registrars\" \"Namecheap\" deepfake`. Most results were unrelated, including different Briver entities. They were not identity evidence. The named-company query was a probe, not prior proof of involvement.\n\nThe complete main body of Ars Technica's16August2024 article was read. It supplies a lead concerning the complaint's requested provider relief; root did not acquire the linked complaint or promote a report of requested relief into an order. The worker independently found the actual Tilak declaration; root's web reader and one ordinary originalGET succeeded. Search and retained-record read only. No provider takedown record was recovered in this bounded lane; that is not proof no provider acted. The separate worker filings.md records the court-gateway limitation and Briver-record search. No further acquisition branch was opened before the forest review.\n", "file": "research/sf-compliance-wave-2026-09-09-provider-disclosure.md" }, { "id": "sf-enforcement-wave-2026-09-09-judgment", "title": "Briver’s signed judgment within People v. Sol Ecom", "description": "Read and captured 9 September 2026. The instrument is a stipulated final judgment against Briver LLC in a multi-defendant case, not a judgment against every website or every captioned defendant. Its court signature and filing stamp are visible. The operative obligations are much more specific than the city announcement, while actual payment and continuing co", "html": "Read and captured 9 September 2026. The acquired instrument is a stipulated final judgment against Briver LLC in a multi-defendant case, not a judgment against every website or every captioned defendant. Its court signature and filing stamp are visible. The operative obligations are much more specific than the city announcement, while actual payment and continuing compliance remain unverified.
\nPrimary sources are the [City Attorney’s 2 June 2025 announcement](http[local research file] and a [copy of the signed judgment republished by Columbia Law School](http[local research file] The present SF.gov page contains no judgment attachment href in either its rendered anchors or retained structured body. The Columbia copy was discovered independently; it is not represented as a direct city or court download. The original PDF is image-only. All 13 PDF pages were read visually from retained renders, including both counterpart signature pages and the service certificate.
\nThe case is People of the State of California, acting by and through San Francisco City Attorney David Chiu v. Sol Ecom, Inc., et al., CGC-24-617237, San Francisco Superior Court. The instrument’s stipulating parties are the People and Briver LLC (PDF 2, opening paragraph and ¶4). Its caption includes other defendants; the announcement expressly says litigation continues against remaining defendants. It separately describes allegations that Briver operated two websites. Neither the ten-site outcome nor the full defendant roster can be assigned to this one judgment.
\nPDF 1 bears a FILED May 30, 2025 court stamp; the proposed designation is crossed out. PDF 12 / printed page 11 bears Judge Ethan P. Schulman’s signature dated May 29, 2025. PDF 13 contains Deputy Clerk Edward Santos’s certificate of electronic service on May 30, 2025 through File & ServeXpress. Paragraph 19 says the court enters judgment in the People’s favor. Together these show a court-signed, filed final-judgment copy, rather than an unsigned proposed instrument. A separate register-of-actions entry was not acquired.
\nThe judge-signature date matters: ¶6 defines the Effective Date as the date the court signs, so the visible operative reference date is 29 May 2025. The city announcement says approval occurred on May 30; preserve that as the office’s wording and distinguish it from signature, filing and service. The two Briver representatives sign “on behalf of Briver”: Yuannan Cai, May 27, 2025, and Xiulu Yuan, May 26, 2025 (PDF 12). The page does not identify their ownership percentages or corporate offices. PDF 11 has blank Briver and judge lines but contains the city counsel’s approved-as-form signature; PDF 12 is the signed counterpart. Paragraph 45 expressly allows counterpart signatures, so the blank lines on PDF 11 do not establish non-execution.
\nThe opening stipulation (PDF 2) states that no proof is taken and the judgment is not evidence of, or an admission concerning, alleged legal or factual issues; the stipulating parties waive appeal. Allegations and the negotiated obligations must remain distinct. Paragraphs 11–12 invoke Business and Professions Code 17200 et seq.; ¶13 preserves continuing jurisdiction under Code of Civil Procedure 664.6. The judgment predates AB 621’s enactment and is not evidence of application of its 2026 notice presumption.
\n| Actor and action | \nOperative scope and locator | \n
|---|---|
| Briver: permanent operating prohibition | \n¶14, PDF 3: Briver must not own, operate, or aid and abet any website using AI models to convert clothed images of identifiable individuals into nude or sexually explicit images of those individuals. The clause has no express consent exception; retain this wording rather than silently narrowing it to the announcement’s NCII label. It concerns this specified image-transformation activity, not every AI service. | \n
| Briver: non-transfer and domain inactivity | \n¶15, PDF 3: for five years after effectiveness, Briver shall not transfer registration for the defined Briver Websites and must maintain registration in inactive hold/lock status so the domains do not resolve in DNS and are unavailable publicly. Read with ¶17(b)–(c) and ¶18, which require no transfer of website ownership and documentary evidence of continued registration to Briver or an affiliate. The clause prohibits transfer; a transfer recipient is not an unresolved term. The five-year period does not expressly terminate ¶14’s permanent prohibition. | \n
| Briver: investigation information | \n¶16, PDF 4: within 30 days of effectiveness, provide all relevant registered website names; information about the image-processing API and affiliate programme, its source, sharing websites and their operators; a summary and explanation of the underlying technology, including the model and any training data; and other known similar-site operators, particularly in California. These are disclosure obligations, not acquired disclosures. | \n
| Briver: recurring sworn proof | \n¶17–18, PDF 4–5: an authorized representative must attest under penalty of perjury within 14 days, and annually for five years, to the operating prohibition, non-transfer of website ownership and continued inactive registrations. Documentary evidence must accompany each attestation. If late, the People must notify Briver; there is a 14-day cure period after receipt before non-submission counts as a violation. | \n
| City Attorney: bounded inspection | \n¶24, PDF 6: on reasonable request, representatives may inspect Briver and affiliated persons’ or entities’ domain-registration and related records, solely to determine compliance with this judgment. This is not an unlimited general data-access entitlement. | \n
| Court: enforcement and modification | \n¶26, PDF 6: a regularly noticed motion and hearing must precede the court’s finding of violation and additional civil penalties, reasonable enforcement/monitoring fees and costs, or further injunctive relief. ¶37–38, PDF 8, preserve modification for materially inaccurate or intentionally false information on a motion supported by adequate proof, and continuing jurisdiction. ¶43, PDF 9: modifications require a court order on noticed motion, or written party consent and court approval. | \n
The definition of Briver in ¶1 includes specified successors, agents, representatives, employees, members, owners, managers, directors, officers and affiliated persons/entities before and after dissolution. The distinct joint-and-several clause ¶25 names Briver and associated persons/entities as of August 14, 2024. Those provisions should not be collapsed into a finding that all case defendants, all future affiliates, or all service providers were adjudged liable.
\nThe defined Briver Websites in ¶2 include the domain strings porngen.art, undresser.ai and nudegenerator.ai, plus further Briver websites identified under ¶16(a). Those are website identifiers, not separate legal defendants. They were not visited. The announcement’s reference to allegations involving two websites and its broader ten-site outcome use different counting frames.
¶19–23, PDF 5: the judgment imposes USD 100,000 in civil penalties. Briver must pay the City Attorney’s Office USD 50,000 within 30 days of effectiveness, then five USD 10,000 instalments dated July 1, August 1, September 1, October 1 and November 1, 2025, by wire under city instructions. The schedule totals USD 100,000; it does not prove receipt. Paragraph 22 provides acceleration upon failure to receive payment within one business day of a deadline, with ten-percent annual interest on unpaid balance. The packet does not establish that default occurred or calculate an actual accrued balance.
\n¶35, PDF 8: except as otherwise specified, the parties waive settlement-related fees and costs. Accordingly, the USD 100,000 is not described as a combined fees/penalty award. Enforcement fees and costs are a separate conditional remedy under ¶26. No payment receipt, satisfaction of judgment or consumer compensation payment was acquired.
\n¶28–30, PDF 6–7: the release is tied to the litigation and effective-date conduct. It does not release the excluded tax/criminal/later-conduct matters, and ¶29 expressly preserves claims of individuals whose images were processed. The city’s settlement is not a release of those individuals’ private claims.
\nThe complete judgment names no independent registrar, registry, host, payment network, bank or authentication company as an actor ordered to terminate service. It directs Briver to disable and maintain its defined domains, provide information and prove compliance. File & ServeXpress appears in the clerk’s service certificate as the document-service mechanism, not as a shutdown target or provider compelled to enforce a content policy.
\nThe announcement reports that, as of June 2, ten targeted websites were offline or no longer accessible in California, and that certain unnamed domain registrars and registries were willing to cooperate during the investigation. It attributes that combined result to the ongoing lawsuit and investigation. This is office-reported outcome evidence, not an independently tested ten-site availability audit, a judgment commanding ten providers, or proof of permanent worldwide disappearance. The signed judgment supplies a concrete coercive instrument for Briver; it does not disclose what each cooperating provider did or under which instruction.
\nThe strongest defensible synthesis is therefore an enforceable company-specific settlement, recurring disclosure/attestation and inspection rights, and a separately reported wider shutdown/cooperation outcome. The missing discriminator is the actual attestation and domain evidence, disclosure response, payment/satisfaction record, or provider-specific instruction and response. No targeted service was visited, no external submission or outreach occurred, and no further acquisition branch was opened.
\nRoot and worker visually reread retained judgment-page-03.png, paragraph 15. The operative phrase is “shall not transfer the registration”. The earlier paraphrase reversed that prohibition and generated a nonexistent transfer-recipient gap. Both are withdrawn and the table now states the non-transfer duty. The original judgment PDF/hash is unchanged; this corrects our interpretation. Pre-correction note and candidate bytes are preserved under sf-integration-wave-2026-09-09/r25-before-domain-correction. The five-year inactivity/reporting distinction and the permanent paragraph 14 prohibition are unchanged.
", "path": "sf-enforcement-wave-2026-09-09/judgment.md", "sourceIds": [ "w25-sf-s-june-announcement", "w25-sf-s-briver-judgment" ], "markdown": "Read and captured 9 September 2026. **The acquired instrument is a stipulated final judgment against Briver LLC in a multi-defendant case, not a judgment against every website or every captioned defendant.** Its court signature and filing stamp are visible. The operative obligations are much more specific than the city announcement, while actual payment and continuing compliance remain unverified.\n\nPrimary sources are the City Attorney’s 2 June 2025 announcement (see source references) and a copy of the signed judgment republished by Columbia Law School (see source references) The present SF.gov page contains no judgment attachment href in either its rendered anchors or retained structured body. The Columbia copy was discovered independently; it is not represented as a direct city or court download. The original PDF is image-only. All **13 PDF pages** were read visually from retained renders, including both counterpart signature pages and the service certificate.\n\n## Identity, signature and legal stage\n\nThe case is **People of the State of California, acting by and through San Francisco City Attorney David Chiu v. Sol Ecom, Inc., et al., CGC-24-617237**, San Francisco Superior Court. The instrument’s stipulating parties are the People and **Briver LLC** (PDF 2, opening paragraph and ¶4). Its caption includes other defendants; the announcement expressly says litigation continues against remaining defendants. It separately describes allegations that Briver operated two websites. Neither the ten-site outcome nor the full defendant roster can be assigned to this one judgment.\n\nPDF 1 bears a **FILED May 30, 2025** court stamp; the proposed designation is crossed out. **PDF 12 / printed page 11** bears Judge **Ethan P. Schulman’s signature dated May 29, 2025**. PDF 13 contains Deputy Clerk Edward Santos’s certificate of electronic service on **May 30, 2025** through File & ServeXpress. Paragraph 19 says the court enters judgment in the People’s favor. Together these show a court-signed, filed final-judgment copy, rather than an unsigned proposed instrument. A separate register-of-actions entry was not acquired.\n\nThe judge-signature date matters: **¶6 defines the Effective Date as the date the court signs**, so the visible operative reference date is **29 May 2025**. The city announcement says approval occurred on May 30; preserve that as the office’s wording and distinguish it from signature, filing and service. The two Briver representatives sign “on behalf of Briver”: **Yuannan Cai, May 27, 2025**, and **Xiulu Yuan, May 26, 2025** (PDF 12). The page does not identify their ownership percentages or corporate offices. PDF 11 has blank Briver and judge lines but contains the city counsel’s approved-as-form signature; PDF 12 is the signed counterpart. Paragraph 45 expressly allows counterpart signatures, so the blank lines on PDF 11 do not establish non-execution.\n\nThe opening stipulation (PDF 2) states that no proof is taken and the judgment is not evidence of, or an admission concerning, alleged legal or factual issues; the stipulating parties waive appeal. Allegations and the negotiated obligations must remain distinct. Paragraphs 11–12 invoke **Business and Professions Code 17200 et seq.**; ¶13 preserves continuing jurisdiction under **Code of Civil Procedure 664.6**. The judgment predates AB 621’s enactment and is not evidence of application of its 2026 notice presumption.\n\n## What the judgment requires\n\n| Actor and action | Operative scope and locator |\n| --- | --- |\n| Briver: permanent operating prohibition | **¶14, PDF 3:** Briver must not own, operate, or aid and abet any website using AI models to convert clothed images of identifiable individuals into nude or sexually explicit images of those individuals. The clause has **no express consent exception**; retain this wording rather than silently narrowing it to the announcement’s NCII label. It concerns this specified image-transformation activity, not every AI service. |\n| Briver: non-transfer and domain inactivity | **¶15, PDF 3:** for five years after effectiveness, Briver shall not transfer registration for the defined Briver Websites and must maintain registration in inactive hold/lock status so the domains do not resolve in DNS and are unavailable publicly. Read with ¶17(b)–(c) and ¶18, which require no transfer of website ownership and documentary evidence of continued registration to Briver or an affiliate. The clause prohibits transfer; a transfer recipient is not an unresolved term. The five-year period does not expressly terminate ¶14’s permanent prohibition. |\n| Briver: investigation information | **¶16, PDF 4:** within 30 days of effectiveness, provide all relevant registered website names; information about the image-processing API and affiliate programme, its source, sharing websites and their operators; a summary and explanation of the underlying technology, including the model and any training data; and other known similar-site operators, particularly in California. These are disclosure obligations, not acquired disclosures. |\n| Briver: recurring sworn proof | **¶17–18, PDF 4–5:** an authorized representative must attest under penalty of perjury within 14 days, and annually for five years, to the operating prohibition, non-transfer of website ownership and continued inactive registrations. Documentary evidence must accompany each attestation. If late, the People must notify Briver; there is a 14-day cure period after receipt before non-submission counts as a violation. |\n| City Attorney: bounded inspection | **¶24, PDF 6:** on reasonable request, representatives may inspect Briver and affiliated persons’ or entities’ domain-registration and related records, solely to determine compliance with this judgment. This is not an unlimited general data-access entitlement. |\n| Court: enforcement and modification | **¶26, PDF 6:** a regularly noticed motion and hearing must precede the court’s finding of violation and additional civil penalties, reasonable enforcement/monitoring fees and costs, or further injunctive relief. **¶37–38, PDF 8**, preserve modification for materially inaccurate or intentionally false information on a motion supported by adequate proof, and continuing jurisdiction. **¶43, PDF 9:** modifications require a court order on noticed motion, or written party consent and court approval. |\n\nThe definition of **Briver** in ¶1 includes specified successors, agents, representatives, employees, members, owners, managers, directors, officers and affiliated persons/entities before and after dissolution. The distinct joint-and-several clause **¶25** names Briver and associated persons/entities **as of August 14, 2024**. Those provisions should not be collapsed into a finding that all case defendants, all future affiliates, or all service providers were adjudged liable.\n\nThe defined **Briver Websites** in ¶2 include the domain strings `porngen.art`, `undresser.ai` and `nudegenerator.ai`, plus further Briver websites identified under ¶16(a). Those are website identifiers, not separate legal defendants. They were not visited. The announcement’s reference to allegations involving two websites and its broader ten-site outcome use different counting frames.\n\n## Money and preserved claims\n\n**¶19–23, PDF 5:** the judgment imposes **USD 100,000 in civil penalties**. Briver must pay the City Attorney’s Office USD 50,000 within 30 days of effectiveness, then five USD 10,000 instalments dated **July 1, August 1, September 1, October 1 and November 1, 2025**, by wire under city instructions. The schedule totals USD 100,000; it does not prove receipt. Paragraph 22 provides acceleration upon failure to receive payment within one business day of a deadline, with ten-percent annual interest on unpaid balance. The packet does not establish that default occurred or calculate an actual accrued balance.\n\n**¶35, PDF 8:** except as otherwise specified, the parties waive settlement-related fees and costs. Accordingly, the USD 100,000 is not described as a combined fees/penalty award. Enforcement fees and costs are a separate conditional remedy under ¶26. No payment receipt, satisfaction of judgment or consumer compensation payment was acquired.\n\n**¶28–30, PDF 6–7:** the release is tied to the litigation and effective-date conduct. It does not release the excluded tax/criminal/later-conduct matters, and **¶29 expressly preserves claims of individuals whose images were processed**. The city’s settlement is not a release of those individuals’ private claims.\n\n## Named service-provider duties and observed outcome\n\nThe complete judgment names **no independent registrar, registry, host, payment network, bank or authentication company as an actor ordered to terminate service**. It directs Briver to disable and maintain its defined domains, provide information and prove compliance. File & ServeXpress appears in the clerk’s service certificate as the document-service mechanism, not as a shutdown target or provider compelled to enforce a content policy.\n\nThe **announcement reports** that, as of June 2, ten targeted websites were offline **or no longer accessible in California**, and that certain unnamed domain registrars and registries were willing to cooperate during the investigation. It attributes that combined result to the ongoing lawsuit and investigation. This is office-reported outcome evidence, not an independently tested ten-site availability audit, a judgment commanding ten providers, or proof of permanent worldwide disappearance. The signed judgment supplies a concrete coercive instrument for Briver; it does not disclose what each cooperating provider did or under which instruction.\n\nThe strongest defensible synthesis is therefore an enforceable company-specific settlement, recurring disclosure/attestation and inspection rights, and a separately reported wider shutdown/cooperation outcome. The missing discriminator is the actual attestation and domain evidence, disclosure response, payment/satisfaction record, or provider-specific instruction and response. No targeted service was visited, no external submission or outreach occurred, and no further acquisition branch was opened.\n\n## Dated correction — 9 September 2026\n\nRoot and worker visually reread retained judgment-page-03.png, paragraph 15. The operative phrase is “shall not transfer the registration”. The earlier paraphrase reversed that prohibition and generated a nonexistent transfer-recipient gap. Both are withdrawn and the table now states the non-transfer duty. The original judgment PDF/hash is unchanged; this corrects our interpretation. Pre-correction note and candidate bytes are preserved under sf-integration-wave-2026-09-09/r25-before-domain-correction. The five-year inactivity/reporting distinction and the permanent paragraph 14 prohibition are unchanged.\n", "file": "research/sf-enforcement-wave-2026-09-09-judgment.md" }, { "id": "sf-notice-wave-2026-09-09-notice-recovery", "title": "San Francisco notices: the original remains a specific gap", "description": "9 September 2026. Root stopped original-notice recovery after the planned credible alternative: ordinary GET of the current official news page and inspection of its server-rendered publication data. Statutory reading and independent notice-comparison are the remaining lanes of this bounded run.", "html": "9 September 2026. Root stopped original-notice recovery after the planned credible alternative: ordinary GET of the current official news page and inspection of its server-rendered publication data. Statutory reading and independent notice-comparison are the remaining lanes of this bounded run.
\nThe [San Francisco Chronicle report](http[local research file] dated17July2026, says the July16 notices cite the47-AG warning as earlier notice. It describes confidential app attachments unavailable to the newspaper, a28-day response request and threatened civil enforcement. This is a reported document connection; root has not acquired those notices or attachments. Root read complete article body121-179, with title/date. The newspaper supplies no original-notice download in the inspected links.
\n[WIRED's report](http[local research file] publishes named spokesperson statements: Google says five identified apps were suspended; Apple says three were removed, developer termination was in progress and four others were required to address violations. These are company accounts to a reporter, not independently observed removals, dated transaction logs or proof of payment disconnection. They are tied to the later San Francisco intervention, not replies to NAAG. Root read the full substantive body65-89 and update time; notice originals were not linked.
\n[TechCrunch](http[local research file] also publishes company responses; its update records comments added17July. This is an interview record, not a public counsel-to-city reply. Root read110-123 and followed its WIRED link. Conflicting comment-availability statements across articles may reflect publication/update times; no substantive contradiction is inferred merely from one outlet initially lacking comment.
\nNo user uploaded images, violating app use, account access or source contact was involved. Specific apps were not sought. Reported download/revenue estimates and the researchers' underlying measurements were not accepted or newly investigated.
\nThe official current news page http[local research file] returned200 through one ordinary GET after the prior lane's web reader returned only a sparse page. Retained official-news.html has244,772bytes and SHA256 d71369491cf1bc11e71ba2cf720113df8e4414203b684176aba09d6403d3fe47. The page's publicly delivered NEXT_DATA contains100 news entries out of a stated252, dated26August2026 through4December2024. Root inspected the structure, searched titles/metadata for Apple/Google/deepfake/nudify and examined dated entries bracketing July2026. No target notice appeared in that delivered subset. This is not a complete site, attachment repository or office-record search.
\nThe subset includes a June2025 website-enforcement announcement, a distinct earlier case not opened. Its existence is a lead, not evidence about the July2026 platform notices. Three targeted web queries for dated Chiu letters, the official media domain and the current official domain did not yield an acquired original. Broad results were largely unrelated; no claim of exhaustive result reading is made. Prior eighteen-query response searches were not repeated.
\nThe public page exposes its backing read API URL, but no further endpoint or private document enumeration was attempted. This run stops recovery here as planned. Missing records are the actual notices, confidential service-specific attachments, proof of delivery, replies and resulting directions. Their absence from the inspected news subset does not mean the office withheld, censored or never published them.
\nRoot read the complete current [California Civil Code1708.86](http[local research file] body56-186. Subdivision(c)(2) supplies a specific route: sufficient evidence from a depicted person or public prosecutor, identifying an enabling service relationship, plus failure to take necessary cessation steps within30days of receipt, with a court-extension provision. The service definition is primary-purpose based. The provision creates a presumption; it does not make a press release a judgment. The28-day reported demand must remain separate from this statutory30-day condition. Detailed elements, exceptions, remedies and enactment source are in statutory-mechanism.md; independent comparison with the earlier general letter is in notice-bridge-review.md.
\nThe evidence now supports separate levels: current statutory design; earlier collective demand; later notices and company responses as attributed reporting; actual merchant/payment outcomes still unverified. No graph edge should silently collapse those levels. Public research/local construction remain authorized; requests for nonpublic correspondence would require further authority.
", "path": "sf-notice-wave-2026-09-09/notice-recovery.md", "sourceIds": [ "w25-sf-s-chronicle", "w25-sf-s-wired", "w25-sf-s-official-news" ], "markdown": "9 September 2026. Root stopped original-notice recovery after the planned credible alternative: ordinary GET of the current official news page and inspection of its server-rendered publication data. Statutory reading and independent notice-comparison are the remaining lanes of this bounded run.\n\n## What was actually read\n\nThe San Francisco Chronicle report (see source references) dated17July2026, says the July16 notices cite the47-AG warning as earlier notice. It describes confidential app attachments unavailable to the newspaper, a28-day response request and threatened civil enforcement. This is a reported document connection; root has not acquired those notices or attachments. Root read complete article body121-179, with title/date. The newspaper supplies no original-notice download in the inspected links.\n\nWIRED's report (see source references) publishes named spokesperson statements: Google says five identified apps were suspended; Apple says three were removed, developer termination was in progress and four others were required to address violations. These are company accounts to a reporter, not independently observed removals, dated transaction logs or proof of payment disconnection. They are tied to the later San Francisco intervention, not replies to NAAG. Root read the full substantive body65-89 and update time; notice originals were not linked.\n\nTechCrunch (see source references) also publishes company responses; its update records comments added17July. This is an interview record, not a public counsel-to-city reply. Root read110-123 and followed its WIRED link. Conflicting comment-availability statements across articles may reflect publication/update times; no substantive contradiction is inferred merely from one outlet initially lacking comment.\n\nNo user uploaded images, violating app use, account access or source contact was involved. Specific apps were not sought. Reported download/revenue estimates and the researchers' underlying measurements were not accepted or newly investigated.\n\n## Official-publication recovery and its limit\n\nThe official current news page http[local research file] returned200 through one ordinary GET after the prior lane's web reader returned only a sparse page. Retained official-news.html has244,772bytes and SHA256 d71369491cf1bc11e71ba2cf720113df8e4414203b684176aba09d6403d3fe47. The page's publicly delivered **NEXT_DATA** contains100 news entries out of a stated252, dated26August2026 through4December2024. Root inspected the structure, searched titles/metadata for Apple/Google/deepfake/nudify and examined dated entries bracketing July2026. No target notice appeared in that delivered subset. This is not a complete site, attachment repository or office-record search.\n\nThe subset includes a June2025 website-enforcement announcement, a distinct earlier case not opened. Its existence is a lead, not evidence about the July2026 platform notices. Three targeted web queries for dated Chiu letters, the official media domain and the current official domain did not yield an acquired original. Broad results were largely unrelated; no claim of exhaustive result reading is made. Prior eighteen-query response searches were not repeated.\n\nThe public page exposes its backing read API URL, but no further endpoint or private document enumeration was attempted. This run stops recovery here as planned. Missing records are the actual notices, confidential service-specific attachments, proof of delivery, replies and resulting directions. Their absence from the inspected news subset does not mean the office withheld, censored or never published them.\n\n## Why the statutory lane matters independently\n\nRoot read the complete current California Civil Code1708.86 (see source references) body56-186. Subdivision(c)(2) supplies a specific route: sufficient evidence from a depicted person or public prosecutor, identifying an enabling service relationship, plus failure to take necessary cessation steps within30days of receipt, with a court-extension provision. The service definition is primary-purpose based. The provision creates a presumption; it does not make a press release a judgment. The28-day reported demand must remain separate from this statutory30-day condition. Detailed elements, exceptions, remedies and enactment source are in statutory-mechanism.md; independent comparison with the earlier general letter is in notice-bridge-review.md.\n\nThe evidence now supports separate levels: current statutory design; earlier collective demand; later notices and company responses as attributed reporting; actual merchant/payment outcomes still unverified. No graph edge should silently collapse those levels. Public research/local construction remain authorized; requests for nonpublic correspondence would require further authority.\n", "file": "research/sf-notice-wave-2026-09-09-notice-recovery.md" }, { "id": "sf-notice-wave-2026-09-09-statutory-mechanism", "title": "California’s third-party deepfake-service liability route", "description": "Research cutoff: 9 September 2026. The independent statutory lane read the complete official current Civil Code section and chaptered bill. No San Francisco letter was read in this lane; the conclusions below identify the instrument and its conditions, not whether a particular notice establishes them.", "html": "Research cutoff: 9 September 2026. The independent statutory lane read the complete official current Civil Code section and chaptered bill. No San Francisco letter was read in this lane; the conclusions below identify the instrument and its conditions, not whether a particular notice establishes them.
\nAB 621 became Chapter 673, Statutes of 2025, approved and filed on 13 October 2025. It amended Civil Code 1708.86, whose current official footer gives 1 January 2026 as its effective date. The official history and current section independently establish the identity; the suggested CalMatters path was not used as evidence. AB 621 section 2 says its provisions clarifying that a deepfake pornography service is subject to section 1708.86 are declaratory of existing law. That limited declaration should not be expanded into a claim that every new facilitation or notice provision previously existed. [Chaptered bill](http[local research file] [official history](http[local research file] [current Civil Code 1708.86](http[local research file]
\nThe substantive rule is 1708.86(b)(3): knowing facilitation or reckless aiding or abetting of the creation/disclosure conduct prohibited by (b)(1)–(2). The separate (c)(2) route creates a conditional presumption of violating that rule. It is not a free-standing administrative removal order, an automatic judgment, or a universal 30-day immunity period for conduct otherwise satisfying (b)(3).
\n| Stage | \nExact statutory condition or consequence | \n
|---|---|
| Covered service and underlying conduct | \nUnder (a)(3), a website, mobile application or other service must have creation of digitized sexually explicit material as its primary purpose. The definition does not itself say its primary purpose must be nonconsensual creation. For the third-party presumption, (c)(2)(A)(i) additionally requires evidence of a service that engages in conduct violating this section. Definitions (a)(4), (6)–(9) concern a depicted natural person and realistic digitization; this is not a blanket designation of every AI service or all pornography. | \n
| Qualified sender and sufficient evidence | \nA depicted individual or a public prosecutor supplies evidence sufficient to demonstrate that the recipient provides services enabling the ongoing operation of the violating deepfake pornography service: (c)(2)(A)(i). A general policy objection or the sender’s official title alone does not establish these predicates. | \n
| Delivery and required particulars | \nUnder (c)(2)(A)(ii), the sender must use a customer-service email or other designated process if prominently displayed on the recipient’s website. The evidence must name the deepfake service, generally describe the enabling services, and give sender contact information. Recipient use of that contact information is restricted to seeking additional report information necessary for the required steps. The three particulars do not replace the sufficient-evidence requirement. | \n
| Recipient’s clock | \nUnder (c)(2)(B)(i), the presumption attaches if the recipient fails to take all necessary steps to stop providing the enabling services within 30 days of receiving that evidence. This is receipt-based, not automatically measured from a press release, posting or letter date. The text says days, not business days. | \n
| Extension and enforcement | \nA court may extend the 30-day period where additional time is needed for an ongoing law-enforcement investigation or operation: (c)(2)(B)(ii). Public prosecutors may bring a civil enforcement action under (f)(2). A notice does not itself adjudicate evidentiary sufficiency, violation, exemptions or penalties. | \n
All table locators refer to the [official current section](http[local research file] Its service-enabling language does not expressly name Apple, Google, app stores, payment networks or banks. Applying it to one of those actors requires the relevant factual connection; the statutory text alone supplies no contract or service-specific finding.
\nThe statute’s exact formulation is a presumption of violation, not merely a presumption that the recipient has read a complaint. Paragraph (c)(1) separately presumes an owner/operator/controller’s knowledge of nonconsent unless express written consent is produced. Paragraph (c)(2) does not set out a separate administrative rebuttal or appeal procedure. Evidence Code 601 distinguishes conclusive and rebuttable presumptions, and two kinds of rebuttable burden; that general text does not by itself establish which burden this particular presumption affects. This packet therefore records the conditional presumption and ordinary civil-action route without claiming to have resolved its detailed litigation treatment or controlling case law. [Evidence Code 601](http[local research file]
\nUnder (a)(14), public prosecutors include the Attorney General, city attorneys, county counsel, district attorneys and other city/county prosecutors. A prevailing public prosecutor may receive injunctive/equitable relief, USD 25,000 per violation, or USD 50,000 for a malicious violation, reasonable attorney’s fees and other appropriate court relief; proof of actual harm to a depicted individual is unnecessary: (f)(2). These are litigation remedies, not amounts automatically assessed when the 30 days expire. A harmed depicted individual has a different remedy menu under (f)(1), including an alternative statutory award of USD 1,500–50,000 for the unauthorized acts concerning any one work, potentially up to USD 250,000 for malice. Do not merge those private damages with the prosecutor’s per-violation penalties. The action limitation is three years from actual or reasonably diligent discovery: (g).
\nMaterial limits remain. Subdivision (d) excludes specified law-enforcement, legal-proceeding, public-concern, news/political and constitutionally protected circumstances; public-figure status alone does not make material newsworthy. Subdivision (i) excludes federally protected conduct, expressly including section 230 protection. Subdivision (j) protects an ISP’s mere transmission, routing or provision of access to third-party content, and preserves the referenced state net-neutrality requirements. Conversely, the disclaimers described in (e) are not defenses. These provisions prevent a blanket reading while leaving application to particular services and conduct unresolved.
\nThe next discriminator belongs to the notice lane: the actual named service, enabling-service allegation, supporting evidence, designated delivery channel, receipt date, any court extension, recipient response and any adjudication. An older generalized letter cannot automatically satisfy the current presumption’s specific predicates. The present packet stops before that application and before legislative-origins research.
\nDirect ordinary GETs successfully retained the official current section, chaptered bill and history HTML. Full operative text was read through the official reader and checked against retained HTML; capture is not a claim that every navigation/script byte was interpreted. Exact metadata and failures are in statutory-captures.json. Current-code HTML SHA-256: 319afffe89143fb923e1bc5de56f727d3da6d80e0b4d47ecd51af04d7a3867fb. Chaptered-bill HTML: 9c3f9f8bd1019fd5c941781001d3bc2083745a3c09c4546768172c6e3d91d3e8. History HTML: 4467ecb0b50d8942d1bcc9366f3015f524f38d4ef8767f844839bf0ca42ea40a.
The analysis index was inspected. Its observed 10 September 2025 Assembly Floor Analysis download form returned HTTP 500 on one ordinary submission; that route was closed without retry and the analysis was not used as substantive evidence. This was a server response, not an administrative access-policy refusal or automatic sandbox approval rejection. No original analysis PDF or statutory bill PDF is claimed. No outreach, canonical case edit or other-law research branch occurred.
", "path": "sf-notice-wave-2026-09-09/statutory-mechanism.md", "sourceIds": [ "w25-sf-s-code", "w25-sf-s-chapter" ], "markdown": "Research cutoff: 9 September 2026. The independent statutory lane read the complete official current Civil Code section and chaptered bill. No San Francisco letter was read in this lane; the conclusions below identify the instrument and its conditions, not whether a particular notice establishes them.\n\nAB 621 became Chapter 673, Statutes of 2025, approved and filed on **13 October 2025**. It amended Civil Code **1708.86**, whose current official footer gives **1 January 2026** as its effective date. The official history and current section independently establish the identity; the suggested CalMatters path was not used as evidence. AB 621 section 2 says its provisions clarifying that a deepfake pornography service is subject to section 1708.86 are declaratory of existing law. That limited declaration should not be expanded into a claim that every new facilitation or notice provision previously existed. Chaptered bill (see source references) official history (see source references) current Civil Code 1708.86 (see source references)\n\nThe substantive rule is **1708.86(b)(3)**: knowing facilitation or reckless aiding or abetting of the creation/disclosure conduct prohibited by (b)(1)–(2). The separate **(c)(2)** route creates a conditional presumption of violating that rule. It is not a free-standing administrative removal order, an automatic judgment, or a universal 30-day immunity period for conduct otherwise satisfying (b)(3).\n\n| Stage | Exact statutory condition or consequence |\n| --- | --- |\n| Covered service and underlying conduct | Under (a)(3), a website, mobile application or other service must have creation of digitized sexually explicit material as its **primary purpose**. The definition does not itself say its primary purpose must be nonconsensual creation. For the third-party presumption, (c)(2)(A)(i) additionally requires evidence of a service that engages in conduct violating this section. Definitions (a)(4), (6)–(9) concern a depicted natural person and realistic digitization; this is not a blanket designation of every AI service or all pornography. |\n| Qualified sender and sufficient evidence | A depicted individual or a public prosecutor supplies evidence sufficient to demonstrate that the recipient provides services enabling the ongoing operation of the violating deepfake pornography service: (c)(2)(A)(i). A general policy objection or the sender’s official title alone does not establish these predicates. |\n| Delivery and required particulars | Under (c)(2)(A)(ii), the sender must use a customer-service email or other designated process **if prominently displayed** on the recipient’s website. The evidence must name the deepfake service, generally describe the enabling services, and give sender contact information. Recipient use of that contact information is restricted to seeking additional report information necessary for the required steps. The three particulars do not replace the sufficient-evidence requirement. |\n| Recipient’s clock | Under (c)(2)(B)(i), the presumption attaches if the recipient fails to take all necessary steps to stop providing the enabling services **within 30 days of receiving that evidence**. This is receipt-based, not automatically measured from a press release, posting or letter date. The text says days, not business days. |\n| Extension and enforcement | A **court** may extend the 30-day period where additional time is needed for an ongoing law-enforcement investigation or operation: (c)(2)(B)(ii). Public prosecutors may bring a civil enforcement action under (f)(2). A notice does not itself adjudicate evidentiary sufficiency, violation, exemptions or penalties. |\n\nAll table locators refer to the official current section (see source references) Its service-enabling language does not expressly name Apple, Google, app stores, payment networks or banks. Applying it to one of those actors requires the relevant factual connection; the statutory text alone supplies no contract or service-specific finding.\n\nThe statute’s exact formulation is a **presumption of violation**, not merely a presumption that the recipient has read a complaint. Paragraph (c)(1) separately presumes an owner/operator/controller’s knowledge of nonconsent unless express written consent is produced. Paragraph (c)(2) does not set out a separate administrative rebuttal or appeal procedure. Evidence Code 601 distinguishes conclusive and rebuttable presumptions, and two kinds of rebuttable burden; that general text does not by itself establish which burden this particular presumption affects. This packet therefore records the conditional presumption and ordinary civil-action route without claiming to have resolved its detailed litigation treatment or controlling case law. Evidence Code 601 (see source references)\n\nUnder **(a)(14)**, public prosecutors include the Attorney General, city attorneys, county counsel, district attorneys and other city/county prosecutors. A prevailing public prosecutor may receive injunctive/equitable relief, **USD 25,000 per violation**, or **USD 50,000 for a malicious violation**, reasonable attorney’s fees and other appropriate court relief; proof of actual harm to a depicted individual is unnecessary: **(f)(2)**. These are litigation remedies, not amounts automatically assessed when the 30 days expire. A harmed depicted individual has a different remedy menu under **(f)(1)**, including an alternative statutory award of USD 1,500–50,000 for the unauthorized acts concerning any one work, potentially up to USD 250,000 for malice. Do not merge those private damages with the prosecutor’s per-violation penalties. The action limitation is three years from actual or reasonably diligent discovery: **(g)**.\n\nMaterial limits remain. Subdivision **(d)** excludes specified law-enforcement, legal-proceeding, public-concern, news/political and constitutionally protected circumstances; public-figure status alone does not make material newsworthy. Subdivision **(i)** excludes federally protected conduct, expressly including section 230 protection. Subdivision **(j)** protects an ISP’s mere transmission, routing or provision of access to third-party content, and preserves the referenced state net-neutrality requirements. Conversely, the disclaimers described in **(e)** are not defenses. These provisions prevent a blanket reading while leaving application to particular services and conduct unresolved.\n\nThe next discriminator belongs to the notice lane: the actual named service, enabling-service allegation, supporting evidence, designated delivery channel, receipt date, any court extension, recipient response and any adjudication. An older generalized letter cannot automatically satisfy the current presumption’s specific predicates. The present packet stops before that application and before legislative-origins research.\n\n## Custody and reading scope\n\nDirect ordinary GETs successfully retained the official current section, chaptered bill and history HTML. Full operative text was read through the official reader and checked against retained HTML; capture is not a claim that every navigation/script byte was interpreted. Exact metadata and failures are in statutory-captures.json. Current-code HTML SHA-256: `319afffe89143fb923e1bc5de56f727d3da6d80e0b4d47ecd51af04d7a3867fb`. Chaptered-bill HTML: `9c3f9f8bd1019fd5c941781001d3bc2083745a3c09c4546768172c6e3d91d3e8`. History HTML: `4467ecb0b50d8942d1bcc9366f3015f524f38d4ef8767f844839bf0ca42ea40a`.\n\nThe analysis index was inspected. Its observed 10 September 2025 Assembly Floor Analysis download form returned HTTP 500 on one ordinary submission; that route was closed without retry and the analysis was not used as substantive evidence. This was a server response, not an administrative access-policy refusal or automatic sandbox approval rejection. No original analysis PDF or statutory bill PDF is claimed. No outreach, canonical case edit or other-law research branch occurred.\n", "file": "research/sf-notice-wave-2026-09-09-statutory-mechanism.md" }, { "id": "signal-protocol-wave-2026-09-09-correction-protocol", "title": "Correction is supported in the client; reversal of consequences remains unverified", "description": "Bounded public-code investigation,9 September2026. Acquisition stopped after six Python source files and repository metadata. This is static inspection of Meta's public reference client at commit 73a742c45c35d711e94c8bf92254df1a2d9c9f01; no package was installed, no code was run, no authentication was used and no private signal API was queried. Lantern's use", "html": "Bounded public-code investigation,9 September2026. Acquisition stopped after six Python source files and repository metadata. This is static inspection of Meta's public reference client at commit 73a742c45c35d711e94c8bf92254df1a2d9c9f01; no package was installed, no code was run, no authentication was used and no private signal API was queried. Lantern's use of Meta's hosted ThreatExchange does not establish that any participant deploys this particular client or commit.
\nThe [update wrapper](http[local research file] recognizes deletion markers. Its file-store implementation removes the matching local entry when applying such an update (lines396–413). Thus the inspected implementation provides a path for removing a previously fetched record, rather than treating received signals as intrinsically permanent.
\nIts checkpoint documentation describes a90-day history for deletion records and an85-day stale threshold (lines218–253). That refers to change-history availability, not Lantern's retention period for underlying signals, personal data or financial records. The hosted API's present retention behaviour was not independently tested. The code also guards against applying a stale incremental update.
\nThe [current API adapter](http[local research file] applies the85-day threshold and maps a deletion marker to an absent record in its delta. It preserves contributor-specific opinions: explicit classifications and selected reactions can produce positive, negative or investigation-seed categories (lines105–180). These generic software categories are not Lantern's legal findings or proof that a receiving company uses a particular threshold.
\nThe fetch command.
\nThe [opinion model](http[local research file] distinguishes negative assessments, investigation leads and positive classifications. Its aggregator can represent conflicting categories as disputed (lines129–171). That capability does not establish that an interface displays the dispute to every investigator, that it suppresses action, or that an earlier account decision is reconsidered.
\nThe [API wrapper](http[local research file] requests updated data including deletion/status fields and exposes reactions (lines271–308,443–520). Nothing was submitted through those methods. Their existence is evidence of the published client interface, not an observed production transaction.
\nThe record now argues against describing the signal exchange as technically incapable of correction. It leaves three separate questions: whether programme obligations require the relevant update path; whether recipients actually process updates correctly and promptly; and whether a corrected signal triggers review of an already completed content, account or payment decision.
\nA local dataset refresh is not itself an account restoration. The inspected code addresses fetched signal state and opinions. It supplies no evidence of Lantern-specific reversal of Cash App closures, Xbox restrictions or any Steam action. This is a boundary on the current evidence, not a repository-wide claim that no related software could implement restoration.
\nThe2024 Lantern report's appeal/removal/TechCoalition-notification requirement remains the programme-level starting point. The next discriminating artifact is an integration contract or anonymized operational trace connecting a correction identifier to recipient acknowledgement, refreshed data, case reopening and disposition. A trace that ends at data deletion would leave the consequential remedy question unresolved.
\ncaptures.json records exact public URLs, commit, bytes and SHA256 for repository-main-ref.json, the complete nontruncated repository-tree.json and six source files. Retained source is not equivalent to full-file review. Root read targeted portions of threat_updates.py1–75,218–343,380–413; fb_threatexchange_api.py1–180,199–217,260–405; api.py271–308,443–520; fetch_cmd.py168–257; fetch_state.py29–51,55–181,271–276. descriptor.py was captured and search-inspected, not substantively reviewed. The tree was used to locate files, not treated as proof of their behaviour. Independent static review is recorded separately when returned.
\nThe Meta deletion documentation and an initially discovered GitHub issue page failed in the web reader. No mirror was used to establish their contents. The ordinary public GitHub repository/API and pinned raw source provided the useful substitute. No repeated attempt, private-access request or live ThreatExchange query occurred. A discovered issue title was not used as evidence of a deployed Lantern defect.
", "path": "signal-protocol-wave-2026-09-09/correction-protocol.md", "sourceIds": [], "markdown": "Bounded public-code investigation,9 September2026. Acquisition stopped after six Python source files and repository metadata. This is static inspection of Meta's public reference client at commit 73a742c45c35d711e94c8bf92254df1a2d9c9f01; no package was installed, no code was run, no authentication was used and no private signal API was queried. Lantern's use of Meta's hosted ThreatExchange does not establish that any participant deploys this particular client or commit.\n\n## What the inspected code establishes\n\nThe update wrapper (see source references) recognizes deletion markers. Its file-store implementation removes the matching local entry when applying such an update (lines396–413). Thus the inspected implementation provides a path for removing a previously fetched record, rather than treating received signals as intrinsically permanent.\n\nIts checkpoint documentation describes a90-day history for deletion records and an85-day stale threshold (lines218–253). That refers to change-history availability, not Lantern's retention period for underlying signals, personal data or financial records. The hosted API's present retention behaviour was not independently tested. The code also guards against applying a stale incremental update.\n\nThe current API adapter (see source references) applies the85-day threshold and maps a deletion marker to an absent record in its delta. It preserves contributor-specific opinions: explicit classifications and selected reactions can produce positive, negative or investigation-seed categories (lines105–180). These generic software categories are not Lantern's legal findings or proof that a receiving company uses a particular threshold.\n\nThe fetch command.\n\nThe opinion model (see source references) distinguishes negative assessments, investigation leads and positive classifications. Its aggregator can represent conflicting categories as disputed (lines129–171). That capability does not establish that an interface displays the dispute to every investigator, that it suppresses action, or that an earlier account decision is reconsidered.\n\nThe API wrapper (see source references) requests updated data including deletion/status fields and exposes reactions (lines271–308,443–520). Nothing was submitted through those methods. Their existence is evidence of the published client interface, not an observed production transaction.\n\n## Meaning for this investigation\n\nThe record now argues against describing the signal exchange as technically incapable of correction. It leaves three separate questions: whether programme obligations require the relevant update path; whether recipients actually process updates correctly and promptly; and whether a corrected signal triggers review of an already completed content, account or payment decision.\n\nA local dataset refresh is not itself an account restoration. The inspected code addresses fetched signal state and opinions. It supplies no evidence of Lantern-specific reversal of Cash App closures, Xbox restrictions or any Steam action. This is a boundary on the current evidence, not a repository-wide claim that no related software could implement restoration.\n\nThe2024 Lantern report's appeal/removal/TechCoalition-notification requirement remains the programme-level starting point. The next discriminating artifact is an integration contract or anonymized operational trace connecting a correction identifier to recipient acknowledgement, refreshed data, case reopening and disposition. A trace that ends at data deletion would leave the consequential remedy question unresolved.\n\n## Custody and scope\n\ncaptures.json records exact public URLs, commit, bytes and SHA256 for repository-main-ref.json, the complete nontruncated repository-tree.json and six source files. Retained source is not equivalent to full-file review. Root read targeted portions of threat_updates.py1–75,218–343,380–413; fb_threatexchange_api.py1–180,199–217,260–405; api.py271–308,443–520; fetch_cmd.py168–257; fetch_state.py29–51,55–181,271–276. descriptor.py was captured and search-inspected, not substantively reviewed. The tree was used to locate files, not treated as proof of their behaviour. Independent static review is recorded separately when returned.\n\nThe Meta deletion documentation and an initially discovered GitHub issue page failed in the web reader. No mirror was used to establish their contents. The ordinary public GitHub repository/API and pinned raw source provided the useful substitute. No repeated attempt, private-access request or live ThreatExchange query occurred. A discovered issue title was not used as evidence of a deployed Lantern defect.\n", "file": "research/signal-protocol-wave-2026-09-09-correction-protocol.md" }, { "id": "signal-protocol-wave-2026-09-09-protocol-independent-review", "title": "Independent technical review: public ThreatExchange update client", "description": "Reviewed 9 September 2026. Static inspection only: no captured module was imported or executed, no API or credentials were used, and no new source was acquired. All six Python source hashes match captures.json at pinned commit 73a742c45c35d711e94c8bf92254df1a2d9c9f01. This review concerns available software paths, not Lantern's deployment or the live server.", "html": "Reviewed 9 September 2026. Static inspection only: no captured module was imported or executed, no API or credentials were used, and no new source was acquired. All six retained Python source hashes match captures.json at pinned commit 73a742c45c35d711e94c8bf92254df1a2d9c9f01. This review concerns available software paths, not Lantern's deployment or the live server.
\nThe proposed narrow conclusion is supported: the public client contains paths for processing deletion/update records and representing negative opinions. ThreatUpdateJSON.should_delete calls a record a tombstone and treats either should_delete or the absence of descriptors as a removal condition. ThreatUpdateFileStore._apply_updates_impl removes the key from its local state for such records, otherwise replaces/adds the record, and serializes the result. The wrapper's apply_updates invokes that implementation for a nonempty delta. This is a concrete local-state deletion path in the source, not a record of any deletion actually delivered to a Lantern participant. See [threat_updates.py lines45-49](http[local research file] lines316-333 and396-413.
\nThe exchange adapter separately maps a tombstone to None, NON_MALICIOUS to a negative category, and a DISAGREE_WITH_TAGS reaction to an implicit negative category under its precedence rules. It retains the resulting opinions in the indicator record. An explicit opinion from the same owner takes precedence over the implicit reaction; these routines do not promise to retain every contradictory statement unchanged. A negative category is metadata, not proof of a false accusation, a successful appeal or account restoration. The adapter fetches threat_updates and emits a delta keyed by threat type and indicator. See [fb_threatexchange_api.py lines105-180](http[local research file] and292-325; fetch_state.py lines79-117 define opinion categories.
\nThe 90-day statement is documentation of deletion-history retention in client comments/docstrings. It is not a verified observation of current backend retention, and especially not Lantern's signal-retention policy. Both checkpoint variants use a strict older-than-85-days threshold, providing a five-day margin against the documented loss of deletion records. The relevant age is the stored last_fetch_time, not each signal's creation date. See threat_updates.py lines218-253 and [fb_threatexchange_api.py lines55-73](http[local research file]
\nThe CLI call chain is present to the storage interface: execute_for_collab obtains a checkpoint through _verify_store_and_checkpoint; if stale, the latter calls store.clear(collab) and returns None. execute_for_collab then passes that None into the exchange adapter's fetch_iter, which requests updates without a starting checkpoint, and merges returned deltas into the store. Thus the code requests clearing and begins a fresh fetch path; do not claim that this review observed a complete successful refetch. The command can encounter failures or limits. See [fetch_cmd.py lines184-257](http[local research file] fb_threatexchange_api.py lines292-325 and api.py lines271-308.
\nA concrete call-path limit remains: the captured fetch_state.py defines abstract merge, flush and clear methods, and the concrete configured CLI storage implementation is not among these six files. The earlier ThreatUpdateFileStore deletion routine and the exchange-adapter/CLI path must not be spliced into one claimed execution trace. This does not erase the demonstrated source-level deletion capability; it limits any statement that the precise deployed CLI persistence path has been independently traced end to end. No broader bug hunt or execution is needed for this inquiry's conclusion.
\nThe missing causal links remain deployment and downstream obligations: whether a particular Lantern participant runs this client/version, ingests the updates promptly, propagates changes to its own exported copies and enforcement systems, or reopens a prior decision. Public client capability cannot establish those facts. The supported refinement is that a public technical mechanism for updates, deletion markers and negative opinions exists; its availability does not demonstrate Lantern-wide correction or remedy performance.
\nInspected scope: threat_updates.py selected lines30-62,195-265,270-413; fb_threatexchange_api.py40-217,265-368; fetch_cmd.py95-295; fetch_state.py1-135,185-280; api.py271-308; descriptor.py scoped status/deletion matches only. Source paths and byte hashes remain in captures.json. No package, case or acquisition record was changed.
", "path": "signal-protocol-wave-2026-09-09/protocol-independent-review.md", "sourceIds": [], "markdown": "Reviewed 9 September 2026. Static inspection only: no captured module was imported or executed, no API or credentials were used, and no new source was acquired. All six retained Python source hashes match captures.json at pinned commit 73a742c45c35d711e94c8bf92254df1a2d9c9f01. This review concerns available software paths, not Lantern's deployment or the live server.\n\nThe proposed narrow conclusion is supported: the public client contains paths for processing deletion/update records and representing negative opinions. ThreatUpdateJSON.should_delete calls a record a tombstone and treats either should_delete or the absence of descriptors as a removal condition. ThreatUpdateFileStore._apply_updates_impl removes the key from its local state for such records, otherwise replaces/adds the record, and serializes the result. The wrapper's apply_updates invokes that implementation for a nonempty delta. This is a concrete local-state deletion path in the source, not a record of any deletion actually delivered to a Lantern participant. See threat_updates.py lines45-49 (see source references) lines316-333 and396-413.\n\nThe exchange adapter separately maps a tombstone to None, NON_MALICIOUS to a negative category, and a DISAGREE_WITH_TAGS reaction to an implicit negative category under its precedence rules. It retains the resulting opinions in the indicator record. An explicit opinion from the same owner takes precedence over the implicit reaction; these routines do not promise to retain every contradictory statement unchanged. A negative category is metadata, not proof of a false accusation, a successful appeal or account restoration. The adapter fetches threat_updates and emits a delta keyed by threat type and indicator. See fb_threatexchange_api.py lines105-180 (see source references) and292-325; fetch_state.py lines79-117 define opinion categories.\n\nThe 90-day statement is documentation of deletion-history retention in client comments/docstrings. It is not a verified observation of current backend retention, and especially not Lantern's signal-retention policy. Both checkpoint variants use a strict older-than-85-days threshold, providing a five-day margin against the documented loss of deletion records. The relevant age is the stored last_fetch_time, not each signal's creation date. See threat_updates.py lines218-253 and fb_threatexchange_api.py lines55-73 (see source references)\n\nThe CLI call chain is present to the storage interface: execute_for_collab obtains a checkpoint through _verify_store_and_checkpoint; if stale, the latter calls store.clear(collab) and returns None. execute_for_collab then passes that None into the exchange adapter's fetch_iter, which requests updates without a starting checkpoint, and merges returned deltas into the store. Thus the code requests clearing and begins a fresh fetch path; do not claim that this review observed a complete successful refetch. The command can encounter failures or limits. See fetch_cmd.py lines184-257 (see source references) fb_threatexchange_api.py lines292-325 and api.py lines271-308.\n\nA concrete call-path limit remains: the captured fetch_state.py defines abstract merge, flush and clear methods, and the concrete configured CLI storage implementation is not among these six files. The earlier ThreatUpdateFileStore deletion routine and the exchange-adapter/CLI path must not be spliced into one claimed execution trace. This does not erase the demonstrated source-level deletion capability; it limits any statement that the precise deployed CLI persistence path has been independently traced end to end. No broader bug hunt or execution is needed for this inquiry's conclusion.\n\nThe missing causal links remain deployment and downstream obligations: whether a particular Lantern participant runs this client/version, ingests the updates promptly, propagates changes to its own exported copies and enforcement systems, or reopens a prior decision. Public client capability cannot establish those facts. The supported refinement is that a public technical mechanism for updates, deletion markers and negative opinions exists; its availability does not demonstrate Lantern-wide correction or remedy performance.\n\nInspected scope: threat_updates.py selected lines30-62,195-265,270-413; fb_threatexchange_api.py40-217,265-368; fetch_cmd.py95-295; fetch_state.py1-135,185-280; api.py271-308; descriptor.py scoped status/deletion matches only. Source paths and byte hashes remain in captures.json. No package, case or acquisition record was changed.\n", "file": "research/signal-protocol-wave-2026-09-09-protocol-independent-review.md" }, { "id": "signal-sharing-wave-2026-09-09-checkstep-sightengine", "title": "Checkstep / Sightengine: classification, enforcement and reuse are separate interfaces", "description": "Bounded public-record packet, 9 September 2026. Six core records, including the existing government-published review. No customer data, accounts or moderation APIs were accessed. These are published specifications and standard terms, not proof of a particular customer's configuration, signed contract or enforcement history.", "html": "Bounded public-record packet, 9 September 2026. Six core records, including the existing government-published review. No customer data, accounts or moderation APIs were accessed. These are published specifications and standard terms, not proof of a particular customer's configuration, signed contract or enforcement history.
\nPUBLIC's example is documented as an integration of moderation vendors. The observed path is a customer submitting content for analysis, Sightengine returning classification scores, Checkstep applying the customer's configured policy, and a decision event returning to that customer's platform. The records do not establish a shared list of restricted people or automatic propagation of one platform's ban to another.
\nThe distinction does not make the interface inconsequential: it can carry instructions that remove content or suspend an account. It also does not establish zero reuse. The published contracts contain two different machine-learning permissions: Checkstep's customer-data permission with an opt-out, and Sightengine's voluntary feedback-data route with a separate controller role and retention exception. Permission and actual use remain separate evidence states.
\n1. PUBLIC's framing and its source. The [Data Landscape Review 2024–2025](http[local research file] §6.1.1 / Case study 4, describes this as signal sharing that lets a moderation provider use classification outputs without acquiring another provider's proprietary datasets. Footnote 35 points directly to Sightengine's coauthored guide, labelled 2025. This is the report's characterization of the vendor example, rather than an independently observed exchange or a cross-platform enforcement agreement. Its phrase about no transfer of raw data must not be expanded into a claim that no submitted content is processed outside the customer's platform.
\n2. The coauthored interface account. [Sightengine, Ultimate Guide to GenAI Moderation—Part 2](http[local research file] “Detecting and moderating…” §§1.1, 1.3, 2.1–2.2, describes media submitted by URL or upload with selected models. Results include AI-generation/deepfake scores between 0 and 1. Sightengine recommends thresholds; users choose how scores map to their policies. Checkstep integrates those outputs, supports multiple scanning providers, and routes possible violations to automated decisions or human review. It returns results and enforcement decisions to the customer. The guide also describes appeals and community reports. Its privacy language about reference-based integration describes an architecture; it is not a sufficient basis for a universal no-storage claim. The current page displays “Guide, 2026” while its introduction anticipates 2025; the report cites it as 2025. No exact original publication date or historical version identity is established here.
\n3. Checkstep's actual integration specification. [Standard Integration](http[local research file] “Sending contents,” “Access delegation” and “Receiving Webhooks,” accepts content IDs, text or media references, author identifiers and optional metadata. Delegated access enables content analysis and display to moderators; examples include cloud-storage reads and streaming bytes. Returned data identifies content, policy, severity and confidence. analysed-content supports provisional withholding while a decision is pending. A decision event instructs the customer's integration to take down, restore or keep content down after an appeal. Human author-decision events can suspend/terminate an account, optionally remove its content, or restore it. The customer must implement the webhook handler for the decision to affect its platform. Thus Sightengine's score is not itself a takedown, but the configured Checkstep workflow can produce an actionable instruction. incident-closed reports appeal outcomes. These are documented capabilities, with no observed customer configuration or successful appeal in this run.
4. Checkstep's reuse, responsibility and retention terms. [Checkstep Terms of Service](http[local research file] definitions and §§5.1–5.5, 13.3(C), 14, retain customer ownership while permitting use of Customer Data for machine learning that develops service features. Section 5.5 treats this as a customer instruction and offers an email opt-out; the read clause does not state an anonymization condition or prove pooled training across customers. The terms contemplate archives/backups. On termination they require disposal of data held, subject to a backup-delivery request made within 30 days and legal retention; 30 days is the request window, not a universal deletion deadline. Section 14 identifies the contracting party as controller and Checkstep as processor, generally processing under agreed instructions. Subprocessors require protective contracts and notification with a five-working-day objection window. Data-subject requests are referred to the contracting party. These published terms do not establish which provider/terms a specific customer accepted or whether it opted out.
\n5. Sightengine's subscriber agreement. [Terms of Service](http[local research file] §§1, 3.5–3.6, 4.1–4.6, 9.1–9.3, identifies the operator as Kozelo SAS. Routine Subscriber Data is licensed for service provision and instructed interoperability. Section 9.3 first grants perpetual, irrevocable use permission for general suggestions/corrections. Its separate nonexclusive licence covers voluntary raw-media submissions, with or without labels, through the Feedback API or similar mechanism for model development, testing and training; the subscriber warrants the necessary rights and consents. The raw-media licence must not inherit the first sentence’s adjectives merely by proximity. This is not proof that normal scanning automatically supplies training media, or that Checkstep sends its customers' data into this route. Sightengine can also send service-compliance notices and suspend its subscriber's API access. That power concerns the provider/customer service relationship and must not be confused with direct control of an end user's account on an unrelated platform.
\n6. Sightengine's feedback exception and correction route. [Data Processing Addendum](http[local research file] PDF 2 / definition, PDF 3–4 / §§2.1–2.3 and PDF 6–7 / §§4–5, normally assigns Sightengine a processor role; a subscriber that is itself a processor must have the controller's authority. For voluntarily submitted Feedback Data, §2.3 instead makes Sightengine an independent controller, subject to privacy law and an appropriate lawful basis. Ordinary deletion instructions are to be completed as soon as reasonably practicable, within 90 days subject to legal retention; expiry provisions also describe a recovery period of up to 30 days. Section 4.4 exempts Feedback Data from those automatic return/deletion provisions and permits perpetual retention subject to its privacy policy and applicable law. Section 5 enables access, correction, restriction, export and assistance with data-subject requests, normally through the subscriber. This is not a finding that privacy rights disappear or that every correction propagates into a trained model.
\n| Interface | \nWhat can travel | \nWhat this packet does not establish | \n
|---|---|---|
| Customer → Checkstep → selected scanner | \nContent/reference access, selected model inputs and identifiers needed for that customer's moderation | \nAn exchange of the scanner's proprietary training corpus | \n
| Sightengine → Checkstep | \nDetection scores/classification results for submitted media | \nA decision binding every other platform or a shared offender identity | \n
| Checkstep → customer platform | \nPolicy-linked content/account decisions, provisional results and appeal outcomes | \nA deployed handler, its actual settings, enforcement against a particular user, or an independent appeal body | \n
| Customer data / voluntary feedback → model development | \nPublished permissions for two distinct reuse routes | \nWhether a customer used either route, whether learning is pooled, the actual training corpus or a resulting cross-customer restriction | \n
The strongest rival to a cross-platform restriction interpretation is ordinary service composition: one vendor supplies classifiers and another supplies policy/workflow software for each customer's site. The explicit contract permissions leave a narrower cross-customer learning possibility open, but that is not the same mechanism as transmitting a ban or restriction signal. Affected people can face automated consequences even where formal policy choice remains with the customer; the discriminating evidence is the configured rule and its execution, not the mere presence of a classifier.
\nOriginal captures and SHA256 values are recorded in cs-captures.json. cs-standard.html, cs-sightengine-guide.html, cs-checkstep-terms.html, cs-sightengine-terms.html and cs-sightengine-dpa.pdf were acquired by ordinary HTTP 200 requests on 9 September 2026. The report uses the existing ../osdi-comparison-wave-2026-09-09/data-landscape-2025.html. Derived text copies support precise reading; cs-sightengine-dpa-extract.txt preserves the inspected PDF pages. Source states are read, with targeted scopes in the manifest; no statement claims every annex was audited.
The DPA's actual redirect target was http[local research file]; SHA256 4d67f249fce7d29c2d6e60bc1af43f9935a61d566131276c1ba03827214a1e7b. The filename identifies a 20260407 version cue. Its legally defined effective date depends on acceptance; the file name is not evidence of the Checkstep integration's historical terms. Current contractual clauses must not be backdated into PUBLIC's 2025 case study.
The auxiliary [Sightengine AI-video documentation](http[local research file] “Moderation result” and FAQ score/threshold sections, was read through the web reader; original bytes were not retained. The public Checkstep home and Sightengine FAQ pages served link discovery. They do not support an additional ownership, customer or deployment claim. No private dashboard, gated guide submission, customer content or test API request was used. No failed access route required a retry.
\nThis example weakens the claim that every programme labelled signal sharing creates a cross-platform restriction system. It strengthens a more specific mechanism: policy choices become machine-readable workflows that can carry real content/account consequences, while contractual reuse can turn some moderation inputs or feedback into model-development resources.
\nReopen on a public customer-specific integration/SOW, settings export or redacted decision-and-appeal audit that identifies the scanner, score thresholds, policy mapping, action handler and correction outcome. For the reuse question, the needed records are a customer opt-out/feedback configuration, the applicable controller/subprocessor terms and a documented training-use pathway. No outreach is authorized or proposed in this packet. Acquisition stops here for the forest review; no graph edits were made.
", "path": "signal-sharing-wave-2026-09-09/checkstep-sightengine.md", "sourceIds": [ "w22-sig-s-standard", "w22-sig-s-guide-current", "w22-sig-s-sightengine-terms", "w22-sig-s-sightengine-dpa", "w22-sig-s-checkstep-terms", "w22-sig-s-guide-cited2025" ], "markdown": "Bounded public-record packet, 9 September 2026. Six core records, including the existing government-published review. No customer data, accounts or moderation APIs were accessed. These are published specifications and standard terms, not proof of a particular customer's configuration, signed contract or enforcement history.\n\n## Result\n\nPUBLIC's example is documented as an integration of moderation vendors. The observed path is a customer submitting content for analysis, Sightengine returning classification scores, Checkstep applying the customer's configured policy, and a decision event returning to that customer's platform. The records do not establish a shared list of restricted people or automatic propagation of one platform's ban to another.\n\nThe distinction does not make the interface inconsequential: it can carry instructions that remove content or suspend an account. It also does not establish zero reuse. The published contracts contain two different machine-learning permissions: Checkstep's customer-data permission with an opt-out, and Sightengine's voluntary feedback-data route with a separate controller role and retention exception. Permission and actual use remain separate evidence states.\n\n## Six decisive records\n\n**1. PUBLIC's framing and its source.** The Data Landscape Review 2024–2025 (see source references) §6.1.1 / Case study 4, describes this as signal sharing that lets a moderation provider use classification outputs without acquiring another provider's proprietary datasets. Footnote 35 points directly to Sightengine's coauthored guide, labelled 2025. This is the report's characterization of the vendor example, rather than an independently observed exchange or a cross-platform enforcement agreement. Its phrase about no transfer of raw data must not be expanded into a claim that no submitted content is processed outside the customer's platform.\n\n**2. The coauthored interface account.** Sightengine, Ultimate Guide to GenAI Moderation—Part 2 (see source references) “Detecting and moderating…” §§1.1, 1.3, 2.1–2.2, describes media submitted by URL or upload with selected models. Results include AI-generation/deepfake scores between 0 and 1. Sightengine recommends thresholds; users choose how scores map to their policies. Checkstep integrates those outputs, supports multiple scanning providers, and routes possible violations to automated decisions or human review. It returns results and enforcement decisions to the customer. The guide also describes appeals and community reports. Its privacy language about reference-based integration describes an architecture; it is not a sufficient basis for a universal no-storage claim. The current page displays “Guide, 2026” while its introduction anticipates 2025; the report cites it as 2025. No exact original publication date or historical version identity is established here.\n\n**3. Checkstep's actual integration specification.** Standard Integration (see source references) “Sending contents,” “Access delegation” and “Receiving Webhooks,” accepts content IDs, text or media references, author identifiers and optional metadata. Delegated access enables content analysis and display to moderators; examples include cloud-storage reads and streaming bytes. Returned data identifies content, policy, severity and confidence. `analysed-content` supports provisional withholding while a decision is pending. A `decision` event instructs the customer's integration to take down, restore or keep content down after an appeal. Human `author-decision` events can suspend/terminate an account, optionally remove its content, or restore it. The customer must implement the webhook handler for the decision to affect its platform. Thus Sightengine's score is not itself a takedown, but the configured Checkstep workflow can produce an actionable instruction. `incident-closed` reports appeal outcomes. These are documented capabilities, with no observed customer configuration or successful appeal in this run.\n\n**4. Checkstep's reuse, responsibility and retention terms.** Checkstep Terms of Service (see source references) definitions and §§5.1–5.5, 13.3(C), 14, retain customer ownership while permitting use of Customer Data for machine learning that develops service features. Section 5.5 treats this as a customer instruction and offers an email opt-out; the read clause does not state an anonymization condition or prove pooled training across customers. The terms contemplate archives/backups. On termination they require disposal of data held, subject to a backup-delivery request made within 30 days and legal retention; 30 days is the request window, not a universal deletion deadline. Section 14 identifies the contracting party as controller and Checkstep as processor, generally processing under agreed instructions. Subprocessors require protective contracts and notification with a five-working-day objection window. Data-subject requests are referred to the contracting party. These published terms do not establish which provider/terms a specific customer accepted or whether it opted out.\n\n**5. Sightengine's subscriber agreement.** Terms of Service (see source references) §§1, 3.5–3.6, 4.1–4.6, 9.1–9.3, identifies the operator as Kozelo SAS. Routine Subscriber Data is licensed for service provision and instructed interoperability. Section 9.3 first grants perpetual, irrevocable use permission for general suggestions/corrections. Its separate nonexclusive licence covers voluntary raw-media submissions, with or without labels, through the Feedback API or similar mechanism for model development, testing and training; the subscriber warrants the necessary rights and consents. The raw-media licence must not inherit the first sentence’s adjectives merely by proximity. This is not proof that normal scanning automatically supplies training media, or that Checkstep sends its customers' data into this route. Sightengine can also send service-compliance notices and suspend its subscriber's API access. That power concerns the provider/customer service relationship and must not be confused with direct control of an end user's account on an unrelated platform.\n\n**6. Sightengine's feedback exception and correction route.** Data Processing Addendum (see source references) PDF 2 / definition, PDF 3–4 / §§2.1–2.3 and PDF 6–7 / §§4–5, normally assigns Sightengine a processor role; a subscriber that is itself a processor must have the controller's authority. For voluntarily submitted Feedback Data, §2.3 instead makes Sightengine an independent controller, subject to privacy law and an appropriate lawful basis. Ordinary deletion instructions are to be completed as soon as reasonably practicable, within 90 days subject to legal retention; expiry provisions also describe a recovery period of up to 30 days. Section 4.4 exempts Feedback Data from those automatic return/deletion provisions and permits perpetual retention subject to its privacy policy and applicable law. Section 5 enables access, correction, restriction, export and assistance with data-subject requests, normally through the subscriber. This is not a finding that privacy rights disappear or that every correction propagates into a trained model.\n\n## What this means for the sharing hypothesis\n\n| Interface | What can travel | What this packet does not establish |\n| --- | --- | --- |\n| Customer → Checkstep → selected scanner | Content/reference access, selected model inputs and identifiers needed for that customer's moderation | An exchange of the scanner's proprietary training corpus |\n| Sightengine → Checkstep | Detection scores/classification results for submitted media | A decision binding every other platform or a shared offender identity |\n| Checkstep → customer platform | Policy-linked content/account decisions, provisional results and appeal outcomes | A deployed handler, its actual settings, enforcement against a particular user, or an independent appeal body |\n| Customer data / voluntary feedback → model development | Published permissions for two distinct reuse routes | Whether a customer used either route, whether learning is pooled, the actual training corpus or a resulting cross-customer restriction |\n\nThe strongest rival to a cross-platform restriction interpretation is ordinary service composition: one vendor supplies classifiers and another supplies policy/workflow software for each customer's site. The explicit contract permissions leave a narrower cross-customer learning possibility open, but that is not the same mechanism as transmitting a ban or restriction signal. Affected people can face automated consequences even where formal policy choice remains with the customer; the discriminating evidence is the configured rule and its execution, not the mere presence of a classifier.\n\n## Custody and version boundaries\n\nOriginal captures and SHA256 values are recorded in `cs-captures.json`. `cs-standard.html`, `cs-sightengine-guide.html`, `cs-checkstep-terms.html`, `cs-sightengine-terms.html` and `cs-sightengine-dpa.pdf` were acquired by ordinary HTTP 200 requests on 9 September 2026. The report uses the existing `../osdi-comparison-wave-2026-09-09/data-landscape-2025.html`. Derived text copies support precise reading; `cs-sightengine-dpa-extract.txt` preserves the inspected PDF pages. Source states are read, with targeted scopes in the manifest; no statement claims every annex was audited.\n\nThe DPA's actual redirect target was `http[local research file]`; SHA256 `4d67f249fce7d29c2d6e60bc1af43f9935a61d566131276c1ba03827214a1e7b`. The filename identifies a 20260407 version cue. Its legally defined effective date depends on acceptance; the file name is not evidence of the Checkstep integration's historical terms. Current contractual clauses must not be backdated into PUBLIC's 2025 case study.\n\nThe auxiliary Sightengine AI-video documentation (see source references) “Moderation result” and FAQ score/threshold sections, was read through the web reader; original bytes were not retained. The public Checkstep home and Sightengine FAQ pages served link discovery. They do not support an additional ownership, customer or deployment claim. No private dashboard, gated guide submission, customer content or test API request was used. No failed access route required a retry.\n\n## Forest implication and next discriminator\n\nThis example weakens the claim that every programme labelled signal sharing creates a cross-platform restriction system. It strengthens a more specific mechanism: policy choices become machine-readable workflows that can carry real content/account consequences, while contractual reuse can turn some moderation inputs or feedback into model-development resources.\n\nReopen on a public customer-specific integration/SOW, settings export or redacted decision-and-appeal audit that identifies the scanner, score thresholds, policy mapping, action handler and correction outcome. For the reuse question, the needed records are a customer opt-out/feedback configuration, the applicable controller/subprocessor terms and a documented training-use pathway. No outreach is authorized or proposed in this packet. Acquisition stops here for the forest review; no graph edits were made.\n", "file": "research/signal-sharing-wave-2026-09-09-checkstep-sightengine.md" }, { "id": "signal-sharing-wave-2026-09-09-lantern-rights-review", "title": "Lantern: safeguards, correction and the limits of public assurance", "description": "Research checked 9 September 2026. Root acquisition stopped after the scoped reads below. This packet distinguishes an assessment, an operator's response, current programme reporting and actual case records. No access to private signals was attempted.", "html": "Research checked 9 September 2026. Root acquisition stopped after the scoped reads below. This packet distinguishes an assessment, an operator's response, current programme reporting and actual case records. No access to private signals was attempted.
\n[BSR's November 2023 HRIA](http[local research file] PDF19–20, describes a multiparty agreement and Tech Coalition admission, compliance and removal powers. Participants must review signals and establish their own policy violation; they must provide appeals. These are described commitments, not an inspected executed agreement.
\nBSR warns that decontextualized or erroneous signals can produce cumulative exclusions, requiring separate appeals to each company (PDF25–30). It identifies resource and legal-pressure risks that can undermine review, and dependence on Meta's infrastructure. These are assessed risks, not proven wrongful cases. Its April–August 2023 work predates later financial integration. The report disclaims audit/certification status.
\nThe assessment therefore supplies a testable distinction: distributed decisions can share an evidential input while remedies remain fragmented. The actual quality of investigation and correction propagation remains unresolved.
\nAccess: primary web reader succeeded (63 pages); one ordinary PDF GET returned403. No original PDF retained; do not retry this closed route. Read disclaimer/executive portions PDF2–9, methodology/stakeholder sections PDF14–18, programme/governance PDF19–30, and commitments/recommendations PDF46–50. Remaining pages were not fully read. Reader page excerpts were inspected; this is not full-document substantive review.
\nThe [7 November 2023 response](http[local research file] says nine of nineteen recommendations had been implemented and ten were being pursued or investigated. It describes retention limits, signal-quality review, participation management and planned transparency. Independent audits or trusted researcher access were still possibilities to explore. This is an implementation claim by the commissioning operator, not independent verification of performance.
\nThe main article was read completely. Retained HTML: lantern-hria-response.html,115875 bytes,SHA2564933c787d9a3ff987e4e521e1aaadda5b1a2b3ec512172b9745928b8746b1262. Original acquisition metadata is in root-hria-captures.json.
\nRoot independently read PDF8–11,14,28–29 of the [2025 transparency report](http[local research file] retained by Hubble. Its case study starts with a law-enforcement request to Meta, followed by Meta's investigation and signals, then Cash App closures and Xbox action. This does not equate to government database membership or direct government uploads.
\nThe report describes commissioned BSR reviews of financial participation and AI-related signals, but not their full findings. It states companies may delete only signals they uploaded, and gives aggregate removals without reasons that separate error correction from retention expiry. Own-policy review is required; shared taxonomy can cover conduct prohibited by platform rules without establishing illegality. The listed cross-platform tags include Steam; a tag is not proof of Steam participation or a particular Steam sanction.
\nOriginal: lantern-report2025.pdf,836595 bytes,SHA256f6667366d192a4582ab5d6e6ea2115a1772ca3312229a4e48bdc18b68d21bc5d. Hubble's lantern.md carries broader report scope and numerical discrepancies. Do not treat both readings as independent evidentiary origins.
\nThe [2024 report](http[local research file] PDF14, supplies a partial remedy requirement: companies must maintain appeals/recourse, remove signals no longer applicable or relevant, and notify Tech Coalition. Root subsequently read that original page. This establishes a stated removal/notification obligation; the public paragraph does not specify notification to every earlier recipient or reopening its decisions.
\nThe high-value missing artifact is a correction and appeal propagation protocol, with an anonymized worked example: when the originating company retracts a signal, who receives notice, whether downstream investigations are reopened, what happens to already-exported data, and whether affected users learn that several decisions share one source.
\nThe current annual totals cannot answer that question. They neither establish a wrongful exclusion nor demonstrate successful end-to-end remedy. A second useful discriminator is the financial-sector HRIA with its recommendation disposition and actual financial onboarding terms.
\nScoped searches: BSR Lantern financial-pilot human-rights review; Tech Coalition Lantern appeals2025; Tech Coalition Lantern financial human-rights. These located a public pilot summary and older/current transparency material, but not the complete later HRIA or an identified public correction case. The [17 July 2025 pilot summary](http[local research file] offers its full report to members/participants through a member resource centre. That access boundary is not authority to enter it; the public summary was read. No outreach or restricted-access acquisition occurred.
\nInterpretation to carry into the forest: admission and taxonomy authority, infrastructure provision, signal origination, downstream enforcement and remedy are five separate functions. A network may distribute exclusions without one actor issuing each final instruction. To claim that it has done so wrongly requires the actual signal-to-decision-to-correction evidence.
", "path": "signal-sharing-wave-2026-09-09/lantern-rights-review.md", "sourceIds": [ "w22-sig-s-bsr2023", "w22-sig-s-hria-response" ], "markdown": "Research checked 9 September 2026. Root acquisition stopped after the scoped reads below. This packet distinguishes an assessment, an operator's response, current programme reporting and actual case records. No access to private signals was attempted.\n\n## The commissioned assessment\n\nBSR's November 2023 HRIA (see source references) PDF19–20, describes a multiparty agreement and Tech Coalition admission, compliance and removal powers. Participants must review signals and establish their own policy violation; they must provide appeals. These are described commitments, not an inspected executed agreement.\n\nBSR warns that decontextualized or erroneous signals can produce cumulative exclusions, requiring separate appeals to each company (PDF25–30). It identifies resource and legal-pressure risks that can undermine review, and dependence on Meta's infrastructure. These are assessed risks, not proven wrongful cases. Its April–August 2023 work predates later financial integration. The report disclaims audit/certification status.\n\nThe assessment therefore supplies a testable distinction: distributed decisions can share an evidential input while remedies remain fragmented. The actual quality of investigation and correction propagation remains unresolved.\n\nAccess: primary web reader succeeded (63 pages); one ordinary PDF GET returned403. No original PDF retained; do not retry this closed route. Read disclaimer/executive portions PDF2–9, methodology/stakeholder sections PDF14–18, programme/governance PDF19–30, and commitments/recommendations PDF46–50. Remaining pages were not fully read. Reader page excerpts were inspected; this is not full-document substantive review.\n\n## The operator's response\n\nThe 7 November 2023 response (see source references) says nine of nineteen recommendations had been implemented and ten were being pursued or investigated. It describes retention limits, signal-quality review, participation management and planned transparency. Independent audits or trusted researcher access were still possibilities to explore. This is an implementation claim by the commissioning operator, not independent verification of performance.\n\nThe main article was read completely. Retained HTML: lantern-hria-response.html,115875 bytes,SHA2564933c787d9a3ff987e4e521e1aaadda5b1a2b3ec512172b9745928b8746b1262. Original acquisition metadata is in root-hria-captures.json.\n\n## Current published operational boundary\n\nRoot independently read PDF8–11,14,28–29 of the 2025 transparency report (see source references) retained by Hubble. Its case study starts with a law-enforcement request to Meta, followed by Meta's investigation and signals, then Cash App closures and Xbox action. This does not equate to government database membership or direct government uploads.\n\nThe report describes commissioned BSR reviews of financial participation and AI-related signals, but not their full findings. It states companies may delete only signals they uploaded, and gives aggregate removals without reasons that separate error correction from retention expiry. Own-policy review is required; shared taxonomy can cover conduct prohibited by platform rules without establishing illegality. The listed cross-platform tags include Steam; a tag is not proof of Steam participation or a particular Steam sanction.\n\nOriginal: lantern-report2025.pdf,836595 bytes,SHA256f6667366d192a4582ab5d6e6ea2115a1772ca3312229a4e48bdc18b68d21bc5d. Hubble's lantern.md carries broader report scope and numerical discrepancies. Do not treat both readings as independent evidentiary origins.\n\n## What the next record should decide\n\nThe 2024 report (see source references) PDF14, supplies a partial remedy requirement: companies must maintain appeals/recourse, remove signals no longer applicable or relevant, and notify Tech Coalition. Root subsequently read that original page. This establishes a stated removal/notification obligation; the public paragraph does not specify notification to every earlier recipient or reopening its decisions.\n\nThe high-value missing artifact is a correction and appeal propagation protocol, with an anonymized worked example: when the originating company retracts a signal, who receives notice, whether downstream investigations are reopened, what happens to already-exported data, and whether affected users learn that several decisions share one source.\n\nThe current annual totals cannot answer that question. They neither establish a wrongful exclusion nor demonstrate successful end-to-end remedy. A second useful discriminator is the financial-sector HRIA with its recommendation disposition and actual financial onboarding terms.\n\nScoped searches: BSR Lantern financial-pilot human-rights review; Tech Coalition Lantern appeals2025; Tech Coalition Lantern financial human-rights. These located a public pilot summary and older/current transparency material, but not the complete later HRIA or an identified public correction case. The 17 July 2025 pilot summary (see source references) offers its full report to members/participants through a member resource centre. That access boundary is not authority to enter it; the public summary was read. No outreach or restricted-access acquisition occurred.\n\nInterpretation to carry into the forest: admission and taxonomy authority, infrastructure provision, signal origination, downstream enforcement and remedy are five separate functions. A network may distribute exclusions without one actor issuing each final instruction. To claim that it has done so wrongly requires the actual signal-to-decision-to-correction evidence.\n", "file": "research/signal-sharing-wave-2026-09-09-lantern-rights-review.md" }, { "id": "signal-sharing-wave-2026-09-09-lantern", "title": "Lantern: operational signals, programme control and financial-account decisions", "description": "Bounded primary-source pass, observed 9 September 2026. This is Tech Coalition's child-safety programme, not a similarly named political vendor. No canonical case changes. Acquisition stopped at this operational-record breakpoint; rights and remedy analysis is separately recorded in lantern-rights-review.md(lantern-rights-review.md).", "html": "Bounded primary-source pass, observed 9 September 2026. This is Tech Coalition's child-safety programme, not a similarly named political vendor. No canonical case changes. Acquisition stopped at this operational-record breakpoint; rights and remedy analysis is separately recorded in lantern-rights-review.md.
\nLantern provides a documented mechanism through which one company's child-safety investigation can inform another company's account or financial-service decision. The operator reports a case involving 18 Cash App account closures and action on Xbox accounts. Block separately confirms that its investigations team uses Lantern signals. These are meaningful programme and participant accounts of operational use, but the specific closures, arrests, accuracy and counterfactual effect have not been independently established in this pass. The public material allocates admission and programme standards to Tech Coalition, hosting to Meta, signal selection and labelling to originating companies, and subsequent investigations and enforcement to receiving companies. It does not establish a single actor directing every resulting decision.
\nThe 2024 transparency report states that participation has no cost, that Tech Coalition and its members fully fund the programme, and that Meta provides technical hosting in kind. Eligible non-members can apply. This supports organizational funding and an in-kind infrastructure relationship; it does not show a monetary budget, participant-by-participant payment, restricted award, or donor approval right. The current programme and membership pages are institutional descriptions rather than financial ledgers. Their funding statements do not establish the legal entity making each payment. [2024 report, PDF 4](http[local research file] [programme page](http[local research file]
\nTech Coalition administers applications and compliance review. The 2024 report describes a formal legal agreement between participants after acceptance; it also describes more granular requirements for properly staffed teams able to investigate and act responsibly. Tech Coalition maintains the Official Program Expectations and reviews them with participating companies. In 2025, Tech Coalition expressly administers vetting and grants database access. Meta's in-kind hosting on ThreatExchange is reaffirmed. Hosting does not by itself establish editorial approval of individual signals, ownership of participant decisions, or a veto over programme rules. The executed agreement, hosting contract and individual voting or veto rights were not acquired. [2024 report, PDF 5–6](http[local research file] [2025 report, PDF 16, 28–29](http[local research file]
\nAdmission standards have a reported consequence: the 2024 report says one unnamed participant could not meet compliance requirements and would not continue in 2025. This is evidence of the operator announcing non-continuation, not independent identification of that company or confirmation of the revocation date. The report counts 26 enrolled companies at the end of 2024, including three financial institutions; the 2025 report counts 31 at year end. The undated current programme page displays 38 participants as observed on 9 September 2026. Those are different observation periods, not interchangeable totals. NGOs, governments, law enforcement and researchers are excluded from programme participation. A law-enforcement request can initiate a company's investigation without making that authority a Lantern member or signal uploader. [2024 report, PDF 4, 7, 9](http[local research file] [2025 report, PDF 9](http[local research file] [programme FAQ](http[local research file]
\nSignals extend beyond hashes of known abuse imagery. Public examples include URLs, account identifiers, keywords or codewords, repeated unsolicited contact with minors, attempts to move conversations between platforms, and payment requests or extortion language. Participants choose what to share and ingest within programme rules, their policies and legal obligations. Published rules require manual review before upload and receiving-company investigation before action. They require a connection to online child sexual exploitation and abuse (OCSEA), necessity and proportionality, consistency with applicable law and public terms or privacy policies, and permitted sharing under the agreement. These are stated operating requirements, not independent proof of compliance in every case. [Programme page, signal sections and FAQ](http[local research file] [2024 report, PDF 14](http[local research file] [2025 report, PDF 28–29](http[local research file]
\nTech Coalition maintains the taxonomy in collaboration with participants, which may propose updates. At least one official taxonomy tag is required; severity and review status accompany signals. The public appendix supplies names and definitions, while the full taxonomy's examples are not published, with sensitivity given as the reason. The reports explicitly distinguish platform-policy classifications from legal determinations: some included behaviour may violate a platform's policy without being criminal CSAM. [2024 report, PDF 14, 19–21](http[local research file] [2025 report, PDF 14](http[local research file]
\nThe scope of individual tags matters. Minor Sexualization extends beyond criminal imagery. Grooming (Inappropriate Contact) concerns inappropriate adult–child trust-building behaviour before an explicit sexual stage; it does not label every adult–child contact. The outrage and humour meme categories require depiction of CSAM or sexualization of minors; they are not generic political outrage or satire categories. Financial Transaction is supplemental and cannot stand alone: it concerns evidence or communication about a completed or proposed transaction possibly connected to abuse. The supplemental self-harm category is likewise connected to OCSEA. These distinctions prevent promotion of a child-safety taxonomy into an unbounded payments, speech or mental-health classification system. [2024 report, Appendix A, PDF 19–21](http[local research file]
\nThe 2025 report describes a new sadistic-online-exploitation tag tied to sexualized suffering, fear or humiliation; an additional protocol for data about minors suspected of perpetration expects extra precautions to verify the basis and necessity of sharing, with Tech Coalition-coordinated one-to-one sharing with the company directly implicated. Platform tags, including Steam, are cross-platform investigation context; they do not establish membership or an enforcement action. Proposed alignment with the Industry Classification in 2026 is forward-looking language in this report, not proof that it occurred. Whitelisted sharing, introduced in 2024, also permits sharing with selected partners rather than every participant. [2025 report, PDF 14–15](http[local research file] [2024 report, PDF 12](http[local research file]
\nThe public pilot account, dated 17 July 2025, identifies Cash App, MEGA, Meta, PayPal, Snap and Western Union as the six participants in an August 2024–first-half-2025 pilot. Tech Coalition reports 1,935 shared signals and 108 financial-institution investigations, some still ongoing, plus reports to US government authorities. It does not supply a count of prosecutions or convictions. Financial participants receive signals but do not upload them. The full pilot report is offered through the Member Resource Center or by contacting Tech Coalition; this pass used the public account and did not authenticate or request access. [Pilot results](http[local research file]
\nOn 14 May 2026, the operator announced eligibility for qualifying financial institutions worldwide, evaluated individually, with priority for peer-to-peer payments. This is an expansion of eligibility, not proof that every bank, payment provider or cryptocurrency service participates. The announcement preserves the receive-only role. [Financial expansion announcement](http[local research file]
\nIn the 2025 transparency report's case study, a law-enforcement request led Meta to conduct its own investigation into alleged CSAM production and distribution, report to NCMEC, and share signals through Lantern. Tech Coalition says Block then identified a network of 18 Cash App accounts suspected of CSAM sales or purchases, closed those accounts, banned users and reported to authorities; Microsoft acted on several Xbox accounts. The report further says law enforcement confirmed a contribution to multiple arrests. No identifiable police release, court docket or underlying order corroborating those arrests was acquired. The evidentiary verb remains “Tech Coalition reports.” This sequence does not show government uploading directly into Lantern. [2025 report, PDF 8](http[local research file]
\nBlock's own September 2025 publication independently establishes what that participant publicly says about its use: Cash App is active in Lantern, and its investigations team uses signals as intelligence and assesses detection rules and product or risk controls. This corroborates participation and claimed operational use from the receiving company, but does not independently verify the particular 18-account case. Its adjacent overall NCMEC reporting figure must not be attributed wholly to Lantern. The same page separately places Cash App in ICMEC's US Financial Coalition Against Child Sexual Exploitation; that is a concrete organizational participation bridge, not evidence that ICMEC controls Lantern. [Block report, PDF 19 / printed 18](http[local research file]
\nThe 2025 report counts 983,602 signals shared and retained for that year. Its cumulative figure is 2,047,982 for 2023–2025, with the retention cut-off stated as 3 March 2026. Retained signals are not all uploads ever made, unique people, or verified criminal cases. The combined action headline also adds accounts, URLs and content units; it is not a count of people. Outcomes are voluntarily supplied by companies, and the operator says its downstream-action measures exclude the original uploader's prior actions. That is a reported accounting rule, not an independently tested causal estimate. [2025 report, PDF 5–7, 16](http[local research file]
\nThere is an unresolved discrepancy inside the captured original: PDF 5 and 7 give 31,504 account actions for 2025, while PDF 16 gives 31,485, a difference of 19. PDF 5 and 16 were visually checked. Do not silently select one or infer wrongdoing from the inconsistency. The report does not establish wrongful-action rates or the proportion of decisions subsequently reversed. The 44,210 removed signals reported on PDF 11 mix reasons that are not publicly disaggregated; deletion is not a measure of confirmed false positives. The separate rights packet covers the original uploader's removal right and the gap concerning propagation of corrections to earlier recipient decisions.
\nSeven originals were captured through ordinary public GETs. lantern-captures.json records exact URLs, byte counts, SHA-256 hashes and reading scopes. Hashes identify original response bytes; extracted text and renders are derivatives. Pages below mean physical PDF pages; printed pages coincide for the two Lantern reports.
\n| Captured source | \nDate and actual inspected scope | \n
|---|---|
| 2025 report | \nReporting year 2025; release landing dated 28 April 2026. Reader text PDF 3–12 and original extraction PDF 13–16, 24–29 read for this lane. PDF 5 and 16 visually checked. Not a comprehensive visual audit of all 30 pages. | \n
| 2024 report | \nReporting year 2024; publication day not established. Original text PDF 4–7, 9, 12–14, 19–21 read; cover/contents inspected. PDF 4 and 20 visually checked. | \n
| Programme page | \nUndated current page, observed 9 September 2026. Main body, signal sections, FAQ and funding footer read. | \n
| Membership page | \nUndated current page, observed 9 September 2026. Main body/footer read as general institutional context; no fee amount or Lantern allocation established. | \n
| Financial expansion | \nMain article dated 14 May 2026, read in full from capture. HTML title names an unrelated Initiate 2026 item; article heading/date/body identify the source used here. | \n
| Financial pilot results | \nMain article dated 17 July 2025, read in full. Restricted full pilot report not accessed. | \n
| Block report | \nCover dated 10 September 2025; cover and PDF 19 / printed 18 read from the original. Not a comprehensive review of the 21-page report. | \n
The old /lantern/ route failed in the reader; the site's actual /programs/lantern/ route succeeded. The financial-expansion reader timed out once; an ordinary GET substitute succeeded. Root's closed BSR PDF GET failure was not retried; its selected reader passages remain under the separate rights packet's custody. No private account or dataset was accessed.
Precise supported relations are: Tech Coalition/member funding of Lantern without allocated amounts; Meta's in-kind hosting; Tech Coalition admission, access and programme-standard administration; participants' manual signal selection and shared taxonomy development; and recipients' own investigation and policy decisions. The 18-account outcome must remain attributed to the operator. No individual member's veto, mandatory recipient sanction, government membership, or donor instruction is established.
\nThe strongest local forest implication is a concrete handoff from platform intelligence to financial-service exclusion, with authority distributed across admission, classification, infrastructure, origination and recipient decision-making. The strongest alternative to a centrally directed sanction system is the published model of legally constrained, company-specific investigations against genuinely cross-platform child exploitation. Public self-reports alone cannot determine how consistently that model operates or how errors travel.
\nThe next discriminating record is the operative participant agreement and a worked, anonymized correction case showing originator retraction, notification to prior recipients, treatment of exported copies, reopening of earlier restrictions, and user notice or remedy. Those records would distinguish a reversible intelligence exchange from a system in which corrected signals can leave lasting distributed consequences. The Tech Coalition programme office and involved participants are the likely custodians; no outreach is authorized or undertaken. A disclosed programme budget or hosting agreement would separately resolve who pays what and any associated rights. This bounded pass stops here.
", "path": "signal-sharing-wave-2026-09-09/lantern.md", "sourceIds": [ "w22-sig-s-report2024", "w22-sig-s-report2025", "w22-sig-s-program", "w22-sig-s-pilot", "w22-sig-s-expansion", "w22-sig-s-block" ], "markdown": "Bounded primary-source pass, observed 9 September 2026. This is Tech Coalition's child-safety programme, not a similarly named political vendor. No canonical case changes. Acquisition stopped at this operational-record breakpoint; rights and remedy analysis is separately recorded in [lantern-rights-review.md](signal-sharing-wave-2026-09-09-lantern-rights-review.md).\n\n## Finding and limits\n\nLantern provides a documented mechanism through which one company's child-safety investigation can inform another company's account or financial-service decision. The operator reports a case involving 18 Cash App account closures and action on Xbox accounts. Block separately confirms that its investigations team uses Lantern signals. These are meaningful programme and participant accounts of operational use, but the specific closures, arrests, accuracy and counterfactual effect have not been independently established in this pass. The public material allocates admission and programme standards to Tech Coalition, hosting to Meta, signal selection and labelling to originating companies, and subsequent investigations and enforcement to receiving companies. It does not establish a single actor directing every resulting decision.\n\n## Money, infrastructure and admission\n\nThe 2024 transparency report states that participation has no cost, that Tech Coalition and its members fully fund the programme, and that Meta provides technical hosting in kind. Eligible non-members can apply. This supports organizational funding and an in-kind infrastructure relationship; it does not show a monetary budget, participant-by-participant payment, restricted award, or donor approval right. The current programme and membership pages are institutional descriptions rather than financial ledgers. Their funding statements do not establish the legal entity making each payment. 2024 report, PDF 4 (see source references) programme page (see source references)\n\nTech Coalition administers applications and compliance review. The 2024 report describes a formal legal agreement between participants after acceptance; it also describes more granular requirements for properly staffed teams able to investigate and act responsibly. Tech Coalition maintains the Official Program Expectations and reviews them with participating companies. In 2025, Tech Coalition expressly administers vetting and grants database access. Meta's in-kind hosting on ThreatExchange is reaffirmed. Hosting does not by itself establish editorial approval of individual signals, ownership of participant decisions, or a veto over programme rules. The executed agreement, hosting contract and individual voting or veto rights were not acquired. 2024 report, PDF 5–6 (see source references) 2025 report, PDF 16, 28–29 (see source references)\n\nAdmission standards have a reported consequence: the 2024 report says one unnamed participant could not meet compliance requirements and would not continue in 2025. This is evidence of the operator announcing non-continuation, not independent identification of that company or confirmation of the revocation date. The report counts 26 enrolled companies at the end of 2024, including three financial institutions; the 2025 report counts 31 at year end. The undated current programme page displays 38 participants as observed on 9 September 2026. Those are different observation periods, not interchangeable totals. NGOs, governments, law enforcement and researchers are excluded from programme participation. A law-enforcement request can initiate a company's investigation without making that authority a Lantern member or signal uploader. 2024 report, PDF 4, 7, 9 (see source references) 2025 report, PDF 9 (see source references) programme FAQ (see source references)\n\n## What enters the system and who labels it\n\nSignals extend beyond hashes of known abuse imagery. Public examples include URLs, account identifiers, keywords or codewords, repeated unsolicited contact with minors, attempts to move conversations between platforms, and payment requests or extortion language. Participants choose what to share and ingest within programme rules, their policies and legal obligations. Published rules require manual review before upload and receiving-company investigation before action. They require a connection to online child sexual exploitation and abuse (OCSEA), necessity and proportionality, consistency with applicable law and public terms or privacy policies, and permitted sharing under the agreement. These are stated operating requirements, not independent proof of compliance in every case. Programme page, signal sections and FAQ (see source references) 2024 report, PDF 14 (see source references) 2025 report, PDF 28–29 (see source references)\n\nTech Coalition maintains the taxonomy in collaboration with participants, which may propose updates. At least one official taxonomy tag is required; severity and review status accompany signals. The public appendix supplies names and definitions, while the full taxonomy's examples are not published, with sensitivity given as the reason. The reports explicitly distinguish platform-policy classifications from legal determinations: some included behaviour may violate a platform's policy without being criminal CSAM. 2024 report, PDF 14, 19–21 (see source references) 2025 report, PDF 14 (see source references)\n\nThe scope of individual tags matters. Minor Sexualization extends beyond criminal imagery. Grooming (Inappropriate Contact) concerns inappropriate adult–child trust-building behaviour before an explicit sexual stage; it does not label every adult–child contact. The outrage and humour meme categories require depiction of CSAM or sexualization of minors; they are not generic political outrage or satire categories. Financial Transaction is supplemental and cannot stand alone: it concerns evidence or communication about a completed or proposed transaction possibly connected to abuse. The supplemental self-harm category is likewise connected to OCSEA. These distinctions prevent promotion of a child-safety taxonomy into an unbounded payments, speech or mental-health classification system. 2024 report, Appendix A, PDF 19–21 (see source references)\n\nThe 2025 report describes a new sadistic-online-exploitation tag tied to sexualized suffering, fear or humiliation; an additional protocol for data about minors suspected of perpetration expects extra precautions to verify the basis and necessity of sharing, with Tech Coalition-coordinated one-to-one sharing with the company directly implicated. Platform tags, including Steam, are cross-platform investigation context; they do not establish membership or an enforcement action. Proposed alignment with the Industry Classification in 2026 is forward-looking language in this report, not proof that it occurred. Whitelisted sharing, introduced in 2024, also permits sharing with selected partners rather than every participant. 2025 report, PDF 14–15 (see source references) 2024 report, PDF 12 (see source references)\n\n## The financial route: pilot, expansion and an attributed case\n\nThe public pilot account, dated 17 July 2025, identifies Cash App, MEGA, Meta, PayPal, Snap and Western Union as the six participants in an August 2024–first-half-2025 pilot. Tech Coalition reports 1,935 shared signals and 108 financial-institution investigations, some still ongoing, plus reports to US government authorities. It does not supply a count of prosecutions or convictions. Financial participants receive signals but do not upload them. The full pilot report is offered through the Member Resource Center or by contacting Tech Coalition; this pass used the public account and did not authenticate or request access. Pilot results (see source references)\n\nOn 14 May 2026, the operator announced eligibility for qualifying financial institutions worldwide, evaluated individually, with priority for peer-to-peer payments. This is an expansion of eligibility, not proof that every bank, payment provider or cryptocurrency service participates. The announcement preserves the receive-only role. Financial expansion announcement (see source references)\n\nIn the 2025 transparency report's case study, a law-enforcement request led Meta to conduct its own investigation into alleged CSAM production and distribution, report to NCMEC, and share signals through Lantern. Tech Coalition says Block then identified a network of 18 Cash App accounts suspected of CSAM sales or purchases, closed those accounts, banned users and reported to authorities; Microsoft acted on several Xbox accounts. The report further says law enforcement confirmed a contribution to multiple arrests. No identifiable police release, court docket or underlying order corroborating those arrests was acquired. The evidentiary verb remains “Tech Coalition reports.” This sequence does not show government uploading directly into Lantern. 2025 report, PDF 8 (see source references)\n\nBlock's own September 2025 publication independently establishes what that participant publicly says about its use: Cash App is active in Lantern, and its investigations team uses signals as intelligence and assesses detection rules and product or risk controls. This corroborates participation and claimed operational use from the receiving company, but does not independently verify the particular 18-account case. Its adjacent overall NCMEC reporting figure must not be attributed wholly to Lantern. The same page separately places Cash App in ICMEC's US Financial Coalition Against Child Sexual Exploitation; that is a concrete organizational participation bridge, not evidence that ICMEC controls Lantern. Block report, PDF 19 / printed 18 (see source references)\n\n## Outcome totals require care\n\nThe 2025 report counts 983,602 signals shared and retained for that year. Its cumulative figure is 2,047,982 for 2023–2025, with the retention cut-off stated as 3 March 2026. Retained signals are not all uploads ever made, unique people, or verified criminal cases. The combined action headline also adds accounts, URLs and content units; it is not a count of people. Outcomes are voluntarily supplied by companies, and the operator says its downstream-action measures exclude the original uploader's prior actions. That is a reported accounting rule, not an independently tested causal estimate. 2025 report, PDF 5–7, 16 (see source references)\n\nThere is an unresolved discrepancy inside the captured original: PDF 5 and 7 give **31,504** account actions for 2025, while PDF 16 gives **31,485**, a difference of 19. PDF 5 and 16 were visually checked. Do not silently select one or infer wrongdoing from the inconsistency. The report does not establish wrongful-action rates or the proportion of decisions subsequently reversed. The 44,210 removed signals reported on PDF 11 mix reasons that are not publicly disaggregated; deletion is not a measure of confirmed false positives. The separate rights packet covers the original uploader's removal right and the gap concerning propagation of corrections to earlier recipient decisions.\n\n## Read scope and custody\n\nSeven originals were captured through ordinary public GETs. lantern-captures.json records exact URLs, byte counts, SHA-256 hashes and reading scopes. Hashes identify original response bytes; extracted text and renders are derivatives. Pages below mean physical PDF pages; printed pages coincide for the two Lantern reports.\n\n| Captured source | Date and actual inspected scope |\n| --- | --- |\n| [2025 report](https://technologycoalition.org/wp-content/uploads/lantern-transparency-report-2025.pdf) | Reporting year 2025; release landing dated 28 April 2026. Reader text PDF 3–12 and original extraction PDF 13–16, 24–29 read for this lane. PDF 5 and 16 visually checked. Not a comprehensive visual audit of all 30 pages. |\n| [2024 report](https://technologycoalition.org/wp-content/uploads/2025/06/Lantern-Transparency-Report-2024.pdf) | Reporting year 2024; publication day not established. Original text PDF 4–7, 9, 12–14, 19–21 read; cover/contents inspected. PDF 4 and 20 visually checked. |\n| [Programme page](https://technologycoalition.org/programs/lantern/) | Undated current page, observed 9 September 2026. Main body, signal sections, FAQ and funding footer read. |\n| Membership page | Undated current page, observed 9 September 2026. Main body/footer read as general institutional context; no fee amount or Lantern allocation established. |\n| [Financial expansion](https://technologycoalition.org/news/lantern-expands-to-financial-institutions-worldwide/) | Main article dated 14 May 2026, read in full from capture. HTML title names an unrelated Initiate 2026 item; article heading/date/body identify the source used here. |\n| [Financial pilot results](https://technologycoalition.org/news/lantern-financial-pilot-results/) | Main article dated 17 July 2025, read in full. Restricted full pilot report not accessed. |\n| [Block report](https://block.xyz/documents/access-and-protection-for-teens-in-the-digital-economy.pdf) | Cover dated 10 September 2025; cover and PDF 19 / printed 18 read from the original. Not a comprehensive review of the 21-page report. |\n\nThe old `/lantern/` route failed in the reader; the site's actual `/programs/lantern/` route succeeded. The financial-expansion reader timed out once; an ordinary GET substitute succeeded. Root's closed BSR PDF GET failure was not retried; its selected reader passages remain under the separate rights packet's custody. No private account or dataset was accessed.\n\n## Integration boundary and next discriminator\n\nPrecise supported relations are: Tech Coalition/member funding of Lantern without allocated amounts; Meta's in-kind hosting; Tech Coalition admission, access and programme-standard administration; participants' manual signal selection and shared taxonomy development; and recipients' own investigation and policy decisions. The 18-account outcome must remain attributed to the operator. No individual member's veto, mandatory recipient sanction, government membership, or donor instruction is established.\n\nThe strongest local forest implication is a concrete handoff from platform intelligence to financial-service exclusion, with authority distributed across admission, classification, infrastructure, origination and recipient decision-making. The strongest alternative to a centrally directed sanction system is the published model of legally constrained, company-specific investigations against genuinely cross-platform child exploitation. Public self-reports alone cannot determine how consistently that model operates or how errors travel.\n\nThe next discriminating record is the operative participant agreement and a worked, anonymized correction case showing originator retraction, notification to prior recipients, treatment of exported copies, reopening of earlier restrictions, and user notice or remedy. Those records would distinguish a reversible intelligence exchange from a system in which corrected signals can leave lasting distributed consequences. The Tech Coalition programme office and involved participants are the likely custodians; no outreach is authorized or undertaken. A disclosed programme budget or hosting agreement would separately resolve who pays what and any associated rights. This bounded pass stops here.\n", "file": "research/signal-sharing-wave-2026-09-09-lantern.md" }, { "id": "stopncii-governance-wave-2026-09-09-operator-resources", "title": "SWGfL, StopNCII and the Revenge Porn Helpline: operator resources and decision rights", "description": "9 September 2026. Acquisition closed. This lane establishes the charity/legal operator, dated restricted resources and constitutional authority. Root owns partner terms/admission; Hubble owns user entry/withdrawal. No personal case, image, hash list, application, account, outreach or paid source was accessed.", "html": "9 September 2026. Acquisition closed. This lane establishes the charity/legal operator, dated restricted resources and constitutional authority. Root owns partner terms/admission; Hubble owns user entry/withdrawal. No personal case, image, hash list, application, account, outreach or paid source was accessed.
\nThe retained [FY2025 filed accounts](http[local research file] identify South West Grid for Learning Trust, company 05589479, charity England/Wales 1120354, Scotland SC051351. The cover and administrative pages identify the numbers; Note1, PDF21/printed19, says it is an English/Welsh company limited by guarantee with no share capital. This is the legal charity behind the SWGfL brand, not Meta's equity subsidiary.
\nThe charity's trustees report discusses the Revenge Porn Helpline as one of its support services (PDF5/printed3) and StopNCII as part of its international work (PDF6/printed4). Its [current institutional page](http[local research file] also lists both under its services. The CEO biography attributes creation of the Helpline in2015 and oversight of StopNCII development to David Wright; Will Earp's current role includes platform development and industry integration. These are operator accounts, not proof of an individually delegated authority over every submitted case. No separate corporation for either named programme was established.
\nThe same current page describes trustee Karuna Nain's former Meta safety-policy role and work with SWGfL on StopNCII's launch. [Companies House's filing history](http[local research file] records her appointment as director on3April2024, filed26April2024. It records Hanna's appointment on11September2024, filed6June2025. The acquired record supports a personnel transition and board role, not Meta ownership, a Meta-nominated seat, or an exercised funder veto. Appointment instruments and their underlying trustee resolutions were not acquired. No wider biography branch was opened.
\nThe accounts' Note19, PDF34–36/printed32–34, contains the following restricted-fund movements. Each listed incoming amount is matched by an equal amount of resources expended, with no opening or closing restricted balance shown for those rows. Amounts are GBP. The FY2024 numbers are comparative figures in this same FY2025 filing; do not add them to FY2025 as one annual budget.
\n| Accounts fund label | \nYear ended31March2024: incoming / expended | \nYear ended31March2025: incoming / expended | \nStated purpose and limit | \n
|---|---|---|---|
| Pershing Grant | \n£78,847 / £78,847 | \n£75,802 / £75,802 | \nPurpose paragraph expressly names StopNCII support. Exact donor legal entity and grant agreement are not identified in the inspected accounts. | \n
| META/Facebook | \n£20,000 / £20,000 | \nNo separate row shown | \nComparative restricted support for combating non-consensual intimate-image sharing and supporting victims. This is broader than a proved StopNCII-only allocation. Omission from the next year's restricted table does not prove that financial, technical or unrestricted support ended. | \n
| Home Office Grant – RP | \n£150,000 / £150,000 | \n£225,000 / £225,000 | \nPurpose stated as support work combating violence against women and girls. Not a demonstrated StopNCII allocation or classification instruction. | \n
Nominet and TikTok funds in the same note have expressly different UK Safer Internet Centre purposes. Their presence in the same charity's accounts does not assign them to StopNCII. Other grants were not traced into new donor networks.
\nThe recognition policy (PDF22/printed20) is consequential: income is recognized when entitlement and performance conditions are met, receipt is probable and value can be measured. Advance income may be deferred; expenditure follows legal/constructive obligation and reliable/probable settlement. These are accounting income/expense states, not acquired bank-payment receipts or evidence that a funder accepted a named technical deliverable. Nor is the Pershing row the total cost of StopNCII. The accounts provide neither a programme cost centre nor a full allocation of unrestricted income, in-kind support and platform fees to the service.
\nThe financial statements carry an independent auditor's positive opinion (PDF13/printed11). That page expressly separates other trustee-report information from its assurance opinion. Programme-impact statements therefore remain the operator's account rather than audited moderation outcomes. Financial-year scope is31March2025; Companies House received this filing31December2025. Some trustee narrative addresses events after the financial year or earlier reporting periods; no impact count is recast here as a clean FY2025 outcome.
\nThe [retained articles](http[local research file] state adoption by special resolution on10October2024 (PDF4/printed1), filed29October2024. This is a concrete governance instrument, not merely a partnership page. The accounts also refer to an earlier January2024 constitutional change; this run does not reconstruct the full earlier transition.
\nArticle8.1 makes the trustees the only members; a trustee becomes a member on appointment. Membership is nontransferable and ends when trusteeship ends (PDF12–13/printed9–10). Article10.7 gives each member one vote. This corporate membership is distinct from a StopNCII platform's programme participation.
\nArticles13–14 put business management with the trustees and allow trustee appointment by trustee resolution (PDF15/printed12). Removal provisions include statutory/charity grounds and an internal resolution process with notice and opportunity for representations (PDF16/printed13). These provisions do not allocate a named donor an appointment right. They also do not prove how a particular appointment was proposed or influenced.
\nArticles19–22 describe a default trustee quorum of one-third unless otherwise fixed, majority decisions at a meeting and a separate75%-of-eligible-trustees agreement route; the chair can have a casting vote (PDF18/printed15). Article23 permits committees including at least one trustee, nontrustee committee participants approved by trustees, and delegation of day-to-day management. Delegation terms must be recorded; committees report to trustees, who can alter or revoke delegation. Terms may restrict expenditure to trustee approval or a previously agreed budget (PDF19/printed16). This is permission to impose those conditions, not a retrieved StopNCII committee budget or minute.
\nArticles24 and26 cover interest declarations/absence, qualified authorization of loyalty conflicts by unconflicted trustees, and trustee appointment of a CEO who reports, advises on policy/budget/funding and implements trustee policy (PDF19–20/printed16–17). Other article5 authorizations were not fully read, so this packet does not claim universal recusal with no exceptions. The trustee report describes Board governance supported by finance/audit, technical, data and communications subgroups (PDF9/printed7); their actual terms, minutes and programme decisions were not acquired.
\nThere is direct evidence of earmarked support reaching the operator's accounts for StopNCII and broader victim-support work, alongside a trustee-governed charity constitution. Funds, corporate membership, service participation and individual trustee biographies are different relationships. The documents do not establish that Meta, Pershing, government or a paying platform selected an entry, dictated a definition, approved a withdrawal or controlled another platform's action.
\nThe next discriminating records are the executed programme grant/partnership schedules, the current StopNCII internal delegation/budget and any actual approval or reporting instrument. Those could reveal renewal/termination rights, deliverable acceptance, IP ownership, policy clearance or operational conditions beyond the purpose restrictions visible in the accounts. A current public donor-side record could resolve the abbreviated Pershing label; this run leaves it unexpanded. Likely custodians are SWGfL and the respective funder/partner. No request was made.
\noperator-captures.json records the two original scanned PDFs, independently published Companies House history HTML, scoped SWGfL primary-reader evidence and failures. PDF text extraction produced no substantive text; local renders, not empty derivatives, supplied the reading. Full-size financial/table/governance pages are retained as operator-accounts2025-p*.png and operator-articles2024-p*.png. The manifest distinguishes full-size reads from contact-sheet navigation and pages merely rendered.
Six focused discovery queries covered SWGfL annual reporting/funding, current accounts, StopNCII supporters and the abbreviated Pershing label. Search-only FY2023 and multilingual supporter passages were not promoted into findings. The Charity Commission accounts landing returned403 and the French StopNCII supporters route a verification challenge. Automatic approval review rejected a proposed bundled ordinary-GET retry before execution; those routes and language variants remained closed. Companies House supplied independently published originals. SWGfL About's original GET timed out; its already successful scoped primary reader remains the evidence, without an original-byte claim. The accounts' web-reader later errored; the existing successful original was read locally, with no retry. No further acquisition follows this packet before root's combined forest assessment.
", "path": "stopncii-governance-wave-2026-09-09/operator-resources.md", "sourceIds": [ "w28-access-s-swgfl-accounts2025", "w28-access-s-swgfl-articles", "w28-access-s-swgfl-about", "w28-access-s-swgfl-ch-history" ], "markdown": "9 September 2026. Acquisition closed. This lane establishes the charity/legal operator, dated restricted resources and constitutional authority. Root owns partner terms/admission; Hubble owns user entry/withdrawal. No personal case, image, hash list, application, account, outreach or paid source was accessed.\n\n## One charity, distinct services\n\nThe retained FY2025 filed accounts (see source references) identify **South West Grid for Learning Trust**, company **05589479**, charity England/Wales **1120354**, Scotland **SC051351**. The cover and administrative pages identify the numbers; Note1, PDF21/printed19, says it is an English/Welsh company limited by guarantee with no share capital. This is the legal charity behind the SWGfL brand, not Meta's equity subsidiary.\n\nThe charity's trustees report discusses the Revenge Porn Helpline as one of its support services (PDF5/printed3) and StopNCII as part of its international work (PDF6/printed4). Its current institutional page (see source references) also lists both under its services. The CEO biography attributes creation of the Helpline in2015 and oversight of StopNCII development to David Wright; Will Earp's current role includes platform development and industry integration. These are operator accounts, not proof of an individually delegated authority over every submitted case. No separate corporation for either named programme was established.\n\nThe same current page describes trustee Karuna Nain's former Meta safety-policy role and work with SWGfL on StopNCII's launch. Companies House's filing history (see source references) records her appointment as director on3April2024, filed26April2024. It records Hanna's appointment on11September2024, filed6June2025. The acquired record supports a personnel transition and board role, not Meta ownership, a Meta-nominated seat, or an exercised funder veto. Appointment instruments and their underlying trustee resolutions were not acquired. No wider biography branch was opened.\n\n## Dated programme money, without turning accounts into a cash ledger\n\nThe accounts' Note19, PDF34–36/printed32–34, contains the following restricted-fund movements. Each listed incoming amount is matched by an equal amount of resources expended, with no opening or closing restricted balance shown for those rows. Amounts are GBP. The FY2024 numbers are comparative figures in this same FY2025 filing; do not add them to FY2025 as one annual budget.\n\n| Accounts fund label | Year ended31March2024: incoming / expended | Year ended31March2025: incoming / expended | Stated purpose and limit |\n| --- | --- | --- | --- |\n| Pershing Grant | £78,847 / £78,847 | £75,802 / £75,802 | Purpose paragraph expressly names StopNCII support. Exact donor legal entity and grant agreement are not identified in the inspected accounts. |\n| META/Facebook | £20,000 / £20,000 | No separate row shown | Comparative restricted support for combating non-consensual intimate-image sharing and supporting victims. This is broader than a proved StopNCII-only allocation. Omission from the next year's restricted table does not prove that financial, technical or unrestricted support ended. |\n| Home Office Grant – RP | £150,000 / £150,000 | £225,000 / £225,000 | Purpose stated as support work combating violence against women and girls. Not a demonstrated StopNCII allocation or classification instruction. |\n\nNominet and TikTok funds in the same note have expressly different UK Safer Internet Centre purposes. Their presence in the same charity's accounts does not assign them to StopNCII. Other grants were not traced into new donor networks.\n\nThe recognition policy (PDF22/printed20) is consequential: income is recognized when entitlement and performance conditions are met, receipt is probable and value can be measured. Advance income may be deferred; expenditure follows legal/constructive obligation and reliable/probable settlement. These are accounting income/expense states, not acquired bank-payment receipts or evidence that a funder accepted a named technical deliverable. Nor is the Pershing row the total cost of StopNCII. The accounts provide neither a programme cost centre nor a full allocation of unrestricted income, in-kind support and platform fees to the service.\n\nThe financial statements carry an independent auditor's positive opinion (PDF13/printed11). That page expressly separates other trustee-report information from its assurance opinion. Programme-impact statements therefore remain the operator's account rather than audited moderation outcomes. Financial-year scope is31March2025; Companies House received this filing31December2025. Some trustee narrative addresses events after the financial year or earlier reporting periods; no impact count is recast here as a clean FY2025 outcome.\n\n## Actual constitutional authority\n\nThe retained articles (see source references) state adoption by special resolution on**10October2024** (PDF4/printed1), filed29October2024. This is a concrete governance instrument, not merely a partnership page. The accounts also refer to an earlier January2024 constitutional change; this run does not reconstruct the full earlier transition.\n\nArticle8.1 makes the trustees the only members; a trustee becomes a member on appointment. Membership is nontransferable and ends when trusteeship ends (PDF12–13/printed9–10). Article10.7 gives each member one vote. This corporate membership is distinct from a StopNCII platform's programme participation.\n\nArticles13–14 put business management with the trustees and allow trustee appointment by trustee resolution (PDF15/printed12). Removal provisions include statutory/charity grounds and an internal resolution process with notice and opportunity for representations (PDF16/printed13). These provisions do not allocate a named donor an appointment right. They also do not prove how a particular appointment was proposed or influenced.\n\nArticles19–22 describe a default trustee quorum of one-third unless otherwise fixed, majority decisions at a meeting and a separate75%-of-eligible-trustees agreement route; the chair can have a casting vote (PDF18/printed15). Article23 permits committees including at least one trustee, nontrustee committee participants approved by trustees, and delegation of day-to-day management. Delegation terms must be recorded; committees report to trustees, who can alter or revoke delegation. Terms may restrict expenditure to trustee approval or a previously agreed budget (PDF19/printed16). This is permission to impose those conditions, not a retrieved StopNCII committee budget or minute.\n\nArticles24 and26 cover interest declarations/absence, qualified authorization of loyalty conflicts by unconflicted trustees, and trustee appointment of a CEO who reports, advises on policy/budget/funding and implements trustee policy (PDF19–20/printed16–17). Other article5 authorizations were not fully read, so this packet does not claim universal recusal with no exceptions. The trustee report describes Board governance supported by finance/audit, technical, data and communications subgroups (PDF9/printed7); their actual terms, minutes and programme decisions were not acquired.\n\n## What is now visible; what remains consequentially open\n\nThere is direct evidence of earmarked support reaching the operator's accounts for StopNCII and broader victim-support work, alongside a trustee-governed charity constitution. Funds, corporate membership, service participation and individual trustee biographies are different relationships. The documents do not establish that Meta, Pershing, government or a paying platform selected an entry, dictated a definition, approved a withdrawal or controlled another platform's action.\n\nThe next discriminating records are the executed programme grant/partnership schedules, the current StopNCII internal delegation/budget and any actual approval or reporting instrument. Those could reveal renewal/termination rights, deliverable acceptance, IP ownership, policy clearance or operational conditions beyond the purpose restrictions visible in the accounts. A current public donor-side record could resolve the abbreviated Pershing label; this run leaves it unexpanded. Likely custodians are SWGfL and the respective funder/partner. No request was made.\n\n## Custody, finite scope and closed routes\n\n`operator-captures.json` records the two original scanned PDFs, independently published Companies House history HTML, scoped SWGfL primary-reader evidence and failures. PDF text extraction produced no substantive text; local renders, not empty derivatives, supplied the reading. Full-size financial/table/governance pages are retained as `operator-accounts2025-p*.png` and `operator-articles2024-p*.png`. The manifest distinguishes full-size reads from contact-sheet navigation and pages merely rendered.\n\nSix focused discovery queries covered SWGfL annual reporting/funding, current accounts, StopNCII supporters and the abbreviated Pershing label. Search-only FY2023 and multilingual supporter passages were not promoted into findings. The Charity Commission accounts landing returned403 and the French StopNCII supporters route a verification challenge. Automatic approval review rejected a proposed bundled ordinary-GET retry before execution; those routes and language variants remained closed. Companies House supplied independently published originals. SWGfL About's original GET timed out; its already successful scoped primary reader remains the evidence, without an original-byte claim. The accounts' web-reader later errored; the existing successful original was read locally, with no retry. No further acquisition follows this packet before root's combined forest assessment.\n", "file": "research/stopncii-governance-wave-2026-09-09-operator-resources.md" }, { "id": "stopncii-governance-wave-2026-09-09-partner-authority", "title": "Partner access, contributions and downstream actions", "description": "9 September 2026. Root acquisition CLOSED. StopNCII exposes several distinct gates: admission and payment, reference contributions and feedback, withdrawal processing, and platform action. Published terms are not an executed partner-specific agreement. All observations below concern public administrative records; no live case or operational API was used.", "html": "9 September 2026. Root acquisition CLOSED. StopNCII exposes several distinct gates: admission and payment, reference contributions and feedback, withdrawal processing, and platform action. Published terms are not an executed partner-specific agreement. All observations below concern public administrative records; no live case or operational API was used.
\nThe [join page](http[local research file] offers Approved Participant and Programme Partner routes. Its comparison table assigns both every listed network/support/development benefit. Yet the [Approved Participant page](http[local research file] expressly denies wider programme/network involvement. This is a real description conflict, not evidence of an undisclosed voting arrangement. The [Programme Partner page](http[local research file] advertises active engagement, working groups and opportunities to influence tools/policies; it grants no stated vote or veto.
\nAll four embedded fee/onboarding diagrams were acquired and visually read. Both routes show registration, verification, a data-sharing agreement, payment and integration. The full programme adds meeting the team and announcement. Published annual GBP fees, in the same revenue bands, are:
\n| Revenue band | \nApproved Participant | \nProgramme Partner | \n
|---|---|---|
| New business | \n1,000 | \n1,000 | \n
| Below 10 million | \n4,000 | \n10,000 | \n
| 10–249 million | \n8,000 | \n20,000 | \n
| 250–499 million | \n12,000 | \n30,000 | \n
| 500–999 million | \n16,000 | \n40,000 | \n
| Over 1 billion | \n20,000 | \n50,000 | \n
These are undated published schedules, not invoices, revenue received or proof every earlier partner paid. The wording leaves exact intervening band boundaries unclarified. The user-facing service being free does not establish free organizational access.
\n[Partner Terms](http[local research file] cover manual and automated admission. They incorporate the unacquired data-sharing agreement and registration form. Sections3–10/schedule specify the lower fee schedule, eligibility/reverification, discretionary refusal, credential controls, security assessment, local caching and access suspension. Redistribution is prohibited; model training has a purpose-limited exception. Permitted purpose depends on the separate agreement. Termination requires cessation/deletion and written certification; annual summaries are due within60days of the access anniversary.
\nThe posting contains conflicting company/charity identifiers, an empty hash-value definition, inconsistent defined-purpose terminology, bracketed implementation/amendment/certification periods and an explicit unfinished implementation note. Do not present those brackets as settled deadlines. Key-compromise notice is24hours; the schedule's general incident notice is48hours—retain their distinct scopes. The five-day certification placeholder concerns partner termination, not each user withdrawal. No signed data-sharing agreement, precedence clause resolving these inconsistencies, executed admission/refusal or payment was acquired. No displayed effective date was established.
\n[SWGfL privacy](http[local research file] Working with Partners, names Deloitte LLP for business verification and Neterium for sanctions/PEP/watchlist/adverse-media screening. This concerns applicant businesses and connected officials, not identification of people submitting intimate-image cases. The notice describes human consideration of adverse results and rejects solely automated refusals. It promises notice where screening influenced an adverse decision, an uninvolved reviewer on request, correction/annotation and relevant corrections passed to vendors. Acknowledgment is2workingdays and substantive response30workingdays. Applicant records are retained during the relationship plus3years, or3years for unsuccessful applications.
\nThose are specific process commitments. The partner terms' general sole-discretion/no-reasons clause remains broader; no acquired priority rule resolves every refusal category. This is not a proved exercise of arbitrary denial or an observed successful appeal. The retained HTML references vendor privacy notices without supplying actual links. Root read the complete partner section; other SWGfL processing purposes are outside this lane.
\nThe [API guide](http[local research file] cover date08/10/2025, describes SWGfL-issued provider credentials, CSP hash submissions, NGO-reviewed Active entries, public-user Received entries and independent provider feedback. It documents change timestamps and asks recipients to record status changes. These are interface descriptions, not proof each function or input category is deployed.
\nPrinted6–10 and14–15 distinguish multiple associated cases, provider contributions, feedback replacement/removal and withdrawal of an organization's own submissions. Deletion markers have limited availability; pagination guidance warns that delayed retrieval can miss deleted records. Printed6's Withdrawn→Deleted timing and printed15's minimum24-hour withdrawal description should not become one exact tested retention guarantee. A batch-management feature is expressly unimplemented. Text/URL type labels and the political-free-speech example are schema/example content, not observed censorship records. Neither generic status feedback nor database withdrawal proves restoration of content/accounts. Ordinary PDF retrieval timed out; the complete16-page reader text was read, without executing examples or obtaining operational data.
\n[SWGfL, 13March2024](http[local research file] announced updated PhotoDNA integration allowing on-device hashing and access by more already-equipped platforms. It also announced Niantic joining. Microsoft provided an attributed supporting statement. This predates the Coalition's January2025 expanded-sublicensing announcement; that later announcement is not the first use of PhotoDNA for adult NCII.
\n[Microsoft, 5September2024](http[local research file] reports donating the capability that March and piloting StopNCII in Bing, with action on268,899 images by the end of August. This is an operator-reported aggregate for image-search results, not an independent audit, unique victims or removal from source websites. Its separate direct-report process describes review against Microsoft's policy and global search delisting. Full substantive reader body read; original GET403 closed.
\n[Google, 17September2025](http[local research file] announced planned hash use over the following months. It is not itself proof of completed implementation. No later completion record was located in the single focused Google follow-up query.
\nThe [eSafety August2025 report, updated February2026](http[local research file] records Snap using StopNCII on Spotlight/Discover/Stories, excluding Withdrawn/Deleted entries, with daily updates. Its Chat row does not list StopNCII. Printed7/11/18 establish the15June–15December2024 reporting period and that this summarizes provider responses, not comprehensive regulator testing. That supplies a dated, legally requested recipient account of withdrawal filtering, not a paired individual deletion/restoration trace or current2026 operating guarantee. Ordinary GET timed out and table screenshot failed; reader table/text scopes were read. No original PDF is claimed.
\n[Free Speech Coalition, 26February2026](http[local research file] urges adoption by adult platforms and explicitly links the case to legal/reputational risk and credibility with banks/payment processors. This is an industry association's recommendation, not a bank mandate or a payment-disconnection record. Its named adopters and volume claims were not independently corroborated here. The full body was read; retained Article metadata dates publication26February2026 and modification27February2026.
\nThe documented architecture permits institutional contributions and differentiated access while preserving platform policy judgments. It includes a described withdrawal channel and a named recipient account of honoring withdrawal states. There is no basis here for claiming all withdrawal requests fail, all recipients obey an automatic command, or banks require this particular product. The most useful missing record remains a current signed data-sharing agreement with its permitted-purpose and correction provisions, followed by a public implementation/audit trace. Wider operator resources and governance are in the sibling finance lane.
\nCustody: partners-captures.json records eight successful HTML originals, four original administrative JPGs, one403 error and two PDF timeouts. HTML text derivatives are separate. Complete substantive reads: join/programme/approved, all partner terms/schedule, PhotoDNA2024, Google2025 and FSC2026; scoped SWGfL partner privacy; all four diagrams visually. Reader-only: full Microsoft2024 body; full16-page API text; eSafety cover/date, report-period/limitations passages and Table3 Snap row. Other eSafety pages/search snippets are not a whole-report reading. The one malformed discovery query is excluded; eight other focused queries were used. A direct agreement query located no separate executed document within this finite search. Sources retain their own dates and claims. No registration form, case-status service, API, private dataset, outreach or paid action was used. Acquisition stopped for review and forest.
", "path": "stopncii-governance-wave-2026-09-09/partner-authority.md", "sourceIds": [ "w28-access-s-stop-join", "w28-access-s-stop-programme", "w28-access-s-stop-approved", "w28-access-s-stop-partner-terms", "w28-access-s-swgfl-partner-privacy", "w28-access-s-approved-fees", "w28-access-s-programme-fees", "w28-access-s-approved-onboarding", "w28-access-s-programme-onboarding", "w28-access-s-stop-api", "w28-access-s-stop-photodna2024", "w28-access-s-bing2024", "w28-access-s-google-stop2025", "w28-access-s-snap-esafety", "w28-access-s-fsc-adoption" ], "markdown": "9 September 2026. Root acquisition CLOSED. StopNCII exposes several distinct gates: admission and payment, reference contributions and feedback, withdrawal processing, and platform action. Published terms are not an executed partner-specific agreement. All observations below concern public administrative records; no live case or operational API was used.\n\n## Access costs and participation\n\nThe join page (see source references) offers Approved Participant and Programme Partner routes. Its comparison table assigns both every listed network/support/development benefit. Yet the Approved Participant page (see source references) expressly denies wider programme/network involvement. This is a real description conflict, not evidence of an undisclosed voting arrangement. The Programme Partner page (see source references) advertises active engagement, working groups and opportunities to influence tools/policies; it grants no stated vote or veto.\n\nAll four embedded fee/onboarding diagrams were acquired and visually read. Both routes show registration, verification, a data-sharing agreement, payment and integration. The full programme adds meeting the team and announcement. Published annual GBP fees, in the same revenue bands, are:\n\n| Revenue band | Approved Participant | Programme Partner |\n| --- | --- | --- |\n| New business | 1,000 | 1,000 |\n| Below 10 million | 4,000 | 10,000 |\n| 10–249 million | 8,000 | 20,000 |\n| 250–499 million | 12,000 | 30,000 |\n| 500–999 million | 16,000 | 40,000 |\n| Over 1 billion | 20,000 | 50,000 |\n\nThese are undated published schedules, not invoices, revenue received or proof every earlier partner paid. The wording leaves exact intervening band boundaries unclarified. The user-facing service being free does not establish free organizational access.\n\n## Posted instrument, with visible drafting limitations\n\nPartner Terms (see source references) cover manual and automated admission. They incorporate the unacquired data-sharing agreement and registration form. Sections3–10/schedule specify the lower fee schedule, eligibility/reverification, discretionary refusal, credential controls, security assessment, local caching and access suspension. Redistribution is prohibited; model training has a purpose-limited exception. Permitted purpose depends on the separate agreement. Termination requires cessation/deletion and written certification; annual summaries are due within60days of the access anniversary.\n\nThe posting contains conflicting company/charity identifiers, an empty hash-value definition, inconsistent defined-purpose terminology, bracketed implementation/amendment/certification periods and an explicit unfinished implementation note. Do not present those brackets as settled deadlines. Key-compromise notice is24hours; the schedule's general incident notice is48hours—retain their distinct scopes. The five-day certification placeholder concerns partner termination, not each user withdrawal. No signed data-sharing agreement, precedence clause resolving these inconsistencies, executed admission/refusal or payment was acquired. No displayed effective date was established.\n\n## Who screens applicants and who reviews a challenge\n\nSWGfL privacy (see source references) Working with Partners, names Deloitte LLP for business verification and Neterium for sanctions/PEP/watchlist/adverse-media screening. This concerns applicant businesses and connected officials, not identification of people submitting intimate-image cases. The notice describes human consideration of adverse results and rejects solely automated refusals. It promises notice where screening influenced an adverse decision, an uninvolved reviewer on request, correction/annotation and relevant corrections passed to vendors. Acknowledgment is2workingdays and substantive response30workingdays. Applicant records are retained during the relationship plus3years, or3years for unsuccessful applications.\n\nThose are specific process commitments. The partner terms' general sole-discretion/no-reasons clause remains broader; no acquired priority rule resolves every refusal category. This is not a proved exercise of arbitrary denial or an observed successful appeal. The retained HTML references vendor privacy notices without supplying actual links. Root read the complete partner section; other SWGfL processing purposes are outside this lane.\n\n## The database has more than a public-user input\n\nThe API guide (see source references) cover date08/10/2025, describes SWGfL-issued provider credentials, CSP hash submissions, NGO-reviewed Active entries, public-user Received entries and independent provider feedback. It documents change timestamps and asks recipients to record status changes. These are interface descriptions, not proof each function or input category is deployed.\n\nPrinted6–10 and14–15 distinguish multiple associated cases, provider contributions, feedback replacement/removal and withdrawal of an organization's own submissions. Deletion markers have limited availability; pagination guidance warns that delayed retrieval can miss deleted records. Printed6's Withdrawn→Deleted timing and printed15's minimum24-hour withdrawal description should not become one exact tested retention guarantee. A batch-management feature is expressly unimplemented. Text/URL type labels and the political-free-speech example are schema/example content, not observed censorship records. Neither generic status feedback nor database withdrawal proves restoration of content/accounts. Ordinary PDF retrieval timed out; the complete16-page reader text was read, without executing examples or obtaining operational data.\n\n## Named operating accounts and timeline\n\nSWGfL, 13March2024 (see source references) announced updated PhotoDNA integration allowing on-device hashing and access by more already-equipped platforms. It also announced Niantic joining. Microsoft provided an attributed supporting statement. This predates the Coalition's January2025 expanded-sublicensing announcement; that later announcement is not the first use of PhotoDNA for adult NCII.\n\nMicrosoft, 5September2024 (see source references) reports donating the capability that March and piloting StopNCII in Bing, with action on268,899 images by the end of August. This is an operator-reported aggregate for image-search results, not an independent audit, unique victims or removal from source websites. Its separate direct-report process describes review against Microsoft's policy and global search delisting. Full substantive reader body read; original GET403 closed.\n\nGoogle, 17September2025 (see source references) announced planned hash use over the following months. It is not itself proof of completed implementation. No later completion record was located in the single focused Google follow-up query.\n\nThe eSafety August2025 report, updated February2026 (see source references) records Snap using StopNCII on Spotlight/Discover/Stories, excluding Withdrawn/Deleted entries, with daily updates. Its Chat row does not list StopNCII. Printed7/11/18 establish the15June–15December2024 reporting period and that this summarizes provider responses, not comprehensive regulator testing. That supplies a dated, legally requested recipient account of withdrawal filtering, not a paired individual deletion/restoration trace or current2026 operating guarantee. Ordinary GET timed out and table screenshot failed; reader table/text scopes were read. No original PDF is claimed.\n\nFree Speech Coalition, 26February2026 (see source references) urges adoption by adult platforms and explicitly links the case to legal/reputational risk and credibility with banks/payment processors. This is an industry association's recommendation, not a bank mandate or a payment-disconnection record. Its named adopters and volume claims were not independently corroborated here. The full body was read; retained Article metadata dates publication26February2026 and modification27February2026.\n\n## What this establishes and what would change the conclusion\n\nThe documented architecture permits institutional contributions and differentiated access while preserving platform policy judgments. It includes a described withdrawal channel and a named recipient account of honoring withdrawal states. There is no basis here for claiming all withdrawal requests fail, all recipients obey an automatic command, or banks require this particular product. The most useful missing record remains a current signed data-sharing agreement with its permitted-purpose and correction provisions, followed by a public implementation/audit trace. Wider operator resources and governance are in the sibling finance lane.\n\nCustody: partners-captures.json records eight successful HTML originals, four original administrative JPGs, one403 error and two PDF timeouts. HTML text derivatives are separate. Complete substantive reads: join/programme/approved, all partner terms/schedule, PhotoDNA2024, Google2025 and FSC2026; scoped SWGfL partner privacy; all four diagrams visually. Reader-only: full Microsoft2024 body; full16-page API text; eSafety cover/date, report-period/limitations passages and Table3 Snap row. Other eSafety pages/search snippets are not a whole-report reading. The one malformed discovery query is excluded; eight other focused queries were used. A direct agreement query located no separate executed document within this finite search. Sources retain their own dates and claims. No registration form, case-status service, API, private dataset, outreach or paid action was used. Acquisition stopped for review and forest.\n", "file": "research/stopncii-governance-wave-2026-09-09-partner-authority.md" }, { "id": "stopncii-governance-wave-2026-09-09-user-authority", "title": "StopNCII public-user authority: entry, withdrawal and continuing platform decisions", "description": "Observed 9 September 2026. Acquisition closed. This packet covers the public-user route; it does not claim that this is the only operational route by which entries reach the database. Root separately owns partner terms, the newly discovered public API reference and named operating accounts. No live case, form, credentials, hash value or media was accessed.", "html": "Observed 9 September 2026. Acquisition closed. This packet covers the public-user route; it does not claim that this is the only operational route by which entries reach the database. Root separately owns partner terms, the newly discovered public API reference and named operating accounts. No live case, form, credentials, hash value or media was accessed.
\nThe significant boundary is between withdrawing a supplied signal and reversing what another organization decided after receiving it. The user documents explicitly address the first; they preserve room for independent platform policy enforcement rather than promising restoration.
\nThe [FAQ](http[local research file] requires the submitter to be depicted, possess the intimate material and be adult (its wording is “over 18” in the image); submission on another person's behalf is refused in this public-user path. It includes depicted-person synthetic images. Status updates may take 3–5 days. Adding further images requires a new case; withdrawal removes the whole case. Losing the case number or password prevents checking or withdrawal, and the FAQ offers no recovery. These are published eligibility and support rules, not evidence of independently verified identity or a legal adjudication of each item. Locators: eligibility questions; synthetic-image, status, adding-images, credential-loss and withdrawal answers.
\nThe [six-step explanation](http[local research file] says the user selects media and generates the hash on their device; the original does not leave that device. Platforms match against their systems and remove matching material if it violates their intimate-image-abuse policies. That conditional policy assessment is a separate decision from the user's selection and submission. The explanation does not establish that every match causes removal. All six steps were read; no workflow was run.
\nThe [StopNCII privacy notice](http[local research file] visibly last updated January 2026, says hashes go to all current and future participating platforms; users cannot choose recipients. The notice identifies SWGfL as controller and describes international sharing. Its “Data retention and security” section requires valid case credentials for withdrawal and promises erasure from SWGfL systems plus instructions to industry partners to erase. It simultaneously preserves partners' ability to continue removing content that violates their policies. The summary uses “ask”; the full notice uses “instruct”. Neither wording proves delivery, compliance or restoration.
\nThe notice separately describes personal-data access, rectification, erasure and objection rights, subject to exceptions, and a privacy complaint route followed by ICO escalation if unresolved. These general rights are not evidence of a working credential-recovery path or an adjudicated appeal against a platform's moderation. It says case numbers and hashed passwords remain while a case is active; this does not establish storage of plaintext passwords. Locators: Parts 1–2, “Who we share your data with”, “Data retention and security”, “Your Rights” and “Complaints”.
\nThe [SWGfL privacy policy](http[local research file] visibly updated May 2026, repeats the erasure/instruction promise and continuing-removal reservation in its StopNCII section. Its recipient sentence mentions platforms signed up at upload; the more specific StopNCII notice and FAQ expressly cover future partners too. The general policy's rectification/erasure/consent sections describe conditional data rights and normally a one-month response to rights requests. They do not specify a recipient hash-deletion acknowledgement or a platform restoration procedure. Read locators: StopNCII section; general “Your Rights”. The shared original was reused, not reacquired.
\nAn express right to keep removing offending content should not be silently converted into permission to retain the withdrawn StopNCII reference unchanged. Conversely, an instruction to erase a reference should not be converted into a promise to restore previously removed content or an affected account. The user documents read here provide neither a worked withdrawal trace nor a recipient acknowledgement deadline. The separate API/partner lane may refine those operational duties; this is not a programme-wide finding that no such protocol exists.
\nThe StopNCII privacy notice incorporates [SWGfL Website Terms of Use](http[local research file] that page expressly lists StopNCII among its sites. These are general website terms, including temporary service access, amendment/suspension rights, truthful registration information and possible registration termination for noncompliance. Its registration clause sets an adult age requirement. It points processing questions to privacy policies and acknowledges additional policies. The entire substantive page was read; no displayed revision date was identified. An original GET timed out; the successful reader extraction is retained instead. These general terms do not supply the separate partner data-sharing agreement or proof of a named user's termination.
\nOne technical-description qualification is necessary: the FAQ collectively calls PDQ/PhotoDNA/MD5 “open-sourced”. That sentence is preserved as publisher wording, not adopted as a verified PhotoDNA licence finding; the earlier Microsoft licensing packet establishes why software and service permissions must be examined separately. No new algorithm or code audit was attempted.
\nThe useful next record is a public, non-personal specification or worked audit connecting a case withdrawal to recipient delta processing, deletion acknowledgement and any separately retained policy basis for later enforcement. A user privacy request, withdrawal instruction and reversal of an enforcement action are different records. No individual case is needed to establish the governing schema, but none was sought here.
\nuser-captures.json records exact URLs, observation times, scopes and hashes. Three original HTML pages were captured here: FAQ, StopNCII privacy and how-it-works. Root's SWGfL privacy original is shared with its exact hash. All four originals pass local hash checks. user-website-terms.reader.json identifies a reader extraction only, not original HTML. The ordinary timeout route is closed. Full substantive reader scopes and narrower retained-text checks are distinguished in the manifest. Two focused searches were used; the indexed creation form was not opened or used. Public API details, partner admission, operator funding and actual platform accounts remain in sibling lanes. No further acquisition is pending in this lane.
9 September 2026. Root completed this bounded acquisition after the instruction forest. No outreach. This packet is not yet integrated into case-r20. Original captures and failures are recorded in captures.json. Public-page statements, contract terms, expenditure and performed outcomes are distinct evidence states.
\nCompanies House identifies THE GLOBAL STRATEGY NETWORK LTD, company 10433562, incorporated 18 October 2016 and currently active. Its public PSC register records Giles Henry Donne Barrett and Matthew Thomas King each with more than 25% and not more than 50% of shares and voting rights, notified respectively 24 June 2021 and 28 October 2020. Richard Barrett's PSC status ceased 24 June 2021. These bands do not establish an exact half-share each, a family relationship, or Richard's current control. Personal addresses and birth information were not extracted into the case.
\nSources: http[local research file] and http[local research file] (direct primary reader).
\nThe current NSOIT privacy notice names Global Strategy Network as its appointed public-social-media analysis contractor following an open tender. Its page's initial publication date does not date the appointment paragraph. The retained privacy capture belongs to the preceding instruction packet. This does not prove Crisp's exact exit date or that the two suppliers never worked in parallel.
\nThe 27 January 2026 ministerial answer to question 106529 separates two functions: GSN provides analysis of online misinformation/disinformation narratives relevant to UK public safety and national security; Faculty Science develops the Counter-Disinformation Data Platform (CDDP). The question itself misspells the supplier names; the answer is the attribution anchor. This is an official role account, not a technical or effectiveness audit.
\nSource: http[local research file]
\nDSIT's September 2025 spending CSV records a payment of GBP 118,000 on 12 September 2025 to The Global Strategy Network Ltd, transaction 719926. Expense: Consultancy Expenditure; area: Security and Online Harms; description: SOH Info Threats and Security Programme. The original CSV has blank contract/project/expenditure-type fields, rendered as Not set in the HTML preview. This is one payment, not a contract ceiling, annual budget or demonstrated invoice scope. VAT basis and service period are not provided. The supplier is named, without a company number in the row; association to the registered company is a name match supported by official supplier naming, not a registry identifier in this payment.
\nOriginal: http[local research file] ; 67,256 bytes; SHA256 2a1765cbb3b334406fd8f8def9897474684520222b909ab9f58d9fe61c39edcc. Captured 9 September 2026. Preview updated 30 July 2026: http[local research file] . Retained HTML SHA256 c04fb5c8ef1af21ed50ffb1d3803ed251fe5a1770d916e16b5d512933a35931f.
\nA government procurement page for CDDP development, hosting and maintenance dates stage-one publication to 11 September 2023 and closing to 25 September 2023. Its stated team locates prioritisation in DSIT: a product manager owns sprint priorities/user research, a programme manager leads strategy, and analysts test as expert users. Proposed working arrangements put sprint goals with DSIT, alongside a prioritised backlog and quarterly strategic workshops. The technical supplier would support the government team using, evaluating and directing the platform, manage AWS hosting, and develop/maintain prototypes. Users are expressly within HMG. These are tender requirements, not proof that every practice occurred.
\nThe page identifies Faculty as incumbent in extended Alpha, describes a March 2024 contract end and procurement for transition. It also gives an October 2023 latest start, residual BEIS branding and test placeholders: preserve those inconsistencies and do not silently normalize the page into an executed contract. Publication records 19 applications; the separate assessment section proposes evaluating five suppliers. Neither is an award result. Stage-two suppliers would sign an NDA before receiving the complete ITT. The listed indicative budget is 500,000–920,000, with a one-year length, six-month extension and six-month break clause; do not aggregate it with contract variations or treat it as expenditure. The dated January 2026 answer still names Faculty, so an actual supplier replacement is not established.
\nSource: http[local research file] . Read context, working arrangements, technical scope, assessment and timeline. Procurement-source wording is not a direct CDDP product inspection.
\nThe official FOI response inventories withheld Discovery Report and Product Operating Model alongside partly disclosed business-case and contract records, including an expert-evaluation procurement, Oliver Wyman evaluation contract, Faculty variations and NewsWhip contract. Its inventory is not equivalent to reading all of those documents.
\nRelease: http[local research file]
\nRetained original: http[local research file] ; 262 PDF pages, 11,778,384 bytes; SHA256 71bd6a522a765d53f9aa79b239134226be64891ea3c67711abc9d788fb87748f. The release path says 2024 while the original filename says 2023; retain both. A full text derivative exists, but only the following pages/passages were substantively inspected in this run. Page numbers below are one-based PDF pages, not always printed pagination.
\nPages 1–2: October 2021 Shared Outcomes Fund business case, DCMS Security and Online Harms, SRO Sarah Connolly; the opening description envisages cross-government data sharing and a common picture of online information threats, with Discovery and Delivery phases. It explicitly includes misinformation, disinformation and manipulative behaviour. Redacted/unreadable participating-department and value fields are not reconstructed. Page 17 begins critical-success criteria; the following detailed evaluation section is reserved for the next run.
\nPages 101–102: Faculty variation 102548 No.2, dated 16 August 2023, effective 17 March 2023, extends to 31 March 2024 with capped fee GBP 510,000 excluding VAT and stated revised total GBP 2,373,873. It coordinates IP terms across the G-Cloud and this call-off, retrospectively and prospectively. Independent visual review confirms this disclosed copy has visibly uncompleted signature, name, title and date fields; execution is not established from this copy.
\nPages 110–112, Appendix C: the parties vary the arrangements so new IP and specially written software shall be supplier-owned and licensed to the buyer under a perpetual licence, not assigned. This does not imply that the earlier ownership allocation was already identical. Permitted third parties can use, adapt and modify the retained IP to supply maintenance, hosting and development to the buyer. They cannot exploit it commercially for other purposes or build a competing commercial product from it. They need no direct agreement with the supplier. The buyer must investigate evidenced misuse and, in sustained breach, terminate the applicable sublicense, subject to procurement rules. Source code and relevant documentation must be accessible and usable to the buyer and its contractors. This splits intellectual-property ownership from technical direction and replacement rights. It does not establish that an accessible repository was actually delivered, that independent maintenance succeeded, or that Faculty dictated the government's policy.
\nPage 114 identifies Faculty Science Limited, registration 08873131, in order 102548. Page 133 contains the 16 February 2022 variation and GBP 250,000 excluding VAT, with a design document and proof-of-concept report among deliverables. Some surrounding deliverable detail is not readable. Page 135, variation 102887 No.1 dated 13 December 2022, says milestone-two MVP completed 30 September 2022 and revises the cap from GBP 1,264,650 to GBP 1,514,650. Page 137, No.2 dated 1 March 2023/effective 16 January, adds GBP 400,000 excluding VAT payable against delivery/buyer acceptance and prints total GBP 1,944,650. That is GBP 30,000 above a simple addition to the previous stated total; intervening terms/reconciliation remain uninspected. No combined-spend total, fraud inference or silent correction is appropriate.
\nPage 143 identifies NewsWhip Media Limited, Irish registration 485016, contract 103186 for real-time social/predictive media news monitoring intelligence, 28 October 2022–27 October 2023 with an optional one-year buyer extension. No charge read. Pages 193–194 contain programme-success and milestone tables with almost no recoverable substantive cells; visual examination is needed before attributing absent text to redaction rather than extraction failure. Pages 109 and 112 mostly blank; page 113 is the next framework cover. The expert-evaluation material was located by text search on pages 34–35, 62–63, 89, 92 and 95 but is not yet substantively read.
\nThe actual partly disclosed Crisp report for the week beginning 24 June 2024 was not located through targeted public searches using the ICO decision, release date and period. This is not proof that no public copy exists. The ICO independently confirms its existence and partial disclosure; no matched referral/removal result was found.
\nThe exact GSN SOW, performance criteria and invoice behind transaction 719926 remain missing. Current appointment, January 2026 official role description and September 2025 payment do not fix the Crisp/GSN handover date. An official 18 November 2025 answer 88698 contains four contract links in the daily parliamentary PDF; the public reader exposed the labels, not underlying URLs. The ordinary PDF download returned 403 and the route is closed. The individual question page may be a later legitimate route if located; do not invent its date.
\nA later Faculty call-off attachment was indexed, but reader/download failed. Its indexed registration field reads 438780, unlike the acquired earlier contract's 08873131. Do not promote that indexed cover into an executed award. The precise failed attachment and attempts are in captures.json; no retries.
\nCompanies House search also surfaced Extrac AI Limited 14532778 with historical GSN corporate control and later personal PSCs. This is an unpromoted lead, not a demonstrated NSOIT supplier, deployed system or inference that all operations belonged to one company. Relevance should be tested before expanding biographies.
\nThe next revealing evidence is ordinary project governance: evaluation commission and resulting assessment, measurable acceptance criteria, and whether the licenced code/replacement rights were actually exercised. The retained bundle makes that a bounded next run rather than speculative searching.
", "path": "supplier-platform-wave-2026-09-09/supplier-platform.md", "sourceIds": [ "w16-supp-s-gsn-register", "w16-supp-s-gsn-psc", "w16-supp-s-roles2026", "w16-supp-s-gsn-payment", "w16-supp-s-procurement2023", "w16-supp-s-contract-links-unavailable", "w16-supp-s-later-faculty-unavailable" ], "markdown": "9 September 2026. Root completed this bounded acquisition after the instruction forest. No outreach. This packet is not yet integrated into case-r20. Original captures and failures are recorded in captures.json. Public-page statements, contract terms, expenditure and performed outcomes are distinct evidence states.\n\n## The present analysis supplier\n\nCompanies House identifies THE GLOBAL STRATEGY NETWORK LTD, company 10433562, incorporated 18 October 2016 and currently active. Its public PSC register records Giles Henry Donne Barrett and Matthew Thomas King each with more than 25% and not more than 50% of shares and voting rights, notified respectively 24 June 2021 and 28 October 2020. Richard Barrett's PSC status ceased 24 June 2021. These bands do not establish an exact half-share each, a family relationship, or Richard's current control. Personal addresses and birth information were not extracted into the case.\n\nSources: http[local research file] and http[local research file] (direct primary reader).\n\nThe current NSOIT privacy notice names Global Strategy Network as its appointed public-social-media analysis contractor following an open tender. Its page's initial publication date does not date the appointment paragraph. The retained privacy capture belongs to the preceding instruction packet. This does not prove Crisp's exact exit date or that the two suppliers never worked in parallel.\n\nThe 27 January 2026 ministerial answer to question 106529 separates two functions: GSN provides analysis of online misinformation/disinformation narratives relevant to UK public safety and national security; Faculty Science develops the Counter-Disinformation Data Platform (CDDP). The question itself misspells the supplier names; the answer is the attribution anchor. This is an official role account, not a technical or effectiveness audit.\n\nSource: http[local research file]\n\nDSIT's September 2025 spending CSV records a payment of GBP 118,000 on 12 September 2025 to The Global Strategy Network Ltd, transaction 719926. Expense: Consultancy Expenditure; area: Security and Online Harms; description: SOH Info Threats and Security Programme. The original CSV has blank contract/project/expenditure-type fields, rendered as Not set in the HTML preview. This is one payment, not a contract ceiling, annual budget or demonstrated invoice scope. VAT basis and service period are not provided. The supplier is named, without a company number in the row; association to the registered company is a name match supported by official supplier naming, not a registry identifier in this payment.\n\nOriginal: http[local research file] ; 67,256 bytes; SHA256 2a1765cbb3b334406fd8f8def9897474684520222b909ab9f58d9fe61c39edcc. Captured 9 September 2026. Preview updated 30 July 2026: http[local research file] . Retained HTML SHA256 c04fb5c8ef1af21ed50ffb1d3803ed251fe5a1770d916e16b5d512933a35931f.\n\n## A separate government-directed technical platform\n\nA government procurement page for CDDP development, hosting and maintenance dates stage-one publication to 11 September 2023 and closing to 25 September 2023. Its stated team locates prioritisation in DSIT: a product manager owns sprint priorities/user research, a programme manager leads strategy, and analysts test as expert users. Proposed working arrangements put sprint goals with DSIT, alongside a prioritised backlog and quarterly strategic workshops. The technical supplier would support the government team using, evaluating and directing the platform, manage AWS hosting, and develop/maintain prototypes. Users are expressly within HMG. These are tender requirements, not proof that every practice occurred.\n\nThe page identifies Faculty as incumbent in extended Alpha, describes a March 2024 contract end and procurement for transition. It also gives an October 2023 latest start, residual BEIS branding and test placeholders: preserve those inconsistencies and do not silently normalize the page into an executed contract. Publication records 19 applications; the separate assessment section proposes evaluating five suppliers. Neither is an award result. Stage-two suppliers would sign an NDA before receiving the complete ITT. The listed indicative budget is 500,000–920,000, with a one-year length, six-month extension and six-month break clause; do not aggregate it with contract variations or treat it as expenditure. The dated January 2026 answer still names Faculty, so an actual supplier replacement is not established.\n\nSource: http[local research file] . Read context, working arrangements, technical scope, assessment and timeline. Procurement-source wording is not a direct CDDP product inspection.\n\n## Original contracts released under FOI\n\nThe official FOI response inventories withheld Discovery Report and Product Operating Model alongside partly disclosed business-case and contract records, including an expert-evaluation procurement, Oliver Wyman evaluation contract, Faculty variations and NewsWhip contract. Its inventory is not equivalent to reading all of those documents.\n\nRelease: http[local research file]\n\nRetained original: http[local research file] ; 262 PDF pages, 11,778,384 bytes; SHA256 71bd6a522a765d53f9aa79b239134226be64891ea3c67711abc9d788fb87748f. The release path says 2024 while the original filename says 2023; retain both. A full text derivative exists, but only the following pages/passages were substantively inspected in this run. Page numbers below are one-based PDF pages, not always printed pagination.\n\nPages 1–2: October 2021 Shared Outcomes Fund business case, DCMS Security and Online Harms, SRO Sarah Connolly; the opening description envisages cross-government data sharing and a common picture of online information threats, with Discovery and Delivery phases. It explicitly includes misinformation, disinformation and manipulative behaviour. Redacted/unreadable participating-department and value fields are not reconstructed. Page 17 begins critical-success criteria; the following detailed evaluation section is reserved for the next run.\n\nPages 101–102: Faculty variation 102548 No.2, dated 16 August 2023, effective 17 March 2023, extends to 31 March 2024 with capped fee GBP 510,000 excluding VAT and stated revised total GBP 2,373,873. It coordinates IP terms across the G-Cloud and this call-off, retrospectively and prospectively. Independent visual review confirms this disclosed copy has visibly uncompleted signature, name, title and date fields; execution is not established from this copy.\n\nPages 110–112, Appendix C: the parties vary the arrangements so new IP and specially written software shall be supplier-owned and licensed to the buyer under a perpetual licence, not assigned. This does not imply that the earlier ownership allocation was already identical. Permitted third parties can use, adapt and modify the retained IP to supply maintenance, hosting and development to the buyer. They cannot exploit it commercially for other purposes or build a competing commercial product from it. They need no direct agreement with the supplier. The buyer must investigate evidenced misuse and, in sustained breach, terminate the applicable sublicense, subject to procurement rules. Source code and relevant documentation must be accessible and usable to the buyer and its contractors. This splits intellectual-property ownership from technical direction and replacement rights. It does not establish that an accessible repository was actually delivered, that independent maintenance succeeded, or that Faculty dictated the government's policy.\n\nPage 114 identifies Faculty Science Limited, registration 08873131, in order 102548. Page 133 contains the 16 February 2022 variation and GBP 250,000 excluding VAT, with a design document and proof-of-concept report among deliverables. Some surrounding deliverable detail is not readable. Page 135, variation 102887 No.1 dated 13 December 2022, says milestone-two MVP completed 30 September 2022 and revises the cap from GBP 1,264,650 to GBP 1,514,650. Page 137, No.2 dated 1 March 2023/effective 16 January, adds GBP 400,000 excluding VAT payable against delivery/buyer acceptance and prints total GBP 1,944,650. That is GBP 30,000 above a simple addition to the previous stated total; intervening terms/reconciliation remain uninspected. No combined-spend total, fraud inference or silent correction is appropriate.\n\nPage 143 identifies NewsWhip Media Limited, Irish registration 485016, contract 103186 for real-time social/predictive media news monitoring intelligence, 28 October 2022–27 October 2023 with an optional one-year buyer extension. No charge read. Pages 193–194 contain programme-success and milestone tables with almost no recoverable substantive cells; visual examination is needed before attributing absent text to redaction rather than extraction failure. Pages 109 and 112 mostly blank; page 113 is the next framework cover. The expert-evaluation material was located by text search on pages 34–35, 62–63, 89, 92 and 95 but is not yet substantively read.\n\n## Gaps retained as specific leads\n\nThe actual partly disclosed Crisp report for the week beginning 24 June 2024 was not located through targeted public searches using the ICO decision, release date and period. This is not proof that no public copy exists. The ICO independently confirms its existence and partial disclosure; no matched referral/removal result was found.\n\nThe exact GSN SOW, performance criteria and invoice behind transaction 719926 remain missing. Current appointment, January 2026 official role description and September 2025 payment do not fix the Crisp/GSN handover date. An official 18 November 2025 answer 88698 contains four contract links in the daily parliamentary PDF; the public reader exposed the labels, not underlying URLs. The ordinary PDF download returned 403 and the route is closed. The individual question page may be a later legitimate route if located; do not invent its date.\n\nA later Faculty call-off attachment was indexed, but reader/download failed. Its indexed registration field reads 438780, unlike the acquired earlier contract's 08873131. Do not promote that indexed cover into an executed award. The precise failed attachment and attempts are in captures.json; no retries.\n\nCompanies House search also surfaced Extrac AI Limited 14532778 with historical GSN corporate control and later personal PSCs. This is an unpromoted lead, not a demonstrated NSOIT supplier, deployed system or inference that all operations belonged to one company. Relevance should be tested before expanding biographies.\n\nThe next revealing evidence is ordinary project governance: evaluation commission and resulting assessment, measurable acceptance criteria, and whether the licenced code/replacement rights were actually exercised. The retained bundle makes that a bounded next run rather than speculative searching.\n", "file": "research/supplier-platform-wave-2026-09-09-supplier-platform.md" }, { "id": "taxonomy-wave-2026-09-09-operative-categories", "title": "From harm vocabulary to proposed Australian legal categories", "description": "Bounded taxonomy comparison, 9 September 2026, after the recorded forest assessment. This note compares express source use with legal choices; the September 2026 document remains an exposure draft. It does not establish that WEF, HRLC or Reset controls Australian drafting or enforcement.", "html": "Bounded taxonomy comparison, 9 September 2026, after the recorded forest assessment. This note compares express source use with legal choices; the September 2026 document remains an exposure draft. It does not establish that WEF, HRLC or Reset controls Australian drafting or enforcement.
\nThe strongest new finding is an independent official route for the WEF vocabulary before HRLC's submission. The government's April 2024 review issues paper already cited the typology. HRLC therefore cannot be treated as the sole intermediary. Independent official route here means an official citation before that HRLC submission; it does not establish causal independence from all advocates, earlier relationships or shared workstreams. The strongest narrowing is the published review's decision to defer a broader consolidation of financial, societal and misinformation harms. The draft has a materially different structure from the typology. Neither conceptual overlap nor omission of a label, by itself, resolves the draft's complete reach.
\n| Record | \nWhat the inspected passage establishes | \nWhat it does not establish | \n
|---|---|---|
| April 2024 government review issues paper | \nPart 3, Overview of online harm, fn17, uses WEF's account of content, contact and conduct and their effects. The technology-facilitated-abuse discussion cites WEF p8 at fn88; the gender-based-violence discussion does so at fn93. | \nThe references inform the issues paper's descriptions. They are not legislative incorporation or a completed response to later NGO proposals. | \n
| HRLC submission155, July 2024 | \nSection 4.3, printed/PDF p9, expressly adopts the WEF framework. Section 5, p10, adopts Reset's five-pillar architecture. Section 5.1.5, p12, proposes complaint/redress categories including misinformation. | \nHRLC's advocacy position is not a government decision; a taxonomy reference does not itself prescribe a ban or particular intervention. | \n
| Current eSafety Safety by Design guidance | \nUnder “The types of online harms,” explicitly uses the WEF typology for overlapping categories. The page displays last updated 06/03/2026. | \nThis is industry guidance, not the exposure draft's definitions or a binding legislative instrument. The update is not an established original publication date. | \n
Sources: [Government issues paper, Part 3/fn17 and later fn88/93](http[local research file] [HRLC155, §§4.3–5.1.5](http[local research file] [eSafety guidance, types of online harms](http[local research file]
\nThe separate [producer packet]([local research file] establishes the workstream and contextual limits from its inspected sources. Its original WEF PDF was not acquired; detailed counterpart definitions must retain the Resolver/explainer attribution. Do not promote that packet into a claim that every cell of the original matrix was visually inspected. This lane adds actual government-side use to the production/NGO path; it does not establish who selected the later statutory clauses.
\nReset's April 2024 Green Paper describes its own proposals as a moment-specific contribution to discussion. Printed p6 explains its focus on online safety and misinformation/disinformation. The five-pillar table, printed p7/PDF9, distinguishes duty, assessment, mitigation, transparency and accountability. Its proposed assessments encompass existing Australian content/abuse categories, scams, electoral processes, public security, rights, public health and environmental risks. It expressly draws comparisons with the EU DSA, UK OSA and proposed Canadian legislation; its commentary criticises plural duties for weakening a systemic focus. This is a documented set of proposed policy choices with identified antecedents, not a WEF-only derivation. [Reset Green Paper, cover/about and printed pp6–7](http[local research file]
\nIn the published updated statutory review, recommendation 5, printed p56, proposes broad families concerning young people, mental/physical wellbeing, harmful practices, national security/social cohesion and other illegality. Its adjoining discussion says broader financial, societal and misinformation harms fit a duty model but are handled by other existing or proposed Australian legislation and regulators. It leaves possible consolidation under an expanded commission to a future review because the current reform is already substantial. This is a deliberate scope choice. It is not a named rejection of WEF or a finding those harms are imaginary. Printed p12 also frames risk assessment through EU/UK models and existing Australian industry codes. [Published review, printed pp12 and 56](http[local research file]
\nThe government's April 2026 response supports recommendation 4 and supports recommendation 5 in principle. Its explanation defines that state as support for intent/merit with alternative delivery or further implementation consideration; it is not acceptance of every proposed term. For recommendation 5 it stresses vulnerable people, criminal activity and public safety alongside expression, and says the duty complements the Criminal Code without changing obligations under existing discrimination statutes. It separately declines to simply lower adult-abuse thresholds, citing operational costs and the distinction between systemic prevention and reactive schemes. [Government response, PDF/printed pp3–6, recommendations 4–5 and adult-abuse discussion](http[local research file]
\nFOI25-247 Document 1, MS24-002100, is a departmental brief seeking acceptance of an updated final report. It records that Rickard sought final improvements at a 1 November 2024 meeting, after delivery on 31 October. Assistant Secretary Andrew Irwin cleared the brief on 8 November. Attachment C, PDF7, reports five additions and 68 total recommendations, with four characterized as already in the text but not called out; it lists new recommendation 5. Its introductory fifth item describes a daily penalty for non-compliance with a duty action, while item g/new recommendation35 describes a $10m company maximum for removal-notice non-compliance and says this was not previously addressed. These descriptions are inconsistent. The later government response says the published review has 67 recommendations. No original total, deleted recommendation or final numbering is reconstructed by subtraction. PDF8 records a terminology-order change placing mental before physical wellbeing. The department describes changes as consistent with the delivered report and not changing findings; this is its characterization, not an acquired complete version comparison.
\nThis establishes a post-delivery recommendation call-out and editorial revision. It does not establish that the substantive harm analysis originated in November, identify the individual drafter of recommendation 5, or show that the minister authored it. The readable approval alternatives/date fields are not proof of executed ministerial acceptance. Use published updated review, and distinguish original delivery from revision, rather than silently dating every public-page formulation to 31 October. [FOI25-247, Document 1 PDF1–3 and Attachment C PDF7–8](http[local research file]
\nThe exposure draft's distinctions are consequential:
\n| Provision | \nScope to preserve | \n
|---|---|
| §§25B–D, printed pp26–29 | \nSeparate serious-harm protection, additional child protection and under-16 social-media feature obligations. Initial lists include targeted harassment, terrorism and harmful practices; no general misinformation/democracy category is named. | \n
| §§25C(2), 25D(2), 27B | \nMinister-added categories require potential serious harm, or potential harm to children respectively; the minister must seek and consider eSafety advice. The child addition test does not expressly require seriousness. | \n
| §§25E, 26A | \nNon-user and combined harms count. Assessment covers all reasonably foreseeable risks, including those relevant to the duty. Label absence does not settle every assessment obligation. | \n
| §§25F–G | \nSpecified features are deemed to have negative behavioural impacts. Ministerial additions have no express serious-harm threshold in §25G itself. | \n
| §§25H, 26(7) | \nReasonable practicability weighs risk, available measures, cost and reasonable privacy expectations. The exception concerns lawful private communications solely between consenting adults; it does not exempt all lawful adult speech or disable every separate Act power. | \n
These are proposed Australian legal mechanisms, not an incorporation of WEF's matrix. False information can intersect listed conduct; future additions remain possible. [Exposure draft, §§25B–26A and 27B, printed pp26–34/37, PDF30–38/41](http[local research file]
\nThe as-made 2021 Act already defined serious physical/mental harm, distinguished serious distress from ordinary emotional reactions, and imposed targeted child/adult abuse tests. Its adult test combined likely intent to cause serious harm with an ordinary-reasonable-person assessment of the material. Sections 8–9 addressed contextual merit/consent and abhorrent violence. Section 106 tied class 1 to the Classification Act 1995 refusal category. These are directly inspected predecessors to the 2023 typology, rather than a presumed earlier neutral vocabulary. They supply a concrete competing source for similarities, while not proving every draft clause was copied from them. [As-made Online Safety Act 2021, §§5–9 and 106](http[local research file]
\nThe April issues paper also discusses Australian research and existing abuse/content schemes beside its WEF references. This is consistent with synthesis from several inputs. The record supports influence through framing and documented advocacy; it does not support an exclusive linear chain in which a private taxonomy became law unchanged. Conversely, the existence of older law and contextual safeguards does not determine how broad future instruments or enforcement would become.
\nPreserve WEF typology → government April issues paper: cited for descriptive framing, independently of WEF typology → HRLC155: expressly adopted framing. Add WEF typology → eSafety guidance: explicitly used for educational categorisation, with the page's update-date boundary. Preserve Reset model → HRLC155: expressly endorsed, and review recommendation 5 → government response: supported in principle. Treat review taxonomy ↔ draft categories as a documented comparison unless a clause-level attribution is obtained. None of these is a donor-instruction, private veto or enacted-ban relation.
\nThe next discriminating record is a category-by-category policy instruction, consultation disposition or drafting table for §§25C–G, showing source, alternatives, threshold selection and ministerial clearance. The likely policy custodian is the department's online-safety reform team; the relevant record family is the 2025–26 duty-of-care drafting/consultation file, including instructions to legislative drafters. It would distinguish accepted wording from shared vocabulary and establish why broader labels were omitted or particular additions delegated. No such instruction or disposition was acquired here. This lane stops at that record-to-decision breakpoint for the forest review.
\nAll current access observations are 9 September 2026. This note contains scoped public-reader/indexed-primary findings. No original PDF bytes, screenshots or original-byte hashes were acquired in this lane. PDF/printed numbering is differentiated where verified. The separate prior authority note supplies a fuller clause discussion, not new visual evidence.
\n| Source alias | \nInspected scope / current working references | \n
|---|---|
| taxonomy-op-s-hrlc155 | \n§4.3 through §5.1.5, printed/PDF pp9–12; reader turn925view0/1 and indexed introductory passage turn926search4. Not every submission page. | \n
| taxonomy-op-s-issues2024 | \nPart 3 overview/fn17; online-hate and technology-facilitated-abuse discussion/fn88/93. Reader turn935view1 and turn937view0/1. Date label April 2024; filename/internal working-title details do not establish a publication day. | \n
| taxonomy-op-s-esafety | \nTypes-of-harms section, following impact paragraphs and displayed update. Reader turn935view2 and turn937view2/3. Current page, no historical revision capture. | \n
| taxonomy-op-s-resetgreen | \nCover/about, printed pp1–3 background and pp6–7 systemic proposal/table. Reader turn940view1, turn941view0, turn942view1. Table is extracted text; no visual matrix claim. | \n
| taxonomy-op-s-review | \nComplete indexed printed p56 recommendation and adjoining scope discussion, plus indexed p12 summary. turn926search0, turn941search12, turn943search12. A reader open returned content-too-large (13,286,332 bytes); no ordinary GET retried. Detailed §5.5 attribution is retained in prior policy-lineage.md, not newly audited here. | \n
| taxonomy-op-s-foi247 | \nDocument 1 PDF1–3, Attachment C PDF7–8, and limited draft-letter context PDF9. turn929view0, turn931view1, turn935view0. No inference from blank/alternative approval text to signed decision. | \n
| taxonomy-op-s-response | \nPDF/printed pp3–6: response-state definitions, recommendations 4–5, related scope and adult-abuse discussion. turn947view0, reusing turn920view0. | \n
| taxonomy-op-s-draft | \n§§25A–26A and 27B; scoped neighbouring compliance text also returned. turn926view0, turn929view1–3, turn931view0, turn942view2, turn943view1, turn944view0. Public text extraction only; closed normal-GET failures not repeated. | \n
| taxonomy-op-s-osa2021 | \nOriginal/as-made ePub §§5–9 and class1 §106. Reached actual dated original through the Register's section link: turn944view1, turn946view0–2. The landing table-of-contents alone was not treated as historical substantive text. A December 2024 compilation was also read for comparison, turn931view2/turn935view3. | \n
No contributor biographies, new donor roster, private accounts, outreach, publication or canonical case edits were added. No inference about unseen draft material or access restrictions follows from these text-only readings.
", "path": "taxonomy-wave-2026-09-09/operative-categories.md", "sourceIds": [ "w14-tax-s-issues2024", "w14-tax-s-esafety-guidance", "w14-tax-s-reset-green", "w14-tax-s-act2021" ], "markdown": "Bounded taxonomy comparison, 9 September 2026, after the recorded forest assessment. This note compares express source use with legal choices; the September 2026 document remains an exposure draft. It does not establish that WEF, HRLC or Reset controls Australian drafting or enforcement.\n\nThe strongest new finding is an **independent official route for the WEF vocabulary before HRLC's submission**. The government's April 2024 review issues paper already cited the typology. HRLC therefore cannot be treated as the sole intermediary. Independent official route here means an official citation before that HRLC submission; it does not establish causal independence from all advocates, earlier relationships or shared workstreams. The strongest narrowing is the published review's decision to defer a broader consolidation of financial, societal and misinformation harms. The draft has a materially different structure from the typology. Neither conceptual overlap nor omission of a label, by itself, resolves the draft's complete reach.\n\n## Express transmission and its limits\n\n| Record | What the inspected passage establishes | What it does not establish |\n| --- | --- | --- |\n| April 2024 government review issues paper | Part 3, Overview of online harm, fn17, uses WEF's account of content, contact and conduct and their effects. The technology-facilitated-abuse discussion cites WEF p8 at fn88; the gender-based-violence discussion does so at fn93. | The references inform the issues paper's descriptions. They are not legislative incorporation or a completed response to later NGO proposals. |\n| HRLC submission155, July 2024 | Section 4.3, printed/PDF p9, expressly adopts the WEF framework. Section 5, p10, adopts Reset's five-pillar architecture. Section 5.1.5, p12, proposes complaint/redress categories including misinformation. | HRLC's advocacy position is not a government decision; a taxonomy reference does not itself prescribe a ban or particular intervention. |\n| Current eSafety Safety by Design guidance | Under “The types of online harms,” explicitly uses the WEF typology for overlapping categories. The page displays last updated 06/03/2026. | This is industry guidance, not the exposure draft's definitions or a binding legislative instrument. The update is not an established original publication date. |\n\nSources: Government issues paper, Part 3/fn17 and later fn88/93 (see source references) HRLC155, §§4.3–5.1.5 (see source references) eSafety guidance, types of online harms (see source references)\n\nThe separate [producer packet]([local research file] establishes the workstream and contextual limits from its inspected sources. Its original WEF PDF was not acquired; detailed counterpart definitions must retain the Resolver/explainer attribution. Do not promote that packet into a claim that every cell of the original matrix was visually inspected. This lane adds actual government-side use to the production/NGO path; it does not establish who selected the later statutory clauses.\n\n## Reset's broader model and the review's narrower choice\n\nReset's April 2024 Green Paper describes its own proposals as a moment-specific contribution to discussion. Printed p6 explains its focus on online safety and misinformation/disinformation. The five-pillar table, printed p7/PDF9, distinguishes duty, assessment, mitigation, transparency and accountability. Its proposed assessments encompass existing Australian content/abuse categories, scams, electoral processes, public security, rights, public health and environmental risks. It expressly draws comparisons with the EU DSA, UK OSA and proposed Canadian legislation; its commentary criticises plural duties for weakening a systemic focus. This is a documented set of proposed policy choices with identified antecedents, not a WEF-only derivation. Reset Green Paper, cover/about and printed pp6–7 (see source references)\n\nIn the published updated statutory review, recommendation 5, printed p56, proposes broad families concerning young people, mental/physical wellbeing, harmful practices, national security/social cohesion and other illegality. Its adjoining discussion says broader financial, societal and misinformation harms fit a duty model but are handled by other existing or proposed Australian legislation and regulators. It leaves possible consolidation under an expanded commission to a future review because the current reform is already substantial. This is a deliberate scope choice. It is **not** a named rejection of WEF or a finding those harms are imaginary. Printed p12 also frames risk assessment through EU/UK models and existing Australian industry codes. Published review, printed pp12 and 56 (see source references)\n\nThe government's April 2026 response supports recommendation 4 and supports recommendation 5 **in principle**. Its explanation defines that state as support for intent/merit with alternative delivery or further implementation consideration; it is not acceptance of every proposed term. For recommendation 5 it stresses vulnerable people, criminal activity and public safety alongside expression, and says the duty complements the Criminal Code without changing obligations under existing discrimination statutes. It separately declines to simply lower adult-abuse thresholds, citing operational costs and the distinction between systemic prevention and reactive schemes. Government response, PDF/printed pp3–6, recommendations 4–5 and adult-abuse discussion (see source references)\n\n## A consequential report-version qualification\n\nFOI25-247 Document 1, MS24-002100, is a departmental brief seeking acceptance of an updated final report. It records that Rickard sought final improvements at a **1 November 2024 meeting**, after delivery on **31 October**. Assistant Secretary **Andrew Irwin** cleared the brief on **8 November**. Attachment C, PDF7, reports five additions and **68 total recommendations**, with four characterized as already in the text but not called out; it lists new recommendation 5. Its introductory fifth item describes a daily penalty for non-compliance with a duty action, while item g/new recommendation35 describes a **$10m company maximum for removal-notice non-compliance** and says this was not previously addressed. These descriptions are inconsistent. The later government response says the published review has **67 recommendations**. No original total, deleted recommendation or final numbering is reconstructed by subtraction. PDF8 records a terminology-order change placing mental before physical wellbeing. The department describes changes as consistent with the delivered report and not changing findings; this is its characterization, not an acquired complete version comparison.\n\nThis establishes a post-delivery recommendation call-out and editorial revision. It does not establish that the substantive harm analysis originated in November, identify the individual drafter of recommendation 5, or show that the minister authored it. The readable approval alternatives/date fields are not proof of executed ministerial acceptance. Use **published updated review**, and distinguish original delivery from revision, rather than silently dating every public-page formulation to 31 October. FOI25-247, Document 1 PDF1–3 and Attachment C PDF7–8 (see source references)\n\n## Operative comparison: initial categories, future additions and adult choice\n\nThe exposure draft's distinctions are consequential:\n\n| Provision | Scope to preserve |\n| --- | --- |\n| §§25B–D, printed pp26–29 | Separate serious-harm protection, additional child protection and under-16 social-media feature obligations. Initial lists include targeted harassment, terrorism and harmful practices; no general misinformation/democracy category is named. |\n| §§25C(2), 25D(2), 27B | Minister-added categories require potential serious harm, or potential harm to children respectively; the minister must seek and consider eSafety advice. The child addition test does not expressly require seriousness. |\n| §§25E, 26A | Non-user and combined harms count. Assessment covers all reasonably foreseeable risks, including those relevant to the duty. Label absence does not settle every assessment obligation. |\n| §§25F–G | Specified features are deemed to have negative behavioural impacts. Ministerial additions have no express serious-harm threshold in §25G itself. |\n| §§25H, 26(7) | Reasonable practicability weighs risk, available measures, cost and reasonable privacy expectations. The exception concerns lawful private communications solely between consenting adults; it does not exempt all lawful adult speech or disable every separate Act power. |\n\nThese are proposed Australian legal mechanisms, not an incorporation of WEF's matrix. False information can intersect listed conduct; future additions remain possible. Exposure draft, §§25B–26A and 27B, printed pp26–34/37, PDF30–38/41 (see source references)\n\n## Earlier Australian legal sources and competing explanations\n\nThe **as-made 2021 Act** already defined serious physical/mental harm, distinguished serious distress from ordinary emotional reactions, and imposed targeted child/adult abuse tests. Its adult test combined likely intent to cause serious harm with an ordinary-reasonable-person assessment of the material. Sections 8–9 addressed contextual merit/consent and abhorrent violence. Section 106 tied class 1 to the Classification Act 1995 refusal category. These are directly inspected predecessors to the 2023 typology, rather than a presumed earlier neutral vocabulary. They supply a concrete competing source for similarities, while not proving every draft clause was copied from them. As-made Online Safety Act 2021, §§5–9 and 106 (see source references)\n\nThe April issues paper also discusses Australian research and existing abuse/content schemes beside its WEF references. This is consistent with synthesis from several inputs. The record supports influence through framing and documented advocacy; it does not support an exclusive linear chain in which a private taxonomy became law unchanged. Conversely, the existence of older law and contextual safeguards does not determine how broad future instruments or enforcement would become.\n\n## Candidate relations and remaining discriminating record\n\nPreserve **WEF typology → government April issues paper: cited for descriptive framing**, independently of **WEF typology → HRLC155: expressly adopted framing**. Add **WEF typology → eSafety guidance: explicitly used for educational categorisation**, with the page's update-date boundary. Preserve **Reset model → HRLC155: expressly endorsed**, and **review recommendation 5 → government response: supported in principle**. Treat **review taxonomy ↔ draft categories** as a documented comparison unless a clause-level attribution is obtained. None of these is a donor-instruction, private veto or enacted-ban relation.\n\nThe next discriminating record is a **category-by-category policy instruction, consultation disposition or drafting table for §§25C–G**, showing source, alternatives, threshold selection and ministerial clearance. The likely policy custodian is the department's online-safety reform team; the relevant record family is the 2025–26 duty-of-care drafting/consultation file, including instructions to legislative drafters. It would distinguish accepted wording from shared vocabulary and establish why broader labels were omitted or particular additions delegated. No such instruction or disposition was acquired here. This lane stops at that record-to-decision breakpoint for the forest review.\n\n## Actual reading scope and access state\n\nAll current access observations are 9 September 2026. This note contains scoped public-reader/indexed-primary findings. **No original PDF bytes, screenshots or original-byte hashes were acquired in this lane.** PDF/printed numbering is differentiated where verified. The separate prior authority note supplies a fuller clause discussion, not new visual evidence.\n\n| Source alias | Inspected scope / current working references |\n| --- | --- |\n| taxonomy-op-s-hrlc155 | §4.3 through §5.1.5, printed/PDF pp9–12; reader turn925view0/1 and indexed introductory passage turn926search4. Not every submission page. |\n| taxonomy-op-s-issues2024 | Part 3 overview/fn17; online-hate and technology-facilitated-abuse discussion/fn88/93. Reader turn935view1 and turn937view0/1. Date label April 2024; filename/internal working-title details do not establish a publication day. |\n| taxonomy-op-s-esafety | Types-of-harms section, following impact paragraphs and displayed update. Reader turn935view2 and turn937view2/3. Current page, no historical revision capture. |\n| taxonomy-op-s-resetgreen | Cover/about, printed pp1–3 background and pp6–7 systemic proposal/table. Reader turn940view1, turn941view0, turn942view1. Table is extracted text; no visual matrix claim. |\n| taxonomy-op-s-review | Complete indexed printed p56 recommendation and adjoining scope discussion, plus indexed p12 summary. turn926search0, turn941search12, turn943search12. A reader open returned content-too-large (13,286,332 bytes); no ordinary GET retried. Detailed §5.5 attribution is retained in prior policy-lineage.md, not newly audited here. |\n| taxonomy-op-s-foi247 | Document 1 PDF1–3, Attachment C PDF7–8, and limited draft-letter context PDF9. turn929view0, turn931view1, turn935view0. No inference from blank/alternative approval text to signed decision. |\n| taxonomy-op-s-response | PDF/printed pp3–6: response-state definitions, recommendations 4–5, related scope and adult-abuse discussion. turn947view0, reusing turn920view0. |\n| taxonomy-op-s-draft | §§25A–26A and 27B; scoped neighbouring compliance text also returned. turn926view0, turn929view1–3, turn931view0, turn942view2, turn943view1, turn944view0. Public text extraction only; closed normal-GET failures not repeated. |\n| taxonomy-op-s-osa2021 | Original/as-made ePub §§5–9 and class1 §106. Reached actual dated original through the Register's section link: turn944view1, turn946view0–2. The landing table-of-contents alone was not treated as historical substantive text. A December 2024 compilation was also read for comparison, turn931view2/turn935view3. |\n\nNo contributor biographies, new donor roster, private accounts, outreach, publication or canonical case edits were added. No inference about unseen draft material or access restrictions follows from these text-only readings.\n", "file": "research/taxonomy-wave-2026-09-09-operative-categories.md" }, { "id": "taxonomy-wave-2026-09-09-taxonomy-producers", "title": "Producers and limits of the WEF 2023 harms typology", "description": "Bounded artifact investigation, 9 September 2026, following the latest forest review. This concerns the August 2023 Toolkit for Digital Safety Design Interventions and Innovations: Typology of Online Harms, which HRLC submission155 expressly adopted as its framing. It is distinct from January 2023 Global Principles on Digital Safety, and from the May 2023 ri", "html": "Bounded artifact investigation, 9 September 2026, following the latest forest review. This concerns the August 2023 Toolkit for Digital Safety Design Interventions and Innovations: Typology of Online Harms, which HRLC submission155 expressly adopted as its framing. It is distinct from January 2023 Global Principles on Digital Safety, and from the May 2023 risk-assessment framework. It is not Australian legislation, a compliance certification or an instruction to suppress everything within its categories.
\nThe useful new mechanism is a regulator/commercial-provider workstream producing a vocabulary that NGOs and officials could use. Primary records establish convening and drafting roles more specifically than membership rosters. They do not establish a private approval right over subsequent legislation. Full PDF credits and financial disclosures remain uninspected because both permitted acquisition routes failed.
\n| Actor / artifact | \nEstablished role and evidence | \nLimit | \n
|---|---|---|
| World Economic Forum Global Coalition for Digital Safety → typology | \nPublisher's landing page dates publication 4 August 2023 and presents it as a coalition output. | \nPublication identifies responsibility for the output, not a contract or vote approving each definition. [Landing page](http[local research file] | \n
| Adam Hildreth and Julie Inman Grant → typology workstream | \nResolver's article dated 24 August 2023 identifies Crisp's founder and Australia's eSafety Commissioner as co-chairs of the workstream compiling the typology. | \nParticipant account; current Resolver branding does not establish the exact 2023 contracting company. [Opening paragraph](http[local research file] | \n
| John-Orr Hanna → typology | \nSWGfL's current biography expressly lists a drafting-committee role for the 2023 typology. | \nRetrospective institutional biography, not a 2023 appointment instrument or proof of final approval. [Trustees, John-Orr Hanna, Recent Publications](http[local research file] | \n
| Four foreword signatories → typology | \nOriginal indexed p3 names Hildreth, Inman Grant, Daniel Dobrygowski and Minos Bantourakis, with Crisp/eSafety/WEF affiliations. | \nForeword authorship is not automatically main-text authorship. [Original PDF, p3](http[local research file] | \n
| Inman Grant, Hildreth and Bantourakis → public launch explanation | \nThey coauthored WEF's 4 August 2023 launch article. It reports collective input from regulators, companies, universities and civil society. | \nArticle byline and general collaboration account do not confer approval on every contributor. [Launch article](http[local research file] | \n
No inspected record identifies a commissioning contract, workstream appointment/selection rules, editorial sign-off procedure, donor restriction, payment amount or project-specific budget. This is an unestablished finance/approval layer, not a finding that production was unpaid or unfunded. The inaccessible PDF's complete contributor, production and acknowledgment sections cannot be described as checked. Secondary bibliographies supplied a possible additional author initial, but it was not converted into a person or authorship edge without the original credit.
\nThe commercial connection matters as production participation. It does not demonstrate that Crisp/Resolver financed the typology, that any official procured its services, or that later statutory implementation requires its product. Resolver's page now includes a product call-to-action; the dated article does not establish when that promotional component was inserted.
\nThe original foreword p3 says the output is intended for cross-jurisdictional discussion, without severity ratings or use as regulatory compliance. The introduction p5 treats harm during production, distribution and consumption as distinct possibilities. A consensually created intimate image can become harmful when distributed without consent; overlapping categories are expected. It acknowledges contextual and regional differences, rather than offering universally precise definitions. These passages support policy development through common language, not automatic legal consequences. [Original, indexed complete p3 and p5 passages](http[local research file]
\nThe coalition's launch article adds that the scheme is not exhaustive and identifies a human-rights approach, including participation and expression. It connects the typology to its separate principles and risk-assessment outputs. This is a coordinated family of tools; the documents have different functions. [Launch article, Content, contact and conduct section and preceding rights paragraphs](http[local research file]
\nSix substantive families sit across content, contact and conduct: personal/community safety; health/wellbeing; hate/discrimination; dignity; privacy; and deception/manipulation. A publisher-hosted September explainer supplies useful boundary examples: graphic material documenting human-rights abuses or condemning violence needs an exception; some age restrictions may appropriately protect users; developmentally inappropriate material concerns children's access. It distinguishes false information spread without the relevant deliberate deception from intentionally misleading disinformation. Those qualifications should travel with the labels. The article is an explanation by Douglas Broom, not a legal enactment or a substitute for the original matrix. [WEF, 1 September 2023, §§1–6](http[local research file]
\nResolver's detailed counterpart additionally makes consent and intent consequential: doxxing combines nonconsensual exposure with intention to place someone at risk; deceptive synthetic media includes an intention to harm. Its misinformation description does not require intent to mislead, while disinformation does. The present packet therefore does not treat every disclosed identity, synthetic image or mistake as satisfying those descriptions. [Resolver, Invasion of privacy and Deception and manipulation](http[local research file]
\nThese are taxonomy/explainer boundaries. They do not establish whether particular speech is lawful in Australia or elsewhere. No universal government-ID requirement, identity-provider preference or legally binding removal rule has been identified in the inspected text. A classification can nevertheless affect what policymakers and platforms consider; the important next handoff is where a decision-maker adopts a category and attaches an enforceable consequence.
\nResolver explicitly says the typology builds on CO:RE's 4Cs research and links the 2021 Livingstone et al. paper. Its short account names content, contact and conduct while referring to 4Cs. Preserve that as a participant's source attribution, not a claim that all prior categories were copied unchanged. The actual linked repository PDF returned an Anubis access denial, so its complete taxonomy and acknowledgments were not read here. [Resolver, What is the Typology of Online Harms?](http[local research file] [observed research link](http[local research file]
\nThe launch authors also name international rights instruments as foundations. That supports a claimed rights lineage; this run did not independently audit each treaty article or reconstruct the original's full Resources list. Neither “WEF invented online harm” nor “WEF merely copied a neutral universal dictionary” is established. Selection, grouping and framing are part of the documented work.
\nThe earlier Australian packet already establishes HRLC's express adoption in submission155 §4.3, PDF/printed p9. A new primary-source search also discovered footnote88 in the government's April 2024 review issues paper citing the typology, before HRLC's July document. That record was handed to the separate comparison lane for passage-level inspection; it is a lead here, not a completed government-adoption claim. This timing could materially challenge a solely HRLC-mediated route. Current eSafety Safety-by-Design guidance also explicitly names the typology; the comparison lane owns that source read. The regulator's production participation and possible independent official use must remain visible alongside NGO uptake.
\n| Candidate ID | \nDirection / type | \nEvidence state | \n
|---|---|---|
| taxonomy-prod-e-published | \nWEF coalition → typology: published | \nObserved publisher statement, 4 August 2023. | \n
| taxonomy-prod-e-cochairs | \nHildreth + Inman Grant → typology workstream: co-chaired compilation | \nReported by participant, dated 24 August 2023; no appointment contract. | \n
| taxonomy-prod-e-hanna | \nHanna → typology: drafting-committee participation | \nReported by current primary biography; no voting/approval inference. | \n
| taxonomy-prod-e-foreword | \nFour named foreword signatories → typology: foreword credited | \nOriginal indexed p3; keep distinct from main authors. | \n
| taxonomy-prod-e-launch | \nThree named launch authors → launch explanation: coauthored | \nObserved byline, 4 August 2023. | \n
| taxonomy-prod-e-core | \nTypology → CO:RE 4Cs: reported intellectual antecedent | \nParticipant attribution; earlier original not read. | \n
| taxonomy-prod-e-hrlc | \nHRLC155 → typology: expressly adopts framing | \nReuse prior disposition edge; no duplication required. | \n
No donor-payment edge, contributor-to-clause-control edge, or typology-to-Australian-law edge is proposed.
\nThe strongest rival to a privately dictated mandate is a jointly produced, explicitly contextual reference tool used by several institutions for their own decisions. The strongest supported concern is narrower and more concrete: regulatory officials and a commercial safety provider helped set the shared vocabulary before later policy debate. Whether that collaboration gave any participant privileged implementation benefits, controlled definitions, or displaced a rejected alternative remains unestablished.
\nThe next discriminating records are the typology's full credits/acknowledgments, workstream terms of reference, category-selection drafts or response log, and any disclosed commissioning/funding instrument. Likely holders are WEF's coalition secretariat and the named co-chair organizations. Reopen through a genuinely obtainable public record, not repeated failed downloads. Return to the forest assessment before another acquisition branch.
\nAll access observations: 9 September 2026. No original bytes or original-byte hash were acquired. This note is the lane output.
\n| Source alias | \nReading scope and ancestry | \nState | \n
|---|---|---|
| taxonomy-prod-s-original | \nExact WEF PDF above. Full indexed foreword p3 and introduction p5 read (reader search records turn928search0, turn924search12). | \nScoped primary indexed text read. PDF web open returned (400) Timeout fetching; one ordinary requests GET substitute returned ReadTimeout ... read timeout=30. Closed after substitute. Full matrix/credits/resources not inspected. | \n
| taxonomy-prod-s-landing | \nOfficial publication landing, publication date and descriptive body (turn923view1). | \nRead HTML through web reader; no local original. | \n
| taxonomy-prod-s-launch | \nOfficial 4 August launch article, byline and complete substantive body (turn927view0). | \nRead HTML through web reader; author account, not participant minutes. | \n
| taxonomy-prod-s-resolver | \nParticipant's 24 August article, opening, prior-source link, categories and closing (turn936view0). | \nRead HTML through web reader. Current branding and page components observed; historical amendment date unknown. | \n
| taxonomy-prod-s-hanna | \nSWGfL institutional page, only John-Orr Hanna biography / publications (turn936view1). | \nRead current primary biography; page itself undated. | \n
| taxonomy-prod-s-explainer | \nDouglas Broom, WEF 1 September article, byline and six categories (turn938view0). | \nRead publisher-hosted explanation; author-view disclaimer present. | \n
| taxonomy-prod-s-core-lead | \nSSOAR PDF reached through Resolver's actual research link. | \nInaccessible: Access Denied: error code 9e4edb5b6b850c41. Page says protected by Anubis. Website refusal, not automatic sandbox-approval review. No retry. | \n
| taxonomy-prod-s-au-issues-lead | \nhttp[local research file] ; indexed fn88 (turn933search17). | \nDiscovered and handed to comparison lane; not a new reader-verified adoption finding here. | \n
| taxonomy-prod-s-esafety-lead | \nhttp[local research file] | \nDiscovered and handed to comparison lane for read; no inference that guidance is statutory duty. | \n
Searches surfaced the separate 2024 metrics paper and an unrelated digital-safety contributors list with a funder acknowledgment. They were excluded from this artifact's finance/credits findings. An acknowledgment on a different WEF output cannot finance this typology by association. No closed ACNC, OPD or Australian draft/review GET route was retried; no outreach, account access, canonical case edit or publication occurred.
", "path": "taxonomy-wave-2026-09-09/taxonomy-producers.md", "sourceIds": [ "w14-tax-s-original", "w14-tax-s-landing", "w14-tax-s-launch", "w14-tax-s-resolver", "w14-tax-s-hanna", "w14-tax-s-explainer", "w14-tax-s-core-inaccessible" ], "markdown": "Bounded artifact investigation, 9 September 2026, following the latest forest review. This concerns the August 2023 **Toolkit for Digital Safety Design Interventions and Innovations: Typology of Online Harms**, which HRLC submission155 expressly adopted as its framing. It is distinct from January 2023 **Global Principles on Digital Safety**, and from the May 2023 risk-assessment framework. It is not Australian legislation, a compliance certification or an instruction to suppress everything within its categories.\n\nThe useful new mechanism is a regulator/commercial-provider workstream producing a vocabulary that NGOs and officials could use. Primary records establish convening and drafting roles more specifically than membership rosters. They do not establish a private approval right over subsequent legislation. Full PDF credits and financial disclosures remain uninspected because both permitted acquisition routes failed.\n\n## Production roles actually established\n\n| Actor / artifact | Established role and evidence | Limit |\n| --- | --- | --- |\n| World Economic Forum Global Coalition for Digital Safety → typology | Publisher's landing page dates publication **4 August 2023** and presents it as a coalition output. | Publication identifies responsibility for the output, not a contract or vote approving each definition. Landing page (see source references) |\n| Adam Hildreth and Julie Inman Grant → typology workstream | Resolver's article dated **24 August 2023** identifies Crisp's founder and Australia's eSafety Commissioner as co-chairs of the workstream compiling the typology. | Participant account; current Resolver branding does not establish the exact 2023 contracting company. Opening paragraph (see source references) |\n| John-Orr Hanna → typology | SWGfL's current biography expressly lists a **drafting-committee** role for the 2023 typology. | Retrospective institutional biography, not a 2023 appointment instrument or proof of final approval. Trustees, John-Orr Hanna, Recent Publications (see source references) |\n| Four foreword signatories → typology | Original indexed p3 names Hildreth, Inman Grant, Daniel Dobrygowski and Minos Bantourakis, with Crisp/eSafety/WEF affiliations. | Foreword authorship is not automatically main-text authorship. Original PDF, p3 (see source references) |\n| Inman Grant, Hildreth and Bantourakis → public launch explanation | They coauthored WEF's **4 August 2023** launch article. It reports collective input from regulators, companies, universities and civil society. | Article byline and general collaboration account do not confer approval on every contributor. Launch article (see source references) |\n\nNo inspected record identifies a commissioning contract, workstream appointment/selection rules, editorial sign-off procedure, donor restriction, payment amount or project-specific budget. This is an **unestablished finance/approval layer**, not a finding that production was unpaid or unfunded. The inaccessible PDF's complete contributor, production and acknowledgment sections cannot be described as checked. Secondary bibliographies supplied a possible additional author initial, but it was not converted into a person or authorship edge without the original credit.\n\nThe commercial connection matters as production participation. It does not demonstrate that Crisp/Resolver financed the typology, that any official procured its services, or that later statutory implementation requires its product. Resolver's page now includes a product call-to-action; the dated article does not establish when that promotional component was inserted.\n\n## What the artifact sets out to do\n\nThe original foreword p3 says the output is intended for cross-jurisdictional discussion, without severity ratings or use as regulatory compliance. The introduction p5 treats harm during production, distribution and consumption as distinct possibilities. A consensually created intimate image can become harmful when distributed without consent; overlapping categories are expected. It acknowledges contextual and regional differences, rather than offering universally precise definitions. These passages support policy development through common language, not automatic legal consequences. Original, indexed complete p3 and p5 passages (see source references)\n\nThe coalition's launch article adds that the scheme is not exhaustive and identifies a human-rights approach, including participation and expression. It connects the typology to its separate principles and risk-assessment outputs. This is a coordinated family of tools; the documents have different functions. Launch article, Content, contact and conduct section and preceding rights paragraphs (see source references)\n\nSix substantive families sit across content, contact and conduct: personal/community safety; health/wellbeing; hate/discrimination; dignity; privacy; and deception/manipulation. A publisher-hosted September explainer supplies useful boundary examples: graphic material documenting human-rights abuses or condemning violence needs an exception; some age restrictions may appropriately protect users; developmentally inappropriate material concerns children's access. It distinguishes false information spread without the relevant deliberate deception from intentionally misleading disinformation. Those qualifications should travel with the labels. The article is an explanation by Douglas Broom, not a legal enactment or a substitute for the original matrix. WEF, 1 September 2023, §§1–6 (see source references)\n\nResolver's detailed counterpart additionally makes consent and intent consequential: doxxing combines nonconsensual exposure with intention to place someone at risk; deceptive synthetic media includes an intention to harm. Its misinformation description does not require intent to mislead, while disinformation does. The present packet therefore does not treat every disclosed identity, synthetic image or mistake as satisfying those descriptions. Resolver, Invasion of privacy and Deception and manipulation (see source references)\n\nThese are taxonomy/explainer boundaries. They do not establish whether particular speech is lawful in Australia or elsewhere. No universal government-ID requirement, identity-provider preference or legally binding removal rule has been identified in the inspected text. A classification can nevertheless affect what policymakers and platforms consider; the important next handoff is where a decision-maker adopts a category and attaches an enforceable consequence.\n\n## Earlier sources and the transmission test\n\nResolver explicitly says the typology builds on **CO:RE's 4Cs research** and links the 2021 Livingstone et al. paper. Its short account names content, contact and conduct while referring to 4Cs. Preserve that as a participant's source attribution, not a claim that all prior categories were copied unchanged. The actual linked repository PDF returned an Anubis access denial, so its complete taxonomy and acknowledgments were not read here. Resolver, What is the Typology of Online Harms? (see source references) observed research link (see source references)\n\nThe launch authors also name international rights instruments as foundations. That supports a claimed rights lineage; this run did not independently audit each treaty article or reconstruct the original's full Resources list. Neither “WEF invented online harm” nor “WEF merely copied a neutral universal dictionary” is established. Selection, grouping and framing are part of the documented work.\n\nThe earlier Australian packet already establishes HRLC's express adoption in submission155 §4.3, PDF/printed p9. A new primary-source search also discovered **footnote88 in the government's April 2024 review issues paper citing the typology**, before HRLC's July document. That record was handed to the separate comparison lane for passage-level inspection; it is a lead here, not a completed government-adoption claim. This timing could materially challenge a solely HRLC-mediated route. Current eSafety Safety-by-Design guidance also explicitly names the typology; the comparison lane owns that source read. The regulator's production participation and possible independent official use must remain visible alongside NGO uptake.\n\n## Candidate relations\n\n| Candidate ID | Direction / type | Evidence state |\n| --- | --- | --- |\n| taxonomy-prod-e-published | WEF coalition → typology: published | Observed publisher statement, 4 August 2023. |\n| taxonomy-prod-e-cochairs | Hildreth + Inman Grant → typology workstream: co-chaired compilation | Reported by participant, dated 24 August 2023; no appointment contract. |\n| taxonomy-prod-e-hanna | Hanna → typology: drafting-committee participation | Reported by current primary biography; no voting/approval inference. |\n| taxonomy-prod-e-foreword | Four named foreword signatories → typology: foreword credited | Original indexed p3; keep distinct from main authors. |\n| taxonomy-prod-e-launch | Three named launch authors → launch explanation: coauthored | Observed byline, 4 August 2023. |\n| taxonomy-prod-e-core | Typology → CO:RE 4Cs: reported intellectual antecedent | Participant attribution; earlier original not read. |\n| taxonomy-prod-e-hrlc | HRLC155 → typology: expressly adopts framing | Reuse prior disposition edge; no duplication required. |\n\nNo donor-payment edge, contributor-to-clause-control edge, or typology-to-Australian-law edge is proposed.\n\nThe strongest rival to a privately dictated mandate is a jointly produced, explicitly contextual reference tool used by several institutions for their own decisions. The strongest supported concern is narrower and more concrete: regulatory officials and a commercial safety provider helped set the shared vocabulary before later policy debate. Whether that collaboration gave any participant privileged implementation benefits, controlled definitions, or displaced a rejected alternative remains unestablished.\n\nThe next discriminating records are the typology's full credits/acknowledgments, workstream terms of reference, category-selection drafts or response log, and any disclosed commissioning/funding instrument. Likely holders are WEF's coalition secretariat and the named co-chair organizations. Reopen through a genuinely obtainable public record, not repeated failed downloads. Return to the forest assessment before another acquisition branch.\n\n## Scope and custody\n\nAll access observations: 9 September 2026. No original bytes or original-byte hash were acquired. This note is the lane output.\n\n| Source alias | Reading scope and ancestry | State |\n| --- | --- | --- |\n| taxonomy-prod-s-original | Exact WEF PDF above. Full indexed foreword p3 and introduction p5 read (reader search records turn928search0, turn924search12). | Scoped primary indexed text read. PDF web open returned `(400) Timeout fetching`; one ordinary requests GET substitute returned `ReadTimeout ... read timeout=30`. Closed after substitute. Full matrix/credits/resources not inspected. |\n| taxonomy-prod-s-landing | Official publication landing, publication date and descriptive body (turn923view1). | Read HTML through web reader; no local original. |\n| taxonomy-prod-s-launch | Official 4 August launch article, byline and complete substantive body (turn927view0). | Read HTML through web reader; author account, not participant minutes. |\n| taxonomy-prod-s-resolver | Participant's 24 August article, opening, prior-source link, categories and closing (turn936view0). | Read HTML through web reader. Current branding and page components observed; historical amendment date unknown. |\n| taxonomy-prod-s-hanna | SWGfL institutional page, only John-Orr Hanna biography / publications (turn936view1). | Read current primary biography; page itself undated. |\n| taxonomy-prod-s-explainer | Douglas Broom, WEF 1 September article, byline and six categories (turn938view0). | Read publisher-hosted explanation; author-view disclaimer present. |\n| taxonomy-prod-s-core-lead | SSOAR PDF reached through Resolver's actual research link. | Inaccessible: `Access Denied: error code 9e4edb5b6b850c41.` Page says protected by Anubis. Website refusal, not automatic sandbox-approval review. No retry. |\n| taxonomy-prod-s-au-issues-lead | http[local research file] ; indexed fn88 (turn933search17). | Discovered and handed to comparison lane; not a new reader-verified adoption finding here. |\n| taxonomy-prod-s-esafety-lead | http[local research file] | Discovered and handed to comparison lane for read; no inference that guidance is statutory duty. |\n\nSearches surfaced the separate 2024 metrics paper and an unrelated digital-safety contributors list with a funder acknowledgment. They were excluded from this artifact's finance/credits findings. An acknowledgment on a different WEF output cannot finance this typology by association. No closed ACNC, OPD or Australian draft/review GET route was retried; no outreach, account access, canonical case edit or publication occurred.\n", "file": "research/taxonomy-wave-2026-09-09-taxonomy-producers.md" }, { "id": "tcap-delivery-wave-2026-09-09-altitude-governance", "title": "Altitude: product governance and correction boundaries", "description": "Observed 9 September 2026. This is a bounded static inspection of the public implementation linked by Tech Against Terrorism (TAT), not an installed or deployed system test. Acquisition is closed. No product/API calls, accounts, forms, content feeds, terrorist-content files, dependency installation or repository test suite were used. Canonical cases, readers", "html": "Observed 9 September 2026. This is a bounded static inspection of the public implementation linked by Tech Against Terrorism (TAT), not an installed or deployed system test. Acquisition is closed. No product/API calls, accounts, forms, content feeds, terrorist-content files, dependency installation or repository test suite were used. Canonical cases, readers and ZIPs were untouched.
\nThe public records establish a real route from privately supplied classifications into a platform's review queue. In the inspected version, TCAP is explicitly treated as a trusted source when calculating confidence for prioritisation. They also establish platform choices over source connections, feedback sharing, authentication and the moderation verdict. These are different decision points: source authority influences attention; the inspected workflow does not make the source's flag an automatic removal order.
\nThere are concrete correction facilities, but they must not be collapsed into an end-user remedy. A moderator can cancel a draft review before publication; the inspected create endpoint can append another review to an existing case. The generic importer contains a source-redaction hook. Optional feedback can return platform verdicts to TCAP. None of these inspected paths establishes that an upstream correction automatically restores removed content, notifies its author, or reopens every affected case.
\nTAT's [22 July 2024 stewardship announcement](http[local research file] says Altitude was built with Google Jigsaw, TAT will lead maintenance and development, and Jigsaw will continue advisory support. It describes prior testing with small and medium platforms, but provides no independently verified deployment, participant list or measured acceptance result in the body read. The announcement and [current product page](http[local research file] offer free access/onboarding through the Bronze support tier, with additional support/features at higher tiers. No financial amount or Canadian grant allocation is stated in these sources.
\nThe product's exact public repository link is [Jigsaw-Code/altitude](http[local research file] Its observed main commit was 496951f7d7a83508e9fac5709d4db2e91c654685, authored/committed 8 August 2024, message “Fix typo.” All retained source files are pinned to that commit. A repository still under the Jigsaw-Code namespace is not proof of Google's present operational veto, and this observation of main does not establish that no other development or deployed version exists.
\nThe [contribution instructions](http[local research file] require a contributor agreement and review of submissions, including those by project members. They say contributors or employers retain copyright. The actual signed CLA, repository permissions, branch protections and named final approvers were not acquired. The retained [Apache 2.0 licence](http[local research file] permits reproduction/modification/distribution subject to its conditions; it disclaims warranty and allows separately accepted support obligations. These software permissions do not grant access to restricted signal databases or establish a support contract's terms.
\nThe [README](http[local research file] requires the platform to obtain TCAP and GIFCT credentials. The product page describes the platform installing the tool, comparing its content to consolidated signals, and a moderator reviewing and selecting action. The [overview](http[local research file] distinguishes signals, the client's target content, cases and human reviews. The inspected public software is therefore not proof that TAT has unrestricted access to every customer dataset or chooses every target.
\nThe [FAQ, lines 90–104](http[local research file] assigns authentication to the integrator and says the tool itself supplies no user accounts. This is a documented deployment responsibility, not a finding that any actual production instance is unauthenticated. No deployed role assignments or separation between administrator and reviewer was inspected.
\nThe [UI service configuration routes, lines 617–707](http[local research file] expose separate enabled and diagnostics-enabled settings for TCAP and GIFCT connections. The [importer creation endpoint, lines 73–133](http[local research file] defaults a new import connection to active and diagnostic sharing to inactive, subject to successful credential pre-check. This establishes configurable connections in the source, not a particular customer's consent setting or the upstream organisations' admission rules.
\nThe [TCAP adapter, lines 141–182](http[local research file] imports source identifiers and report dates, associated organisation text, personal-information and extreme-content flags, a URL and its checked availability status. Those are input metadata, not an independent adjudication by Altitude.
\nThe [priority implementation, lines 24–135](http[local research file] has TRUSTED_SOURCES containing TCAP. A signal from that source qualifies for the highest confidence component (3), as can multiple source names or sufficient trust metadata. Severity is calculated separately from tags and features. TCAP's confidence boost does not alone guarantee the highest overall queue priority, and neither score is a moderation verdict. The term confidence here denotes the implemented heuristic; no calibration study was inspected.
The [FAQ, lines 29–64](http[local research file] says cases are ranked by combined confidence/severity, with severity weighted and absent flagger data shown as N/A. Its example table is illustrative. The inspected scoring choices are source-code constants; no customer-facing editor for those rules was established. Apache modification rights make adaptation possible, but do not show that a customer changed these defaults or that TAT approves every local change.
\nThe review endpoint, lines 47–87.
\nThe [publication/delivery code, lines 522–604](http[local research file] publishes the draft and sends client context, APPROVE/BLOCK and decision time to a configured action receiver. Without that receiver, the inspected branch writes a local verdict log. A successful receiver response, or that local log branch, produces the code's ACCEPTED delivery status; it is not proof of removal from a live service. The README's general statement that all decisions are logged should therefore not replace this conditional implementation detail. A platform must implement what receipt of a verdict does.
\nThe [FAQ, lines 106–113](http[local research file] explicitly allows “Remove” to be renamed for platforms whose action differs. The retained client config maps APPROVE to “Leave up” and BLOCK to “Remove.” Labels are not evidence of execution. The FAQ also says uploaded content remains archived locally after platform removal and needs separate database deletion (lines 22–27).
\nThe task exporter, lines 701–738.
\nThe [TCAP exporter, lines 227–254](http[local research file] sends URL plus decision, translating BLOCK to REMOVE and APPROVE to APPROVE. It demonstrates a feedback mechanism, not that TCAP accepts a disagreement, changes a label or propagates a correction. The FAQ's claims about improved accuracy and effectiveness are product claims, not measured results in this review. GIFCT's adapter payload was not inspected; its existence must not inherit TCAP's exact payload semantics.
\nThe generic importer, lines 156–195.
\naltitude-captures.json records 24 retained public originals, exact URLs, retrieval times, byte lengths, SHA-256 values, pinned commit and per-file reading scope. Five GitHub tree captures were used only for exact path discovery. Announcement/product bodies, README, overview, FAQ, CONTRIBUTING, the full priority/review/importer/TCAP-adapter/base-importer files, and the specified task/UI route passages were read. Licence sections 2–9 were read. No fixtures, image examples or underlying harmful content were opened. The GitHub docs-directory reader cache miss was closed; exact linked static files were available. Successful ordinary GETs were retained without claiming a full repository audit.
The next discriminator is a named deployment's operative configuration and acceptance/correction receipt: version and changes, source and diagnostics choices, reviewer authentication/authority, action-receiver behaviour, and an upstream withdrawal carried through to a logged decision and restored or retained content. TAT/platform custodians would hold those records. The historical funder's accepted output or code-version record would separately connect grant-funded delivery to this implementation. No Canadian allocation, donor instruction, universal platform execution or live correction result is inferred from this public code.
", "path": "tcap-delivery-wave-2026-09-09/altitude-governance.md", "sourceIds": [ "w26-ver-s-altitude-announcement", "w26-ver-s-altitude-product", "w26-ver-s-altitude-readme", "w26-ver-s-altitude-license", "w26-ver-s-altitude-contribution", "w26-ver-s-altitude-faq", "w26-ver-s-altitude-priority", "w26-ver-s-altitude-review", "w26-ver-s-altitude-import-config", "w26-ver-s-altitude-tasks", "w26-ver-s-altitude-tcap-adapter", "w26-ver-s-altitude-base-importer" ], "markdown": "Observed 9 September 2026. This is a bounded static inspection of the public implementation linked by Tech Against Terrorism (TAT), not an installed or deployed system test. Acquisition is closed. No product/API calls, accounts, forms, content feeds, terrorist-content files, dependency installation or repository test suite were used. Canonical cases, readers and ZIPs were untouched.\n\n## Result\n\nThe public records establish a real route from privately supplied classifications into a platform's review queue. In the inspected version, TCAP is explicitly treated as a trusted source when calculating confidence for prioritisation. They also establish platform choices over source connections, feedback sharing, authentication and the moderation verdict. These are different decision points: source authority influences attention; the inspected workflow does not make the source's flag an automatic removal order.\n\nThere are concrete correction facilities, but they must not be collapsed into an end-user remedy. A moderator can cancel a draft review before publication; the inspected create endpoint can append another review to an existing case. The generic importer contains a source-redaction hook. Optional feedback can return platform verdicts to TCAP. None of these inspected paths establishes that an upstream correction automatically restores removed content, notifies its author, or reopens every affected case.\n\n## Public identity, maintenance and version\n\nTAT's 22 July 2024 stewardship announcement (see source references) says Altitude was built with Google Jigsaw, TAT will lead maintenance and development, and Jigsaw will continue advisory support. It describes prior testing with small and medium platforms, but provides no independently verified deployment, participant list or measured acceptance result in the body read. The announcement and current product page (see source references) offer free access/onboarding through the Bronze support tier, with additional support/features at higher tiers. No financial amount or Canadian grant allocation is stated in these sources.\n\nThe product's exact public repository link is Jigsaw-Code/altitude (see source references) Its observed main commit was **496951f7d7a83508e9fac5709d4db2e91c654685**, authored/committed **8 August 2024**, message “Fix typo.” All retained source files are pinned to that commit. A repository still under the Jigsaw-Code namespace is not proof of Google's present operational veto, and this observation of main does not establish that no other development or deployed version exists.\n\nThe contribution instructions (see source references) require a contributor agreement and review of submissions, including those by project members. They say contributors or employers retain copyright. The actual signed CLA, repository permissions, branch protections and named final approvers were not acquired. The retained Apache 2.0 licence (see source references) permits reproduction/modification/distribution subject to its conditions; it disclaims warranty and allows separately accepted support obligations. These software permissions do not grant access to restricted signal databases or establish a support contract's terms.\n\n## Who configures access and source use\n\nThe README (see source references) requires the platform to obtain TCAP and GIFCT credentials. The product page describes the platform installing the tool, comparing its content to consolidated signals, and a moderator reviewing and selecting action. The overview (see source references) distinguishes signals, the client's target content, cases and human reviews. The inspected public software is therefore not proof that TAT has unrestricted access to every customer dataset or chooses every target.\n\nThe FAQ, lines 90–104 (see source references) assigns authentication to the integrator and says the tool itself supplies no user accounts. This is a documented deployment responsibility, not a finding that any actual production instance is unauthenticated. No deployed role assignments or separation between administrator and reviewer was inspected.\n\nThe UI service configuration routes, lines 617–707 (see source references) expose separate enabled and diagnostics-enabled settings for TCAP and GIFCT connections. The importer creation endpoint, lines 73–133 (see source references) defaults a new import connection to active and **diagnostic sharing to inactive**, subject to successful credential pre-check. This establishes configurable connections in the source, not a particular customer's consent setting or the upstream organisations' admission rules.\n\n## Classification and priority: a specific implementation of trust\n\nThe TCAP adapter, lines 141–182 (see source references) imports source identifiers and report dates, associated organisation text, personal-information and extreme-content flags, a URL and its checked availability status. Those are input metadata, not an independent adjudication by Altitude.\n\nThe priority implementation, lines 24–135 (see source references) has `TRUSTED_SOURCES` containing TCAP. A signal from that source qualifies for the highest confidence component (3), as can multiple source names or sufficient trust metadata. Severity is calculated separately from tags and features. TCAP's confidence boost **does not alone guarantee the highest overall queue priority**, and neither score is a moderation verdict. The term confidence here denotes the implemented heuristic; no calibration study was inspected.\n\nThe FAQ, lines 29–64 (see source references) says cases are ranked by combined confidence/severity, with severity weighted and absent flagger data shown as N/A. Its example table is illustrative. The inspected scoring choices are source-code constants; no customer-facing editor for those rules was established. Apache modification rights make adaptation possible, but do not show that a customer changed these defaults or that TAT approves every local change.\n\n## Verdict, delivery and undo are separate\n\nThe review endpoint, lines 47–87.\n\nThe publication/delivery code, lines 522–604 (see source references) publishes the draft and sends client context, APPROVE/BLOCK and decision time to a configured action receiver. **Without that receiver, the inspected branch writes a local verdict log.** A successful receiver response, or that local log branch, produces the code's ACCEPTED delivery status; it is not proof of removal from a live service. The README's general statement that all decisions are logged should therefore not replace this conditional implementation detail. A platform must implement what receipt of a verdict does.\n\nThe FAQ, lines 106–113 (see source references) explicitly allows “Remove” to be renamed for platforms whose action differs. The retained client config maps APPROVE to “Leave up” and BLOCK to “Remove.” Labels are not evidence of execution. The FAQ also says uploaded content remains archived locally after platform removal and needs separate database deletion (lines 22–27).\n\n## Feedback and upstream correction: partial paths\n\nThe task exporter, lines 701–738.\n\nThe TCAP exporter, lines 227–254 (see source references) sends URL plus decision, translating BLOCK to REMOVE and APPROVE to APPROVE. It demonstrates a feedback mechanism, not that TCAP accepts a disagreement, changes a label or propagates a correction. The FAQ's claims about improved accuracy and effectiveness are product claims, not measured results in this review. GIFCT's adapter payload was not inspected; its existence must not inherit TCAP's exact payload semantics.\n\nThe generic importer, lines 156–195.\n\n## Custody, reading scope and finite frontier\n\n`altitude-captures.json` records 24 retained public originals, exact URLs, retrieval times, byte lengths, SHA-256 values, pinned commit and per-file reading scope. Five GitHub tree captures were used only for exact path discovery. Announcement/product bodies, README, overview, FAQ, CONTRIBUTING, the full priority/review/importer/TCAP-adapter/base-importer files, and the specified task/UI route passages were read. Licence sections 2–9 were read. No fixtures, image examples or underlying harmful content were opened. The GitHub docs-directory reader cache miss was closed; exact linked static files were available. Successful ordinary GETs were retained without claiming a full repository audit.\n\nThe next discriminator is a **named deployment's operative configuration and acceptance/correction receipt**: version and changes, source and diagnostics choices, reviewer authentication/authority, action-receiver behaviour, and an upstream withdrawal carried through to a logged decision and restored or retained content. TAT/platform custodians would hold those records. The historical funder's accepted output or code-version record would separately connect grant-funded delivery to this implementation. No Canadian allocation, donor instruction, universal platform execution or live correction result is inferred from this public code.\n", "file": "research/tcap-delivery-wave-2026-09-09-altitude-governance.md" }, { "id": "tcap-delivery-wave-2026-09-09-archive-access", "title": "TCAP Archive, Hash API and Trusted Flagger access", "description": "9 September 2026. Bounded public-document pass complete. No account, policy-access application, content database, live API endpoint, underlying media, demo or contact request was accessed or submitted. Product documentation is evidence of the published interface and intended workflow, not runtime verification or an executed customer agreement. No publication", "html": "9 September 2026. Bounded public-document pass complete. No account, policy-access application, content database, live API endpoint, underlying media, demo or contact request was accessed or submitted. Product documentation is evidence of the published interface and intended workflow, not runtime verification or an executed customer agreement. No publication or deployment date is inferred from a 2025 documentation copyright, a 2026 website footer or illustrative timestamps.
\nThe useful new finding is a documented item-level correction interface and a deletion signal intended for recipient caches. The remaining gap is narrower than an absence of correction machinery: actual handling of disputed hashes, recipient synchronization and reconsideration of already imposed restrictions remain unobserved.
\n| Surface | \nPublished actor and capability | \nAuthority/evidence boundary | \n
|---|---|---|
| Archive Research Portal | \nTAT advertises media access for platforms and researchers, with automated/human sourcing and classification by AI models and OSINT specialists. | \nThe page is a product explanation and contact route. It does not publish an admission rubric, download licence, redistribution permission or current government-access rule. | \n
| Archive Hash List API | \nPlatforms can ingest fingerprints in bulk, filter them and cache updates. | \nThis concerns fingerprints; it is not permission to download or republish the underlying media. Recipient implementation and contractual reuse terms were not acquired. | \n
| Trusted Flagger Platform | \nTAT describes trusted partners submitting categorized URLs for removal referrals, supported by its analysts and human/AI classifiers. | \nParticipation is applied for. Detailed selection/permission rules and a statutory designation of these participants were not acquired. The product name alone supplies neither. | \n
| Hash feedback | \nOnboarded users can report individual-hash problems and see analyst responses and decisions on their own feedback. | \nThis is an internal review interface, not an established appeal right for every affected content creator or a demonstrated independent adjudication process. | \n
Research source: [Archive Database Research Portal](http[local research file] complete substantive product body; retained archive-research.txt lines 1–19. Hash product source: [Archive Database Hash List API](http[local research file] complete body; archive-hash-product.txt lines 1–16. Both are current observed pages with unknown revision dates. The hash page advertises daily updates and filters by file type, algorithm and ideology; its contact link is not an unrestricted public-download offer.
The [Trusted Flagger product page](http[local research file] describes tiered URL submissions, a permission-controlled review environment, TAT analyst assessment and metadata supplied to recipient platforms. It advertises referrals to 125-plus platforms, retained historic requests, platform feedback and expansion of the archive/hash lists. That is an explicit advertised contribution path from referrals to a reusable collection, not proof that every submitted URL is admitted or every recipient acts. Its accuracy claim concerns online/offline status classification, not a measured terrorist-label accuracy rate. Complete substantive body read, archive-trusted-flagger.txt lines 1–56, especially 4–16 and 30–50. “Apply” on this page returned the same public page through the reader; no application was submitted.
The earlier accepted classification summary says TAT in-house experts assess source and content; the detailed verification policy remains behind an access-request form. The earlier accepted tier-policy/FAQ findings are reused from tat.md, not reacquired. In particular, the old FAQ's no-download and no-government-access statements cannot settle the current Archive product's rights: that FAQ still describes future 2021/2022 development. The new TFP page's five-tier capability likewise does not establish when an older forthcoming tier became operational or what its current rule text says.
\n[Content Hash List Feedback documentation](http[local research file] describes feedback from hash-list users for false positives, classification errors, duplicates and other issues. A report identifies the hash and category; an explanation is optional. Users can retrieve their submitted feedback, analyst responses and decisions. The decision states include removal, non-removal and pending, with response-sent status. Analysts aim to reply by onboarding email within three to five working days. This is an aim, not a demonstrated deadline met in a real case. Locators: Overview; Request Body; Feedback Status; Query Parameters/Feedback Object; Response notifications. Retained archive-feedback-docs-prose.txt lines 1–6, 18–30 and 35–48.
The example JSON is illustrative: fictional names and placeholder prose appear in it. Its counts, decisions and timestamps are not actual complaint volumes, decided appeals or deployment evidence. No request was executed. Neither this page nor the other inspected public access material establishes an independent appeal body, a second-level review route or a rule requiring restoration after a successful challenge.
\n[Content Hash List documentation](http[local research file] exposes fingerprints plus classification/file metadata, a deletion marker and last-update time. It describes checkpointed incremental updates and recommends local caching and monitoring deprecated entries. Soft deletion retains hash history. Thus it documents a technical means for a recipient to learn of a withdrawal while updating a local cache; it does not prove that recipients implemented it or erased prior copies. Locators: Query Parameters, Checkpointing, Hash Object Fields, Implementation Notes and Best Practices; archive-hash-docs-prose.txt lines 15–39.
The same page documents a ThreatExchange client configuration for fetching/comparing TAT hashes (lines 40–48). This is an interoperability instruction, not a new governance, funding or compulsory-database-membership relation. No client was installed or run. The inspected documentation does not specify a recipient-wide push notification, correction acknowledgement, compulsory rescan or reopening of prior moderation decisions. Those are acquisition gaps, not evidence they never occur.
\nThe [Authentication documentation](http[local research file] introduction and service list, expressly presupposes an already onboarded TCAP or Archive user with credentials. archive-auth-docs.txt lines 1–18; web-reader introduction and service lists also read. That establishes an access prerequisite, not the identity of the onboarding decision-maker or eligibility contract. Token examples and later implementation details were captured but not audited or executed.
The public [Policies index](http[local research file] describes its Archive destination as containing terms of use, an access policy and taxonomy. Its actual [Archive link](http[local research file] returned a web-reader internal error. The sole ordinary GET substitute succeeded, but the substantive HTML body contained only the heading Archiving Policies. Inspection found no iframe, embedded document or body terms/access-policy link. This is an obtainable page with no substantive policy text, not an administrative denial or evidence that no policy exists. Original retained as archive-archive-policy.html.
The index's [legacy Research Portal feature](http[local research file] failed in the web reader with “Cache miss”; one ordinary GET succeeded with an empty substantive body. Its [legacy hash feature](http[local research file] supplied navigation and an empty heading in the reader. These routes were then closed. No guessed private document path, gated full-policy request or account route was used. The latter reader-only observation is retained in this note; no original hash-feature HTML was acquired.
\nConsequently this pass does not establish current media-download rights, onward hash redistribution/licensing, customer-model training permission, accepted-flagger criteria, cross-user visibility of archived requests, or an affected nonmember's formal challenge route. TAT/OHF would hold the current access agreement, permissions matrix and review procedure. Participating recipients would hold their cache-update and reconsideration records. Asking for those records would require a separately authorized outreach branch; none occurred here.
\nThe plausible mechanism is distributed classification with controlled access and recipient discretion. The advertised Archive/TFP relationship can enlarge a shared collection; documented review and soft-delete fields can also correct it. A claim that there is no item-level challenge or withdrawal mechanism is now too strong. A claim that such a mechanism guarantees downstream remedy is also too strong. The strongest rival to a correction-failure theory is that ordinary recipient implementations consume deletion updates promptly and independently prevent or reverse mistakes; the public interface is consistent with that account.
\nThe discriminating record is a redacted completed hash challenge linked to its analyst decision, resulting update/deletion record, recipient synchronization acknowledgement and any reconsideration of prior action. An executed access/usage agreement would separately establish mandatory handling and reuse limits. These records are not replaced by a documentation example or URL-availability statistics. Acquisition stops here for the parent's forest review.
\narchive-captures.json contains exact discovered URLs, final URLs, UTC acquisition times, hashes, lengths, reading scopes and source ancestry. Eight ordinary HTML captures succeeded; originals are response bytes and text files are derivatives. Seven new useful substantive/product or boundary sources were read/inspected plus the empty legacy Research Portal capture. The API home and Policies index were discovery/reader sources, not separately downloaded. Source IDs use the local archive-s- prefix; they are not canonical-case insertions.
The hash and feedback substantive descriptions/tables were read from both the public reader and newly retained HTML-derived prose. Example JSON was inspected in the reader for its illustrative nature; code snippets were not tested. Research/TFP/hash product bodies were read in full. Authentication scope is its prerequisite and service list only. Archive HTML was inspected for visible body content, links and document embeds. No full policy, live service behavior, extremist original or private customer dataset is claimed as read.
\nDecisive original SHA-256: hash docs e8e4c49c4b8bb5ee3c4be1fade240bb09d2f60834e3204a6b0a4d3f6e1c78df4; feedback docs e454d1fadc7944f8e115cef5edd73412da40831b4f4e72275b0ce7a89a58b71f; Archive policy shell d5c6e1fdafdb822b89a9bedd1997a468641e69a1d3dd04f9fb902eca43176c02. Other hashes and exact scopes are in the manifest. No canonical case, previous packet or ZIP was changed.
| From → to | \nTyped relation and evidence stage | \nSource pointer | \n
|---|---|---|
| Trusted Flagger participants → TAT review/referral system | \nSubmission permission/capability, as advertised; no actual submission inferred. | \narchive-s-trusted-flagger, body lines 4–14, 30–50. | \n
| TAT classification/Archive → onboarded hash-list recipient | \nDocumented fingerprint and update interface; recipient implementation not observed. | \narchive-s-hash-docs, Hash Object Fields and Implementation Notes; archive-s-auth-docs, introduction. | \n
| Hash-list user → TAT analyst | \nDocumented error/inclusion feedback route, including response/decision visibility for the submitting user. | \narchive-s-feedback-docs, Overview, Feedback Status, Response notifications. | \n
| TAT hash record → recipient cache | \nDocumented deletion/update signal plus recommended incremental synchronization; not proven delivery, erasure or restoration. | \narchive-s-hash-docs, Implementation Notes and Best Practices. | \n
All are current public-document observations on 9 September 2026 with unknown exact deployment dates. No donation, procurement mandate, DSA appointment or government item-selection edge is supported by this packet.
", "path": "tcap-delivery-wave-2026-09-09/archive-access.md", "sourceIds": [ "w26-ver-s-archive-product", "w26-ver-s-archive-hash-api", "w26-ver-s-archive-feedback", "w26-ver-s-archive-auth", "w26-ver-s-archive-policy-shell", "w26-ver-s-tfp-product" ], "markdown": "9 September 2026. Bounded public-document pass complete. No account, policy-access application, content database, live API endpoint, underlying media, demo or contact request was accessed or submitted. Product documentation is evidence of the published interface and intended workflow, not runtime verification or an executed customer agreement. No publication or deployment date is inferred from a 2025 documentation copyright, a 2026 website footer or illustrative timestamps.\n\nThe useful new finding is a documented **item-level correction interface and a deletion signal intended for recipient caches**. The remaining gap is narrower than an absence of correction machinery: actual handling of disputed hashes, recipient synchronization and reconsideration of already imposed restrictions remain unobserved.\n\n## Who has which published capability\n\n| Surface | Published actor and capability | Authority/evidence boundary |\n| --- | --- | --- |\n| Archive Research Portal | TAT advertises media access for platforms and researchers, with automated/human sourcing and classification by AI models and OSINT specialists. | The page is a product explanation and contact route. It does not publish an admission rubric, download licence, redistribution permission or current government-access rule. |\n| Archive Hash List API | Platforms can ingest fingerprints in bulk, filter them and cache updates. | This concerns fingerprints; it is not permission to download or republish the underlying media. Recipient implementation and contractual reuse terms were not acquired. |\n| Trusted Flagger Platform | TAT describes trusted partners submitting categorized URLs for removal referrals, supported by its analysts and human/AI classifiers. | Participation is applied for. Detailed selection/permission rules and a statutory designation of these participants were not acquired. The product name alone supplies neither. |\n| Hash feedback | Onboarded users can report individual-hash problems and see analyst responses and decisions on their own feedback. | This is an internal review interface, not an established appeal right for every affected content creator or a demonstrated independent adjudication process. |\n\nResearch source: Archive Database Research Portal (see source references) complete substantive product body; retained `archive-research.txt` lines 1–19. Hash product source: Archive Database Hash List API (see source references) complete body; `archive-hash-product.txt` lines 1–16. Both are current observed pages with unknown revision dates. The hash page advertises daily updates and filters by file type, algorithm and ideology; its contact link is not an unrestricted public-download offer.\n\n## Submission, review and routing are separate steps\n\nThe Trusted Flagger product page (see source references) describes tiered URL submissions, a permission-controlled review environment, TAT analyst assessment and metadata supplied to recipient platforms. It advertises referrals to 125-plus platforms, retained historic requests, platform feedback and expansion of the archive/hash lists. That is an explicit advertised contribution path from referrals to a reusable collection, not proof that every submitted URL is admitted or every recipient acts. Its accuracy claim concerns online/offline status classification, not a measured terrorist-label accuracy rate. Complete substantive body read, `archive-trusted-flagger.txt` lines 1–56, especially 4–16 and 30–50. “Apply” on this page returned the same public page through the reader; no application was submitted.\n\nThe earlier accepted classification summary says TAT in-house experts assess source and content; the detailed verification policy remains behind an access-request form. The earlier accepted tier-policy/FAQ findings are reused from [tat.md](private-verification-wave-2026-09-09-tat.md), not reacquired. In particular, the old FAQ's no-download and no-government-access statements cannot settle the current Archive product's rights: that FAQ still describes future 2021/2022 development. The new TFP page's five-tier capability likewise does not establish when an older forthcoming tier became operational or what its current rule text says.\n\n## The public correction interface\n\nContent Hash List Feedback documentation (see source references) describes feedback from hash-list users for false positives, classification errors, duplicates and other issues. A report identifies the hash and category; an explanation is optional. Users can retrieve their submitted feedback, analyst responses and decisions. The decision states include removal, non-removal and pending, with response-sent status. Analysts aim to reply by onboarding email within three to five working days. This is an aim, not a demonstrated deadline met in a real case. Locators: Overview; Request Body; Feedback Status; Query Parameters/Feedback Object; Response notifications. Retained `archive-feedback-docs-prose.txt` lines 1–6, 18–30 and 35–48.\n\nThe example JSON is illustrative: fictional names and placeholder prose appear in it. Its counts, decisions and timestamps are **not** actual complaint volumes, decided appeals or deployment evidence. No request was executed. Neither this page nor the other inspected public access material establishes an independent appeal body, a second-level review route or a rule requiring restoration after a successful challenge.\n\n## What can reach another system after correction\n\nContent Hash List documentation (see source references) exposes fingerprints plus classification/file metadata, a deletion marker and last-update time. It describes checkpointed incremental updates and recommends local caching and monitoring deprecated entries. Soft deletion retains hash history. Thus it documents a technical means for a recipient to learn of a withdrawal while updating a local cache; it does not prove that recipients implemented it or erased prior copies. Locators: Query Parameters, Checkpointing, Hash Object Fields, Implementation Notes and Best Practices; `archive-hash-docs-prose.txt` lines 15–39.\n\nThe same page documents a ThreatExchange client configuration for fetching/comparing TAT hashes (lines 40–48). This is an interoperability instruction, not a new governance, funding or compulsory-database-membership relation. No client was installed or run. The inspected documentation does not specify a recipient-wide push notification, correction acknowledgement, compulsory rescan or reopening of prior moderation decisions. Those are acquisition gaps, not evidence they never occur.\n\nThe Authentication documentation (see source references) introduction and service list, expressly presupposes an already onboarded TCAP or Archive user with credentials. `archive-auth-docs.txt` lines 1–18; web-reader introduction and service lists also read. That establishes an access prerequisite, not the identity of the onboarding decision-maker or eligibility contract. Token examples and later implementation details were captured but not audited or executed.\n\n## Why current Archive terms remain an identified gap\n\nThe public Policies index (see source references) describes its Archive destination as containing terms of use, an access policy and taxonomy. Its actual Archive link (see source references) returned a web-reader internal error. The sole ordinary GET substitute succeeded, but the substantive HTML body contained only the heading **Archiving Policies**. Inspection found no iframe, embedded document or body terms/access-policy link. This is an obtainable page with no substantive policy text, not an administrative denial or evidence that no policy exists. Original retained as `archive-archive-policy.html`.\n\nThe index's legacy Research Portal feature (see source references) failed in the web reader with “Cache miss”; one ordinary GET succeeded with an empty substantive body. Its legacy hash feature (see source references) supplied navigation and an empty heading in the reader. These routes were then closed. No guessed private document path, gated full-policy request or account route was used. The latter reader-only observation is retained in this note; no original hash-feature HTML was acquired.\n\nConsequently this pass does not establish current media-download rights, onward hash redistribution/licensing, customer-model training permission, accepted-flagger criteria, cross-user visibility of archived requests, or an affected nonmember's formal challenge route. TAT/OHF would hold the current access agreement, permissions matrix and review procedure. Participating recipients would hold their cache-update and reconsideration records. Asking for those records would require a separately authorized outreach branch; none occurred here.\n\n## Forest implication and next discriminator\n\nThe plausible mechanism is distributed classification with controlled access and recipient discretion. The advertised Archive/TFP relationship can enlarge a shared collection; documented review and soft-delete fields can also correct it. A claim that there is no item-level challenge or withdrawal mechanism is now too strong. A claim that such a mechanism guarantees downstream remedy is also too strong. The strongest rival to a correction-failure theory is that ordinary recipient implementations consume deletion updates promptly and independently prevent or reverse mistakes; the public interface is consistent with that account.\n\nThe discriminating record is a redacted completed hash challenge linked to its analyst decision, resulting update/deletion record, recipient synchronization acknowledgement and any reconsideration of prior action. An executed access/usage agreement would separately establish mandatory handling and reuse limits. These records are not replaced by a documentation example or URL-availability statistics. Acquisition stops here for the parent's forest review.\n\n## Custody and scope\n\n`archive-captures.json` contains exact discovered URLs, final URLs, UTC acquisition times, hashes, lengths, reading scopes and source ancestry. Eight ordinary HTML captures succeeded; originals are response bytes and text files are derivatives. Seven new useful substantive/product or boundary sources were read/inspected plus the empty legacy Research Portal capture. The API home and Policies index were discovery/reader sources, not separately downloaded. Source IDs use the local `archive-s-` prefix; they are not canonical-case insertions.\n\nThe hash and feedback substantive descriptions/tables were read from both the public reader and newly retained HTML-derived prose. Example JSON was inspected in the reader for its illustrative nature; code snippets were not tested. Research/TFP/hash product bodies were read in full. Authentication scope is its prerequisite and service list only. Archive HTML was inspected for visible body content, links and document embeds. No full policy, live service behavior, extremist original or private customer dataset is claimed as read.\n\nDecisive original SHA-256: hash docs `e8e4c49c4b8bb5ee3c4be1fade240bb09d2f60834e3204a6b0a4d3f6e1c78df4`; feedback docs `e454d1fadc7944f8e115cef5edd73412da40831b4f4e72275b0ce7a89a58b71f`; Archive policy shell `d5c6e1fdafdb822b89a9bedd1997a468641e69a1d3dd04f9fb902eca43176c02`. Other hashes and exact scopes are in the manifest. No canonical case, previous packet or ZIP was changed.\n\n### Candidate typed relations for a later accepted integration\n\n| From → to | Typed relation and evidence stage | Source pointer |\n| --- | --- | --- |\n| Trusted Flagger participants → TAT review/referral system | Submission permission/capability, as advertised; no actual submission inferred. | `archive-s-trusted-flagger`, body lines 4–14, 30–50. |\n| TAT classification/Archive → onboarded hash-list recipient | Documented fingerprint and update interface; recipient implementation not observed. | `archive-s-hash-docs`, Hash Object Fields and Implementation Notes; `archive-s-auth-docs`, introduction. |\n| Hash-list user → TAT analyst | Documented error/inclusion feedback route, including response/decision visibility for the submitting user. | `archive-s-feedback-docs`, Overview, Feedback Status, Response notifications. |\n| TAT hash record → recipient cache | Documented deletion/update signal plus recommended incremental synchronization; not proven delivery, erasure or restoration. | `archive-s-hash-docs`, Implementation Notes and Best Practices. |\n\nAll are current public-document observations on 9 September 2026 with unknown exact deployment dates. No donation, procurement mandate, DSA appointment or government item-selection edge is supported by this packet.\n", "file": "research/tcap-delivery-wave-2026-09-09-archive-access.md" }, { "id": "tcap-delivery-wave-2026-09-09-canadian-acceptance", "title": "Canadian TCAP delivery and sponsor acceptance — bounded return", "description": "Acquisition closed: 9 September 2026. This is a candidate research note, not a canonical-case edit. It builds on the accepted ../quantspark-finance-wave-2026-09-09/canadian-awards.md, read before this inquiry. Original custody and hashes are in canadian-acceptance-captures.json.", "html": "Acquisition closed: 9 September 2026. This is a candidate research note, not a canonical-case edit. It builds on the accepted ../quantspark-finance-wave-2026-09-09/canadian-awards.md, read before this inquiry. Original custody and hashes are in canadian-acceptance-captures.json.
A publicly released departmental status document records Canada partnering with Tech Against Terrorism to formally launch TCAP 2.0 at the Canada Centre's event of 1–5 May 2023. This earns a dated government participation/delivery relation. It does not establish acceptance of either grant workstream, completion of Phase 2, a contractual milestone, or authorization/payment of a particular instalment.
\nThe same October 2023 document records Canada Centre–US National Security Council coordination on engagement with TAT, and reports a US funding process for TAT linked to Christchurch commitments. These are substantive governmental coordination and resource-mobilization relations, alongside the launch. The acquired Canadian record supplies the attribution; it does not identify a completed US award, paying agency, amount, conditions or payment.
\nThe public structure therefore includes intergovernmental engagement around TAT and a reported US funding process, as well as the Canadian agreement → stated workstreams → joint-launch chain. The Canadian sponsor's specific acceptance and financial certification record is still missing. A launch near the beginning of the disclosed Phase 2 term can concern an initial release while further development continues.
\nOriginal: [Public Safety Leader-Level Commitments and Roadmap Items, CAN033380](http[local research file] Internal document updated 5 October 2023, visually confirmed; publicly released in PIFI Canada Release 033, 12 August 2024. The old Protected B marking remains on the copy, alongside the public-release marking. It is an openly published commission exhibit.
\nRead scope: first four of 22 PDF pages, including table context; visual inspection of all four pages, with page 2 added during the local refinement. No claim to reading later pages. Page 1 establishes the Background / Current Status / Next Steps columns. Pages 3–4, rightmost Next Steps column, report the completed joint launch at the event of 1–5 May 2023. Page 2's general “On track” status concerns broader commitment 36 and is not TCAP grant acceptance. The adjoining Phase 2 capability description remains prospective. Do not attribute grant-approval authority to officials merely listed in the broader coordination row.
\nInstitutional context, scoped to this 5 October 2023 internal status account: PDF2's Background column, Commitment 36 / Prevention, says the Canada Centre worked with the US National Security Council on engagement with TAT concerning support for small social-media platforms addressing terrorist exploitation. The visual column alignment has been checked. This reports coordination around a named intermediary, without documenting who drafted particular classifications or could order a platform decision. Its separate OECD bullet does not make OECD part of the TCAP award.
\nPDF4's Next Steps column, continuing the joint-launch paragraph from PDF3, reports that the US was in the process of granting funding to TAT and links that funding to advancing Christchurch Call commitments. It earns a Canadian-department-reported US funding process as of October 2023, not an executed or paid US award. The passage does not name the US awarding agency, agreement, amount, exact application/approval stage, conditions or payment. The NSC engagement and US funding statements should remain separate relations; the latter does not identify the NSC as the funder.
\nRetained: canadian-can033380.pdf; SHA-256 51e45f81d9fa135c043e29e73c2071ea8fb5eb058c85d847a6839cf1d3381e16. Scoped text and page images are derivatives of these exact bytes. OCR rendered the update date incorrectly; the stated 5 October date comes from the image.
Original: [Public Safety Canada backgrounder, 3 May 2023](http[local research file]
\nRead scope: title/date/overview and the complete Recipient: Online Harms Foundation / TCAP Phase 2 entry; adjacent entries only as extraction boundaries. It names CAD1,945,904 and describes a protected archive, broader classification coverage and development of a moderation workflow tool. The prospective wording does not certify delivery. Its promotional database-size characterization is a departmental assertion, not an acquired independent performance measurement.
\nThe exact amount and recipient corroborate the previously retained disclosure for agreement 22841; this is another source for that same award, not another allocation. The entry does not split the award between workstreams, specify acceptance tests, attach a reporting schedule, or disclose paid cash.
\nRetained: canadian-may2023-announcement.html; SHA-256 951c5c680c4322f64f1cbef0fc4afd6b133c674f5e45f3234081e1ad82eb2d6a.
Original: [Public Safety Departmental Results Report 2019–20](http[local research file]
\nRead scope: complete Countering Radicalization to Violence subsection, including the TAT box and funding-agreement paragraph; report title, errata notice and current footer date. This is not a full departmental report audit. The current page says modified 9 April 2025; its reporting period remains 2019–20, and this inquiry did not establish its original online publication date.
\nThe department reports six signed funding agreements in 2019–20 and includes a rounded CAD1m TAT agreement. This supports attributed government reporting of execution during that fiscal year. It does not supply the executed instrument, exact signature date, payment receipt, or evidence that the repository had already been completed. The precise contribution value and legal recipient remain sourced to the earlier disclosure records. Do not convert the rounded number into a competing exact total.
\nRetained: canadian-drr2019-20.html; SHA-256 2ad60e90104fe46c2fda489fb0dc4f0f7502d0f8da4f966415c098128d9f199f. canadian-drr2019-20-scoped-extract.txt is a derivative of its named subsection.
Original: [Contributions du gouvernement du Canada à l'Appel de Christchurch 2022–2023](http[local research file] Full substantive French body read. Current footer modified 26 June 2025; the title's 2022–23 period is distinct from that maintenance date.
\nThe account recalls the 2022 TAT funding announcement and the subsequent Canadian conference and financing announcement. It provides no project evaluation, milestone certification, acceptance standard, or payment authorization. Its broader digital-citizenship and law-enforcement discussion was not converted into TCAP funding or operational control. This is a government contribution narrative, not a programme evaluation or project audit.
\nRetained: canadian-christchurch2022-23-fr.html; SHA-256 a5784a73ab6713af139526ee33218a0958f2233deb52655b3fa20f1ec3f093a1.
Agreement 21415: QuantSpark Foundation, initial CAD1,155,622.44; amended total CAD1,256,688.09, with the prior packet preserving the one-day start-date discrepancy and the 22 November 2021 amendment. Agreement 22841: Online Harms Foundation, CAD1,945,904, disclosed term 1 April 2023–31 March 2026. The exact detail records and amendment are already retained; they were not reacquired. Legal aliasing to company 11656320 depends on the separate registry evidence, not the Canadian records themselves.
\nThe earlier packet's current CRF terms show a reporting/payment framework but are modified 25 May 2026. They must not be silently imported as the actual 2020 or 2023 executed agreement terms. No new current-programme-rule edge is warranted here.
\n| Relation | Evidence/action state | What this permits and excludes |
\n|---|---|---|
\n| Canada Centre ↔ US National Security Council | Canadian-department-reported coordination on engagement with TAT; S1 PDF2 Background; status account updated 5 October 2023 | Named institutional coordination and subject matter; meeting date, individual decision rights and operational instructions unestablished. |\n| Canada Centre / US NSC engagement → TAT support for small social-media platforms | Reported engagement purpose; same S1 PDF2 passage | Preserve the named intermediary and purpose, without inventing a joint legal body or a government classification mandate. |\n| US government → TAT funding process | Canadian-department-reported process underway; S1 PDF4 Next Steps; as of 5 October 2023 | Christchurch-linked intended funding; awarding agency, agreement, amount, execution and payment unknown. Do not substitute NSC as the funder. |\n| Public Safety Canada / Canada Centre → TCAP 2.0 launch, 1–5 May 2023 | Government-reported joint launch; S1 PDF3–4 | Dated participation in launch, not award acceptance or complete Phase 2 delivery. Link TAT as the named partner separately from OHF's legal-recipient role. |
\n| Public Safety Canada → agreement 21415 / original TCAP funding | Government-reported signed agreement during FY2019–20; S3 subsection | Adds source ancestry supporting reported execution; exact amount and identity stay with disclosure. No cash state. |
\n| Public Safety Canada → agreement 22841 / two workstreams | Official announced purpose, 3 May 2023; S2 recipient entry | Corroborates already-mapped award and intended deliverables; no duplicate amount or new award. |
\nNo edge from government to an accepted milestone, performance score, specific content decision, platform takedown requirement, or disbursement is earned by this return.
\nThis inquiry used seven focused queries, rather than reopening all award routes: site:publicsafety.gc.ca "TCAP" "results"; site:canada.ca "22841" "Terrorist"; site:publicsafety.gc.ca "Tech Against Terrorism" evaluation; "21415" "Terrorist Content" report acceptance; site:publicsafety.gc.ca ("21415" OR "22841"); site:canada.ca "Terrorist Content Analytics Platform" (evaluation OR milestones OR acceptance OR results); site:publicsafety.gc.ca "Tech Against Terrorism" (evaluation OR "final report" OR "performance").
Results supplied the released status exhibit and official reporting above, existing award-disclosure duplicates, announcements, and tangential records. Paginated disclosure snippets were discovery only; no inference from their surrounding unrelated entries. The 2022–23 departmental report and February 2026 briefing appeared in search results but were not opened or treated as independent acquired acceptance evidence. Parliamentary testimony and recipient reports were left to the delivery lane, because recipient performance assertions do not themselves establish sponsor approval. No programme evaluation was promoted to project acceptance.
\nThe French online-extremism page supplied the Christchurch report link. An initial combined tool result was wholly truncated and counted as unread; a single re-open of that link succeeded. The report's English-language link returned a web internal error; the French original and full substantive text sufficed, so there was no retry or proxy. Four ordinary public GET captures succeeded with status 200. Previously closed award routes were not repeated.
\nPrecise frontier: executed agreements 21415 and 22841 and their schedules/amendments; the recipient's dated interim/final workstream reports; the department's comments, acceptance or deficiency decision; and any linked payment certification. The likely record holders are Public Safety Canada's Canada Centre/CRF grants administration and the recipient for its counterpart copies. No request has been sent. Public non-location in this finite pass is not proof those records do not exist or that the project failed.
\nA discriminating next record would match an actual archived-data or workflow-tool deliverable to a specified agreement milestone and show the responsible official accepting, conditionally accepting or rejecting it. Until then the rival explanations remain ordinary staged development after launch, departmental endorsement without published contractual acceptance, or further review/conditions whose records are not public in this corpus. The present evidence does not select between those explanations.
\nOn 9 September 2026, the prior packet was preserved as canadian-acceptance-before-institutional-refinement.md before adding the institutional context already present in retained CAN033380. PDF2 was rendered and visually read; PDF3–4 had already been visually checked. The refinement adds no external acquisition, US award search or new source. See canadian-acceptance-refinement-receipt.json for before/after hashes. Acquisition remains closed pending the parent forest assessment.
9 September 2026. Root delivery-account lane closed. This packet reads operator statements and the previously retained transparency report; Canadian sponsor evidence and product governance are separate lanes. No live product, content database, extremist material, account or access request was opened.
\nThe [5 May 2023 launch account](http[local research file] dates the Ottawa presentation to 1 May. TAT described Phase 2 as underway: building an archive and hashes, with new crisis-response/threat-to-life APIs. It attributes a supportive statement to Canada Centre executive Robert Burley and connects the programme to Canada's Christchurch Call commitments. This is an event account with attributed government remarks, not a milestone sign-off or payment record. Claimed totals were over 44,000 unique URLs and 25,000 alerts to 100 platforms. The embedded video was not inspected.
\nThe [21 December 2023 review](http[local research file] authored by Hadley, connects TCAP development to Public Safety Canada support, describes a Microsoft AI-detection pilot, and says Altitude was being built with Google Jigsaw. It also reports an implemented tiered system and alerts covering unofficial supporter-media outlets and attack-preparation material. This supplies an operator account of expanded coverage, rather than making every prospective tier operational. Its claims about policy participation and successful disruption were not independently validated here.
\nThe [20 December 2024 review](http[local research file] reports an archive exceeding 46,000 files, over 182,000 hashes, and media/URL hash-sharing APIs. It describes a developed LLM for TCAP identification/classification with Public Safety Canada support, and Altitude co-development with Jigsaw. It reports a trusted-flagger programme enabling selected subject specialists to alert designated propaganda. These are attributed implementation claims; the article supplies no model-validation results, project-cost allocation, executed licence or government acceptance. Reported cumulative URLs, alerts, entities and archive counts describe different units and must not be added. The separate Australian-supported crisis team and European programme are not recast as Canadian-funded work.
\nThe organisation's [31 December 2025 update](http[local research file] reports a March 2025 interface relaunch, 10,349 added archive files, one 30-day period with 10,000 alerted URLs, and hash/URL sharing with approximately six major technology companies. It describes systems as auditable but supplies no completed audit. Its account of trained specialist flaggers adds a claimed submission channel. Statements about further AI-company engagement and identifying potentially dangerous individuals are forward-looking priorities, not proof of a deployed population-surveillance system. No recipient roster or case-level removal/restoration evidence is supplied.
\nThe short [9 January 2023 Jigsaw announcement](http[local research file] describes early collaboration on a prioritisation tool ingesting TCAP URLs. The primary indexed reader delivered the announcement; the ordinary GET returned404. Retained HTML is the error response, not an original of the announcement. The later2023/2024 accounts independently establish what TAT subsequently said about this collaboration; none is independent of TAT as an institution. Actual Altitude stewardship/code is examined in the separate product lane.
\nThe retained [second TCAP transparency report](http[local research file] PDF28–29, describes collection, human verification/classification, database submission, hashing/archiving, advisory email alerts and automated availability checks. Provider moderation is expressly discretionary. Its offline metric tests availability at a submitted source; it is not a direct observation of why the item disappeared. PDF33 distinguishes official group material in the then-policy from supporter content; later expansion should be dated rather than silently projected backward. PDF34 reports public consultation and a legal review commissioned by TAT; the full legal review is privileged and the shorter version requires a request. Neither was acquired. Root additionally read PDF5–6 programme/statistical context in this lane. This limited read does not include every chart or all40 pages.
\nThese accounts create a time-ordered delivery trail that fits the disclosed archive/classification/workflow workstreams. That fit is an inference, not a contractual allocation: no acquired budget annex shows which of Altitude, the LLM, the archive or an API was paid from agreement22841. There is no need to invent one fully integrated product to explain the record. A shared archive can supply signals to multiple systems; a workflow tool can mediate provider decisions; government support can fund development while particular data and enforcement decisions remain private. The public product and Canadian lanes must resolve their own parts.
\nThe highest-value missing evidence is a specific delivery/milestone schedule joined to acceptance and costs, plus the operative rules for classifier selection, access, redistribution and correction. A funding announcement, an operator's product statistic and a subsequent offline URL each answer a different question. None alone supplies the intervening command chain.
\nroot-delivery-captures.json records three successful original article responses and the launch response, plus the unsuccessful Jigsaw response. The prior report remains in private-verification-wave with SHA2568e41db80e8a9619aa5fccf252b7a831508e34c82c02c924f58c68e79ee5c8d40. No originals were overwritten. Five targeted queries sought later TCAP reporting, Phase2 outputs and Jigsaw/March2025 developments; the first mistakenly used site.techagainstterrorism.org rather than site: and is not counted as successful domain restriction. The initial combined search output truncated; relevant sources were subsequently opened/read in full substantive scope. One mistaken link click returned the same category page; its actual launch link was then read. ISD/ResearchGate snippets were discovery only, not findings. No new broad institutional branch opened.
", "path": "tcap-delivery-wave-2026-09-09/delivery-accounts.md", "sourceIds": [ "w26-ver-s-tat-review2023", "w26-ver-s-tat-review2024", "w26-ver-s-tat-review2025" ], "markdown": "9 September 2026. Root delivery-account lane closed. This packet reads operator statements and the previously retained transparency report; Canadian sponsor evidence and product governance are separate lanes. No live product, content database, extremist material, account or access request was opened.\n\n## Dated delivery statements\n\nThe 5 May 2023 launch account (see source references) dates the Ottawa presentation to 1 May. TAT described Phase 2 as underway: building an archive and hashes, with new crisis-response/threat-to-life APIs. It attributes a supportive statement to Canada Centre executive Robert Burley and connects the programme to Canada's Christchurch Call commitments. This is an event account with attributed government remarks, not a milestone sign-off or payment record. Claimed totals were over 44,000 unique URLs and 25,000 alerts to 100 platforms. The embedded video was not inspected.\n\nThe 21 December 2023 review (see source references) authored by Hadley, connects TCAP development to Public Safety Canada support, describes a Microsoft AI-detection pilot, and says Altitude was being built with Google Jigsaw. It also reports an implemented tiered system and alerts covering unofficial supporter-media outlets and attack-preparation material. This supplies an operator account of expanded coverage, rather than making every prospective tier operational. Its claims about policy participation and successful disruption were not independently validated here.\n\nThe 20 December 2024 review (see source references) reports an archive exceeding 46,000 files, over 182,000 hashes, and media/URL hash-sharing APIs. It describes a developed LLM for TCAP identification/classification with Public Safety Canada support, and Altitude co-development with Jigsaw. It reports a trusted-flagger programme enabling selected subject specialists to alert designated propaganda. These are attributed implementation claims; the article supplies no model-validation results, project-cost allocation, executed licence or government acceptance. Reported cumulative URLs, alerts, entities and archive counts describe different units and must not be added. The separate Australian-supported crisis team and European programme are not recast as Canadian-funded work.\n\nThe organisation's 31 December 2025 update (see source references) reports a March 2025 interface relaunch, 10,349 added archive files, one 30-day period with 10,000 alerted URLs, and hash/URL sharing with approximately six major technology companies. It describes systems as auditable but supplies no completed audit. Its account of trained specialist flaggers adds a claimed submission channel. Statements about further AI-company engagement and identifying potentially dangerous individuals are forward-looking priorities, not proof of a deployed population-surveillance system. No recipient roster or case-level removal/restoration evidence is supplied.\n\nThe short 9 January 2023 Jigsaw announcement (see source references) describes early collaboration on a prioritisation tool ingesting TCAP URLs. The primary indexed reader delivered the announcement; the ordinary GET returned404. Retained HTML is the error response, not an original of the announcement. The later2023/2024 accounts independently establish what TAT subsequently said about this collaboration; none is independent of TAT as an institution. Actual Altitude stewardship/code is examined in the separate product lane.\n\n## What the earlier process report actually explains\n\nThe retained second TCAP transparency report (see source references) PDF28–29, describes collection, human verification/classification, database submission, hashing/archiving, advisory email alerts and automated availability checks. Provider moderation is expressly discretionary. Its offline metric tests availability at a submitted source; it is not a direct observation of why the item disappeared. PDF33 distinguishes official group material in the then-policy from supporter content; later expansion should be dated rather than silently projected backward. PDF34 reports public consultation and a legal review commissioned by TAT; the full legal review is privileged and the shorter version requires a request. Neither was acquired. Root additionally read PDF5–6 programme/statistical context in this lane. This limited read does not include every chart or all40 pages.\n\n## What changes, and what remains unjoined\n\nThese accounts create a time-ordered delivery trail that fits the disclosed archive/classification/workflow workstreams. That fit is an inference, not a contractual allocation: no acquired budget annex shows which of Altitude, the LLM, the archive or an API was paid from agreement22841. There is no need to invent one fully integrated product to explain the record. A shared archive can supply signals to multiple systems; a workflow tool can mediate provider decisions; government support can fund development while particular data and enforcement decisions remain private. The public product and Canadian lanes must resolve their own parts.\n\nThe highest-value missing evidence is a specific delivery/milestone schedule joined to acceptance and costs, plus the operative rules for classifier selection, access, redistribution and correction. A funding announcement, an operator's product statistic and a subsequent offline URL each answer a different question. None alone supplies the intervening command chain.\n\n## Custody and finite search\n\nroot-delivery-captures.json records three successful original article responses and the launch response, plus the unsuccessful Jigsaw response. The prior report remains in private-verification-wave with SHA2568e41db80e8a9619aa5fccf252b7a831508e34c82c02c924f58c68e79ee5c8d40. No originals were overwritten. Five targeted queries sought later TCAP reporting, Phase2 outputs and Jigsaw/March2025 developments; the first mistakenly used site.techagainstterrorism.org rather than site: and is not counted as successful domain restriction. The initial combined search output truncated; relevant sources were subsequently opened/read in full substantive scope. One mistaken link click returned the same category page; its actual launch link was then read. ISD/ResearchGate snippets were discovery only, not findings. No new broad institutional branch opened.\n", "file": "research/tcap-delivery-wave-2026-09-09-delivery-accounts.md" }, { "id": "uk-age-wave-2026-09-09-advocacy-findings", "title": "UK age assurance: specific advocacy contributions and decisions", "description": "The clearest route is a named contributor working on a specific government amendment, followed by a recorded parliamentary decision. The regulator record supplies a second route: an NGO's identifiable proposal and Ofcom's explicit refusal of its requested criterion architecture. Later 5Rights advocacy links implementation to standards and certification. Thes", "html": "Prepared 9 September 2026. Bounded candidate only; case-r10 unchanged.
\nThe clearest route is a named contributor working on a specific government amendment, followed by a recorded parliamentary decision. The regulator record supplies a second route: an NGO's identifiable proposal and Ofcom's explicit refusal of its requested criterion architecture. Later 5Rights advocacy links implementation to standards and certification. These are contribution and decision records; they do not identify a common sponsor directing the package, a project-specific funding allocation, or a compulsory identity vendor.
\nOn 6 July 2023, Lord Parkinson thanked Kidron, Harding and Bethell for months of collaboration on age assurance and specifically credited Kidron's work on Amendment 124. Kidron had declared her chair role at 5Rights in the same day's debate. This is a contemporaneous NGO/parliamentary bridge and ministerial attribution, stronger than a shared conference appearance. It does not establish sole authorship or which contributor drafted each clause. [Hansard, L1600, L4422, L4436](http[local research file]
\nThe [official Amendment 124 record](http[local research file] names Parkinson as lead member, Kidron and Clement-Jones as sponsors, and records agreement. Its principles include age/age-range effectiveness, AADC standards, proportionality, interoperability and ease of use; it permits references to industry/technical standards without naming a mandatory certifier. [The 10 July chamber record, L4267–4293](http[local research file] confirms Amendment 124 agreed and Amendment 125 not moved.
\nKidron's [Amendment 125](http[local research file] proposed purpose limits and deletion/secure-storage wording. The minister declined that addition, citing existing privacy/data law and concern about legal confusion. Not moved is not a vote defeating it. Kidron also distinguished identifying children through assurance from designing a service safe for children; this does not support describing all her proposals as universal government-ID requirements. [10 July debate, L129, L136–140, L342–371, L405](http[local research file]
\nThe actual [5Rights Part 5 response, cover March 2024](http[local research file] requested a privacy-preserving criterion, minimal collection, limited retention and security. Its proposed on-device/result-only approach sought to avoid a central identity database linked to online activity (PDF 1–2, 5–7). The response's March cover date differs from the consultation page's later listing date; no exact submission or upload day is inferred.
\nOfcom explicitly attributed the request to 5Rights in January 2025 statement 3.277–3.278 (PDF/printed 62). It declined the separate criterion because data-protection requirements already apply (3.288, PDF/printed 64), while clarifying design/minimisation guidance and saying both regimes are mandatory (3.289–3.290, PDF/printed 65). The clarification is not attributed solely to 5Rights. Shared primary source: w7-policy-s-statement2025; independently read here from policy-reader-closure1.json and policy-reader-current1.json. Token attribution is retained and independently read in policy-reader-token-attribution.json.
\nThe 17 July 2024 children response.
\nThe October 2025 additional-measures response.
\nThis earns an explicit advocacy-to-certification reference. It establishes neither subsequent Ofcom adoption nor a 5Rights–ACCS procurement, payment or clearance agreement. The market lane independently owns legal control and standards roles. Keep ACCS's historical umbrella distinct from the newer exact legal nodes. The response does not identify which ACCS legal person would receive business.
\nThe market lane's acquired [euCONSENT project update, 9 January 2025](http[local research file] reports commissioning a child/human-rights review through team academics Simone van der Hof, Sonia Livingstone and Abhilash Nair, with Emma Day and Sabine Witting of Tech Legality as intended co-leads (paragraph beginning 'We have also commissioned'; captured text line 71). Day, Witting and Livingstone reuse the exact people IDs already documented as authors of the CSEA principles report cited by eSafety.
\nThis is an actual attributed commission linking named work across the two lanes. It is not a finding that eSafety commissioned AgeAware, that all team academics personally contracted, or that either report determined the regulator's text. The review, fee, contract, clearance terms and completed delivery are unknown. 'Independent' is the commissioning participant's description. The commissioning actor remains euCONSENT in the article's later-ASBL project context; no original 2021 consortium grant is reassigned.
\nSource w7-market-s-ageaware-update reuses the market lane's original HTML receipt and hash. Its substantive commission paragraph was independently read here; this supplemental source is distinct from the advocacy lane's reader-only acquisitions.
\nThe current Kidron interests register was read through the public reader and upgrades the existing source src-576f1b1b6fa2 from discovered to scoped read. Its current business, parliamentary-assistance and travel entries do not establish funding of the 2023 amendment work. Existing funding edges must not be earmarked to this UK package without a programme record.
\nThis bounded pass did not earn new NSPCC, Internet Matters, WeProtect, Fabian or WEF contribution edges. That is a limit of this pass, not evidence they had no relevant role. The policy disagreement and privacy demands challenge a flattening of the record into a single universal-identity campaign; they do not disprove broader power through implementation requirements.
\nFive ordinary public GETs returned HTTP 403: both targeted Hansard pages, the July 2024 and October 2025 responses, and the current register. No unchanged-route retries occurred. Public web-reader text was the substitute. Nine reader derivatives preserve scoped text and locators; no original HTML/PDF bytes or original-source hashes are claimed. The March response and official amendment records were likewise read through that public reader. See advocacy-capture-manifest.json and advocacy-derivative-manifest.json. Scopes are attached to each source in advocacy-candidate.json; the findings do not imply full PDF visual review.
\nStop here unless a materially new record appears: contributor/government amendment redlines and clearance records; a programme-specific grant/commission and allocation; or a substantive disposition of the October certification recommendation. No outreach, submission, payment, access workaround or canonical edit occurred. Candidate: seven new nodes, seventeen new edges, seven new sources plus the existing-register access update and one shared market source; three parked, custodian-specific leads. Shared Ofcom source is supplied by the policy lane.
", "path": "uk-age-wave-2026-09-09/advocacy-findings.md", "sourceIds": [ "src-576f1b1b6fa2", "w7-adv-s-hansard0706", "w7-adv-s-hansard0710", "w7-adv-s-amend124", "w7-adv-s-amend125", "w7-adv-s-5rights-part5", "w7-adv-s-5rights-child2024", "w7-adv-s-5rights-additional2025" ], "markdown": "Prepared 9 September 2026. Bounded candidate only; case-r10 unchanged.\n\nThe clearest route is a named contributor working on a specific government amendment, followed by a recorded parliamentary decision. The regulator record supplies a second route: an NGO's identifiable proposal and Ofcom's explicit refusal of its requested criterion architecture. Later 5Rights advocacy links implementation to standards and certification. These are contribution and decision records; they do not identify a common sponsor directing the package, a project-specific funding allocation, or a compulsory identity vendor.\n\n## Parliament: credited work, an agreed amendment, and a limit\n\nOn 6 July 2023, Lord Parkinson thanked Kidron, Harding and Bethell for months of collaboration on age assurance and specifically credited Kidron's work on Amendment 124. Kidron had declared her chair role at 5Rights in the same day's debate. This is a contemporaneous NGO/parliamentary bridge and ministerial attribution, stronger than a shared conference appearance. It does not establish sole authorship or which contributor drafted each clause. Hansard, L1600, L4422, L4436 (see source references)\n\nThe official Amendment 124 record (see source references) names Parkinson as lead member, Kidron and Clement-Jones as sponsors, and records agreement. Its principles include age/age-range effectiveness, AADC standards, proportionality, interoperability and ease of use; it permits references to industry/technical standards without naming a mandatory certifier. The 10 July chamber record, L4267–4293 (see source references) confirms Amendment 124 agreed and Amendment 125 not moved.\n\nKidron's Amendment 125 (see source references) proposed purpose limits and deletion/secure-storage wording. The minister declined that addition, citing existing privacy/data law and concern about legal confusion. Not moved is not a vote defeating it. Kidron also distinguished identifying children through assurance from designing a service safe for children; this does not support describing all her proposals as universal government-ID requirements. 10 July debate, L129, L136–140, L342–371, L405 (see source references)\n\n## Ofcom: a traceable request and a mixed disposition\n\nThe actual 5Rights Part 5 response, cover March 2024 (see source references) requested a privacy-preserving criterion, minimal collection, limited retention and security. Its proposed on-device/result-only approach sought to avoid a central identity database linked to online activity (PDF 1–2, 5–7). The response's March cover date differs from the consultation page's later listing date; no exact submission or upload day is inferred.\n\nOfcom explicitly attributed the request to 5Rights in January 2025 statement 3.277–3.278 (PDF/printed 62). It declined the separate criterion because data-protection requirements already apply (3.288, PDF/printed 64), while clarifying design/minimisation guidance and saying both regimes are mandatory (3.289–3.290, PDF/printed 65). The clarification is not attributed solely to 5Rights. Shared primary source: w7-policy-s-statement2025; independently read here from policy-reader-closure1.json and policy-reader-current1.json. Token attribution is retained and independently read in policy-reader-token-attribution.json.\n\nThe 17 July 2024 children response.\n\n## The standards/certification seam\n\nThe October 2025 additional-measures response.\n\nThis earns an explicit advocacy-to-certification reference. It establishes neither subsequent Ofcom adoption nor a 5Rights–ACCS procurement, payment or clearance agreement. The market lane independently owns legal control and standards roles. Keep ACCS's historical umbrella distinct from the newer exact legal nodes. The response does not identify which ACCS legal person would receive business.\n\n## A commissioned-work bridge to the Australian inquiry\n\nThe market lane's acquired euCONSENT project update, 9 January 2025 (see source references) reports commissioning a child/human-rights review through team academics Simone van der Hof, Sonia Livingstone and Abhilash Nair, with Emma Day and Sabine Witting of Tech Legality as intended co-leads (paragraph beginning 'We have also commissioned'; captured text line 71). Day, Witting and Livingstone reuse the exact people IDs already documented as authors of the CSEA principles report cited by eSafety.\n\nThis is an actual attributed commission linking named work across the two lanes. It is not a finding that eSafety commissioned AgeAware, that all team academics personally contracted, or that either report determined the regulator's text. The review, fee, contract, clearance terms and completed delivery are unknown. 'Independent' is the commissioning participant's description. The commissioning actor remains euCONSENT in the article's later-ASBL project context; no original 2021 consortium grant is reassigned.\n\nSource w7-market-s-ageaware-update reuses the market lane's original HTML receipt and hash. Its substantive commission paragraph was independently read here; this supplemental source is distinct from the advocacy lane's reader-only acquisitions.\n\n## Funding, useful negatives and custody\n\nThe current Kidron interests register was read through the public reader and upgrades the existing source src-576f1b1b6fa2 from discovered to scoped read. Its current business, parliamentary-assistance and travel entries do not establish funding of the 2023 amendment work. Existing funding edges must not be earmarked to this UK package without a programme record.\n\nThis bounded pass did not earn new NSPCC, Internet Matters, WeProtect, Fabian or WEF contribution edges. That is a limit of this pass, not evidence they had no relevant role. The policy disagreement and privacy demands challenge a flattening of the record into a single universal-identity campaign; they do not disprove broader power through implementation requirements.\n\nFive ordinary public GETs returned HTTP 403: both targeted Hansard pages, the July 2024 and October 2025 responses, and the current register. No unchanged-route retries occurred. Public web-reader text was the substitute. Nine reader derivatives preserve scoped text and locators; no original HTML/PDF bytes or original-source hashes are claimed. The March response and official amendment records were likewise read through that public reader. See advocacy-capture-manifest.json and advocacy-derivative-manifest.json. Scopes are attached to each source in advocacy-candidate.json; the findings do not imply full PDF visual review.\n\nStop here unless a materially new record appears: contributor/government amendment redlines and clearance records; a programme-specific grant/commission and allocation; or a substantive disposition of the October certification recommendation. No outreach, submission, payment, access workaround or canonical edit occurred. Candidate: seven new nodes, seventeen new edges, seven new sources plus the existing-register access update and one shared market source; three parked, custodian-specific leads. Shared Ofcom source is supplied by the policy lane.\n", "file": "research/uk-age-wave-2026-09-09-advocacy-findings.md" }, { "id": "uk-age-wave-2026-09-09-implementation-findings", "title": "UK implementation: platform choice, overlapping regulators and a cancelled national programme", "description": "Read on9September2026. This is a bounded public-document account, not a live verification-flow experiment, source-code audit or proof of actual deletion.", "html": "Read on9September2026. This is a bounded public-document account, not a live verification-flow experiment, source-code audit or proof of actual deletion.
\nThe March2026 Cabinet Office digital-ID consultation proposed online age proof as a use for younger teenagers and said the credential would not be the only permitted method (PDF51–52). It also considered a cross-government identifier (PDF33–34), with boundaries around government use. These are dated proposals. The House of Commons Library briefing dated4September2026 says the national scheme was cancelled21July2026, while OneLogin, Wallet and the private-sector DVS framework are separate. A ministerial answer dated8September explicitly confirms cancellation and says a full business case had not been finalised. Do not project the Marchproposal forward as current policy or claim cancellation removes online-age obligations. The recorded linkage is historical; current private and service-specific routes need separate tracing.
\nOriginal91page consultation PDF captured; scoped read PDF32–34,39–40,51–52,81–83. Library and minister pages read through web reader; originalGET403, no retry, derivative observations only. The Library statement about continuing other systems is an institutional summary, not verification of every component's live deployment.
\nCurrent Reddit Help, updated24June2026, names Persona for ID/selfie verification, describes account-signal and third-party email inference, and lists app-store age signals depending on region/device. This generic page does not establish that every option is available to every UK user. Reddit says it stores verified age/status and receives no verification photos. The customer-specific Persona FAQ identifies Reddit as controller and Persona as processor under Reddit's instructions; it says ID-route birthdate or selfie-route estimated age is returned and personal information deleted within three days. These are published provider representations; contractual schedules and deletion evidence were not obtained. The date of birth return should not be reduced to a universal binary over18 token.
\nPersona's original FAQ HTML was captured; Reddit ordinaryGET403, readable web substitute retained as derivative. No guessed email-inference supplier, fee or link to VerifyMy is earned. Current provider summaries must not be backdated to2025 retention arrangements.
\nThe ICO's signed penalty notice23February2026 imposes GBP14,472,500 for specified UKGDPR infringements. Its processing period ends8July2025; its DPIA period ends31December2024. Paragraph74/PDF40 records Reddit's UK NSFW age-verification through Persona starting14July2025, after the relevant period. Paragraph301(a)/PDF152 expressly rejects treating later OSA compliance work as mitigation of those earlier infringements. Thus the fine was not an OSA fine for deploying Persona incorrectly, and the notice does not direct procurement of Persona.
\nThe ICO's1April2026 update records an appeal. No tribunal disposition was established in the scoped current-source review. No receipt of penalty payment is claimed. The source includes a dispute over pseudonymity and additional data in paragraphs71–74/PDF38–40. This is an actual opposing platform position, not analyst speculation.
\nThe ICO's separate FOIresponse24March2026 says it holds regular Ofcom meetings, varying by team from fortnightly to quarterly, while refusing a broad multi-year request on burden grounds. That establishes a reported coordination routine, not that every meeting concerned Reddit. It identifies normally generated records worth seeking with a much narrower time/topic boundary. The refusal is not evidence of concealed misconduct.
\nOriginal209page penalty and8pageFOI PDFs retained. Scoped penalty reading: PDF4–11 introduction/periods,38–40 age-gating and competing position,42–43 controller/period discussion,134 and152 mitigation. No full annex, statistical or tax/financial audit. ICOpressHTML captured and read for penalty and appeal update. FOIpp1–7read request, response, meeting description and assistance; individual requested categories are not all confirmed held.
\nThe UK-specific help page updated28July2026 names k-ID and describes age checks to unblur flagged media, access age-restricted spaces and change certain settings, including MessageRequests. It says those declining retain account, DMs, voice, friends and servers, with the stated restrictions. It describes on-device facial estimation, document checks and a GoogleWallet test; methods are not interchangeable disclosures. This refines existing Discord/k-ID/Wallet edges rather than inventing another deployment. Its privacy and deletion statements are company claims, not tested properties.
\nOriginalGET403; web-reader body read, derivative only. Existing general Discord source and historic age-appeal support exposure remain separate. Do not treat the historic support breach as a new compromise of the normal UK verification path.
\nReddit–Persona processing schedule, age-output schema and retention/deletion settings would identify the actual controller instructions and retained fields. The UK rollout/configuration record would establish which alternative routes work there. A dated ICO–Ofcom age-assurance meeting agenda/action log near July2025 would test specific coordination without requesting years of unrelated material. Discord's UK change/requirement mapping would show which feature restrictions were platform choices and how these relate to statutory duties. No outreach, submission, paid access or logged-in flow was undertaken.
", "path": "uk-age-wave-2026-09-09/implementation-findings.md", "sourceIds": [ "w7-impl-s-reddit", "w7-impl-s-persona", "w7-impl-s-discord", "w7-impl-s-penalty", "w7-impl-s-appeal", "w7-impl-s-foi", "w7-impl-s-consultation", "w7-impl-s-library", "w7-impl-s-minister" ], "markdown": "Read on9September2026. This is a bounded public-document account, not a live verification-flow experiment, source-code audit or proof of actual deletion.\n\n## Currentness correction\n\nThe March2026 Cabinet Office digital-ID consultation proposed online age proof as a use for younger teenagers and said the credential would not be the only permitted method (PDF51–52). It also considered a cross-government identifier (PDF33–34), with boundaries around government use. These are dated proposals. The House of Commons Library briefing dated4September2026 says the national scheme was cancelled21July2026, while OneLogin, Wallet and the private-sector DVS framework are separate. A ministerial answer dated8September explicitly confirms cancellation and says a full business case had not been finalised. Do not project the Marchproposal forward as current policy or claim cancellation removes online-age obligations. The recorded linkage is historical; current private and service-specific routes need separate tracing.\n\nOriginal91page consultation PDF captured; scoped read PDF32–34,39–40,51–52,81–83. Library and minister pages read through web reader; originalGET403, no retry, derivative observations only. The Library statement about continuing other systems is an institutional summary, not verification of every component's live deployment.\n\n## Reddit: named processor and constrained claims\n\nCurrent Reddit Help, updated24June2026, names Persona for ID/selfie verification, describes account-signal and third-party email inference, and lists app-store age signals depending on region/device. This generic page does not establish that every option is available to every UK user. Reddit says it stores verified age/status and receives no verification photos. The customer-specific Persona FAQ identifies Reddit as controller and Persona as processor under Reddit's instructions; it says ID-route birthdate or selfie-route estimated age is returned and personal information deleted within three days. These are published provider representations; contractual schedules and deletion evidence were not obtained. The date of birth return should not be reduced to a universal binary over18 token.\n\nPersona's original FAQ HTML was captured; Reddit ordinaryGET403, readable web substitute retained as derivative. No guessed email-inference supplier, fee or link to VerifyMy is earned. Current provider summaries must not be backdated to2025 retention arrangements.\n\n## A second regulatory mechanism, with a specific period\n\nThe ICO's signed penalty notice23February2026 imposes GBP14,472,500 for specified UKGDPR infringements. Its processing period ends8July2025; its DPIA period ends31December2024. Paragraph74/PDF40 records Reddit's UK NSFW age-verification through Persona starting14July2025, after the relevant period. Paragraph301(a)/PDF152 expressly rejects treating later OSA compliance work as mitigation of those earlier infringements. Thus the fine was not an OSA fine for deploying Persona incorrectly, and the notice does not direct procurement of Persona.\n\nThe ICO's1April2026 update records an appeal. No tribunal disposition was established in the scoped current-source review. No receipt of penalty payment is claimed. The source includes a dispute over pseudonymity and additional data in paragraphs71–74/PDF38–40. This is an actual opposing platform position, not analyst speculation.\n\nThe ICO's separate FOIresponse24March2026 says it holds regular Ofcom meetings, varying by team from fortnightly to quarterly, while refusing a broad multi-year request on burden grounds. That establishes a reported coordination routine, not that every meeting concerned Reddit. It identifies normally generated records worth seeking with a much narrower time/topic boundary. The refusal is not evidence of concealed misconduct.\n\nOriginal209page penalty and8pageFOI PDFs retained. Scoped penalty reading: PDF4–11 introduction/periods,38–40 age-gating and competing position,42–43 controller/period discussion,134 and152 mitigation. No full annex, statistical or tax/financial audit. ICOpressHTML captured and read for penalty and appeal update. FOIpp1–7read request, response, meeting description and assistance; individual requested categories are not all confirmed held.\n\n## Discord: feature gates rather than account-wide identity entry\n\nThe UK-specific help page updated28July2026 names k-ID and describes age checks to unblur flagged media, access age-restricted spaces and change certain settings, including MessageRequests. It says those declining retain account, DMs, voice, friends and servers, with the stated restrictions. It describes on-device facial estimation, document checks and a GoogleWallet test; methods are not interchangeable disclosures. This refines existing Discord/k-ID/Wallet edges rather than inventing another deployment. Its privacy and deletion statements are company claims, not tested properties.\n\nOriginalGET403; web-reader body read, derivative only. Existing general Discord source and historic age-appeal support exposure remain separate. Do not treat the historic support breach as a new compromise of the normal UK verification path.\n\n## Maximum-value missing records\n\nReddit–Persona processing schedule, age-output schema and retention/deletion settings would identify the actual controller instructions and retained fields. The UK rollout/configuration record would establish which alternative routes work there. A dated ICO–Ofcom age-assurance meeting agenda/action log near July2025 would test specific coordination without requesting years of unrelated material. Discord's UK change/requirement mapping would show which feature restrictions were platform choices and how these relate to statutory duties. No outreach, submission, paid access or logged-in flow was undertaken.\n", "file": "research/uk-age-wave-2026-09-09-implementation-findings.md" }, { "id": "uk-age-wave-2026-09-09-market-findings", "title": "UK age-assurance market: instruments, interests and limits", "description": "Prepared 9 September 2026 (America/Chicago). Candidate acquisition only; case-r10 unchanged.", "html": "Prepared 9 September 2026 (America/Chicago). Candidate acquisition only; case-r10 unchanged.
\nThe strongest new structure is not a speculative relationship between national digital ID and age checks. It is a documented commercial certification business, its control over the separately incorporated nonprofit scheme, its principal's standards-editing role, and a funded interoperability project's explicit proposed commercial and admission rules. The source record supports these distinct relationships. It does not establish regulatory capture, compulsory use of ACCS, a monopoly, or deployment of every proposed AgeAware feature.
\nThe existing node accs combines a brand, scheme and service provider. The current ACCS governance page identifies Age Check Certification Scheme Ltd (11493870), a nonprofit company limited by guarantee, and says audit, certification and test purchasing are supplied under a Service Level Agreement by AVID Certification Services Limited (14865982). These are not interchangeable legal persons.
\nCompanies House's current scheme-company PSC page lists AVID as the active corporate controller, notified 1 June 2023, with director appointment/removal rights. AVID's PSC page lists Tony Iain Allen, notified 12 May 2023, with more than 50% but less than 75% of shares and voting rights and director appointment/removal rights; Nicholas John Bate, notified 14 July 2023, holds more than 25% but no more than 50% of shares and votes. These are disclosed control bands, not exact stakes. The register also contains an earlier ceased Nick Bate entry. We have not inspected the change filings; do not count that entry as another independent owner or turn it into an unexplained beneficial-owner departure.
\nThis earns an ownership/control path: Allen -> AVID -> scheme-company directors. It does not establish Allen personally decided any named certificate, Australian trial result, standard clause or regulatory disposition. The SLA's payment, indemnity, renewal and termination provisions remain unacquired.
\nSources: market-avid-psc.html, current active entries; market-scheme-psc.html, active AVID entry; market-accs-governance.html, Financial support and legal footer. Exact originals, URLs and SHA-256 are in market-captures.jsonl and market-candidate.json.
\nThe 2026 General Scheme Rules are an actual published operating instrument. Sections 6.24-6.27 and 7 (PDF20) put the decision with a certification officer employed by the assessment body and separate from the evaluator; the decision cannot be outsourced. Section14 (PDF33) sets individually supplied fee schedules, invoices due within 28 days and annual fee review. Failure to pay is a listed nonconformity, alongside technical and audit failures (10.8, PDF25). Withdrawal entails loss of the certificate, registry entry and mark, plus notifications to relevant regulators (10.9). Recertification is at most two years apart (4.7, PDF10). The governance page says certification fees are the main income source.
\nThe same rules provide meaningful limits: an independent appeal panel can uphold, remit or revoke decisions (11.18-11.20, PDF29); an impartiality panel includes the chief executive and at least two independent members (12.4, PDF30). The assessment body must not design or provide the certified product/service, clients must not manage certification, association membership cannot be a condition, and decision personnel face a two-year prior-client/consultancy restriction (12.6-12.11, PDF30-31). These are stated safeguards, not independently tested implementation. No executed customer agreement, individual invoice or certifier profit allocation was obtained.
\nThe source is headed ACCS0:2026 but retains an old 0:2020 header fragment. We treat the downloaded 2026 instrument as current-source evidence, not proof these exact clauses governed a 2024 or July2025 decision. Its own numerical classifications are scheme criteria, not automatically Ofcom's thresholds.
\nBSI's event page for 23 October2025 identifies Tony Allen as ACCS Executive Director, lead editor of ISO/IEC27566-1 and chair of BSI panel IST/33/5/5 (Tony Allen speaker biography). The standard's commercial certification is now advertised on ACCS's own site. BSI's 3 February2026 announcement describes UK editorial and committee input and a technology-neutral framework that does not always require full identity verification.
\nThis is a specific overlap of commercial and standard-setting roles. It does not mean a sole author, unilateral standards approval or a statutory vendor appointment. The ISO record lists publication in December2025; its ordinary public GET returned403. BSI's official announcement was the single substitute retained. The full paid standard was not purchased/read. Do not backdate the final standard to the July2025 UK implementation deadline.
\nVerifyMy's current product page names its legal operator as VerifyMy Limited, 12050874. It sells email-age estimation using an email, ownership checks and digital-footprint analysis, with background account-creation/checkout and explicit age-gate variants. It reports ACCS certification of its platform under ACCS1:2025 incorporating ISO27566-1. This is a provider's certification claim, not independent inspection of its certificate or algorithm. It supplies no price, profit, customer list or assignment to Reddit.
\nPolicy lane separately owns the Ofcom January2025 disposition crediting AVPA/VerifyMy evidence for adding email estimation. That is actual method-level uptake; it is not evidence VerifyMy supplies every platform using the method. Ofcom's refusal to make independent certification compulsory or automatically sufficient must accompany any combined narrative.
\nThe acquired original is the AgeAware Specification Consultation Document, WP1 Business Requirements, version1.0, dated5 August2024, 25pages. Cover authors are Iain Corby, Alastair Graham (design authority pro tem) and Ben Gower (technical architect). PDF2's document-control table names T Allen of ACCS as the required approver and R Kiddle of ACCS as reviewer. The Action column identifies Approver/Reviewer; completion Date cells are blank and no signed approval is shown. It is an intended approval chain, not evidence that Allen signed or approved the final implementation. AVPA, Yoti, AgeChecked and VerifyMy are named proposed reviewers. The January9,2025 euCONSENT account subsequently says the three providers worked on detailed design; these are separate evidence states.
\nThe proposal is sponsored by Safe Online and delivered by the Belgian nonprofit euCONSENT ASBL (cover and1.9, PDF3). It builds on an earlier European Commission-funded consortium project; do not backdate ASBL as the original 2021 grant recipient. The reported project-level funding purpose is established. The executed grant agreement, amount, subproject/vendor allocation, ultimate donor allocation and clearance/acceptance rights remain unacquired; a separately encountered $2million figure belongs to a ten-project portfolio and is not assigned here.
\nIts proposed operating provisions are concrete:\n* The ASBL board sets admission/certification requirements, applies a fit-and-proper test and is elected by members, then limited to original project participants (5.1, PDF9-10).\n* The secretariat collects cost-recovery licensing fees; a commercial steering group may recommend restrictions on below-cost or negative pricing, expressly subject to competition law (5.3/5.5, PDF10-11).\n* Usage tallies support provider billing and cross-charging while withholding individual commercial prices/commissions from the tally system. AgeAware controls tally access and can pause/remove participants or blacklist tokens (6.4-6.5, PDF15-16).\n* Limited funds motivate negotiated initial exclusivity for core-component suppliers and a proposed three-year exclusive independent commercial scheme, followed by contestability (14.3-14.5, PDF23).\n* Proposed contracts limit third-party and nonprofit liability; further provisions need committee input and board approval (15.2, PDF24).
\nThese provisions locate design-stage discretion and intended revenue, not actual invoices, procurement awards or legally tested competition effects. They also state rival public-interest purposes: repeatable low-cost checks, provider choice, anonymity, security and long-run competition. Those objectives and the privacy design have not been independently technically audited in this pass.
\nJanuary2025's project update reports completed issuer/key/relying-party/tally services, with anonymisation/app and scaling still in progress. Current AVPA text says development continued until2025 and that euCONSENT's demonstration mission is fulfilled with private-sector alternatives available. Thus no candidate claims AgeAware currently operates as a mandatory gate. Its design also included regulatory sandboxes and Australia trial consideration; an offered/tendered solution is not an award.
\nThe 2024 draft naming an ACCS approver and using ACCS as an example auditor creates a precise question when read beside the 2026 scheme's prohibition on designing what it certifies. It is not evidence of a breach: dates differ, approval was not shown completed, the ultimate delivered product/assessor is unknown, and independent assessment could have been arranged. The missing records are the final approved specification, actual implementation/procurement contracts and assessor appointment, the applicable dated impartiality assessment/recusal record, and any final certification decision.
\nAll retained successful originals have ordinary public GET receipts, timestamps and hashes. HTML text derivatives remove script/style content; PDF derivatives use pypdf and page markers. See each candidate source's reading_extent: this is scoped substantive reading, not an assertion that every navigation item, technical security requirement or filing history was reviewed.
\nClosed access branches: the ISO original page returned403; official BSI was used once instead. The separate ACCS impartiality-page web route failed; the accessible full scheme rules supplied the operative provisions. A search for a directly accessible EU executed grant did not yield the award instrument; the captured recipient/trade-association funding acknowledgements remain attributed. No current acquirer roster, national-ID dependency, MindGeek continuity or investor speculation was imported.
\nHigh-value records still with identifiable custodians: AVID/scheme SLA and current certified-client contract template; dated impartiality risk register and relevant determinations; euCONSENT/Safe Online grant schedule and acceptance conditions; final post-consultation AgeAware specification and signed approval table; actual exclusivity award, fee schedules and operating/termination history. Public interest can be served by ordinary publication of these instruments; this lane performed no outreach.
\nKeep accs as a historical umbrella and add exact legal nodes; do not duplicate the existing A$3,830,277 Australia award as new money or retarget older ambiguous ACCS assertions without the original named entity. Preserve the original consortium/ASBL chronology in euconsent. Yoti and AgeChecked here are credited provider brands, not newly researched legal subsidiaries. Reported self-certification, proposed approval, actual published rules and registered control must remain distinct. Current national digital-ID policy is outside this lane and is not a prerequisite for any relationship claimed here.
", "path": "uk-age-wave-2026-09-09/market-findings.md", "sourceIds": [ "w7-market-s-avid-psc", "w7-market-s-scheme-psc", "w7-market-s-governance", "w7-market-s-rules", "w7-market-s-bsi-roles", "w7-market-s-bsi-standard", "w7-market-s-iso-service", "w7-market-s-ageaware-spec", "w7-market-s-ageaware-update", "w7-market-s-avpa-project", "w7-market-s-verifymy" ], "markdown": "Prepared 9 September 2026 (America/Chicago). Candidate acquisition only; case-r10 unchanged.\n\n## What this pass adds\n\nThe strongest new structure is not a speculative relationship between national digital ID and age checks. It is a documented commercial certification business, its control over the separately incorporated nonprofit scheme, its principal's standards-editing role, and a funded interoperability project's explicit proposed commercial and admission rules. The source record supports these distinct relationships. It does not establish regulatory capture, compulsory use of ACCS, a monopoly, or deployment of every proposed AgeAware feature.\n\n## 1. The legal control arrow matters\n\nThe existing node accs combines a brand, scheme and service provider. The current ACCS governance page identifies Age Check Certification Scheme Ltd (11493870), a nonprofit company limited by guarantee, and says audit, certification and test purchasing are supplied under a Service Level Agreement by AVID Certification Services Limited (14865982). These are not interchangeable legal persons.\n\nCompanies House's current scheme-company PSC page lists AVID as the active corporate controller, notified 1 June 2023, with director appointment/removal rights. AVID's PSC page lists Tony Iain Allen, notified 12 May 2023, with more than 50% but less than 75% of shares and voting rights and director appointment/removal rights; Nicholas John Bate, notified 14 July 2023, holds more than 25% but no more than 50% of shares and votes. These are disclosed control bands, not exact stakes. The register also contains an earlier ceased Nick Bate entry. We have not inspected the change filings; do not count that entry as another independent owner or turn it into an unexplained beneficial-owner departure.\n\nThis earns an ownership/control path: Allen -> AVID -> scheme-company directors. It does not establish Allen personally decided any named certificate, Australian trial result, standard clause or regulatory disposition. The SLA's payment, indemnity, renewal and termination provisions remain unacquired.\n\nSources: market-avid-psc.html, current active entries; market-scheme-psc.html, active AVID entry; market-accs-governance.html, Financial support and legal footer. Exact originals, URLs and SHA-256 are in market-captures.jsonl and market-candidate.json.\n\n## 2. Certification creates repeat commercial relationships, with safeguards\n\nThe 2026 General Scheme Rules are an actual published operating instrument. Sections 6.24-6.27 and 7 (PDF20) put the decision with a certification officer employed by the assessment body and separate from the evaluator; the decision cannot be outsourced. Section14 (PDF33) sets individually supplied fee schedules, invoices due within 28 days and annual fee review. Failure to pay is a listed nonconformity, alongside technical and audit failures (10.8, PDF25). Withdrawal entails loss of the certificate, registry entry and mark, plus notifications to relevant regulators (10.9). Recertification is at most two years apart (4.7, PDF10). The governance page says certification fees are the main income source.\n\nThe same rules provide meaningful limits: an independent appeal panel can uphold, remit or revoke decisions (11.18-11.20, PDF29); an impartiality panel includes the chief executive and at least two independent members (12.4, PDF30). The assessment body must not design or provide the certified product/service, clients must not manage certification, association membership cannot be a condition, and decision personnel face a two-year prior-client/consultancy restriction (12.6-12.11, PDF30-31). These are stated safeguards, not independently tested implementation. No executed customer agreement, individual invoice or certifier profit allocation was obtained.\n\nThe source is headed ACCS0:2026 but retains an old 0:2020 header fragment. We treat the downloaded 2026 instrument as current-source evidence, not proof these exact clauses governed a 2024 or July2025 decision. Its own numerical classifications are scheme criteria, not automatically Ofcom's thresholds.\n\n## 3. A standards editor is also a certification-business principal\n\nBSI's event page for 23 October2025 identifies Tony Allen as ACCS Executive Director, lead editor of ISO/IEC27566-1 and chair of BSI panel IST/33/5/5 (Tony Allen speaker biography). The standard's commercial certification is now advertised on ACCS's own site. BSI's 3 February2026 announcement describes UK editorial and committee input and a technology-neutral framework that does not always require full identity verification.\n\nThis is a specific overlap of commercial and standard-setting roles. It does not mean a sole author, unilateral standards approval or a statutory vendor appointment. The ISO record lists publication in December2025; its ordinary public GET returned403. BSI's official announcement was the single substitute retained. The full paid standard was not purchased/read. Do not backdate the final standard to the July2025 UK implementation deadline.\n\nVerifyMy's current product page names its legal operator as VerifyMy Limited, 12050874. It sells email-age estimation using an email, ownership checks and digital-footprint analysis, with background account-creation/checkout and explicit age-gate variants. It reports ACCS certification of its platform under ACCS1:2025 incorporating ISO27566-1. This is a provider's certification claim, not independent inspection of its certificate or algorithm. It supplies no price, profit, customer list or assignment to Reddit.\n\nPolicy lane separately owns the Ofcom January2025 disposition crediting AVPA/VerifyMy evidence for adding email estimation. That is actual method-level uptake; it is not evidence VerifyMy supplies every platform using the method. Ofcom's refusal to make independent certification compulsory or automatically sufficient must accompany any combined narrative.\n\n## 4. AgeAware proposed a market constitution, not just a token format\n\nThe acquired original is the AgeAware Specification Consultation Document, WP1 Business Requirements, version1.0, dated5 August2024, 25pages. Cover authors are Iain Corby, Alastair Graham (design authority pro tem) and Ben Gower (technical architect). PDF2's document-control table names T Allen of ACCS as the required approver and R Kiddle of ACCS as reviewer. The Action column identifies Approver/Reviewer; completion Date cells are blank and no signed approval is shown. It is an intended approval chain, not evidence that Allen signed or approved the final implementation. AVPA, Yoti, AgeChecked and VerifyMy are named proposed reviewers. The January9,2025 euCONSENT account subsequently says the three providers worked on detailed design; these are separate evidence states.\n\nThe proposal is sponsored by Safe Online and delivered by the Belgian nonprofit euCONSENT ASBL (cover and1.9, PDF3). It builds on an earlier European Commission-funded consortium project; do not backdate ASBL as the original 2021 grant recipient. The reported project-level funding purpose is established. The executed grant agreement, amount, subproject/vendor allocation, ultimate donor allocation and clearance/acceptance rights remain unacquired; a separately encountered $2million figure belongs to a ten-project portfolio and is not assigned here.\n\nIts proposed operating provisions are concrete:\n* The ASBL board sets admission/certification requirements, applies a fit-and-proper test and is elected by members, then limited to original project participants (5.1, PDF9-10).\n* The secretariat collects cost-recovery licensing fees; a commercial steering group may recommend restrictions on below-cost or negative pricing, expressly subject to competition law (5.3/5.5, PDF10-11).\n* Usage tallies support provider billing and cross-charging while withholding individual commercial prices/commissions from the tally system. AgeAware controls tally access and can pause/remove participants or blacklist tokens (6.4-6.5, PDF15-16).\n* Limited funds motivate negotiated initial exclusivity for core-component suppliers and a proposed three-year exclusive independent commercial scheme, followed by contestability (14.3-14.5, PDF23).\n* Proposed contracts limit third-party and nonprofit liability; further provisions need committee input and board approval (15.2, PDF24).\n\nThese provisions locate design-stage discretion and intended revenue, not actual invoices, procurement awards or legally tested competition effects. They also state rival public-interest purposes: repeatable low-cost checks, provider choice, anonymity, security and long-run competition. Those objectives and the privacy design have not been independently technically audited in this pass.\n\nJanuary2025's project update reports completed issuer/key/relying-party/tally services, with anonymisation/app and scaling still in progress. Current AVPA text says development continued until2025 and that euCONSENT's demonstration mission is fulfilled with private-sector alternatives available. Thus no candidate claims AgeAware currently operates as a mandatory gate. Its design also included regulatory sandboxes and Australia trial consideration; an offered/tendered solution is not an award.\n\n## The most useful unresolved junction\n\nThe 2024 draft naming an ACCS approver and using ACCS as an example auditor creates a precise question when read beside the 2026 scheme's prohibition on designing what it certifies. It is not evidence of a breach: dates differ, approval was not shown completed, the ultimate delivered product/assessor is unknown, and independent assessment could have been arranged. The missing records are the final approved specification, actual implementation/procurement contracts and assessor appointment, the applicable dated impartiality assessment/recusal record, and any final certification decision.\n\n## Source custody, boundaries and next records\n\nAll retained successful originals have ordinary public GET receipts, timestamps and hashes. HTML text derivatives remove script/style content; PDF derivatives use pypdf and page markers. See each candidate source's reading_extent: this is scoped substantive reading, not an assertion that every navigation item, technical security requirement or filing history was reviewed.\n\nClosed access branches: the ISO original page returned403; official BSI was used once instead. The separate ACCS impartiality-page web route failed; the accessible full scheme rules supplied the operative provisions. A search for a directly accessible EU executed grant did not yield the award instrument; the captured recipient/trade-association funding acknowledgements remain attributed. No current acquirer roster, national-ID dependency, MindGeek continuity or investor speculation was imported.\n\nHigh-value records still with identifiable custodians: AVID/scheme SLA and current certified-client contract template; dated impartiality risk register and relevant determinations; euCONSENT/Safe Online grant schedule and acceptance conditions; final post-consultation AgeAware specification and signed approval table; actual exclusivity award, fee schedules and operating/termination history. Public interest can be served by ordinary publication of these instruments; this lane performed no outreach.\n\n## Integration instructions\n\nKeep accs as a historical umbrella and add exact legal nodes; do not duplicate the existing A$3,830,277 Australia award as new money or retarget older ambiguous ACCS assertions without the original named entity. Preserve the original consortium/ASBL chronology in euconsent. Yoti and AgeChecked here are credited provider brands, not newly researched legal subsidiaries. Reported self-certification, proposed approval, actual published rules and registered control must remain distinct. Current national digital-ID policy is outside this lane and is not a prerequisite for any relationship claimed here.\n", "file": "research/uk-age-wave-2026-09-09-market-findings.md" }, { "id": "uk-age-wave-2026-09-09-policy-findings", "title": "UK age assurance: statutory chain and consequential choices", "description": "As observed 9 September 2026, the Online Safety Act duties and Ofcom's implementation require highly effective age assurance (HEAA) for defined services and harmful-content risks. They do not create a universal identification requirement for everyone using the internet. The consequential influence record is unusually explicit: Ofcom credited AVPA and VerifyM", "html": "As observed 9 September 2026, the Online Safety Act duties and Ofcom's implementation require highly effective age assurance (HEAA) for defined services and harmful-content risks. They do not create a universal identification requirement for everyone using the internet. The consequential influence record is unusually explicit: Ofcom credited AVPA and VerifyMy evidence when adding email-based age estimation to its examples of methods capable of HEAA. That is acceptance of a method-level contribution, not approval of the company, procurement, demonstrated deployment performance or evidence of a payment.
\nThe [Online Safety Act 2023](http[local research file] assented to on 26 October 2023, defines a child as under 18 (section 236, PDF 218/printed 204). Section 230 (PDF 212/printed 198) distinguishes verifying exact age from estimating age or age-range; a bare self-declaration is neither. Identity-document matching is one possible method. Facial age estimation is another. The final Part 5 guidance explains that open banking can pass an over-18 result without passing date of birth or other information to the relying service (paragraph 4.12, PDF 17). A token can communicate a completed age check; it is not itself the underlying age-assurance method.
\n| Decision or duty | \nPrecise scope and timing | \n
|---|---|
| OSA section 81, [current official text](http[local research file] | \nPart 5 provider-published pornography: use verification/estimation so children are not normally able to encounter it; method and deployment must be highly effective. Also record methods/privacy consideration and publish a summary. Official commencement annotation: 17 January 2025, SI 2024/1333 regulation 2(1)(a). | \n
| OSA section 12, [current official text](http[local research file] | \nUser-to-user services likely to be accessed by children: proportionate protection duties. Sections 12(4)–(6) require HEAA for identified primary-priority content unless that kind is prohibited for all users by the terms. A prohibition does not remove the broader protection duties. This covers more than pornography: section 61 includes suicide, self-injury and eating-disorder encouragement/instructions. | \n
| OSA sections 35–36 | \nAll Part 3 user-to-user and search services assess children's access. The statutory test examines actual use/likely attraction, not just the intended audience. Claiming children cannot access requires verification/estimation that actually excludes them. Initial access-assessment deadline: 16 April 2025. | \n
| OSA section 82 and Schedule 4 paragraph 12 | \nOfcom must give Part 5 guidance with examples and compliance principles; the children's codes must recommend HEAA for the section 12(3)(a) duty and have regard to that guidance. Statutory guidance obligation is distinct from certification of vendors. | \n
| Final Protection of Children Code | \nStatement/guidance 24 April 2025; issued code cover dated 4 July 2025 and in force 25 July 2025. Initial children's risk-assessment deadline 24 July 2025. See [issued user-to-user code](http[local research file] | \n
The section 12 commencement annotation says 10 January 2024. That should not be confused with the operational compliance deadline after the statutory assessment/code process: Ofcom's dated implementation table and enforcement case expressly identify 25 July 2025 for these children's safety duties. The [current implementation dates table](http[local research file] preserves the sequence.
\nThe code's recommendations are a statutory route to compliance, not every word a freestanding legislative prohibition. OSA section 49(1) treats relevant recommended measures as compliance;49(5)–(6) accommodates alternative measures subject to scrutiny and privacy/expression safeguards. Section 50(1) says failure to follow a code provision does not itself create court/tribunal liability. The underlying duties remain mandatory.
\nThe issued code, PDF 20–27 (printed 21–28), distinguishes service-wide adult access controls for services principally hosting primary-priority/priority harmful content (PCU B2/B3) from content/feed protections (B4–B7). For B4–B7 it allows appropriate restrictions to remain on users who do not elect to prove adulthood, rather than demanding a document from every user. Each measure has its own content/risk/functionality conditions; B7 excludes bullying from its priority-content trigger. Search services have children's-access and protection duties, but not the same HEAA duty as user-to-user services to prevent access to pornography. Ofcom's July 2026 report paragraph 4.34, PDF 31, makes that distinction expressly.
\nThe [16 January 2025 statement](http[local research file] provides the decision record. Its currently linked attachment carries a 16 May 2025 listing date, while the PDF cover says16 January 2025; no new decision date is inferred from the media listing or URL query.
\nParagraphs 3.85–3.87, PDF 27: AVPA/VerifyMy evidence supported adding email-based estimation, conditional on effective implementation. Paragraphs 3.103–3.105, PDF 30–31: token references were added, retaining responsibility for the underlying check and sharing process. The preceding response summary, paragraphs 3.66–3.68 and footnote 94 on PDF 24, attributes the token recommendation to 5Rights, Yoti and the ACT App Association. Paragraph 3.108, PDF 31, and 3.350–3.353, PDF 76–77: certification can evidence compliance but is neither compulsory nor automatic; Ofcom did not adopt Yoti's request to mandate independent testing/certification. Paragraphs 3.243–3.259, PDF 54–57: requested numerical thresholds were declined at that stage, citing inadequate comparable evidence/testing and market-development concerns. Paragraphs 3.277–3.290, PDF 62–65: 5Rights' separate privacy-criterion proposal was declined because existing data-protection law applied; design/minimisation guidance was strengthened. None of these outcomes proves a private bargain or sole authorship.
\nThe [April 2025 children's statement Volume 4](http[local research file] paragraphs 13.343–13.357, PDF 204–207, separately handles behavioural age inference. It rejected adding inference to the HEAA examples: evidence of efficacy was insufficient and inference needs a period of on-service activity. That is incompatible with keeping children out of a harmful-content service from entry. For services children may use, Ofcom left an evidential route for a wider process, requiring proof of accuracy, timing, protection during the inference interval, reliability and fairness. Email-based estimation from other established service uses is not interchangeable with this behavioural on-platform inference.
\nA short official [FOI response2053425](http[local research file] dated 26 August 2025 despite its October folder, is particularly useful. PDF 2 Q3 says Ofcom does not certify/approve specific vendors and had no vendor contracts or MOUs. Q4 says Ofcom has no access to data users submit to those vendors; it expressly does not answer for other authorities. These are dated institutional statements, not a current independent audit of data flows or proof that platforms and vendors cannot access data. Q5 rejects a general internet-wide age-check requirement.
\nThe [July 2026 statutory report](http[local research file] cover 15 July 2026, describes early findings predominantly from the first six months after July 2025 duties, not final causal proof. Paragraphs 5.30–5.35, PDF 42–43, retain the inference concerns; its sampled service evidence does not justify claiming every listed method works in every deployment. The 2 September resource follows these identified implementation shortcomings.
\nThe [Xgroovy case update](http[local research file] case CW/01318/09/25, distinguishes a 16 June 2026 provisional notice from a 3 September confirmation decision, announced 4 September. Ofcom reports a £700,000 section 12 penalty and a separate £30,000 section 102(8)(a) information-request penalty, plus a conditional daily information penalty. The underlying non-confidential confirmation decision is still promised for later publication. Record these as reported imposed penalties, not money collected or a court judgment. The case identifies the provider by service, not a resolved legal-company name.
\nThe strongest rival to a vendor-capture reading is disclosed technical consultation within a technology-neutral regime: one commercial method was accepted, while commercial certification and threshold requests were not adopted wholesale. The missing discriminating records are the complete technical evidence submitted for email-based estimation and independent test/evaluation records showing why it met each criterion; any actual procurement/payment or contractual link needs its own evidence. The market lane owns the vendor counterpart route; the advocacy lane owns original 5Rights submissions and should reuse this statement source for disposition.
\nFurther public acquisition stops at this coherent packet. Reopen the enforcement branch when the non-confidential Xgroovy confirmation decision is published, and the prospective 16+ branch when the actual assessment, legislative instrument and commencement arrangements appear. Nothing here assesses the separate national digital-ID project's status; that belongs to root's currentness check. Do not turn its cancellation or continuation into a conclusion about OSA age assurance or the separate DVS framework.
\nOriginal bytes retained: enacted Act and current section 12/81 HTML, with hashes. Ordinary Ofcom GETs returned 403; the public web reader was the successful substitute. Ofcom captures are scoped derivative reader outputs with extract hashes only—no original PDF/HTML hash is claimed. Source-level locators and actual read extents are in policy-candidate.json; manifests distinguish error-response hashes from originals. The old /illegal-and-harmful-content/age-assurance path returned 404, while the actual current linked /protecting-children/age-assurance route was read. No paid access, outreach, restricted-route workaround or canonical edits.
", "path": "uk-age-wave-2026-09-09/policy-findings.md", "sourceIds": [ "w7-policy-s-osa-enacted", "w7-policy-s-s12-current", "w7-policy-s-s81-current", "w7-policy-s-statement2025", "w7-policy-s-part3", "w7-policy-s-part5", "w7-policy-s-child-code", "w7-policy-s-children-decisions", "w7-policy-s-report2026", "w7-policy-s-current-page", "w7-policy-s-vet-vendor", "w7-policy-s-foi2053425", "w7-policy-s-xgroovy", "w7-policy-s-dates" ], "markdown": "As observed 9 September 2026, the Online Safety Act duties and Ofcom's implementation require highly effective age assurance (HEAA) for defined services and harmful-content risks. They do not create a universal identification requirement for everyone using the internet. The consequential influence record is unusually explicit: Ofcom credited AVPA and VerifyMy evidence when adding email-based age estimation to its examples of methods capable of HEAA. That is acceptance of a method-level contribution, not approval of the company, procurement, demonstrated deployment performance or evidence of a payment.\n\n## The legal chain\n\nThe Online Safety Act 2023 (see source references) assented to on 26 October 2023, defines a child as under 18 (section 236, PDF 218/printed 204). Section 230 (PDF 212/printed 198) distinguishes verifying exact age from estimating age or age-range; a bare self-declaration is neither. Identity-document matching is one possible method. Facial age estimation is another. The final Part 5 guidance explains that open banking can pass an over-18 result without passing date of birth or other information to the relying service (paragraph 4.12, PDF 17). A token can communicate a completed age check; it is not itself the underlying age-assurance method.\n\n| Decision or duty | Precise scope and timing |\n| --- | --- |\n| OSA section 81, current official text (see source references) | Part 5 provider-published pornography: use verification/estimation so children are not normally able to encounter it; method and deployment must be highly effective. Also record methods/privacy consideration and publish a summary. Official commencement annotation: 17 January 2025, SI 2024/1333 regulation 2(1)(a). |\n| OSA section 12, current official text (see source references) | User-to-user services likely to be accessed by children: proportionate protection duties. Sections 12(4)–(6) require HEAA for identified primary-priority content unless that kind is prohibited for all users by the terms. A prohibition does not remove the broader protection duties. This covers more than pornography: section 61 includes suicide, self-injury and eating-disorder encouragement/instructions. |\n| OSA sections 35–36 | All Part 3 user-to-user and search services assess children's access. The statutory test examines actual use/likely attraction, not just the intended audience. Claiming children cannot access requires verification/estimation that actually excludes them. Initial access-assessment deadline: 16 April 2025. |\n| OSA section 82 and Schedule 4 paragraph 12 | Ofcom must give Part 5 guidance with examples and compliance principles; the children's codes must recommend HEAA for the section 12(3)(a) duty and have regard to that guidance. Statutory guidance obligation is distinct from certification of vendors. |\n| Final Protection of Children Code | Statement/guidance 24 April 2025; issued code cover dated 4 July 2025 and in force 25 July 2025. Initial children's risk-assessment deadline 24 July 2025. See issued user-to-user code (see source references) |\n\nThe section 12 commencement annotation says 10 January 2024. That should not be confused with the operational compliance deadline after the statutory assessment/code process: Ofcom's dated implementation table and enforcement case expressly identify 25 July 2025 for these children's safety duties. The current implementation dates table (see source references) preserves the sequence.\n\nThe code's recommendations are a statutory route to compliance, not every word a freestanding legislative prohibition. OSA section 49(1) treats relevant recommended measures as compliance;49(5)–(6) accommodates alternative measures subject to scrutiny and privacy/expression safeguards. Section 50(1) says failure to follow a code provision does not itself create court/tribunal liability. The underlying duties remain mandatory.\n\nThe issued code, PDF 20–27 (printed 21–28), distinguishes service-wide adult access controls for services principally hosting primary-priority/priority harmful content (PCU B2/B3) from content/feed protections (B4–B7). For B4–B7 it allows appropriate restrictions to remain on users who do not elect to prove adulthood, rather than demanding a document from every user. Each measure has its own content/risk/functionality conditions; B7 excludes bullying from its priority-content trigger. Search services have children's-access and protection duties, but not the same HEAA duty as user-to-user services to prevent access to pornography. Ofcom's July 2026 report paragraph 4.34, PDF 31, makes that distinction expressly.\n\n## What consultation actually changed\n\nThe 16 January 2025 statement (see source references) provides the decision record. Its currently linked attachment carries a 16 May 2025 listing date, while the PDF cover says16 January 2025; no new decision date is inferred from the media listing or URL query.\n\nParagraphs 3.85–3.87, PDF 27: AVPA/VerifyMy evidence supported adding email-based estimation, conditional on effective implementation. Paragraphs 3.103–3.105, PDF 30–31: token references were added, retaining responsibility for the underlying check and sharing process. The preceding response summary, paragraphs 3.66–3.68 and footnote 94 on PDF 24, attributes the token recommendation to 5Rights, Yoti and the ACT App Association. Paragraph 3.108, PDF 31, and 3.350–3.353, PDF 76–77: certification can evidence compliance but is neither compulsory nor automatic; Ofcom did not adopt Yoti's request to mandate independent testing/certification. Paragraphs 3.243–3.259, PDF 54–57: requested numerical thresholds were declined at that stage, citing inadequate comparable evidence/testing and market-development concerns. Paragraphs 3.277–3.290, PDF 62–65: 5Rights' separate privacy-criterion proposal was declined because existing data-protection law applied; design/minimisation guidance was strengthened. None of these outcomes proves a private bargain or sole authorship.\n\nThe April 2025 children's statement Volume 4 (see source references) paragraphs 13.343–13.357, PDF 204–207, separately handles behavioural age inference. It rejected adding inference to the HEAA examples: evidence of efficacy was insufficient and inference needs a period of on-service activity. That is incompatible with keeping children out of a harmful-content service from entry. For services children may use, Ofcom left an evidential route for a wider process, requiring proof of accuracy, timing, protection during the inference interval, reliability and fairness. Email-based estimation from other established service uses is not interchangeable with this behavioural on-platform inference.\n\n## Currentness, certification and enforcement\n\nThe 2 September 2026 update.\n\nA short official FOI response2053425 (see source references) dated 26 August 2025 despite its October folder, is particularly useful. PDF 2 Q3 says Ofcom does not certify/approve specific vendors and had no vendor contracts or MOUs. Q4 says Ofcom has no access to data users submit to those vendors; it expressly does not answer for other authorities. These are dated institutional statements, not a current independent audit of data flows or proof that platforms and vendors cannot access data. Q5 rejects a general internet-wide age-check requirement.\n\nThe July 2026 statutory report (see source references) cover 15 July 2026, describes early findings predominantly from the first six months after July 2025 duties, not final causal proof. Paragraphs 5.30–5.35, PDF 42–43, retain the inference concerns; its sampled service evidence does not justify claiming every listed method works in every deployment. The 2 September resource follows these identified implementation shortcomings.\n\nThe Xgroovy case update (see source references) case CW/01318/09/25, distinguishes a 16 June 2026 provisional notice from a 3 September confirmation decision, announced 4 September. Ofcom reports a £700,000 section 12 penalty and a separate £30,000 section 102(8)(a) information-request penalty, plus a conditional daily information penalty. The underlying non-confidential confirmation decision is still promised for later publication. Record these as reported imposed penalties, not money collected or a court judgment. The case identifies the provider by service, not a resolved legal-company name.\n\n## Boundaries and missing discriminating records\n\nThe strongest rival to a vendor-capture reading is disclosed technical consultation within a technology-neutral regime: one commercial method was accepted, while commercial certification and threshold requests were not adopted wholesale. The missing discriminating records are the complete technical evidence submitted for email-based estimation and independent test/evaluation records showing why it met each criterion; any actual procurement/payment or contractual link needs its own evidence. The market lane owns the vendor counterpart route; the advocacy lane owns original 5Rights submissions and should reuse this statement source for disposition.\n\nFurther public acquisition stops at this coherent packet. Reopen the enforcement branch when the non-confidential Xgroovy confirmation decision is published, and the prospective 16+ branch when the actual assessment, legislative instrument and commencement arrangements appear. Nothing here assesses the separate national digital-ID project's status; that belongs to root's currentness check. Do not turn its cancellation or continuation into a conclusion about OSA age assurance or the separate DVS framework.\n\nOriginal bytes retained: enacted Act and current section 12/81 HTML, with hashes. Ordinary Ofcom GETs returned 403; the public web reader was the successful substitute. Ofcom captures are scoped derivative reader outputs with extract hashes only—no original PDF/HTML hash is claimed. Source-level locators and actual read extents are in policy-candidate.json; manifests distinguish error-response hashes from originals. The old /illegal-and-harmful-content/age-assurance path returned 404, while the actual current linked /protecting-children/age-assurance route was read. No paid access, outreach, restricted-route workaround or canonical edits.\n", "file": "research/uk-age-wave-2026-09-09-policy-findings.md" }, { "id": "wipo-payment-wave-2026-09-09-participant-authority", "title": "WIPO ALERT-PAY: participant confirmation and provider decision rights", "description": "9 September 2026. Bounded acquisition closed. Two PayPal policy originals were retained; Mastercard's complete public policy was read through the successful official reader after its ordinary GET failed. No report form, account, live list or payment operation was used. This is a scoped comparison of public rules, not confirmation of a signed WIPO-provider ag", "html": "9 September 2026. Bounded acquisition closed. Two PayPal policy originals were retained; Mastercard's complete public policy was read through the successful official reader after its ordinary GET failed. No report form, account, live list or payment operation was used. This is a scoped comparison of public rules, not confirmation of a signed WIPO-provider agreement.
\nThe acquired provider policies show that a notification can enter a provider-controlled investigation and contract process. They do not establish exactly how ALERT-PAY notifications are classified within that process. In particular, Mastercard's public rules distinguish stopping payments for an infringing product from terminating the merchant and from wider notice through MATCH. Neither an alert nor an infringing-product finding proves that all three occurred.
\nRoot retained the official [WIPO/ACE/18/40 presentation](http[local research file] whose [document index](http[local research file] records publication on 16 June 2026. This worker read original slides 4–7. Slide 4 names Mastercard and PayPal among payment service providers in the described development and says PSPs choose what action, if any, to take under their policies. Slide 5 describes WIPO checking formalities, administering notifications/review and aggregating lists. Slide 7 reports an anonymous six-rightsholder/two-PSP pilot; those numbers alone do not identify the two pilot firms.
\nThe finite participant-owned searches did not locate a PayPal or Mastercard ALERT-PAY announcement, signed participation terms, programme-specific challenge rule or correction agreement. That is a public-source gap, not evidence that either company did not participate. WIPO fee-payment pages, Mastercard's own website-content copyright terms, and unrelated industry initiatives were excluded. No identity was inferred from adjacent programmes, logos or illustrative application screens.
\nRoot's separate indexed-transcript review qualifies slide 6's test-purchase box: according to that reading, a purchase depends on the PSP and none was needed during the pilot. This worker's pinpoint attempt to reopen that transcript returned a zero-line page, so this is not an independently reread transcript claim here. Do not treat slide 6 as proof of a universal purchase requirement or a contradiction with the provider policy below. Root owns the reconciled programme chronology and transcript scope.
\n[Mastercard Anti-Piracy Policy](http[local research file] undated US public page observed 9 September 2026, numbered sections 1–4, allocates these roles:
\n| Stage | \nPublic policy allocation | \n
|---|---|
| Intake | \nLaw-enforcement evidence, or a rightsholder/authorized trade-association referral supported by specified evidence and attestations | \n
| Investigation | \nMastercard identifies the acquiring relationship and requires the acquirer to investigate/report: two business days for the law-enforcement route, five for the private-referral route; extensions are discretionary | \n
| Finding | \nAn affirmative finding requires ending Mastercard acceptance for the infringing product; a negative finding requires compelling evidence | \n
| Termination | \nIf the acquirer terminates the merchant, Mastercard requires MATCH listing where applicable, notifying acquirers of termination and its reason code | \n
| Enforcement | \nMastercard retains interpretation/enforcement authority and may restrict or sanction a noncompliant acquirer's membership | \n
Private referrals require ownership evidence, specific alleged products/locations, notice materials and payment-method evidence; a test transaction is optional. Submitters consent to disclosure and cooperation and may face indemnity requirements. The private-referral route includes reporting investigation results back to the rightsholder/association. Section 3 also addresses restricting cross-border acquiring where the product's legality differs by country.
\nThis is the provider's general policy, not an acquired ALERT-PAY addendum. WIPO's institutional status does not establish that its alert enters the law-enforcement route. Do not add Mastercard's two/five-business-day clocks to WIPO's notification clocks. A merits response, merchant termination and MATCH entry remain distinct. No MATCH deletion/restoration rule or completed appeal was acquired in this lane; that does not establish that none exists elsewhere.
\nThe policy's opening BRAM context also reserves discretion over transactions considered damaging to Mastercard's goodwill. That broader context should not be silently turned into an ALERT-PAY eligibility criterion. The inspected programme concerns counterfeit goods and subscription piracy; the underlying provider contract may have other grounds, whose application requires separate evidence.
\nThe [UK Infringement Report Policy](http[local research file] is labelled 15 July 2026; the [US page](http[local research file] is labelled 6 May 2021. Both retained originals give the same operative policy paragraph: PayPal may take appropriate action where necessary concerning claimed infringement and invites IP rights owners to submit its report. The UK page separately identifies PayPal UK Ltd, company 14741686. That locale disclosure does not identify the legal PSP party to WIPO's pilot or contract.
\nThese short pages do not state a fixed action deadline, evidential adjudication procedure, counter-notice timetable, automatic account termination or ALERT-PAY-specific correction rule. The blank report and broader account contracts were not opened. Absence from this paragraph is not absence of wider PayPal remedies. The different update dates cannot be treated as dates when ALERT-PAY was adopted or when the operative paragraph changed.
\nThe strongest supported interpretation is a notification channel feeding institutions that retain their own decision rights, with Mastercard's acquirer relationship adding another decision step. The strongest rival to a simple automatic cutoff model is an investigation that rejects the allegation, targets only the disputed product, or requires another response without terminating the merchant. The public rules permit that distinction; they do not reveal actual frequencies or outcomes. Conversely, retaining discretion and a negative-finding route does not prove that challenges succeed or that corrections propagate reliably.
\nCandidate relations for later review: WIPO to named PSPs, programme description only; rightsholder/authorized association to Mastercard, general referral right; Mastercard to acquirer, investigation/reporting and policy-enforcement duties; acquirer to infringing-product acceptance, conditional cessation; terminated merchant to MATCH, conditional subsequent notice; PayPal to IP rights owner, public reporting invitation. There is no acquired agreement edge binding either provider to act on every WIPO entry and no exercised payment-cutoff edge from these policies.
\nThe next revealing record is the actual participation agreement or a public provider explanation tying an ALERT-PAY notification to a particular internal route, together with a redacted completed case showing review, reasoned outcome, correction and any merchant/MATCH restoration. Likely holders are WIPO's Building Respect for IP Division and the relevant PSP, with an acquirer involved for Mastercard. This names a record family and accountable actors, not a claim that a particular unpublished document exists. No broader bank, cryptocurrency or donor branch was opened.
\nparticipant-captures.json records URLs, dates, source states, queries, originals, reader ancestry and the failed transcript pinpoint. The complete provider-reader output is retained in participant-policies.reader.txt. Mastercard original GET returned 403; its error hash is not policy custody, and neither that route nor a PDF variant was retried.
PayPal UK original SHA256: 9e0f6c2c882c7a8b3ec12d2bb65c0740297b97a01f06fb45bb1c3b15d309ba97. PayPal US original SHA256: ff31dfe0fe37189bb4321f48c0688eb5c9ce47239537dfd5d8c6560bf3305edd. Both complete short substantive policies were read from original HTML and through the primary reader; status/footer context was kept separate. No form or image was inspected.
Mastercard scope: entire substantive introductory policy and sections 1–4, reader lines 156–184, including both routes and the failure-to-comply section. Its linked forms, issuer process and general Rules/MATCH manuals were not read. WIPO scope: root's original slides 4–7 only, SHA256 e4afb3bdce8e4d4266a5d992abe117b95b38cf57ff27979ccf6d70af4b7e11a4, with a separate local extract; the presentation was not redownloaded. Root's programme packet supplies its fuller primary reading.
Acquisition closed for independent review and the combined forest. No case, atlas, current-work, forest, ZIP or visible UI changed.
", "path": "wipo-payment-wave-2026-09-09/participant-authority.md", "sourceIds": [ "w30-class-s-mc-antipiracy", "w30-class-s-paypal-uk", "w30-class-s-paypal-us" ], "markdown": "9 September 2026. Bounded acquisition closed. Two PayPal policy originals were retained; Mastercard's complete public policy was read through the successful official reader after its ordinary GET failed. No report form, account, live list or payment operation was used. This is a scoped comparison of public rules, not confirmation of a signed WIPO-provider agreement.\n\nThe acquired provider policies show that a notification can enter a provider-controlled investigation and contract process. They do not establish exactly how ALERT-PAY notifications are classified within that process. In particular, Mastercard's public rules distinguish stopping payments for an infringing product from terminating the merchant and from wider notice through MATCH. Neither an alert nor an infringing-product finding proves that all three occurred.\n\n## What is confirmed about programme identity\n\nRoot retained the official WIPO/ACE/18/40 presentation (see source references) whose document index (see source references) records publication on 16 June 2026. This worker read original slides 4–7. Slide 4 names Mastercard and PayPal among payment service providers in the described development and says PSPs choose what action, if any, to take under their policies. Slide 5 describes WIPO checking formalities, administering notifications/review and aggregating lists. Slide 7 reports an anonymous six-rightsholder/two-PSP pilot; those numbers alone do not identify the two pilot firms.\n\nThe finite participant-owned searches did not locate a PayPal or Mastercard ALERT-PAY announcement, signed participation terms, programme-specific challenge rule or correction agreement. That is a public-source gap, not evidence that either company did not participate. WIPO fee-payment pages, Mastercard's own website-content copyright terms, and unrelated industry initiatives were excluded. No identity was inferred from adjacent programmes, logos or illustrative application screens.\n\nRoot's separate indexed-transcript review qualifies slide 6's test-purchase box: according to that reading, a purchase depends on the PSP and none was needed during the pilot. This worker's pinpoint attempt to reopen that transcript returned a zero-line page, so this is not an independently reread transcript claim here. Do not treat slide 6 as proof of a universal purchase requirement or a contradiction with the provider policy below. Root owns the reconciled programme chronology and transcript scope.\n\n## Mastercard: investigation, product-level action, and conditional wider notice\n\nMastercard Anti-Piracy Policy (see source references) undated US public page observed 9 September 2026, numbered sections 1–4, allocates these roles:\n\n| Stage | Public policy allocation |\n| --- | --- |\n| Intake | Law-enforcement evidence, or a rightsholder/authorized trade-association referral supported by specified evidence and attestations |\n| Investigation | Mastercard identifies the acquiring relationship and requires the acquirer to investigate/report: two business days for the law-enforcement route, five for the private-referral route; extensions are discretionary |\n| Finding | An affirmative finding requires ending Mastercard acceptance for the infringing product; a negative finding requires compelling evidence |\n| Termination | If the acquirer terminates the merchant, Mastercard requires MATCH listing where applicable, notifying acquirers of termination and its reason code |\n| Enforcement | Mastercard retains interpretation/enforcement authority and may restrict or sanction a noncompliant acquirer's membership |\n\nPrivate referrals require ownership evidence, specific alleged products/locations, notice materials and payment-method evidence; a test transaction is optional. Submitters consent to disclosure and cooperation and may face indemnity requirements. The private-referral route includes reporting investigation results back to the rightsholder/association. Section 3 also addresses restricting cross-border acquiring where the product's legality differs by country.\n\nThis is the provider's general policy, not an acquired ALERT-PAY addendum. WIPO's institutional status does not establish that its alert enters the law-enforcement route. Do not add Mastercard's two/five-business-day clocks to WIPO's notification clocks. A merits response, merchant termination and MATCH entry remain distinct. No MATCH deletion/restoration rule or completed appeal was acquired in this lane; that does not establish that none exists elsewhere.\n\nThe policy's opening BRAM context also reserves discretion over transactions considered damaging to Mastercard's goodwill. That broader context should not be silently turned into an ALERT-PAY eligibility criterion. The inspected programme concerns counterfeit goods and subscription piracy; the underlying provider contract may have other grounds, whose application requires separate evidence.\n\n## PayPal: a short reporting policy, with distinct locale dates\n\nThe UK Infringement Report Policy (see source references) is labelled **15 July 2026**; the US page (see source references) is labelled **6 May 2021**. Both retained originals give the same operative policy paragraph: PayPal may take appropriate action where necessary concerning claimed infringement and invites IP rights owners to submit its report. The UK page separately identifies PayPal UK Ltd, company 14741686. That locale disclosure does not identify the legal PSP party to WIPO's pilot or contract.\n\nThese short pages do not state a fixed action deadline, evidential adjudication procedure, counter-notice timetable, automatic account termination or ALERT-PAY-specific correction rule. The blank report and broader account contracts were not opened. Absence from this paragraph is not absence of wider PayPal remedies. The different update dates cannot be treated as dates when ALERT-PAY was adopted or when the operative paragraph changed.\n\n## Evidence strength, useful relations and the next discriminator\n\nThe strongest supported interpretation is a notification channel feeding institutions that retain their own decision rights, with Mastercard's acquirer relationship adding another decision step. The strongest rival to a simple automatic cutoff model is an investigation that rejects the allegation, targets only the disputed product, or requires another response without terminating the merchant. The public rules permit that distinction; they do not reveal actual frequencies or outcomes. Conversely, retaining discretion and a negative-finding route does not prove that challenges succeed or that corrections propagate reliably.\n\nCandidate relations for later review: WIPO to named PSPs, programme description only; rightsholder/authorized association to Mastercard, general referral right; Mastercard to acquirer, investigation/reporting and policy-enforcement duties; acquirer to infringing-product acceptance, conditional cessation; terminated merchant to MATCH, conditional subsequent notice; PayPal to IP rights owner, public reporting invitation. There is no acquired agreement edge binding either provider to act on every WIPO entry and no exercised payment-cutoff edge from these policies.\n\nThe next revealing record is the actual participation agreement or a public provider explanation tying an ALERT-PAY notification to a particular internal route, together with a redacted completed case showing review, reasoned outcome, correction and any merchant/MATCH restoration. Likely holders are WIPO's Building Respect for IP Division and the relevant PSP, with an acquirer involved for Mastercard. This names a record family and accountable actors, not a claim that a particular unpublished document exists. No broader bank, cryptocurrency or donor branch was opened.\n\n## Custody and reading scope\n\n`participant-captures.json` records URLs, dates, source states, queries, originals, reader ancestry and the failed transcript pinpoint. The complete provider-reader output is retained in `participant-policies.reader.txt`. Mastercard original GET returned 403; its error hash is not policy custody, and neither that route nor a PDF variant was retried.\n\nPayPal UK original SHA256: `9e0f6c2c882c7a8b3ec12d2bb65c0740297b97a01f06fb45bb1c3b15d309ba97`. PayPal US original SHA256: `ff31dfe0fe37189bb4321f48c0688eb5c9ce47239537dfd5d8c6560bf3305edd`. Both complete short substantive policies were read from original HTML and through the primary reader; status/footer context was kept separate. No form or image was inspected.\n\nMastercard scope: entire substantive introductory policy and sections 1–4, reader lines 156–184, including both routes and the failure-to-comply section. Its linked forms, issuer process and general Rules/MATCH manuals were not read. WIPO scope: root's original slides 4–7 only, SHA256 `e4afb3bdce8e4d4266a5d992abe117b95b38cf57ff27979ccf6d70af4b7e11a4`, with a separate local extract; the presentation was not redownloaded. Root's programme packet supplies its fuller primary reading.\n\nAcquisition closed for independent review and the combined forest. No case, atlas, current-work, forest, ZIP or visible UI changed.\n", "file": "research/wipo-payment-wave-2026-09-09-participant-authority.md" }, { "id": "wipo-payment-wave-2026-09-09-programme-process", "title": "WIPO ALERT-PAY: the programme, the described process, and what its pilot measures", "description": "9 September 2026. Root core lane closed for independent review. This is a separate payment-information programme, not WIPO Pay for office fees, WIPO ALERT advertising, or Brazil's domestic distribution to payment intermediaries. Source descriptions and reported outcomes are kept at their dates. No account, list, payment operation or video playback was perfor", "html": "9 September 2026. Root core lane closed for independent review. This is a separate payment-information programme, not WIPO Pay for office fees, WIPO ALERT advertising, or Brazil's domestic distribution to payment intermediaries. Source descriptions and reported outcomes are kept at their dates. No account, list, payment operation or video playback was performed. Original speech has not been watched or authenticated against the automatic transcript.
\nThe retained official presentation is WIPO/ACE/18/40/PRESENTATION, linked by the official document page published16June2026. The deck identifies the ACE eighteenth session,2–4June2026. Its direct original is http[local research file] ; retained core-presentation.pdf SHA256 e4afb3bdce8e4d4266a5d992abe117b95b38cf57ff27979ccf6d70af4b7e11a4. All18pages of extractable text read. Root actually visually inspected full slides6 and9–17, rendered locally with Poppler from these exact bytes. PDF screenshots from the web tool were unusable (one empty result/three Cachemiss errors); local rendering was the successful substitute. Slides4,5,7,8 are text-read, not falsely labelled image-inspected. Interface images are illustrative material in a development presentation, not live system observation.
\nThe official Secretariat activities report is WIPO/ACE/18/2/REV., dated28May2026 and covering January2025–April2026. Original http[local research file] ; retained core-activities2026.pdf SHA25641e5b746daa77847b90733af50c8f4c9c4d7ff0f48083ef16544f6cdf204e4bc. Root read title/introductory scope in the official reader and all originalPDFpage8, especially§13; the full13pages are retained/extracted but not all substantively read. Other programme counts, budgets and collaboration rosters are not adopted as ALERT-PAY facts.
\nThe4June2026afternoon WIPO webcast page supplies an indexed automatic transcript: http[local research file] . Root substantively read the ToddReeves update, development/interface explanation and subsequent Brazil/Secretariat exchange through web search reader turn1478search0. Root's direct200HTML is only the app shell and contains no transcript. The indexed text explicitly warns that it is generated automatically and only original speech is the official proceedings record. core-transcript-scope.md is root paraphrase and locators, NOT verbatim transcript or an audiovisual original. There is no claim of complete-session reading, transcription fidelity, or original speech inspection.
\nActivities§13 reports BRIP finished a manual pilot in2025 with a small rights-holder group and two major payment providers, then began online-platform development in2026. Deck7 supplies the more exact window November2024–August2025, six rights holders, two PSPs,17actions and35sites. The broader statement 'pilot conducted in2025' on deck4 is consistent with this spanning window; it must not erase the2024start.
\nDeck4 names Mastercard and PayPal, alongside other PSPs as the intended/global cooperation class. Deck7 does not individually identify the two pilot PSPs or six rights holders. The automatic transcript says the interface currently under development offers PayPal and Mastercard. None of these independently names the two historical pilot providers. Illustrative Adidas-labelled screens likewise do not prove Adidas joined the pilot or signed a programme contract.
\nThe target described by§13 and deck3–4 is websites predominantly offering/selling counterfeit goods or subscription services offering pirated content. This is a programme scope claim, not a validated finding about any particular site, a consumer-identification duty, or a universal restriction on lawful purchasing. WIPO receives right-holder claims and checks completeness/required attestations. Deck5 assigns WIPO hosting, notices and requests for review, aggregation and access for PSPs. Deck4/§13 leave whether to act and whether merchant-account termination is warranted to each PSP under its existing policies. No WIPO adjudication order or instruction to terminate every listed merchant is established.
\nThe scheme reduces the coordination work needed for allegations to reach payment providers. That infrastructure/resource mechanism is supported by the described roles; an inference that WIPO thereby controls every provider decision is not. The general Mastercard/PayPal policies investigated in a parallel lane must not be silently treated as incorporated ALERT-PAY terms.
\nVisually inspected deck6 describes an authorized right holder identifying a site, simulating a purchase and sending a first merchant/alternative-contact notice. After3workingdays the right holder applies to WIPO. WIPO checks completeness and sends a second notice; after another3workingdays it includes the site for PSP assessment. These are successive design stages, not a proved six-day maximum or an observed case timeline. There is no specified deadline for WIPO's completeness work in this diagram.
\nThe automatic transcript materially qualifies the abbreviated diagram: it says inclusion follows the second3workingdays without a response. It describes actual or simulated purchases as depending on the PSP's requirements, and says actual test transactions were not conducted during the pilot because providers said they were unnecessary. Preserve this as the presenter's automatically transcribed account, not an independently inspected transaction record. The demonstration's transaction checkbox and optional-looking financial fields are consistent with variable collection, but do not prove backend validation rules.
\nDeck4 promises notice/review to protect operators; deck5 makes WIPO the administrator of review requests; deck16 shows a review-requested state. The presentation does not disclose who resolves a contested infringement on its merits, evidential burden, representation/translation provisions, proof of delivery, listing status while review is pending, review deadlines, liability allocation, or binding recipient correction/restoration rules. Those are specific unacquired instruments, not evidence that no safeguards exist. No imported advertising-only limitation or Brazil60dayreview clock applies automatically.
\nSlides9–14 display steps for URL/provider selection; existing or newly registered IP rights with jurisdiction, registration and certificate image; descriptions/evidence of goods; seller and registrar contact routes; transaction/checkout evidence; and a recorded right-holder notice with copy upload. Slide13's financial-account field requests redacted/encrypted information. It establishes a displayed data-handling instruction, not verified encryption or retention practice. The seven-step header includes a declaration stage, but its actual attestation text is not shown in these inspected screens.
\nSlides15–17 show formalities status/timeline, review/waiting/ready states, PSP feedback, pending feedback and recipient outcome categories. Illustrative fake/test domains and sample identities/dates must not become actual incidents, contributors or account actions on the map. In particular, same-day dates in an illustrative screen do not prove that actual cases bypassed the staged waiting periods. No listed demonstration domain was opened. No underlying platform, deployed revision, access rules or runtime behavior was inspected.
\nDeck8 reports71percent removed, defined beneath the headline as flagged sites offline OR goods removed. Deck7's17actions/35sites is not a count of17merchant terminations. The automatic transcript similarly describes flagged sites or infringing listings removed, and reports some false branding: provider logos did not always correspond to an actual payment service. Thus an apparent logo does not establish a payment relationship or a route that can be cut off.
\nNeither acquired source supplies the numerator/rounding method, per-site dates, denominator consistency, classification accuracy, permanent removal, individual PSP decisions, the number of merchant accounts terminated, revenue prevented, a counterfactual, or an independently audited outcome. The public aggregate cannot isolate voluntary operator response to notices from registrar action or PSP enforcement. The presenter's report that participants considered the pilot useful is participant feedback, not an evaluation independent of the programme. The deck's prediction that broad adoption could be disruptive is a projection, not measured global reach.
\nDeck8 says the system is under development at the June presentation. The automatic transcript attributes interface construction to an on-staff software engineer working for the preceding2–3months and says completion/availability is hoped for soon. This supplies an attributed staffing/resource input, not a name, employer contract, budget or completed launch. The parallel resources lane seeks later dated status and specific funding evidence.
\nThe indexed automatic transcript records Brazil objecting that document18/40 was unavailable before the discussion, limiting preparation, and a Secretariat apology/promise to publish the presentation. The official16June publication metadata confirms later public posting of this document; it does not prove deliberate suppression, privately shared alternative terms, or the exact reason/timing for the delay. Brazil's broader AI/enforcement criticism is a separate intervention and is not adopted as a factual finding about ALERT-PAY.
\nHighest-value gaps are the versioned right-holder/PSP participation instruments, authorized-contributor eligibility and attestations, contested-case review rules, pilot case-level action/outcome table, exact participant roster and any later public go-live/acceptance record. A redacted worked notice, response, decision and recipient correction would distinguish notice-induced compliance from account exclusion and show who can reverse which stage. No claim of a PIPCU input, Steam intervention, bank-owner instruction, stablecoin infrastructure or shared command centre follows from this packet.
", "path": "wipo-payment-wave-2026-09-09/programme-process.md", "sourceIds": [ "w30-class-s-pay-deck", "w30-class-s-pay-activities" ], "markdown": "9 September 2026. Root core lane closed for independent review. This is a separate payment-information programme, not WIPO Pay for office fees, WIPO ALERT advertising, or Brazil's domestic distribution to payment intermediaries. Source descriptions and reported outcomes are kept at their dates. No account, list, payment operation or video playback was performed. Original speech has not been watched or authenticated against the automatic transcript.\n\n## Primary records and scope\n\nThe retained official presentation is WIPO/ACE/18/40/PRESENTATION, linked by the official document page published16June2026. The deck identifies the ACE eighteenth session,2–4June2026. Its direct original is http[local research file] ; retained core-presentation.pdf SHA256 e4afb3bdce8e4d4266a5d992abe117b95b38cf57ff27979ccf6d70af4b7e11a4. All18pages of extractable text read. Root actually visually inspected full slides6 and9–17, rendered locally with Poppler from these exact bytes. PDF screenshots from the web tool were unusable (one empty result/three Cachemiss errors); local rendering was the successful substitute. Slides4,5,7,8 are text-read, not falsely labelled image-inspected. Interface images are illustrative material in a development presentation, not live system observation.\n\nThe official Secretariat activities report is WIPO/ACE/18/2/REV., dated28May2026 and covering January2025–April2026. Original http[local research file] ; retained core-activities2026.pdf SHA25641e5b746daa77847b90733af50c8f4c9c4d7ff0f48083ef16544f6cdf204e4bc. Root read title/introductory scope in the official reader and all originalPDFpage8, especially§13; the full13pages are retained/extracted but not all substantively read. Other programme counts, budgets and collaboration rosters are not adopted as ALERT-PAY facts.\n\nThe4June2026afternoon WIPO webcast page supplies an indexed automatic transcript: http[local research file] . Root substantively read the ToddReeves update, development/interface explanation and subsequent Brazil/Secretariat exchange through web search reader turn1478search0. Root's direct200HTML is only the app shell and contains no transcript. The indexed text explicitly warns that it is generated automatically and only original speech is the official proceedings record. core-transcript-scope.md is root paraphrase and locators, NOT verbatim transcript or an audiovisual original. There is no claim of complete-session reading, transcription fidelity, or original speech inspection.\n\n## Identity, chronology and decision rights\n\nActivities§13 reports BRIP finished a manual pilot in2025 with a small rights-holder group and two major payment providers, then began online-platform development in2026. Deck7 supplies the more exact window November2024–August2025, six rights holders, two PSPs,17actions and35sites. The broader statement 'pilot conducted in2025' on deck4 is consistent with this spanning window; it must not erase the2024start.\n\nDeck4 names Mastercard and PayPal, alongside other PSPs as the intended/global cooperation class. Deck7 does not individually identify the two pilot PSPs or six rights holders. The automatic transcript says the interface currently under development offers PayPal and Mastercard. None of these independently names the two historical pilot providers. Illustrative Adidas-labelled screens likewise do not prove Adidas joined the pilot or signed a programme contract.\n\nThe target described by§13 and deck3–4 is websites predominantly offering/selling counterfeit goods or subscription services offering pirated content. This is a programme scope claim, not a validated finding about any particular site, a consumer-identification duty, or a universal restriction on lawful purchasing. WIPO receives right-holder claims and checks completeness/required attestations. Deck5 assigns WIPO hosting, notices and requests for review, aggregation and access for PSPs. Deck4/§13 leave whether to act and whether merchant-account termination is warranted to each PSP under its existing policies. No WIPO adjudication order or instruction to terminate every listed merchant is established.\n\nThe scheme reduces the coordination work needed for allegations to reach payment providers. That infrastructure/resource mechanism is supported by the described roles; an inference that WIPO thereby controls every provider decision is not. The general Mastercard/PayPal policies investigated in a parallel lane must not be silently treated as incorporated ALERT-PAY terms.\n\n## The notice chain and its unresolved branch\n\nVisually inspected deck6 describes an authorized right holder identifying a site, simulating a purchase and sending a first merchant/alternative-contact notice. After3workingdays the right holder applies to WIPO. WIPO checks completeness and sends a second notice; after another3workingdays it includes the site for PSP assessment. These are successive design stages, not a proved six-day maximum or an observed case timeline. There is no specified deadline for WIPO's completeness work in this diagram.\n\nThe automatic transcript materially qualifies the abbreviated diagram: it says inclusion follows the second3workingdays without a response. It describes actual or simulated purchases as depending on the PSP's requirements, and says actual test transactions were not conducted during the pilot because providers said they were unnecessary. Preserve this as the presenter's automatically transcribed account, not an independently inspected transaction record. The demonstration's transaction checkbox and optional-looking financial fields are consistent with variable collection, but do not prove backend validation rules.\n\nDeck4 promises notice/review to protect operators; deck5 makes WIPO the administrator of review requests; deck16 shows a review-requested state. The presentation does not disclose who resolves a contested infringement on its merits, evidential burden, representation/translation provisions, proof of delivery, listing status while review is pending, review deadlines, liability allocation, or binding recipient correction/restoration rules. Those are specific unacquired instruments, not evidence that no safeguards exist. No imported advertising-only limitation or Brazil60dayreview clock applies automatically.\n\n## What the application images actually show\n\nSlides9–14 display steps for URL/provider selection; existing or newly registered IP rights with jurisdiction, registration and certificate image; descriptions/evidence of goods; seller and registrar contact routes; transaction/checkout evidence; and a recorded right-holder notice with copy upload. Slide13's financial-account field requests redacted/encrypted information. It establishes a displayed data-handling instruction, not verified encryption or retention practice. The seven-step header includes a declaration stage, but its actual attestation text is not shown in these inspected screens.\n\nSlides15–17 show formalities status/timeline, review/waiting/ready states, PSP feedback, pending feedback and recipient outcome categories. Illustrative fake/test domains and sample identities/dates must not become actual incidents, contributors or account actions on the map. In particular, same-day dates in an illustrative screen do not prove that actual cases bypassed the staged waiting periods. No listed demonstration domain was opened. No underlying platform, deployed revision, access rules or runtime behavior was inspected.\n\n## Pilot outcomes: removal is a composite endpoint\n\nDeck8 reports71percent removed, defined beneath the headline as flagged sites offline OR goods removed. Deck7's17actions/35sites is not a count of17merchant terminations. The automatic transcript similarly describes flagged sites or infringing listings removed, and reports some false branding: provider logos did not always correspond to an actual payment service. Thus an apparent logo does not establish a payment relationship or a route that can be cut off.\n\nNeither acquired source supplies the numerator/rounding method, per-site dates, denominator consistency, classification accuracy, permanent removal, individual PSP decisions, the number of merchant accounts terminated, revenue prevented, a counterfactual, or an independently audited outcome. The public aggregate cannot isolate voluntary operator response to notices from registrar action or PSP enforcement. The presenter's report that participants considered the pilot useful is participant feedback, not an evaluation independent of the programme. The deck's prediction that broad adoption could be disruptive is a projection, not measured global reach.\n\n## Development and the publication gap\n\nDeck8 says the system is under development at the June presentation. The automatic transcript attributes interface construction to an on-staff software engineer working for the preceding2–3months and says completion/availability is hoped for soon. This supplies an attributed staffing/resource input, not a name, employer contract, budget or completed launch. The parallel resources lane seeks later dated status and specific funding evidence.\n\nThe indexed automatic transcript records Brazil objecting that document18/40 was unavailable before the discussion, limiting preparation, and a Secretariat apology/promise to publish the presentation. The official16June publication metadata confirms later public posting of this document; it does not prove deliberate suppression, privately shared alternative terms, or the exact reason/timing for the delay. Brazil's broader AI/enforcement criticism is a separate intervention and is not adopted as a factual finding about ALERT-PAY.\n\n## Discriminating next records within the run\n\nHighest-value gaps are the versioned right-holder/PSP participation instruments, authorized-contributor eligibility and attestations, contested-case review rules, pilot case-level action/outcome table, exact participant roster and any later public go-live/acceptance record. A redacted worked notice, response, decision and recipient correction would distinguish notice-induced compliance from account exclusion and show who can reverse which stage. No claim of a PIPCU input, Steam intervention, bank-owner instruction, stablecoin infrastructure or shared command centre follows from this packet.\n", "file": "research/wipo-payment-wave-2026-09-09-programme-process.md" }, { "id": "wipo-payment-wave-2026-09-09-resources-delivery", "title": "WIPO ALERT-PAY: resources and delivery boundary", "description": "Acquisition closed on 9 September 2026. This lane read public WIPO planning, reporting and announcement records. It did not operate the system, inspect flagged sites, identify customers, or contact anyone. The core presentation, activity report and indexed webcast are owned by the root lane and were not rehere.", "html": "Acquisition closed on 9 September 2026. This lane read public WIPO planning, reporting and announcement records. It did not operate the system, inspect flagged sites, identify customers, or contact anyone. The core presentation, activity report and indexed webcast are owned by the root lane and were not reacquired here.
\nALERT-PAY is a named deliverable in WIPO's approved work programme, rather than only an informal proposal. The current official budget page says Member States approved the 2026/27 programme in July 2025 and describes the Director General's submission and annual performance-reporting route. The same page links the exact budget PDF retained here. About 96% of WIPO's projected income comes from its IP-system service fees; this is organization-wide financing, not an ALERT-PAY earmark. [Official budget and accountability page](http[local research file] substantive body from its opening July-approval paragraph through Reporting and monitoring; original resources-budget-approval.html.
The programme's Global Challenges and Partnerships section places ALERT-PAY deployment among Building Respect for IP priorities for 2026/27. The budget allocates CHF2.858 million to the broader Expected Result 2.3 (international dialogue/cooperation on Building Respect for IP) and CHF31.319 million to the whole sector. Neither is a disclosed ALERT-PAY allocation. Its ER2.3 performance indicator concerns implementation of the ACE agenda, not a numerical payment-account closure target. The adjacent trust-fund estimates are also not an attribution of funding to this product; the annex expressly qualifies estimated contributions. [Approved programme](http[local research file] PDF54-55 / printed53-54 (priority), PDF57-59 / printed56-58 (indicator/resources), PDF114 / printed113 (estimate qualification). Source-derived text scope, not an expenditure audit.
\nThe current linked 2024/25 Performance Report supplies an actual expenditure state at a broader level: CHF3.207 million for ER2.3 and CHF26.887 million for the GCP sector. Its expenditure note defines figures as pre-IPSAS actual expenditure. These are not Pay project costs or bank transfers. The ER2.3 narrative and dashboard document policy dialogue, studies, awareness and training, but do not separately certify the Pay pilot, its payment-provider actions or software acceptance. [Performance Report 2024/25](http[local research file] PDF10-11 / printed9-10; PDF25 / printed24; PDF28 / printed27 ER2.3 row; PDF70 / printed69. Its publication date was not independently established; the title identifies the reporting period. This is a WIPO self-report to Member States, not outside validation of particular outcomes.
\nThe July 2024 Director General report already announced the forthcoming Pay module. Its adjacent count of more than 8,000 live sites describes advertising ALERT, not Pay. This is historical programme intention. [Director General report, Assemblies July 9-17, 2024](http[local research file] heading and the Building Respect for IP/ALERT paragraph only; the whole speech was not substantively reviewed.
\nThe later official ACE recap was published 5 August 2026 and explicitly describes the June 2-4 meeting. It says the Secretariat shared an update on the Pay initiative. It supplies neither an online launch date nor new post-pilot outcome figures. Do not recast the recap's publication date as a deployment event. [ACE/18 recap](http[local research file] full substantive article and publication date read from retained original, resources-ace-aug2026.html.
The root lane's retained June2026 presentation/activity documents are the relevant evidence for the manual pilot and development phase. This lane does not independently corroborate payment suspensions, the identities of the two pilot providers, or the causal mechanism behind the reported removal percentage. No later launch or independently measured outcome was located in the finite search below. That is a search boundary, not a conclusion that launch never occurred.
\nThe supported institutional path is Member State approval of a programme -> WIPO sector work priority -> Secretariat delivery and reporting. It is not evidence of an earmarked PSP-financed programme, a procurement vendor's decision rights, or direct government control over merchant decisions. WIPO's general service-fee financing weakens a simplistic assumption that the presence of payment companies proves they financed this particular tool. It does not exclude a specific contribution or contract that the acquired records do not disclose.
\nA public delivery commitment can precede completion. A development demo, pilot, budget allocation and later public recap cannot be pooled into a single claim of an operating global cutoff system. Conversely, the absence of a Pay line in this high-level performance narrative is too weak to suggest abandonment: the more specific June programme records establish work that this annual document does not separately itemize.
\nThe highest-value missing resource instrument is an ALERT-PAY project authorization/workplan or procurement/engagement record identifying the cost centre, named delivery role, milestones, acceptance owner and recurrent operating allocation. For exercised effect, the useful counterpart is an attributable provider outcome statement or aggregate pilot evaluation separating warnings, withdrawal of goods, site disappearance, false logos and actual merchant restrictions. No private case file or harmful list contents are needed to state that missing distinction.
\nSix discovery queries were run in two batches: site:wipo.int "ALERT-PAY"; site:wipo.int "ALERT PAY" funding budget procurement; site:ungm.org "ALERT" "WIPO"; "WIPO ALERT-PAY" launch August September 2026 procurement; "WIPO ALERT PAY" tender software contract budget; site:wipo.int "Program of Work and Budget 2026/27" approved Assemblies July 2025. Returned candidates were assessed; no exact public Pay contract, award notice or post-June launch instrument was located. This was not an exhaustive procurement-archive audit. WIPOPay service-fee material, advertising ALERT and Brazilian domestic lists were excluded as different instruments. A later keynote and webcast were discovery candidates only; the root already owns the decisive webcast scope, so they were not separately promoted here.
Five ordinary unauthenticated GETs succeeded with HTTP200. No acquisition route failed or was bypassed. The original PDF text was machine-extracted for targeted reading; neither PDF was visually audited in this lane. A case-insensitive scan of the 142-page performance-report text found no ALERT product-name occurrence; the relevant ER2.3 narrative/table were then actually read. This is not a full report read and is not a claim about text inside unextracted images. Exact byte hashes, dates and scoped-read extents are in resources-captures.json. Supporting extracts are derived convenience files and never replace their originals.
WIPO Member States -> WIPO 2026/27 programme: approved July2025 (source-record statement on official budget page).
\nWIPO 2026/27 programme -> ALERT-PAY: planned deployment priority (programme decision; no project amount assigned).
\nWIPO Director General -> Member States: presents biennial programme; WIPO -> Member States: annual performance reporting (public accountability route; no product acceptance right inferred).
\nWIPO GCP ER2.3 -> reported2024/25 expenditure: CHF3.207m, broad result only. Keep this context in notes if a separate budget node would misleadingly suggest a Pay allocation.
\nWIPO -> August2026 ACE recap: retrospective announcement of a June update (not launch or operational outcome).
", "path": "wipo-payment-wave-2026-09-09/resources-delivery.md", "sourceIds": [ "w30-class-s-wipo-budget", "w30-class-s-wipo-recap", "w30-class-s-wipo-approval", "w30-class-s-wipo-spend" ], "markdown": "Acquisition closed on 9 September 2026. This lane read public WIPO planning, reporting and announcement records. It did not operate the system, inspect flagged sites, identify customers, or contact anyone. The core presentation, activity report and indexed webcast are owned by the root lane and were not reacquired here.\n\n## What the resource record establishes\n\nALERT-PAY is a named deliverable in WIPO's approved work programme, rather than only an informal proposal. The current official budget page says Member States approved the 2026/27 programme in July 2025 and describes the Director General's submission and annual performance-reporting route. The same page links the exact budget PDF retained here. About 96% of WIPO's projected income comes from its IP-system service fees; this is organization-wide financing, not an ALERT-PAY earmark. Official budget and accountability page (see source references) substantive body from its opening July-approval paragraph through Reporting and monitoring; original `resources-budget-approval.html`.\n\nThe programme's Global Challenges and Partnerships section places ALERT-PAY deployment among Building Respect for IP priorities for 2026/27. The budget allocates CHF2.858 million to the broader Expected Result 2.3 (international dialogue/cooperation on Building Respect for IP) and CHF31.319 million to the whole sector. Neither is a disclosed ALERT-PAY allocation. Its ER2.3 performance indicator concerns implementation of the ACE agenda, not a numerical payment-account closure target. The adjacent trust-fund estimates are also not an attribution of funding to this product; the annex expressly qualifies estimated contributions. Approved programme (see source references) PDF54-55 / printed53-54 (priority), PDF57-59 / printed56-58 (indicator/resources), PDF114 / printed113 (estimate qualification). Source-derived text scope, not an expenditure audit.\n\nThe current linked 2024/25 Performance Report supplies an actual expenditure state at a broader level: CHF3.207 million for ER2.3 and CHF26.887 million for the GCP sector. Its expenditure note defines figures as pre-IPSAS actual expenditure. These are not Pay project costs or bank transfers. The ER2.3 narrative and dashboard document policy dialogue, studies, awareness and training, but do not separately certify the Pay pilot, its payment-provider actions or software acceptance. Performance Report 2024/25 (see source references) PDF10-11 / printed9-10; PDF25 / printed24; PDF28 / printed27 ER2.3 row; PDF70 / printed69. Its publication date was not independently established; the title identifies the reporting period. This is a WIPO self-report to Member States, not outside validation of particular outcomes.\n\n## What later announcements do and do not establish\n\nThe July 2024 Director General report already announced the forthcoming Pay module. Its adjacent count of more than 8,000 live sites describes advertising ALERT, not Pay. This is historical programme intention. Director General report, Assemblies July 9-17, 2024 (see source references) heading and the Building Respect for IP/ALERT paragraph only; the whole speech was not substantively reviewed.\n\nThe later official ACE recap was published 5 August 2026 and explicitly describes the June 2-4 meeting. It says the Secretariat shared an update on the Pay initiative. It supplies neither an online launch date nor new post-pilot outcome figures. Do not recast the recap's publication date as a deployment event. ACE/18 recap (see source references) full substantive article and publication date read from retained original, `resources-ace-aug2026.html`.\n\nThe root lane's retained June2026 presentation/activity documents are the relevant evidence for the manual pilot and development phase. This lane does not independently corroborate payment suspensions, the identities of the two pilot providers, or the causal mechanism behind the reported removal percentage. No later launch or independently measured outcome was located in the finite search below. That is a search boundary, not a conclusion that launch never occurred.\n\n## Explanatory significance and competing interpretation\n\nThe supported institutional path is Member State approval of a programme -> WIPO sector work priority -> Secretariat delivery and reporting. It is not evidence of an earmarked PSP-financed programme, a procurement vendor's decision rights, or direct government control over merchant decisions. WIPO's general service-fee financing weakens a simplistic assumption that the presence of payment companies proves they financed this particular tool. It does not exclude a specific contribution or contract that the acquired records do not disclose.\n\nA public delivery commitment can precede completion. A development demo, pilot, budget allocation and later public recap cannot be pooled into a single claim of an operating global cutoff system. Conversely, the absence of a Pay line in this high-level performance narrative is too weak to suggest abandonment: the more specific June programme records establish work that this annual document does not separately itemize.\n\nThe highest-value missing resource instrument is an ALERT-PAY project authorization/workplan or procurement/engagement record identifying the cost centre, named delivery role, milestones, acceptance owner and recurrent operating allocation. For exercised effect, the useful counterpart is an attributable provider outcome statement or aggregate pilot evaluation separating warnings, withdrawal of goods, site disappearance, false logos and actual merchant restrictions. No private case file or harmful list contents are needed to state that missing distinction.\n\n## Finite search and custody scope\n\nSix discovery queries were run in two batches: `site:wipo.int \"ALERT-PAY\"`; `site:wipo.int \"ALERT PAY\" funding budget procurement`; `site:ungm.org \"ALERT\" \"WIPO\"`; `\"WIPO ALERT-PAY\" launch August September 2026 procurement`; `\"WIPO ALERT PAY\" tender software contract budget`; `site:wipo.int \"Program of Work and Budget 2026/27\" approved Assemblies July 2025`. Returned candidates were assessed; no exact public Pay contract, award notice or post-June launch instrument was located. This was not an exhaustive procurement-archive audit. WIPOPay service-fee material, advertising ALERT and Brazilian domestic lists were excluded as different instruments. A later keynote and webcast were discovery candidates only; the root already owns the decisive webcast scope, so they were not separately promoted here.\n\nFive ordinary unauthenticated GETs succeeded with HTTP200. No acquisition route failed or was bypassed. The original PDF text was machine-extracted for targeted reading; neither PDF was visually audited in this lane. A case-insensitive scan of the 142-page performance-report text found no ALERT product-name occurrence; the relevant ER2.3 narrative/table were then actually read. This is not a full report read and is not a claim about text inside unextracted images. Exact byte hashes, dates and scoped-read extents are in `resources-captures.json`. Supporting extracts are derived convenience files and never replace their originals.\n\n## Candidate relations, not canonical edits\n\nWIPO Member States -> WIPO 2026/27 programme: approved July2025 (source-record statement on official budget page).\n\nWIPO 2026/27 programme -> ALERT-PAY: planned deployment priority (programme decision; no project amount assigned).\n\nWIPO Director General -> Member States: presents biennial programme; WIPO -> Member States: annual performance reporting (public accountability route; no product acceptance right inferred).\n\nWIPO GCP ER2.3 -> reported2024/25 expenditure: CHF3.207m, broad result only. Keep this context in notes if a separate budget node would misleadingly suggest a Pay allocation.\n\nWIPO -> August2026 ACE recap: retrospective announcement of a June update (not launch or operational outcome).\n", "file": "research/wipo-payment-wave-2026-09-09-resources-delivery.md" }, { "id": "consolidation-wave-2026-09-09-synthesis-claim-index", "title": "Claim and evidence index", "description": "This index accompanies the current consolidated account. Stable IDs below are checked against accepted r33, including the six additional packets. Scoped note anchors retain the evidence and read limits behind those graph relationships. Several edges can depend on one source. These are traceable evidence bundles, not independent witness counts. The account an", "html": "This index accompanies the current consolidated account. Stable IDs below are checked against accepted r33, including the six additional packets. Scoped note anchors retain the evidence and read limits behind those graph relationships. Several edges can depend on one source. These are traceable evidence bundles, not independent witness counts. The account and index use the retained 9 September 2026 cutoff.
\n\nCase relationships: e44, e46, w2-fli-e-lightcone25.
Source records: src-e4e7930e8216, w2-fli-s-audit25.
Case relationships: e74, w1-creator-e-pf-terms, w1-creator-e-output-fli, w1-creator-e-frame-ltff, w1-creator-e-frame-aitof, w1-creator-e-rational-partnership.
Source records: src-65df3f2c31b3, src-c33ab812f862, w1-creator-s-controlai-rational, w1-creator-s-frame-public-ui-script, w1-creator-s-mateus, w1-creator-s-rational.
Case relationships: w1-controlai-e-uk-formal-authority, w1-controlai-e-us-payable.
Source records: w1-controlai-s-uk-articles, w1-controlai-s-us-return.
Case relationships: w8-red-e-work-segerie, w8-red-e-work-niki, w8-red-e-work-controlai, w8-red-e-work-fli, w8-red-e-signature-hubinger, w8-red-e-ctrl-interpretation.
Source records: w8-red-s-ack, w8-red-s-controlai, w8-red-s-signatories.
Case relationships: w31-fa-e-oakcma, w31-fa-e-oakstatewatch, w31-fa-e-oakdff, w31-fa-e-swcsaposition, w32-gdi-e-libertyfiled.
Source records: src-12639f108d96, src-5046aaa630e7, w31-fa-s-swcsa, w32-gdi-s-judgment, w32-gdi-s-libertyfiling.
Case relationships: w32-gdi-e-peerselection, w32-gdi-e-staffsupport, w32-gdi-e-terms, w32-gdi-e-juneselection.
Source records: w31-fa-s-dffround, w32-gdi-s-dffadoption, w32-gdi-s-dffpeer.
Case relationships: w10-duty-e-carnegie-review, w10-duty-e-reset-review, w10-duty-e-review-draft, w10-duty-e-research-narrower, w11-fund-e-carnegie-programme, w11-fund-e-mckinnon-functioning.
Source records: w10-duty-s-announcement, w10-duty-s-carnegie2022, w10-duty-s-draft, w10-duty-s-reset70, w10-duty-s-review, w11-fund-s-carnegie2021, w11-fund-s-carnegie2023, w11-fund-s-functioning-report, w11-fund-s-mckinnon-commissions.
Case relationships: w7-policy-e-email-accepted, w31-fa-e-checks, w31-fa-e-clearance, w31-fa-e-ffinput, w31-fa-e-noticerecommendation, w31-fa-e-declinemonetization.
Source records: w31-fa-s-cmasub, w31-fa-s-crisis, w31-fa-s-fullfact, w7-policy-s-part5, w7-policy-s-statement2025.
Case relationships: w22-sig-e-meta-case, w22-sig-e-case-signals, w22-sig-e-xbox-outcome.
Source records: w22-sig-s-report2025.
Case relationships: w26-ver-e-qs-controls-ohf, w26-ver-e-hash-api-deletions, w26-ver-e-tat-analyst-feedback, w26-ver-e-altitude-source-input.
Source records: w26-ver-s-altitude-tcap-adapter, w26-ver-s-archive-feedback, w26-ver-s-archive-hash-api, w26-ver-s-ohf-control.
Case relationships: w30-class-e-match-inquiry, w30-class-e-stripe-policy, w30-class-e-stripe-match-effect, w30-class-e-filmon-to-egumball, w30-class-e-egumball-slapp-disposition.
Source records: w30-class-s-egumball2025, w30-class-s-filmon-opinion, w30-class-s-match-rules, w30-class-s-stripe-match.
Case relationships: w23-cash-e-restriction-duty, w23-cash-e-historical-recourse, w24-ag-e-july-assent, w24-ag-e-july-redress, w24-ag-e-credits-cfpb, w24-ag-e-fallback-approval, w24-ag-e-naag-reimbursement.
Source records: w23-cash-s-cfpb-order, w23-cash-s-issuer2026q2, w24-ag-s-july-ca, w24-ag-s-july-or.
Case relationships: w29-enf-e-iwf-historical-membership, w29-enf-e-pilot-design-ipo, w30-class-e-creative-tactics.
Source records: w29-enf-s-funding2026, w29-enf-s-iwf2013, w30-class-s-creative-guidance.
Case relationships: w30-class-e-senate-tag-pressure, w30-class-e-default-announced-hi-category, w30-class-e-default-reported-hi-category, w30-class-e-tag-determination, w30-class-e-complaint-cma-allegation.
Source records: w30-class-s-dv-complaint, w30-class-s-dv-defaults, w30-class-s-dv-progress, w30-class-s-senate-tag, w30-class-s-tag-release, w30-class-s-tag-report.
Scoped retained notes are the evidence anchors; no supplemental graph relationship is invented.
\nStablecoin control check 2026 09 09
\n\nCase relationships: w32-gdi-e-oraclepackages, w32-gdi-e-oracleend, w32-gdi-e-iasdiscontinues, w32-gdi-e-appealstanding, w32-gdi-e-ratingrevision.
Source records: w32-gdi-s-fonevidence, w32-gdi-s-iasexit, w32-gdi-s-oracle, w32-gdi-s-oracleend.
Case relationships: w32-gdi-e-fundedremit, w32-gdi-e-ecgerman, w32-gdi-e-gdicontribution, w32-gdi-e-warningmethod, w32-gdi-e-humanvalidation, w32-gdi-e-requestedplatformrole.
Source records: w32-gdi-s-cdmd, w32-gdi-s-cordis, w32-gdi-s-d21, w32-gdi-s-fcdoanswer, w32-gdi-s-finalreport, w32-gdi-s-gdiresearch.
Case relationships: w33-con-e-vetting, w33-con-e-fundingdisclosure, w33-con-e-publicaccess, w33-con-e-priority, w33-con-e-publish.
Source records: w33-con-s-definition, w33-con-s-delegated, w33-con-s-dsa.
Case relationships: w33-con-e-nominatepay, w33-con-e-approveankura, w33-con-e-mandate, w33-con-e-monitor, w33-con-e-trustreport.
Source records: w33-con-s-decision, w33-con-s-trustee.
Case relationships: w33-con-e-ftiopinion, w33-con-e-noapprovals, w33-con-e-legalprocess.
Source records: w33-con-s-audit.
Case relationships: w33-con-e-trackeroutcomes, w33-con-e-alientry, w33-con-e-reasons, w33-con-e-mercator, w33-con-e-bmftr, w33-con-e-berlin.
Source records: w33-con-s-entry, w33-con-s-paper, w33-con-s-tracker.
Analytical comparison of the preceding evidence. No universal cross-domain pipeline is asserted.
\n\nCase relationships: w33-con-e-owcommission, w33-con-e-evalscope, w33-con-e-treasurygate, w33-con-e-satisfaction, w33-con-e-pilotfunding.
Source records: w16-supp-s-foi-bundle, w16-supp-s-foi-release, w33-con-s-round2.
Case relationships: w33-con-e-chairs, w33-con-e-woodgroup, w33-con-e-paevaluate, w33-con-e-publicosdi, w33-con-e-sandboxdelivery, w33-con-e-methodreuse, w33-con-e-publiccddp, w33-con-e-subapproval.
Source records: w16-supp-s-foi-bundle, w33-con-s-cdei, w33-con-s-osdiabout, w33-con-s-pa, w33-con-s-publicclaim, w33-con-s-sandbox, w33-con-s-stcf.
This synthesis uses accepted case records and retained research, with complete reading of the decisive newer notes and selected earlier scopes. It is not a fresh read of every original, all 365 audit pages or every historical branch. Source records retain access states, reading scopes and custody distinctions. The original three-part worker narrative and independent review were delivered in the task when the child hosts could not persist files. Root authored the final account, added the older evaluation/OSDI evidence and checked these references.
", "path": "consolidation-wave-2026-09-09/synthesis/claim-index.md", "sourceIds": [], "markdown": "This index accompanies the current consolidated account. Stable IDs below are checked against accepted r33, including the six additional packets. Scoped note anchors retain the evidence and read limits behind those graph relationships. Several edges can depend on one source. These are traceable evidence bundles, not independent witness counts. The account and index use the retained 9 September 2026 cutoff.\n\n## A1 — FLI resources\n\nCase relationships: `e44`, `e46`, `w2-fli-e-lightcone25`.\n\nSource records: `src-e4e7930e8216`, `w2-fli-s-audit25`.\n\n[Fli findings](red-threads-wave-2026-09-08-fli-findings.md)\n\n[Tegmark lightcone followup](tegmark-lightcone-followup.md)\n\n## A2 — Creator mechanisms\n\nCase relationships: `e74`, `w1-creator-e-pf-terms`, `w1-creator-e-output-fli`, `w1-creator-e-frame-ltff`, `w1-creator-e-frame-aitof`, `w1-creator-e-rational-partnership`.\n\nSource records: `src-65df3f2c31b3`, `src-c33ab812f862`, `w1-creator-s-controlai-rational`, `w1-creator-s-frame-public-ui-script`, `w1-creator-s-mateus`, `w1-creator-s-rational`.\n\n[Creator findings](red-threads-wave-2026-09-08-creator-findings.md)\n\n## A3 — ControlAI authority\n\nCase relationships: `w1-controlai-e-uk-formal-authority`, `w1-controlai-e-us-payable`.\n\nSource records: `w1-controlai-s-uk-articles`, `w1-controlai-s-us-return`.\n\n[Controlai findings](red-threads-wave-2026-09-08-controlai-findings.md)\n\n## A4 — Red Lines production\n\nCase relationships: `w8-red-e-work-segerie`, `w8-red-e-work-niki`, `w8-red-e-work-controlai`, `w8-red-e-work-fli`, `w8-red-e-signature-hubinger`, `w8-red-e-ctrl-interpretation`.\n\nSource records: `w8-red-s-ack`, `w8-red-s-controlai`, `w8-red-s-signatories`.\n\n[Campaign formation](red-lines-wave-2026-09-09-campaign-formation.md)\n\n## A5 — Common donor and divergent interventions\n\nCase relationships: `w31-fa-e-oakcma`, `w31-fa-e-oakstatewatch`, `w31-fa-e-oakdff`, `w31-fa-e-swcsaposition`, `w32-gdi-e-libertyfiled`.\n\nSource records: `src-12639f108d96`, `src-5046aaa630e7`, `w31-fa-s-swcsa`, `w32-gdi-s-judgment`, `w32-gdi-s-libertyfiling`.\n\n[Statewatch divergence](oak-digital-rights-divergence-wave-2026-09-09-statewatch-divergence.md)\n\n[Proceedings](liberty-tcn-proceedings-wave-2026-09-09-proceedings.md)\n\n## A6 — Selection and execution\n\nCase relationships: `w32-gdi-e-peerselection`, `w32-gdi-e-staffsupport`, `w32-gdi-e-terms`, `w32-gdi-e-juneselection`.\n\nSource records: `w31-fa-s-dffround`, `w32-gdi-s-dffadoption`, `w32-gdi-s-dffpeer`.\n\n[Allocation governance](dff-allocation-governance-wave-2026-09-09-allocation-governance.md)\n\n## A7 — Duty model and narrowing\n\nCase relationships: `w10-duty-e-carnegie-review`, `w10-duty-e-reset-review`, `w10-duty-e-review-draft`, `w10-duty-e-research-narrower`, `w11-fund-e-carnegie-programme`, `w11-fund-e-mckinnon-functioning`.\n\nSource records: `w10-duty-s-announcement`, `w10-duty-s-carnegie2022`, `w10-duty-s-draft`, `w10-duty-s-reset70`, `w10-duty-s-review`, `w11-fund-s-carnegie2021`, `w11-fund-s-carnegie2023`, `w11-fund-s-functioning-report`, `w11-fund-s-mckinnon-commissions`.\n\n[Operative authority](duty-care-wave-2026-09-09-operative-authority.md)\n\n[Policy lineage](duty-care-wave-2026-09-09-policy-lineage.md)\n\n## A8 — UK method choice and selective uptake\n\nCase relationships: `w7-policy-e-email-accepted`, `w31-fa-e-checks`, `w31-fa-e-clearance`, `w31-fa-e-ffinput`, `w31-fa-e-noticerecommendation`, `w31-fa-e-declinemonetization`.\n\nSource records: `w31-fa-s-cmasub`, `w31-fa-s-crisis`, `w31-fa-s-fullfact`, `w7-policy-s-part5`, `w7-policy-s-statement2025`.\n\n[Policy findings](uk-age-wave-2026-09-09-policy-findings.md)\n\n[Submission findings](cma-ofcom-monetization-wave-2026-09-09-submission-findings.md)\n\n[Disposition status](ofcom-crisis-authority-wave-2026-09-09-disposition-status.md)\n\n[Fullfact origin](ministerial-notice-origins-wave-2026-09-09-fullfact-origin.md)\n\n## A9 — Shared signals\n\nCase relationships: `w22-sig-e-meta-case`, `w22-sig-e-case-signals`, `w22-sig-e-xbox-outcome`.\n\nSource records: `w22-sig-s-report2025`.\n\n[Lantern](signal-sharing-wave-2026-09-09-lantern.md)\n\n[Lantern rights review](signal-sharing-wave-2026-09-09-lantern-rights-review.md)\n\n## A10 — Interfaces and correction\n\nCase relationships: `w26-ver-e-qs-controls-ohf`, `w26-ver-e-hash-api-deletions`, `w26-ver-e-tat-analyst-feedback`, `w26-ver-e-altitude-source-input`.\n\nSource records: `w26-ver-s-altitude-tcap-adapter`, `w26-ver-s-archive-feedback`, `w26-ver-s-archive-hash-api`, `w26-ver-s-ohf-control`.\n\n[Archive access](tcap-delivery-wave-2026-09-09-archive-access.md)\n\n[Correction protocol](signal-protocol-wave-2026-09-09-correction-protocol.md)\n\n[Protocol independent review](signal-protocol-wave-2026-09-09-protocol-independent-review.md)\n\n## A11 — MATCH and judicial bridge\n\nCase relationships: `w30-class-e-match-inquiry`, `w30-class-e-stripe-policy`, `w30-class-e-stripe-match-effect`, `w30-class-e-filmon-to-egumball`, `w30-class-e-egumball-slapp-disposition`.\n\nSource records: `w30-class-s-egumball2025`, `w30-class-s-filmon-opinion`, `w30-class-s-match-rules`, `w30-class-s-stripe-match`.\n\n[Acquirer implementation](payment-termination-wave-2026-09-09-acquirer-implementation.md)\n\n[Later disposition](egumball-disposition-wave-2026-09-09-later-disposition.md)\n\n## A12 — Customer remedies\n\nCase relationships: `w23-cash-e-restriction-duty`, `w23-cash-e-historical-recourse`, `w24-ag-e-july-assent`, `w24-ag-e-july-redress`, `w24-ag-e-credits-cfpb`, `w24-ag-e-fallback-approval`, `w24-ag-e-naag-reimbursement`.\n\nSource records: `w23-cash-s-cfpb-order`, `w23-cash-s-issuer2026q2`, `w24-ag-s-july-ca`, `w24-ag-s-july-or`.\n\n[Cfpb](cashapp-regulator-wave-2026-09-09-cfpb.md)\n\n## A13 — Police and infrastructure\n\nCase relationships: `w29-enf-e-iwf-historical-membership`, `w29-enf-e-pilot-design-ipo`, `w30-class-e-creative-tactics`.\n\nSource records: `w29-enf-s-funding2026`, `w29-enf-s-iwf2013`, `w30-class-s-creative-guidance`.\n\n[Policy authority](nominet-domain-authority-wave-2026-09-09-policy-authority.md)\n\n[Funding resources](pipcu-resources-wave-2026-09-09-funding-resources.md)\n\n[List authority](creative-advertising-wave-2026-09-09-list-authority.md)\n\n[Origins transmission](kybc-policy-wave-2026-09-09-origins-transmission.md)\n\n## A14 — DV and counterpressure\n\nCase relationships: `w30-class-e-senate-tag-pressure`, `w30-class-e-default-announced-hi-category`, `w30-class-e-default-reported-hi-category`, `w30-class-e-tag-determination`, `w30-class-e-complaint-cma-allegation`.\n\nSource records: `w30-class-s-dv-complaint`, `w30-class-s-dv-defaults`, `w30-class-s-dv-progress`, `w30-class-s-senate-tag`, `w30-class-s-tag-release`, `w30-class-s-tag-report`.\n\n[Policy pressure](dv-policy-pressure-wave-2026-09-09-policy-pressure.md)\n\n[Category authority](ncmec-advertising-input-wave-2026-09-09-category-authority.md)\n\n[Current product](filmon-classification-wave-2026-09-09-current-product.md)\n\n## A15 — Dated stablecoin supplement\n\nScoped retained notes are the evidence anchors; no supplemental graph relationship is invented.\n\n[Stablecoin control check 2026 09 09](stablecoin-control-check-2026-09-09.md)\n\n## A16 — GDI historical discontinuity and recourse\n\nCase relationships: `w32-gdi-e-oraclepackages`, `w32-gdi-e-oracleend`, `w32-gdi-e-iasdiscontinues`, `w32-gdi-e-appealstanding`, `w32-gdi-e-ratingrevision`.\n\nSource records: `w32-gdi-s-fonevidence`, `w32-gdi-s-iasexit`, `w32-gdi-s-oracle`, `w32-gdi-s-oracleend`.\n\n[Gdi mechanism](gdi-resource-to-advertising-wave-2026-09-09-gdi-mechanism.md)\n\n[Recourse findings](gdi-classification-recourse-wave-2026-09-09-recourse-findings.md)\n\n## A17 — Research funding and proposed use\n\nCase relationships: `w32-gdi-e-fundedremit`, `w32-gdi-e-ecgerman`, `w32-gdi-e-gdicontribution`, `w32-gdi-e-warningmethod`, `w32-gdi-e-humanvalidation`, `w32-gdi-e-requestedplatformrole`.\n\nSource records: `w32-gdi-s-cdmd`, `w32-gdi-s-cordis`, `w32-gdi-s-d21`, `w32-gdi-s-fcdoanswer`, `w32-gdi-s-finalreport`, `w32-gdi-s-gdiresearch`.\n\n[Fcdo findings](gdi-fcdo-remit-wave-2026-09-09-fcdo-findings.md)\n\n[Horizon findings](gdi-horizon-role-wave-2026-09-09-horizon-findings.md)\n\n[Classification findings](ai4trust-classification-decisions-wave-2026-09-09-classification-findings.md)\n\n## A18 — Access gates\n\nCase relationships: `w33-con-e-vetting`, `w33-con-e-fundingdisclosure`, `w33-con-e-publicaccess`, `w33-con-e-priority`, `w33-con-e-publish`.\n\nSource records: `w33-con-s-definition`, `w33-con-s-delegated`, `w33-con-s-dsa`.\n\n[Root law findings](dsa-research-access-gates-wave-2026-09-09-root-law-findings.md)\n\n[Gates findings](dsa-research-access-gates-wave-2026-09-09-gates-findings.md)\n\n## A19 — Binding access oversight\n\nCase relationships: `w33-con-e-nominatepay`, `w33-con-e-approveankura`, `w33-con-e-mandate`, `w33-con-e-monitor`, `w33-con-e-trustreport`.\n\nSource records: `w33-con-s-decision`, `w33-con-s-trustee`.\n\n[Commitment mechanism](dsa-access-enforcement-wave-2026-09-09-commitment-mechanism.md)\n\n[Root enforcement findings](dsa-access-enforcement-wave-2026-09-09-root-enforcement-findings.md)\n\n## A20 — Positive readiness and no approvals\n\nCase relationships: `w33-con-e-ftiopinion`, `w33-con-e-noapprovals`, `w33-con-e-legalprocess`.\n\nSource records: `w33-con-s-audit`.\n\n[Root audit findings](ali-oversight-evidence-wave-2026-09-09-root-audit-findings.md)\n\n[Oversight findings](ali-oversight-evidence-wave-2026-09-09-oversight-findings.md)\n\n## A21 — Voluntary reports and unresolved counts\n\nCase relationships: `w33-con-e-trackeroutcomes`, `w33-con-e-alientry`, `w33-con-e-reasons`, `w33-con-e-mercator`, `w33-con-e-bmftr`, `w33-con-e-berlin`.\n\nSource records: `w33-con-s-entry`, `w33-con-s-paper`, `w33-con-s-tracker`.\n\n[Root reading](dsa-application-decisions-wave-2026-09-09-root-reading.md)\n\n[Application findings](dsa-application-decisions-wave-2026-09-09-application-findings.md)\n\n## A22 — Cross-domain interpretation\n\nAnalytical comparison of the preceding evidence. No universal cross-domain pipeline is asserted.\n\n## A23 — CDDP evaluation and funding gates\n\nCase relationships: `w33-con-e-owcommission`, `w33-con-e-evalscope`, `w33-con-e-treasurygate`, `w33-con-e-satisfaction`, `w33-con-e-pilotfunding`.\n\nSource records: `w16-supp-s-foi-bundle`, `w16-supp-s-foi-release`, `w33-con-s-round2`.\n\n[Evaluation and acceptance](evaluation-wave-2026-09-09-evaluation-and-acceptance.md)\n\n## A24 — OSDI governance, evaluation reuse and PUBLIC participation\n\nCase relationships: `w33-con-e-chairs`, `w33-con-e-woodgroup`, `w33-con-e-paevaluate`, `w33-con-e-publicosdi`, `w33-con-e-sandboxdelivery`, `w33-con-e-methodreuse`, `w33-con-e-publiccddp`, `w33-con-e-subapproval`.\n\nSource records: `w16-supp-s-foi-bundle`, `w33-con-s-cdei`, `w33-con-s-osdiabout`, `w33-con-s-pa`, `w33-con-s-publicclaim`, `w33-con-s-sandbox`, `w33-con-s-stcf`.\n\n[Osdi governance market](osdi-comparison-wave-2026-09-09-osdi-governance-market.md)\n\n[Osdi evaluation output](osdi-comparison-wave-2026-09-09-osdi-evaluation-output.md)\n\n[Public cddp](osdi-comparison-wave-2026-09-09-public-cddp.md)\n\n## Reading limits\n\nThis synthesis uses accepted case records and retained research, with complete reading of the decisive newer notes and selected earlier scopes. It is not a fresh read of every original, all 365 audit pages or every historical branch. Source records retain access states, reading scopes and custody distinctions. The original three-part worker narrative and independent review were delivered in the task when the child hosts could not persist files. Root authored the final account, added the older evaluation/OSDI evidence and checked these references.\n", "file": "research/consolidation-wave-2026-09-09-synthesis-claim-index.md" }, { "id": "coxon-resignation-check-2026-09-09-findings", "title": "Jacob Coxon: bounded resignation and network check", "description": "9 September 2026 local campaign date. User requested a look at Coxon's viral resignation before proceeding. Fresh check confined to public research credentials, the statement, directly relevant funding/advocacy links and chronology. No outreach, graph edit, ZIP rebuild or visible UI. r33 contains no Coxon or hilbertspaess record; these findings remain a sepa", "html": "9 September 2026 local campaign date. User requested a look at Coxon's viral resignation before proceeding. Fresh check confined to public research credentials, the statement, directly relevant funding/advocacy links and chronology. No outreach, graph edit, ZIP rebuild or visible UI. r33 contains no Coxon or hilbertspaess record; these findings remain a separate supplement.
\nCoxon has a verifiable technical publication record. OpenAI's GPT-4o system card names him among contributors. The six-author paper Weight-sparse transformers have interpretable circuits credits him with optimization and circuit-pruning work, tasks and a circuit visualizer. Reading scope: author list and contribution statement, not full technical replication or verification of all employment dates.
\n[GPT-4o system card](http[local research file] PDF26; [sparse-transformer paper](http[local research file] PDF1 and32. OpenAI's emergent-misalignment paper thanks both Coxon and Evan Hubinger for feedback; shared acknowledgment does not itself establish direct collaboration between them.
\nHis seven-post statement reports resignation, attributes private fears to lab colleagues, forecasts dangerous capabilities and advocates coordinated pacing with a possible temporary capability-development ban. These are distinct testimony, forecasts and policy positions. The thread supplies no underlying internal documents or quantitative extinction-risk derivation. Native X failed; all seven posts were read through the [Thread Reader mirror](http[local research file] September8 US evening versus September9 UTC is consistent with the displayed date difference. Exact native rendering was not inspected.
\nThe ProPublica-rendered IRS return for Good Ventures Foundation, EIN46-1008520, TAX YEAR2022, covering1July2022 to30June2023 and submitted15May2024, names JACOB COXON in PartXV3a Paid during the year. Amount USD20,159; purpose Long-Term Future Scholarship Program - Research Scholarship. Exact-name scholarship recipient match is verified in the filing; independent confirmation that this is the researcher has not been acquired. The earlier fiscal-year-endingJune2022 return had no Coxon text match. It is inaccurate to silently treat the TY2022 label as a proved calendar2022 payment date.
\nThe same return names Cari Tuna as president/director and Dustin Moskovitz as secretary/treasurer/director. This connects the named scholarship to the foundation's governance. It supplies no present campaign compensation, scholarship conditions, continuing obligation or instruction concerning this resignation. Personal recipient address is deliberately omitted from this note; the public filing remains in the separate original capture.
\n[Full filing landing page](http[local research file] [rendered IRS990PF](http[local research file] Ordinary public GET supplied the readable return after the web reader could not fetch the embedded rendering. Scoped reading: header/tax period, named grant row and paid-section label, officer rows; not full financial audit.
\nAndrea Miotti's9September ControlAI article leads with Coxon's resignation, then describes ControlAI's UK/US legislative work. This establishes use of the statement in advocacy. Miotti claims ControlAI supplied the UK bill, consulted on the Sanders-Casar framework and drafted a government kill-switch amendment. These role claims are attributed to ControlAI. The article distinguishes its narrower precursor/ASI approach from the US proposal's broader temporary advanced-AI pause. It establishes neither sponsorship of Coxon's departure nor advance media coordination.
\n[Miotti article](http[local research file] Full substantive article read via ordinary publisher GET. Government chronology checked separately: Sanders' official3September announcement describes forthcoming legislation, permanent ASI prohibition, a temporary advanced-development pause and a new cabinet-level regulator/advisory board. That announcement predates Coxon's resignation. Announcement is not enactment or a complete bill-text analysis.
\n[Sanders announcement](http[local research file]
\nABC reports Hubinger publicly agreed with the danger concerns while saying Anthropic was trying its best and lacked a settled superintelligence-alignment plan. His personal greater-than10percent next-decade estimate is not an established probability. Native X reply was not fetched; [ABC's report](http[local research file] supplies the accessible attributed text. Existing publication overlap is a separate and weaker relationship than this response.
\nCoxon told Axios he left before Anthropic equity vested, retains OpenAI equity, and had not observed Anthropic compromising safety. [Axios interview](http[local research file]
\nThe supported structural finding is public translation of a technically credentialed researcher's warning into an existing legislative campaign. Scholarship history, sincere conviction, publication relationships, media planning and advocacy use can coexist, but are distinct propositions. No acquired evidence establishes paid resignation, covert direction, an NGO employment role or private coordination of the rollout. Claims about first quote-post order and precise amplification timing encountered in social commentary were not verified and are not promoted.
\nMost informative next records, if pursued: recipient confirmation and scholarship terms; actual advance briefing/communications arrangements; Coxon's next institutional role or funding; and whether policymakers explicitly use his testimony to justify particular operative powers. Broad acquisition stops at this bounded check. The existing consolidated explorer remains frozen.
", "path": "coxon-resignation-check-2026-09-09/findings.md", "sourceIds": [], "markdown": "9 September 2026 local campaign date. User requested a look at Coxon's viral resignation before proceeding. Fresh check confined to public research credentials, the statement, directly relevant funding/advocacy links and chronology. No outreach, graph edit, ZIP rebuild or visible UI. r33 contains no Coxon or hilbertspaess record; these findings remain a separate supplement.\n\n## Assessment\n\nCoxon has a verifiable technical publication record. OpenAI's GPT-4o system card names him among contributors. The six-author paper Weight-sparse transformers have interpretable circuits credits him with optimization and circuit-pruning work, tasks and a circuit visualizer. Reading scope: author list and contribution statement, not full technical replication or verification of all employment dates.\n\nGPT-4o system card (see source references) PDF26; sparse-transformer paper (see source references) PDF1 and32. OpenAI's emergent-misalignment paper thanks both Coxon and Evan Hubinger for feedback; shared acknowledgment does not itself establish direct collaboration between them.\n\nHis seven-post statement reports resignation, attributes private fears to lab colleagues, forecasts dangerous capabilities and advocates coordinated pacing with a possible temporary capability-development ban. These are distinct testimony, forecasts and policy positions. The thread supplies no underlying internal documents or quantitative extinction-risk derivation. Native X failed; all seven posts were read through the Thread Reader mirror (see source references) September8 US evening versus September9 UTC is consistent with the displayed date difference. Exact native rendering was not inspected.\n\n## Financial record\n\nThe ProPublica-rendered IRS return for Good Ventures Foundation, EIN46-1008520, TAX YEAR2022, covering1July2022 to30June2023 and submitted15May2024, names JACOB COXON in PartXV3a Paid during the year. Amount USD20,159; purpose Long-Term Future Scholarship Program - Research Scholarship. Exact-name scholarship recipient match is verified in the filing; independent confirmation that this is the researcher has not been acquired. The earlier fiscal-year-endingJune2022 return had no Coxon text match. It is inaccurate to silently treat the TY2022 label as a proved calendar2022 payment date.\n\nThe same return names Cari Tuna as president/director and Dustin Moskovitz as secretary/treasurer/director. This connects the named scholarship to the foundation's governance. It supplies no present campaign compensation, scholarship conditions, continuing obligation or instruction concerning this resignation. Personal recipient address is deliberately omitted from this note; the public filing remains in the separate original capture.\n\nFull filing landing page (see source references) rendered IRS990PF (see source references) Ordinary public GET supplied the readable return after the web reader could not fetch the embedded rendering. Scoped reading: header/tax period, named grant row and paid-section label, officer rows; not full financial audit.\n\n## Advocacy and institutional response\n\nAndrea Miotti's9September ControlAI article leads with Coxon's resignation, then describes ControlAI's UK/US legislative work. This establishes use of the statement in advocacy. Miotti claims ControlAI supplied the UK bill, consulted on the Sanders-Casar framework and drafted a government kill-switch amendment. These role claims are attributed to ControlAI. The article distinguishes its narrower precursor/ASI approach from the US proposal's broader temporary advanced-AI pause. It establishes neither sponsorship of Coxon's departure nor advance media coordination.\n\nMiotti article (see source references) Full substantive article read via ordinary publisher GET. Government chronology checked separately: Sanders' official3September announcement describes forthcoming legislation, permanent ASI prohibition, a temporary advanced-development pause and a new cabinet-level regulator/advisory board. That announcement predates Coxon's resignation. Announcement is not enactment or a complete bill-text analysis.\n\nSanders announcement (see source references)\n\nABC reports Hubinger publicly agreed with the danger concerns while saying Anthropic was trying its best and lacked a settled superintelligence-alignment plan. His personal greater-than10percent next-decade estimate is not an established probability. Native X reply was not fetched; ABC's report (see source references) supplies the accessible attributed text. Existing publication overlap is a separate and weaker relationship than this response.\n\nCoxon told Axios he left before Anthropic equity vested, retains OpenAI equity, and had not observed Anthropic compromising safety. Axios interview (see source references)\n\n## Forest and next discriminator\n\nThe supported structural finding is public translation of a technically credentialed researcher's warning into an existing legislative campaign. Scholarship history, sincere conviction, publication relationships, media planning and advocacy use can coexist, but are distinct propositions. No acquired evidence establishes paid resignation, covert direction, an NGO employment role or private coordination of the rollout. Claims about first quote-post order and precise amplification timing encountered in social commentary were not verified and are not promoted.\n\nMost informative next records, if pursued: recipient confirmation and scholarship terms; actual advance briefing/communications arrangements; Coxon's next institutional role or funding; and whether policymakers explicitly use his testimony to justify particular operative powers. Broad acquisition stops at this bounded check. The existing consolidated explorer remains frozen.\n", "file": "research/coxon-resignation-check-2026-09-09-findings.md" } ], "maps": [ { "id": "oakfunding", "title": "Oak’s grants to five organisations", "description": "Oak reports three-year support to these five recipients. Each amount belongs to its own grant year and purpose. The arrows have equal weight; they do not measure financial influence.", "boundary": "Reported awards are not annual expenditure or payment receipts. Contracting Oak entities and grant-specific terms remain limited. No traced allocation from Oak to the later Liberty case is established.", "topic": "Resources & divergence", "nodes": [ { "id": "oak", "title": "Oak Foundation", "sub": "Three-year awards; cash unverified", "pos": [ 3, 1, 1 ] }, { "id": "5rights", "title": "5Rights", "sub": "USD 2,000,000 · 3 years · 2023 award", "pos": [ 1, 2, 1 ] }, { "id": "weprotect", "title": "WeProtect", "sub": "USD 2,368,421 · 3 years · 2025 report", "pos": [ 2, 2, 1 ] }, { "id": "w30-class-cma", "title": "Check My Ads", "sub": "USD 600,000 · 3 years · 2025 report", "pos": [ 3, 2, 1 ] }, { "id": "w31-fa-statewatch", "title": "Statewatch", "sub": "USD 303,622 · 3 years · 2025 report", "pos": [ 4, 2, 1 ] }, { "id": "w31-fa-dff", "title": "Digital Freedom Fund", "sub": "USD 631,579 · 3 years · 2025 report", "pos": [ 5, 2, 1 ] } ], "edgeIds": [ "e160", "e161", "w31-fa-e-oakcma", "w31-fa-e-oakstatewatch", "w31-fa-e-oakdff" ] }, { "id": "divergence", "title": "Where Oak-funded organisations disagree", "description": "Statewatch endorsed a CSA objection and an OSA safeguards letter. The WeProtect Secretariat supported the CSA initiative with safeguards. DFF separately reports funding Liberty’s TCN challenge.", "boundary": "The CSA positions concern different dates/drafts. Statewatch’s 2023 objection predates the 2025 Oak grant. TCN litigation concerns IPA powers, not an OSA/age-ID case. Later governance and procedural records are now accessible in the DFF and court views; no merits or grant-payment result is inferred.", "topic": "Resources & divergence", "nodes": [ { "id": "wp-secretariat", "title": "WeProtect Secretariat", "sub": "Supports CSA initiative; Secretariat only", "pos": [ 1, 1, 1 ] }, { "id": "eu-csa-proposal", "title": "EU CSA proposal", "sub": "Different dates and draft contexts", "pos": [ 2, 1, 1 ] }, { "id": "w31-fa-swcsa", "title": "CSA objection · 2023", "sub": "Rights/security safeguards requested", "pos": [ 3, 1, 1 ] }, { "id": "w31-fa-statewatch", "title": "Statewatch", "sub": "Dated objections and safeguards", "pos": [ 4, 1, 1 ] }, { "id": "w31-fa-dff", "title": "Digital Freedom Fund", "sub": "Core grant; separate case allocation", "pos": [ 1, 3, 1 ] }, { "id": "w31-fa-liberty", "title": "Liberty", "sub": "Reported €36,000 litigation grant", "pos": [ 2, 3, 1 ] }, { "id": "w31-fa-tcn", "title": "TCN challenge", "sub": "IPA regime; separate from OSA", "pos": [ 3, 3, 1 ] }, { "id": "w31-fa-swletter", "title": "OSA safeguards letter", "sub": "Published 4 August 2025", "pos": [ 4, 2, 1 ] }, { "id": "ofcom", "title": "Ofcom", "sub": "Selects measures within its remit", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "n25", "w31-fa-e-swcsaendorse", "w31-fa-e-swcsaposition", "w31-fa-e-swlettersign", "w31-fa-e-swrequest", "w31-fa-e-libertyaward", "w31-fa-e-libertychallenge" ] }, { "id": "fundingselection", "title": "Who selects and authorises grants", "description": "Named grants establish resources; KR’s statutes establish internal selection and appointment rights. CMA’s two funding charts describe a changed mix without explaining its cause.", "boundary": "Ford approval, KR awards and Democracy Fund paid-schedule reports are different money states. Democracy row amounts are withheld here pending visual alignment. Chart percentages are not a dollar budget, cash balance or proof that a donor chose a tactic.", "topic": "Resources & divergence", "nodes": [ { "id": "w31-fa-villum", "title": "VILLUM FONDEN", "sub": "Up to two board appointments", "pos": [ 1, 1, 1 ] }, { "id": "w31-fa-krboard", "title": "KR board", "sub": "Distribution and grant decisions", "pos": [ 2, 1, 1 ] }, { "id": "w31-fa-kr", "title": "KR Foundation", "sub": "Dated climate/election grant", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-cma", "title": "Check My Ads", "sub": "Funding and advocacy are distinct", "pos": [ 4, 1, 1 ] }, { "id": "w31-fa-ford", "title": "Ford Foundation", "sub": "Approved general support, 2022", "pos": [ 1, 2, 1 ] }, { "id": "w31-fa-democracy", "title": "Democracy Fund, Inc.", "sub": "Reported paid rows; amounts unresolved", "pos": [ 2, 2, 1 ] }, { "id": "w31-fa-mix2024", "title": "CMA funding · 2024", "sub": "59 / 30 / 11 chart values", "pos": [ 3, 2, 1 ] }, { "id": "w31-fa-mix2025", "title": "CMA funding · 2025", "sub": "82.8% foundation grants", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w31-fa-e-krappointments", "w31-fa-e-krdistribution", "w31-fa-e-kr2023", "w31-fa-e-ford2022", "w31-fa-e-democracy2024", "w31-fa-e-mix2024", "w31-fa-e-mix2025", "w31-fa-e-democracy2023", "w31-fa-e-kr2025", "w31-fa-e-krelections" ] }, { "id": "cmaasks", "title": "Check My Ads: requests and responses", "description": "CMA asks for creator/advertiser checks, platform review before monetization and sanctions. Its crisis revenue/refund proposal has an explicit regulator response: specific monetization practices were not recommended.", "boundary": "These checks concern monetized creators and advertisers, not every viewer. The disposition of other recommendations is unresolved in the accepted source wave; no enacted universal ID, actual ban, refund or automated decision is inferred.", "topic": "Requests & dispositions", "nodes": [ { "id": "w30-class-cma", "title": "Check My Ads", "sub": "Funding and advocacy are distinct", "pos": [ 1, 1, 1 ] }, { "id": "w31-fa-cmasub", "title": "CMA response · 2025", "sub": "Listed December; submission day unknown", "pos": [ 2, 1, 1 ] }, { "id": "ofcom", "title": "Ofcom", "sub": "Selects measures within its remit", "pos": [ 3, 1, 1 ] }, { "id": "w31-fa-crisis-money", "title": "Crisis revenue / refunds", "sub": "Specific practices not recommended", "pos": [ 4, 1, 1 ] }, { "id": "w31-fa-creatorchecks", "title": "Creator / advertiser checks", "sub": "Proposal; not universal viewer ID", "pos": [ 1, 2, 1 ] }, { "id": "w31-fa-premonetization", "title": "Pre-monetization review", "sub": "Platform assessment before clearance", "pos": [ 2, 2, 1 ] }, { "id": "w31-fa-sanctions", "title": "Bans and clawbacks", "sub": "Requested; implementation unverified", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w31-fa-e-cmapublish", "w31-fa-e-cmasubmission", "w31-fa-e-checks", "w31-fa-e-clearance", "w31-fa-e-sanctionrequest", "w31-fa-e-crisismoney", "w31-fa-e-declinemonetization", "w31-fa-e-remainingforecast" ] }, { "id": "ffasks", "title": "Full Fact: requested access and official responses", "description": "Full Fact seeks broader crisis scope, meaningful data access for accredited fact-checkers and stronger supervision. Ofcom retains a narrower formal channel and declines central coordination and routine submission of every analysis.", "boundary": "Full Fact acknowledges Ofcom cannot mandate fact-checker partnerships. Class-level protocol decisions are not rejection of every data-access request, nor a new right for Full Fact to direct a platform.", "topic": "Requests & dispositions", "nodes": [ { "id": "w31-fa-fullfact", "title": "Full Fact", "sub": "Respondent seeking scope and access", "pos": [ 1, 1, 1 ] }, { "id": "w31-fa-ffsub", "title": "Full Fact response", "sub": "Dated 17 October 2025", "pos": [ 2, 1, 1 ] }, { "id": "ofcom", "title": "Ofcom", "sub": "Selects measures within its remit", "pos": [ 3, 1, 1 ] }, { "id": "w31-fa-protocol", "title": "Crisis protocol", "sub": "Decision published; commencement unresolved", "pos": [ 4, 1, 1 ] }, { "id": "w31-fa-ffscope", "title": "Broader crisis scope", "sub": "Request, with safeguards", "pos": [ 1, 2, 1 ] }, { "id": "w31-fa-ffaccess", "title": "Fact-checker data access", "sub": "Requested role; not a granted right", "pos": [ 2, 2, 1 ] }, { "id": "w31-fa-ffoversight", "title": "Stronger supervision", "sub": "Requested records and coordination", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w31-fa-e-ffpublish", "w31-fa-e-ffsubmission", "w31-fa-e-ffscope", "w31-fa-e-ffaccess", "w31-fa-e-ffoversight", "w31-fa-e-declineformalngo", "w31-fa-e-declinecentral", "w31-fa-e-protocoladoption" ] }, { "id": "noticeinputs", "title": "How Full Fact’s input changed a notice indicator", "description": "Ofcom attributes the notice-indicator amendment to Full Fact’s response and questions at an earlier DSIT meeting. Its FOI answer separately reports initial planning and a May 2026 approval window.", "boundary": "The response is dated October 2025; the meeting was in August. No Full Fact attendance or initiation is inferred. Record existence and an approval window do not reveal the underlying minute, draft edits or approver’s terms.", "topic": "Requests & dispositions", "nodes": [ { "id": "w31-fa-ffsub", "title": "Full Fact response", "sub": "Dated 17 October 2025", "pos": [ 1, 1, 1 ] }, { "id": "w31-fa-noticechange", "title": "Notice indicator amendment", "sub": "Ofcom attributes two input routes", "pos": [ 2, 1, 1 ] }, { "id": "w31-fa-protocol", "title": "Crisis protocol", "sub": "Decision published; commencement unresolved", "pos": [ 3, 1, 1 ] }, { "id": "w31-fa-dsitmeeting", "title": "Ofcom / DSIT meeting", "sub": "4 August 2025; minute outside this wave", "pos": [ 2, 2, 1 ] }, { "id": "ofcom", "title": "Ofcom", "sub": "Selects measures within its remit", "pos": [ 3, 2, 1 ] }, { "id": "w31-fa-planning", "title": "Initial planning record", "sub": "Existence and chronology confirmed", "pos": [ 4, 1, 1 ] }, { "id": "w31-fa-approval", "title": "Approval window", "sub": "21–26 May 2026, agency reported", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w31-fa-e-ffinput", "w31-fa-e-meetinginput", "w31-fa-e-noticerecommendation", "w31-fa-e-ofcommeeting", "w31-fa-e-planningreported", "w31-fa-e-approvalreported", "w31-fa-e-ffsubmission", "w31-fa-e-protocoladoption", "w31-fa-e-declineformalngo", "w31-fa-e-declinecentral" ] }, { "id": "ministerial", "title": "Ministers, Ofcom and crisis notices", "description": "Section 175 gives the Secretary a direction route through Ofcom. A notice can require a public statement and information. The provider’s criteria-based crisis determination remains a separate decision.", "boundary": "This enacted-text source is not a fresh consolidated-law/commencement audit. No particular notice or provider response is shown. It does not establish a direct power to order removal of any specified lawful post.", "topic": "Authority & records", "nodes": [ { "id": "w31-fa-secretary", "title": "Secretary of State", "sub": "Section 175 statutory role", "pos": [ 1, 1, 1 ] }, { "id": "ofcom", "title": "Ofcom", "sub": "Selects measures within its remit", "pos": [ 2, 1, 1 ] }, { "id": "w31-fa-notice", "title": "Public statement notice", "sub": "Explanation and information route", "pos": [ 3, 1, 1 ] }, { "id": "w31-fa-providers", "title": "In-scope providers", "sub": "Determine crisis using criteria", "pos": [ 4, 1, 1 ] }, { "id": "w31-fa-protocol", "title": "Crisis protocol", "sub": "Decision published; commencement unresolved", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w31-fa-e-ministernotice", "w31-fa-e-ofcomnotice", "w31-fa-e-publicaccount", "w31-fa-e-information", "w31-fa-e-providerdecision", "w31-fa-e-protocoladoption", "w31-fa-e-recommendprovider", "w31-fa-e-ministerliteracy", "w31-fa-e-reasonspublish", "w31-fa-e-directionrevoke", "w31-fa-e-declineformalngo", "w31-fa-e-declinecentral" ] }, { "id": "providerrecords", "title": "Who acts on a crisis—and who records it", "description": "Signals inform the provider, which determines activation and assembles its team. Large services have a specific police-channel recommendation. Providers retain analyses; Ofcom can seek them.", "boundary": "The channel does not replace legal data-request processes. Existing appeals, accuracy, privacy and expression safeguards remain. A 90-day protocol/analysis clock is not a promise to restore removed posts or accounts.", "topic": "Authority & records", "nodes": [ { "id": "w31-fa-indicators", "title": "External / internal signals", "sub": "Inputs; no automatic trigger", "pos": [ 1, 1, 1 ] }, { "id": "w31-fa-providers", "title": "In-scope providers", "sub": "Determine crisis using criteria", "pos": [ 2, 1, 1 ] }, { "id": "w31-fa-protocol", "title": "Crisis protocol", "sub": "Decision published; commencement unresolved", "pos": [ 3, 1, 1 ] }, { "id": "w31-fa-safeguards", "title": "Existing safeguards", "sub": "Appeals, expression and privacy", "pos": [ 4, 1, 1 ] }, { "id": "w31-fa-team", "title": "Provider crisis team", "sub": "Senior cross-functional responsibility", "pos": [ 1, 2, 1 ] }, { "id": "w31-fa-lechannel", "title": "Police crisis channel", "sub": "Large-service scope; existing law applies", "pos": [ 2, 2, 1 ] }, { "id": "w31-fa-analysis", "title": "Analysis and retained records", "sub": "Ofcom can ask; not all routinely public", "pos": [ 3, 2, 1 ] }, { "id": "ofcom", "title": "Ofcom", "sub": "Selects measures within its remit", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w31-fa-e-inputs", "w31-fa-e-providerdecision", "w31-fa-e-teamselection", "w31-fa-e-channeloperation", "w31-fa-e-analysisrecords", "w31-fa-e-analysisrequest", "w31-fa-e-retainappeals", "w31-fa-e-protocoladoption", "w31-fa-e-recommendprovider", "w31-fa-e-information", "w31-fa-e-policechannel", "w31-fa-e-declineformalngo", "w31-fa-e-declinecentral" ] }, { "id": "codestatus", "title": "From published code to legal effect", "description": "The crisis statement was published in June and children-code amendments were laid. Their source record does not resolve final crisis commencement. September’s NCII announcement supplies a different measure’s date.", "boundary": "The draft’s analysis clock is not a restoration guarantee. September 30 belongs to the NCII measure, not crisis rules. StopNCII is a recommended option with alternatives; no sole-supplier purchase requirement is established. Public partnership terms are in the earlier access reader; this measure does not prove execution.", "topic": "Authority & records", "nodes": [ { "id": "ofcom", "title": "Ofcom", "sub": "Selects measures within its remit", "pos": [ 2, 1, 1 ] }, { "id": "w31-fa-protocol", "title": "Crisis protocol", "sub": "Decision published; commencement unresolved", "pos": [ 1, 1, 1 ] }, { "id": "w31-fa-childdraft", "title": "Children-code draft", "sub": "Submitted 12 June; laid 18 June 2026", "pos": [ 3, 1, 1 ] }, { "id": "w31-fa-analysis", "title": "Analysis and retained records", "sub": "Ofcom can ask; not all routinely public", "pos": [ 4, 1, 1 ] }, { "id": "w31-fa-nciimeasure", "title": "Separate NCII measure", "sub": "30 September 2026 date; not crisis", "pos": [ 1, 2, 1 ] }, { "id": "w28-access-stopncii", "title": "StopNCII", "sub": "Recommended option; alternatives allowed", "pos": [ 2, 2, 1 ] }, { "id": "w28-access-swgfl", "title": "SWGfL", "sub": "Expertise; public terms separately retained", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w31-fa-e-protocoladoption", "w31-fa-e-laidchild", "w31-fa-e-draftanalysistime", "w31-fa-e-nciidate", "w31-fa-e-nciirecommend", "w31-fa-e-swgflexpertise", "w31-fa-e-declineformalngo", "w31-fa-e-declinecentral", "w31-fa-e-analysisrequest" ] }, { "id": "creative", "title": "From evidence to police and advertiser decisions", "description": "Operation Creative separates rightsholder referrals, police assessment and distribution to advertising recipients. Operators can challenge action; police removal has a stated evidential condition.", "boundary": "The embedded police guidance is dated 2019, released in 2021. A request to disrupt services is not a court order. Removal from a list does not prove downstream advertising restoration.", "topic": "Advertising lists", "nodes": [ { "id": "w30-class-creative-referrers", "title": "Rightsholder referrers", "sub": "Evidence enters police review", "pos": [ 1, 1, 1 ] }, { "id": "w29-enf-pipcu", "title": "PIPCU police unit", "sub": "Police assessment and list review", "pos": [ 2, 1, 1 ] }, { "id": "w30-class-iwl", "title": "Infringing Website List", "sub": "Police list; not a universal ban", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-ad-recipients", "title": "Advertising recipients", "sub": "Apply their own terms and choices", "pos": [ 4, 1, 1 ] }, { "id": "w30-class-site-operators", "title": "Website operators", "sub": "Notice and challenge role", "pos": [ 1, 2, 1 ] }, { "id": "w30-class-creative-head", "title": "Creative Head of Unit", "sub": "Chooses Creative tactics case by case", "pos": [ 2, 2, 1 ] }, { "id": "w30-class-creative", "title": "Operation Creative", "sub": "Dated procedure; portal design separate", "pos": [ 3, 2, 1 ] }, { "id": "w30-class-prs", "title": "PRS for Music", "sub": "138 referrals reported; period unspecified", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w30-class-e-creative-evidence", "w30-class-e-creative-screen", "w30-class-e-creative-tactics", "w30-class-e-creative-notice", "w30-class-e-creative-challenge", "w30-class-e-creative-removal", "w30-class-e-prs-referrals", "w30-class-e-iwl-distribution" ] }, { "id": "distribution", "title": "International sharing of advertising warnings", "description": "WIPO ALERT carries national assessments to approved advertising users. Contributors retain responsibilities and access choices; users choose which lists to use. Review begins in the relevant national process.", "boundary": "PIPCU is cited as a precedent, not a proved operational feed. Cross-list deletion notification is not automatic worldwide removal or advertising restoration.", "topic": "Advertising lists", "nodes": [ { "id": "w30-class-contributors", "title": "National contributors", "sub": "National process and list control", "pos": [ 1, 1, 1 ] }, { "id": "w30-class-wipo-alert", "title": "WIPO ALERT", "sub": "Advertising distribution only", "pos": [ 2, 1, 1 ] }, { "id": "w30-class-alert-users", "title": "Authorized advertising users", "sub": "Choose permitted lists", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-whitebullet", "title": "White Bullet", "sub": "Official lists and proprietary scoring", "pos": [ 4, 1, 1 ] }, { "id": "w30-class-site-operators", "title": "Website operators", "sub": "Notice and challenge role", "pos": [ 1, 2, 1 ] }, { "id": "w30-class-wipo", "title": "WIPO", "sub": "Administers access and distribution", "pos": [ 3, 2, 1 ] }, { "id": "w30-class-iwl", "title": "Infringing Website List", "sub": "Police list; not a universal ban", "pos": [ 2, 2, 1 ] } ], "edgeIds": [ "w30-class-e-wipo-precedent", "w30-class-e-wipo-host", "w30-class-e-wipo-classification", "w30-class-e-wipo-user-approval", "w30-class-e-wipo-contributor-permission", "w30-class-e-wipo-user-selection", "w30-class-e-wipo-purpose", "w30-class-e-wipo-national-review", "w30-class-e-wipo-cross-list-notification", "w30-class-e-wb-wipo-use" ] }, { "id": "commercial", "title": "Classifications, licences and business decisions", "description": "White Bullet describes an official-list index and a separate proprietary risk index. Peer39 markets an exclusion offering. Gambling Commission duties give a distinct contractual route for licensees.", "boundary": "Vendor descriptions and announcements do not prove a customer activated the service. LCCP contractual requirements are not a mandate to treat every IWL entry as a universal prohibition.", "topic": "Advertising lists", "nodes": [ { "id": "w30-class-gpi", "title": "Official-list index", "sub": "Official findings combined", "pos": [ 1, 1, 1 ] }, { "id": "w30-class-whitebullet", "title": "White Bullet", "sub": "Official lists and proprietary scoring", "pos": [ 2, 1, 1 ] }, { "id": "w30-class-ipi", "title": "Proprietary risk index", "sub": "Vendor assessment, not judgment", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-peer39", "title": "Peer39", "sub": "Markets exclusion; activation unverified", "pos": [ 4, 1, 1 ] }, { "id": "w30-class-gambling-commission", "title": "Gambling Commission", "sub": "Licence duties; IWL guidance", "pos": [ 1, 2, 1 ] }, { "id": "w30-class-gambling-condition", "title": "LCCP 16.1.1 condition", "sub": "Contract terms and reasonable steps", "pos": [ 2, 2, 1 ] }, { "id": "w30-class-gambling-licensees", "title": "Gambling licensees", "sub": "Specified licence scope, not all advertisers", "pos": [ 3, 2, 1 ] }, { "id": "w30-class-iwl", "title": "Infringing Website List", "sub": "Police list; not a universal ban", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w30-class-e-gambling-regulation", "w30-class-e-gambling-contract-right", "w30-class-e-gambling-iwl", "w30-class-e-wb-official-index", "w30-class-e-wb-proprietary-index", "w30-class-e-wb-peer-plan", "w30-class-e-peer-offering" ] }, { "id": "matching", "title": "How merchant records affect payment access", "description": "Acquirers report qualifying terminations; authorized users must inquire. Mastercard permits admission of a listed merchant, while Stripe publishes a more restrictive practice with exceptions. Correction is a distinct route.", "boundary": "August 2026 public network rules are not backdated to earlier disputes. Inquiry, response, early removal and five-year purge have different clocks. No automatic admission or restoration is shown.", "topic": "Payment restrictions", "nodes": [ { "id": "mc-acquirers", "title": "Mastercard acquirers", "sub": "Report, inquire, decide and correct", "pos": [ 1, 1, 1 ] }, { "id": "w30-class-match", "title": "MATCH Pro", "sub": "Inquiry is mandatory; exclusion is not", "pos": [ 2, 1, 1 ] }, { "id": "w30-class-match-users", "title": "Authorized MATCH users", "sub": "Access is approved; inquiry required", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-stripe-acceptance", "title": "Stripe acceptance practice", "sub": "General exclusion, with exceptions", "pos": [ 4, 1, 1 ] }, { "id": "w30-class-merchants", "title": "Merchants & owners", "sub": "Applicable remedy routes", "pos": [ 1, 2, 1 ] }, { "id": "mastercard", "title": "Mastercard", "sub": "Network standards and limited remedies", "pos": [ 2, 2, 1 ] }, { "id": "stripe", "title": "Stripe", "sub": "Publishes a separate acceptance policy", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w30-class-e-mc-investigation", "w30-class-e-mc-product-cessation", "w30-class-e-mc-termination-match", "w30-class-e-mc-enforcement", "w30-class-e-match-access", "w30-class-e-match-entry", "w30-class-e-match-inquiry", "w30-class-e-match-admission", "w30-class-e-match-accuracy", "w30-class-e-match-retroactive", "w30-class-e-match-purge", "w30-class-e-match-correction", "w30-class-e-match-personal-rights", "w30-class-e-stripe-policy", "w30-class-e-stripe-match-effect" ] }, { "id": "alertpay", "title": "How notices reach payment companies", "description": "The proposed process carries authorized claims through WIPO formalities to payment providers. WIPO reports a manual pilot; its programme separately plans deployment.", "boundary": "Pilot: 17 actions, 35 sites; 71% went offline OR removed infringing goods. These are reported composite site outcomes, not account-closure counts. Named development partners are not proved to be the two unnamed pilot providers. ER2.3 is a broader budget, not a Pay allocation.", "topic": "Payment restrictions", "nodes": [ { "id": "w30-class-pay-rightsholders", "title": "Authorized rightsholders", "sub": "Claimants in the proposed workflow", "pos": [ 1, 1, 1 ] }, { "id": "w30-class-pay", "title": "WIPO ALERT-PAY", "sub": "Design and pilot, separate from ALERT", "pos": [ 2, 1, 1 ] }, { "id": "w30-class-payment-providers", "title": "Payment providers", "sub": "Own-policy decisions; pilot firms unnamed", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-pay-pilot", "title": "Manual pilot", "sub": "71% composite outcome, reported", "pos": [ 4, 1, 1 ] }, { "id": "w30-class-wipo", "title": "WIPO", "sub": "Administers access and distribution", "pos": [ 2, 2, 1 ] }, { "id": "w30-class-wipo-programme", "title": "Approved 2026/27 programme", "sub": "Planned Pay deployment within broad work", "pos": [ 1, 2, 1 ] }, { "id": "w30-class-wipo-er23", "title": "Broader ER2.3 budget", "sub": "Not an ALERT-PAY allocation", "pos": [ 1, 3, 1 ] } ], "edgeIds": [ "w30-class-e-pay-hosting", "w30-class-e-pay-first-notice", "w30-class-e-pay-provider-discretion", "w30-class-e-pay-pilot-outcome", "w30-class-e-wipo-priority", "w30-class-e-wipo-er-budget", "w30-class-e-wipo-er-spend" ] }, { "id": "courts", "title": "Court challenges and the limits of restoration", "description": "eGumball sought emergency delisting relief. The 2023 appeal reversed that provisional relief; the 2025 appeal rejected anti-SLAPP protection at its first step. FilmOn supplied an expressly applied legal framework.", "boundary": "The rulings do not establish merchant liability, actual delisting or new processing access. The May 2025 modifying order corrects an inverted sentence without changing judgment. Current MATCH rules are a separate source period.", "topic": "Payment restrictions", "nodes": [ { "id": "w30-class-egumball", "title": "eGumball, Inc.", "sub": "Challenges listing; restoration unproved", "pos": [ 1, 1, 1 ] }, { "id": "w30-class-egumball-superior", "title": "Superior Court", "sub": "Provisional 2021 removal-request order", "pos": [ 2, 1, 1 ] }, { "id": "w30-class-merrick", "title": "Merrick Bank Corporation", "sub": "Historical termination and listing role", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-match", "title": "MATCH Pro", "sub": "Inquiry is mandatory; exclusion is not", "pos": [ 4, 1, 1 ] }, { "id": "w30-class-appeal-fourth", "title": "Court of Appeal", "sub": "Separate 2023 and 2025 procedural rulings", "pos": [ 1, 2, 1 ] }, { "id": "w30-class-egumball2023", "title": "2023 TRO appeal", "sub": "Emergency relief reversed; merits open", "pos": [ 2, 2, 1 ] }, { "id": "w30-class-egumball2025", "title": "2025 anti-SLAPP appeal", "sub": "Procedural ruling; no restoration", "pos": [ 3, 2, 1 ] }, { "id": "w30-class-filmon-opinion", "title": "FilmOn precedent · 2019", "sub": "Legal precedent, not shared data", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w30-class-e-merrick-match", "w30-class-e-egumball-challenge", "w30-class-e-egumball-tro", "w30-class-e-egumball-tro-reversal", "w30-class-e-merrick-slapp", "w30-class-e-egumball-slapp-disposition", "w30-class-e-egumball-version", "w30-class-e-egumball-public-brief", "w30-class-e-filmon-to-egumball" ] }, { "id": "inquiry", "title": "Political pressure and three different responses", "description": "Adalytics research and CMA’s claimed advocacy precede the February 2025 requests to DV, MRC and TAG. TAG reports no investigated BSC violations; DV disputes interpretation and reports additional safeguards.", "boundary": "The letters are requests, not subpoenas. CMA’s claimed contribution is not proof it drafted them. TAG’s determination is its own reported result, not independent exoneration; no MRC revocation is proved.", "topic": "Standards pressure", "nodes": [ { "id": "w30-class-adalytics", "title": "Adalytics", "sub": "Research source; later named defendant", "pos": [ 1, 1, 1 ] }, { "id": "w30-class-senate-inquiry", "title": "February 2025 inquiry", "sub": "Requests, not mandatory orders", "pos": [ 2, 1, 1 ] }, { "id": "w30-class-dv", "title": "DoubleVerify brand", "sub": "Brand/product role; legal entities separate", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-tag", "title": "TAG", "sub": "Certification body; reports its review", "pos": [ 4, 1, 1 ] }, { "id": "w30-class-cma", "title": "Check My Ads", "sub": "Claims advocacy; receives legal demand", "pos": [ 1, 2, 1 ] }, { "id": "w30-class-mrc", "title": "Media Rating Council", "sub": "Audit and revocation questions received", "pos": [ 3, 2, 1 ] }, { "id": "w30-class-tag-response", "title": "TAG reported finding", "sub": "No BSC violations, TAG reports", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w30-class-e-adalytics-early-access", "w30-class-e-cma-contribution", "w30-class-e-senate-dv-requests", "w30-class-e-senate-mrc-pressure", "w30-class-e-senate-tag-pressure", "w30-class-e-dv-critical-response", "w30-class-e-tag-determination", "w30-class-e-dv-rebuttal-april" ] }, { "id": "inputs", "title": "How outside reports become vendor policy", "description": "DV describes selecting public NCMEC material, conducting further assessment and introducing a Highly Illicit category. IWF’s assessed URL input is separate from DV’s membership and contribution band.", "boundary": "Incoming reports differ from outbound notices; the exact files used by DV remain unknown. Neither a public-data source nor the over-£25,000 contribution band establishes approval rights. Planned integration and later reported implementation are distinct.", "topic": "Standards pressure", "nodes": [ { "id": "w28-access-ncmec", "title": "NCMEC", "sub": "Public aggregates, distinct meanings", "pos": [ 1, 1, 1 ] }, { "id": "w30-class-dv-ncmec-input", "title": "DV-selected public input", "sub": "Exact files and versions unknown", "pos": [ 2, 1, 1 ] }, { "id": "w30-class-hi-category", "title": "Highly Illicit category", "sub": "DV policy and additional judgment", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-dv-platforms", "title": "Platform partners", "sub": "Distribution ≠ activated implementation", "pos": [ 4, 1, 1 ] }, { "id": "w30-class-dv", "title": "DoubleVerify brand", "sub": "Brand/product role; legal entities separate", "pos": [ 2, 2, 1 ] }, { "id": "iwf", "title": "Internet Watch Foundation", "sub": "URL assessment and membership resources", "pos": [ 1, 3, 1 ] }, { "id": "w30-class-iwf-url", "title": "IWF assessed URL List", "sub": "Specialist input, distinct from fees", "pos": [ 3, 2, 1 ] }, { "id": "w30-class-ncmec-outbound", "title": "Outbound notice tables", "sub": "Not a domain guilt list", "pos": [ 1, 2, 1 ] } ], "edgeIds": [ "w30-class-e-default-announced-hi-category", "w30-class-e-default-reported-hi-category", "w30-class-e-dv-category-distribution", "w30-class-e-dv-adoption", "w30-class-e-ncmec-publishes-outbound", "w30-class-e-dv-ncmec-origin", "w30-class-e-dv-ncmec-transformation", "w30-class-e-dv-input-category", "w30-class-e-iwf-maintenance", "w30-class-e-iwf-dv-planned", "w30-class-e-iwf-dv-testimonial", "w30-class-e-dv-iwf-membership", "w30-class-e-dv-iwf-band" ] }, { "id": "returnpaths", "title": "Preservation requests and legal counterpressure", "description": "DV Holdings’ counsel issued a preservation demand, and CMA publicly objected. The later complaint names both DV legal entities as plaintiffs and Adalytics as defendant; it separately alleges coordination with CMA.", "boundary": "A preservation demand is not a court order or immediate production requirement. Objection is not proof of compliance or chilling. Requested damages and corrective advertising are not granted relief, and CMA is not the named defendant.", "topic": "Standards pressure", "nodes": [ { "id": "w30-class-dv-holdings", "title": "DoubleVerify Holdings", "sub": "Issuer and named plaintiff", "pos": [ 1, 1, 1 ] }, { "id": "w30-class-preservation-letter", "title": "Preservation demand", "sub": "Request to CMA; not a court order", "pos": [ 2, 1, 1 ] }, { "id": "w30-class-cma", "title": "Check My Ads", "sub": "Claims advocacy; receives legal demand", "pos": [ 3, 1, 1 ] }, { "id": "w30-class-bondi", "title": "Bradley Bondi", "sub": "Signed counsel; no fee data acquired", "pos": [ 4, 1, 1 ] }, { "id": "w30-class-dv-inc", "title": "DoubleVerify Inc.", "sub": "Historical litigant and named plaintiff", "pos": [ 1, 2, 1 ] }, { "id": "w30-class-dv-complaint", "title": "Filed complaint", "sub": "Adalytics is named defendant", "pos": [ 2, 2, 1 ] }, { "id": "w30-class-adalytics", "title": "Adalytics", "sub": "Research source; later named defendant", "pos": [ 3, 2, 1 ] }, { "id": "w30-class-paul-hastings", "title": "Paul Hastings", "sub": "Firm named on demand and complaint", "pos": [ 4, 2, 1 ] } ], "edgeIds": [ "w30-class-e-bondi-representation", "w30-class-e-firm-representation", "w30-class-e-dv-preservation", "w30-class-e-preservation-recipient", "w30-class-e-cma-objection", "w30-class-e-complaint-plaintiff-dv-holdings", "w30-class-e-complaint-plaintiff-dv-inc", "w30-class-e-complaint-defendant", "w30-class-e-complaint-cma-allegation" ] }, { "id": "research", "title": "Who financed the policy research", "description": "All five disclosed co-funders connect to the same corrected March 2021 report. Publication, authored recommendations and MPA’s later parliamentary citation remain distinct routes.", "boundary": "No grant amount, contract, editorial veto or donor instruction was acquired. The report says validation workshops excluded funders; that is a reported method. MPA quotes the executive summary, not identical numbered recommendation 4.", "topic": "Earlier research", "nodes": [ { "id": "w29-enf-ipo", "title": "Intellectual Property Office", "sub": "Amount and approval rights unknown", "pos": [ 1, 1, 1 ] }, { "id": "w29-enf-alliance", "title": "Alliance for IP", "sub": "Amount and approval rights unknown", "pos": [ 2, 1, 1 ] }, { "id": "w1-icmec-mpa", "title": "Motion Picture Association", "sub": "Amount and approval rights unknown", "pos": [ 3, 1, 1 ] }, { "id": "w29-enf-industry-trust", "title": "Industry Trust", "sub": "Amount and approval rights unknown", "pos": [ 4, 1, 1 ] }, { "id": "w29-enf-premier-league", "title": "Premier League", "sub": "Amount and approval rights unknown", "pos": [ 5, 1, 1 ] }, { "id": "w29-enf-rusi", "title": "RUSI", "sub": "Publisher of corrected report", "pos": [ 1, 3, 1 ] }, { "id": "w29-enf-rusi-report2021", "title": "Taking the Profit Out of IP Crime", "sub": "March 2021 corrected report", "pos": [ 3, 2, 1 ] }, { "id": "w29-enf-rusi-recommendations", "title": "Research recommendations", "sub": "Recommendations 4, 14 and 15", "pos": [ 5, 3, 1 ] }, { "id": "w29-enf-mpa-evidence2022", "title": "MPA parliamentary evidence", "sub": "Internal date · 10 May 2022", "pos": [ 3, 3, 1 ] } ], "edgeIds": [ "w29-enf-e-rusi-publication", "w29-enf-e-rusi-funder-ipo", "w29-enf-e-rusi-funder-alliance", "w29-enf-e-rusi-funder-mpa", "w29-enf-e-rusi-funder-industry-trust", "w29-enf-e-rusi-funder-premier-league", "w29-enf-e-rusi-recommendations", "w29-enf-e-mpa-rusi-citation" ] }, { "id": "gdi-awards", "title": "Oak’s grants to GDI and rights organisations", "description": "Oak’s 2023 report records four three-year awards. Each arrow retains its own amount and purpose. No grant is allocated to the later complaint or a named classification decision.", "boundary": "Awards are not bank receipts or annual spending. The UK operator and German research participant remain separate roles; common branding does not establish corporate ownership.", "topic": "Grant resources & selection", "nodes": [ { "id": "oak", "title": "Oak Foundation", "sub": "Reported awards; cash unknown", "pos": [ 2, 1, 1 ] }, { "id": "w32-gdi-gdi", "title": "Global Disinformation Index", "sub": "USD 903,614 · core capacity", "pos": [ 1, 2, 1 ] }, { "id": "w31-fa-liberty", "title": "Liberty", "sub": "USD 783,133 · human-rights work", "pos": [ 2, 2, 1 ] }, { "id": "w32-gdi-org", "title": "Open Rights Group", "sub": "USD 437,349 · migration / data", "pos": [ 3, 2, 1 ] }, { "id": "w32-gdi-pi", "title": "Privacy International", "sub": "USD 1,138,554 · privacy safeguards", "pos": [ 2, 3, 1 ] } ], "edgeIds": [ "w32-gdi-e-oakgdi", "w32-gdi-e-oakliberty", "w32-gdi-e-oakorg", "w32-gdi-e-oakpi" ] }, { "id": "peer-selection", "title": "Who chooses Digital Freedom Fund projects", "description": "DFF describes staged 2025 selection rounds, March 2026 board adoption and peer selection of the June round. Staff and the published terms support that process.", "boundary": "These are attributed governance accounts. No original board resolution, complete recusal rules, named Liberty vote, grant signatory or payment is acquired. The March 2025 complaint predates the June 2026 grant.", "topic": "Grant resources & selection", "nodes": [ { "id": "w32-gdi-dffboard", "title": "DFF board", "sub": "March 2026 process adoption", "pos": [ 1, 1, 1 ] }, { "id": "w32-gdi-peerprocess", "title": "DFF selection process", "sub": "Staged pilot; later adoption", "pos": [ 2, 1, 1 ] }, { "id": "w32-gdi-peer", "title": "DFF Peer Group", "sub": "Reported project selection", "pos": [ 3, 1, 1 ] }, { "id": "w32-gdi-dffstaff", "title": "DFF staff", "sub": "Information and facilitation", "pos": [ 1, 2, 1 ] }, { "id": "w32-gdi-peerterms", "title": "Peer Terms of Reference", "sub": "Conflict and confidentiality commitments", "pos": [ 2, 2, 1 ] }, { "id": "w31-fa-dffround", "title": "June 2026 grant round", "sub": "Attributed peer selection; no payment proof", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w32-gdi-e-boardadoption", "w32-gdi-e-peerselection", "w32-gdi-e-staffsupport", "w32-gdi-e-terms", "w32-gdi-e-juneselection" ] }, { "id": "public-funding", "title": "GDI: public grants and EU research funding", "description": "FCDO reports support to the named GDI operation. The Commission records AI4TRUST project funding and a contribution for its German participant. These are two documented channels.", "boundary": "The FCDO invoice is not an additional transfer above its aggregates. Programme totals are not recipient income; the German contribution is not a UK-company payment. No ownership or common-grant bridge is drawn.", "topic": "Grant resources & selection", "nodes": [ { "id": "w32-gdi-fcdo", "title": "FCDO", "sub": "Dated aggregate funding account", "pos": [ 1, 1, 1 ] }, { "id": "w32-gdi-gdi", "title": "Global Disinformation Index", "sub": "Named operation; UK/German roles differ", "pos": [ 2, 1, 1 ] }, { "id": "w32-gdi-cdmd", "title": "CDMD programme", "sub": "Programme envelope ≠ recipient income", "pos": [ 3, 1, 1 ] }, { "id": "european-commission", "title": "European Commission", "sub": "Specified authority; no inferred case outcome", "pos": [ 1, 2, 1 ] }, { "id": "w32-gdi-ai4trust", "title": "AI4TRUST", "sub": "EU-funded research project", "pos": [ 2, 2, 1 ] }, { "id": "w32-gdi-gdide", "title": "GDI gUG · Germany", "sub": "Named EU research participant", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w32-gdi-e-fcdo2019", "w32-gdi-e-fcdo2022", "w32-gdi-e-ledger", "w32-gdi-e-programmepartnership", "w32-gdi-e-ecproject", "w32-gdi-e-ecgerman" ] }, { "id": "methodology", "title": "From research terms to classification decisions", "description": "GDI contributed existing keyword resources; FBK edited D2.1 and recorded internal coordinator review. The document’s sampling and methodological choices are inspectable.", "boundary": "Coordinator accepted / Ready for submission is an internal designation, not Commission acceptance. Keywords select material for research; they do not establish truth, illegality or deployed enforcement.", "topic": "Research inputs & evaluation", "nodes": [ { "id": "w32-gdi-gdi", "title": "Global Disinformation Index", "sub": "Named operation; UK/German roles differ", "pos": [ 1, 1, 1 ] }, { "id": "w32-gdi-d21", "title": "D2.1 methodology", "sub": "Internal acceptance ≠ EC acceptance", "pos": [ 2, 1, 1 ] }, { "id": "w32-gdi-fbk", "title": "Fondazione Bruno Kessler", "sub": "Project coordinator / editor", "pos": [ 3, 1, 1 ] }, { "id": "w32-gdi-ucamreview", "title": "UCAM review role", "sub": "D2.1 quality assurance", "pos": [ 1, 2, 1 ] }, { "id": "w32-gdi-seeds", "title": "Retrieval seeds", "sub": "Sampling inputs, not truth verdicts", "pos": [ 2, 2, 1 ] }, { "id": "w32-gdi-framing", "title": "Methodological frame", "sub": "Explicitly adopted analytical approach", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w32-gdi-e-gdicontribution", "w32-gdi-e-d21editor", "w32-gdi-e-qualityreview", "w32-gdi-e-sampling", "w32-gdi-e-framing" ] }, { "id": "component-results", "title": "Test scores versus real-world results", "description": "AI4TRUST reports a final warning system, a textual component, a human-validation dashboard and user evaluation. Each finding belongs to its stated component and test.", "boundary": "A textual weighted F1 near 56% is not aggregate system accuracy. Reported users and annotations do not establish production deployment, automatic sanctions or independent replication.", "topic": "Research inputs & evaluation", "nodes": [ { "id": "w32-gdi-ai4trust", "title": "AI4TRUST", "sub": "EU-funded research project", "pos": [ 2, 1, 1 ] }, { "id": "w32-gdi-warning", "title": "Warning system", "sub": "Reported final ensemble", "pos": [ 1, 2, 1 ] }, { "id": "w32-gdi-textsignals", "title": "Textual component", "sub": "About 56% weighted F1; component only", "pos": [ 2, 2, 1 ] }, { "id": "w32-gdi-validation", "title": "Human-validation dashboard", "sub": "Reported tool; not automatic action", "pos": [ 3, 2, 1 ] }, { "id": "w32-gdi-evaluation", "title": "User evaluation", "sub": "206 users / three rounds reported", "pos": [ 2, 3, 1 ] } ], "edgeIds": [ "w32-gdi-e-warningmethod", "w32-gdi-e-componentmetric", "w32-gdi-e-humanvalidation", "w32-gdi-e-userevaluation" ] }, { "id": "advertising-exits", "title": "GDI’s advertising relationships—and their endings", "description": "The historical Oracle segment and IAS input connect GDI ratings to buyer choices. Both named intermediaries later announced exits from the relevant service or input.", "boundary": "Service announcements are not observed cache deletion or publisher restoration. Buyer configuration and actual revenue effects remain not obtained; these are historical channels, not proof of current use.", "topic": "Distribution, exits & review", "nodes": [ { "id": "w32-gdi-gdi", "title": "Global Disinformation Index", "sub": "Named operation; UK/German roles differ", "pos": [ 1, 1, 1 ] }, { "id": "w32-gdi-segment", "title": "Oracle GDI segment", "sub": "Selectable ratings input", "pos": [ 2, 1, 1 ] }, { "id": "w32-gdi-oracle", "title": "Oracle", "sub": "Historical segment; service exit", "pos": [ 3, 1, 1 ] }, { "id": "w32-gdi-iasproduct", "title": "IAS Context Control", "sub": "Customer category / risk choices", "pos": [ 1, 2, 1 ] }, { "id": "w32-gdi-buyers", "title": "Advertising buyers", "sub": "Choose targeting / exclusions", "pos": [ 2, 2, 1 ] }, { "id": "w32-gdi-ias", "title": "Integral Ad Science", "sub": "Announced input discontinuation", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w32-gdi-e-ratingsupply", "w32-gdi-e-oraclepackages", "w32-gdi-e-buyerchoice", "w32-gdi-e-oracleend", "w32-gdi-e-iashistorical", "w32-gdi-e-iasdiscontinues", "w32-gdi-e-iaschoices" ] }, { "id": "review-routes", "title": "How an advertising classification can be challenged", "description": "GDI describes publisher rating appeals and reassessment. Its UK operator’s policy separately offers qualified consideration of concerns about research outputs.", "boundary": "Neither process supplies an recipient-correction instruction, successful appeal or duty for buyers to resume advertising. Privacy deletion is not the same as withdrawing a rating.", "topic": "Distribution, exits & review", "nodes": [ { "id": "w32-gdi-publishers", "title": "Affected publishers", "sub": "Described appeal standing", "pos": [ 1, 1, 1 ] }, { "id": "w32-gdi-ratingreview", "title": "GDI rating review", "sub": "Described reassessment; outcome unknown", "pos": [ 2, 1, 1 ] }, { "id": "w32-gdi-gdi", "title": "Global Disinformation Index", "sub": "Named operation; UK/German roles differ", "pos": [ 3, 1, 1 ] }, { "id": "w32-gdi-gdiuk", "title": "Disinformation Index Ltd", "sub": "UK website operator / controller", "pos": [ 1, 2, 1 ] }, { "id": "w32-gdi-researchcomplaints", "title": "Research representations", "sub": "Qualified review, not restoration", "pos": [ 2, 2, 1 ] } ], "edgeIds": [ "w32-gdi-e-appealstanding", "w32-gdi-e-ratingrevision", "w32-gdi-e-researchreview", "w32-gdi-e-researchlimits" ] }, { "id": "procedural-return", "title": "Apple, Liberty and a partly private court case", "description": "Liberty and Privacy International filed a separate complaint and submitted arguments. The tribunal denied concealment of basic Apple case details while keeping the March hearing private.", "boundary": "This is a procedural decision, not a TCN merits ruling, a universal open-hearing right or an advertising remedy. The later identified DFF grant cannot be assumed to have initiated the earlier filing.", "topic": "Distribution, exits & review", "nodes": [ { "id": "w31-fa-liberty", "title": "Liberty", "sub": "Core award; separate procedural action", "pos": [ 1, 1, 1 ] }, { "id": "w32-gdi-complaint", "title": "Liberty / PI complaint", "sub": "Filed 13 March 2025", "pos": [ 2, 1, 1 ] }, { "id": "w32-gdi-pi", "title": "Privacy International", "sub": "Privacy and surveillance safeguards", "pos": [ 3, 1, 1 ] }, { "id": "w32-gdi-ipt", "title": "Investigatory Powers Tribunal", "sub": "Decides procedural applications", "pos": [ 1, 2, 1 ] }, { "id": "w32-gdi-secrecyapplication", "title": "Basic case details", "sub": "Application to withhold", "pos": [ 2, 2, 1 ] }, { "id": "w32-gdi-hearing", "title": "March hearing", "sub": "Remained private", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w32-gdi-e-libertyfiled", "w32-gdi-e-pifiled", "w32-gdi-e-applicationdenied", "w32-gdi-e-privatehearing" ] }, { "id": "access-gates", "title": "Who can access platform data", "description": "The provider-establishment DSC decides project vetting and request conditions. The researcher supplies eligibility and funding information; providers have a bounded amendment route.", "boundary": "This selected Article 40(4) path is separate from public-data access under 40(12) and notice priority under Article 22. An 80-working-day request/refusal clock is not a delivery promise. No GDI application is matched here.", "topic": "Researcher access decisions", "nodes": [ { "id": "w33-con-homedsc", "title": "Researcher-home DSC", "sub": "Initial assessment; not final decision", "pos": [ 1, 1, 1 ] }, { "id": "w33-con-estdsc", "title": "Provider-establishment DSC", "sub": "Final project vetting / request", "pos": [ 2, 1, 1 ] }, { "id": "w33-con-request", "title": "Reasoned access request", "sub": "Scope, conditions and time period", "pos": [ 3, 1, 1 ] }, { "id": "w33-con-vetted", "title": "Vetted researchers", "sub": "Project-specific eligibility", "pos": [ 1, 2, 1 ] }, { "id": "w33-con-providers", "title": "VLOP / VLOSE providers", "sub": "Article 40 data-provider role", "pos": [ 2, 2, 1 ] } ], "edgeIds": [ "w33-con-e-homeassessment", "w33-con-e-vetting", "w33-con-e-fundingdisclosure", "w33-con-e-modalities", "w33-con-e-provide", "w33-con-e-amend" ] }, { "id": "ali-oversight", "title": "Who pays and who supervises AliExpress’s trustee", "description": "AliExpress must fund the trustee; the Commission approved Ankura and controls its mandate. QA and monitoring reports should travel back through specified channels.", "boundary": "Appointment is acquired; mandate, fee payments and actual monitoring reports are not. The trustee may recommend and report, but cannot make product, business or technical decisions. These commitments concern Article 40(12).", "topic": "Researcher access decisions", "nodes": [ { "id": "european-commission", "title": "European Commission", "sub": "Specified authority; no inferred case outcome", "pos": [ 1, 1, 1 ] }, { "id": "w33-con-ankura", "title": "Ankura Europe", "sub": "Approved trustee; no business veto", "pos": [ 2, 1, 1 ] }, { "id": "w33-con-ali", "title": "AliExpress", "sub": "Examined service; duties ≠ performance", "pos": [ 3, 1, 1 ] }, { "id": "w33-con-qa", "title": "AliExpress access QA", "sub": "Tests / semiannual metrics required", "pos": [ 2, 2, 1 ] } ], "edgeIds": [ "w33-con-e-approveankura", "w33-con-e-nominatepay", "w33-con-e-mandate", "w33-con-e-monitor", "w33-con-e-qareport", "w33-con-e-trustreport" ] }, { "id": "audit-and-tracker", "title": "A positive audit and no approved requests", "description": "FTI’s 2024/25 audit concluded positively on readiness, reported zero approvals and found the inspected applications incomplete or invalid. A later voluntary tracker provides a separate, unmatched account.", "boundary": "The two groups are not linked by identity or time. Tracker 70-headline / 68-table and 61-homepage / 60-table public counts remain unresolved. Complete registration is not legal validity; accepted is not delivered.", "topic": "Researcher access decisions", "nodes": [ { "id": "w33-con-fti", "title": "FTI Consulting LLP", "sub": "Named 2024/25 auditor", "pos": [ 1, 1, 1 ] }, { "id": "w33-con-ftiaudit", "title": "FTI audit opinion", "sub": "Positive readiness; not delivery", "pos": [ 2, 1, 1 ] }, { "id": "w33-con-appcohort", "title": "2024/25 applications", "sub": "Zero approvals; count redacted", "pos": [ 3, 1, 1 ] }, { "id": "w33-con-collab", "title": "DSA40 Collaboratory", "sub": "Application support / voluntary reports", "pos": [ 1, 2, 1 ] }, { "id": "w33-con-tracker", "title": "Application tracker", "sub": "70 headline / 68 table; not census", "pos": [ 2, 2, 1 ] }, { "id": "w33-con-aliopen", "title": "One AliExpress entry", "sub": "Open; identity and date unknown", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w33-con-e-ftiopinion", "w33-con-e-noapprovals", "w33-con-e-runs", "w33-con-e-alientry" ] }, { "id": "evaluation-gate", "title": "Who commissioned the £656,000 evaluation", "description": "The released CDDP material identifies Oliver Wyman and a £656,000 ex-VAT price. Payment depends on satisfactory deliverables; a later phase has an intended Treasury approval gate.", "boundary": "Price is not payment. Actual reports, acceptance certificate, paid invoice and exercised Treasury decision are not obtained. The released schedule has date inconsistencies and empty signature fields.", "topic": "Assurance versus results", "nodes": [ { "id": "w16-supp-dcms", "title": "DCMS", "sub": "Buyer / programme role", "pos": [ 1, 1, 1 ] }, { "id": "w33-con-oliver", "title": "Oliver Wyman", "sub": "Inventory-identified evaluator", "pos": [ 2, 1, 1 ] }, { "id": "w33-con-evalcontract", "title": "CDDP evaluation 102384", "sub": "£656,000 ex VAT price; not payment", "pos": [ 3, 1, 1 ] }, { "id": "w33-con-evaloutput", "title": "Evaluation / acceptance", "sub": "Commissioned; outputs not obtained", "pos": [ 1, 2, 1 ] }, { "id": "w33-con-fundgate", "title": "Phase continuation", "sub": "Approval specified; decision unknown", "pos": [ 2, 2, 1 ] }, { "id": "w33-con-treasury", "title": "HM Treasury", "sub": "Intended later-phase funding gate", "pos": [ 3, 2, 1 ] } ], "edgeIds": [ "w33-con-e-owcommission", "w33-con-e-satisfaction", "w33-con-e-evalscope", "w33-con-e-gatecondition", "w33-con-e-treasurygate" ] }, { "id": "osdi-assurance", "title": "From safety prototypes to reused evaluation methods", "description": "DCMS appointed PA for OSDI assurance. PUBLIC reports prototype testing and later credits OSDI practices in a separate STCF evaluation. A dated expert group supplied advice.", "boundary": "PA’s report is not acquired. PUBLIC’s TRL4/testing account is not production acceptance, nor proof it was the redacted CDDP subcontractor. Poppy Wood’s advisory role is not procurement veto power.", "topic": "Assurance versus results", "nodes": [ { "id": "w16-supp-dcms", "title": "DCMS", "sub": "Buyer / programme role", "pos": [ 1, 1, 1 ] }, { "id": "pa-consulting", "title": "PA Consulting", "sub": "External programme assurance", "pos": [ 2, 1, 1 ] }, { "id": "w33-con-osdi", "title": "OSDI", "sub": "Separate online-safety data initiative", "pos": [ 3, 1, 1 ] }, { "id": "w33-con-public", "title": "PUBLIC", "sub": "Supplier / programme brand", "pos": [ 1, 2, 1 ] }, { "id": "w33-con-sandbox", "title": "PUBLIC sandbox", "sub": "Reported TRL4 prototype / testing", "pos": [ 2, 2, 1 ] }, { "id": "w33-con-stcf", "title": "STCF Round 2", "sub": "Separate 2024 evaluation", "pos": [ 3, 2, 1 ] }, { "id": "w33-con-expertgroup", "title": "OSDI expert group", "sub": "Dated advice; no procurement veto", "pos": [ 1, 3, 1 ] }, { "id": "w12-gov-wood", "title": "Poppy Wood", "sub": "July 2021 Reset Tech advisory role", "pos": [ 2, 3, 1 ] } ], "edgeIds": [ "w33-con-e-paevaluate", "w33-con-e-jointscope", "w33-con-e-publicosdi", "w33-con-e-sandboxdelivery", "w33-con-e-methodreuse", "w33-con-e-advises", "w33-con-e-woodgroup" ] }, { "id": "overview-risk", "title": "AI foundations & creators", "description": "Follow the named relationships between the main participants in this part of the investigation.", "boundary": "Dates and relationship details are shown below the diagram. A historical connection is not necessarily a current one.", "topic": "The main branches", "nodes": [ { "id": "fli", "title": "Future of Life Institute", "sub": "EIN 47-1052538. Grantmaking, research and advocacy; affiliated asset transfers must not be added to persuasion spending. Named co-signatory of the June 2025 AI moratorium opposition letter.", "pos": [ 1, 1, 1 ] }, { "id": "vitalik", "title": "Vitalik Buterin", "sub": "2021 SHIB support confirmed. Total gift or liquidation value not established; FLI says gift unconditional and no decision role.", "pos": [ 2, 1, 1 ] }, { "id": "lightcone", "title": "Lightcone Foundation", "sub": "Nevada charity EIN 92-0636259. Not Lightcone Infrastructure/LessWrong or other similarly named entities.", "pos": [ 3, 1, 1 ] }, { "id": "tegmark", "title": "Max Erik Tegmark", "sub": "FLI founder/chair in current September 2026 biography; overlapping 2024 Lightcone governance.", "pos": [ 1, 2, 1 ] }, { "id": "cais", "title": "Center for AI Safety", "sub": "CAIS EIN 88-1751310. Separate from its Action Fund.", "pos": [ 2, 2, 1 ] }, { "id": "tallinn", "title": "Jaan Tallinn", "sub": "FLI and CAIS board bridge. Former DeepMind investor-director.", "pos": [ 3, 2, 1 ] }, { "id": "goodventures", "title": "Good Ventures", "sub": "Historical CAIS award inventory ends with October 2023 exit grant. Not established current sponsor.", "pos": [ 1, 3, 1 ] }, { "id": "frame", "title": "Frame Fellowship", "sub": "Shared creator training and placement infrastructure; independent and amplifier tracks.", "pos": [ 2, 3, 1 ] }, { "id": "talos", "title": "Talos Network", "sub": "Membership and donor disclosure; gifts not earmarked to Eva Behrens on the retrieved record.", "pos": [ 3, 3, 1 ] } ], "edgeIds": [ "e42", "e43", "e44", "e45", "e48", "e49", "e54", "e55", "e65", "e76", "e117", "e63a", "e63b", "e63c", "e63d", "w2-fli-e-lightcone25" ] }, { "id": "overview-industry", "title": "Industry & electoral advocacy", "description": "Follow the named relationships between the main participants in this part of the investigation.", "boundary": "Dates and relationship details are shown below the diagram. A historical connection is not necessarily a current one.", "topic": "The main branches", "nodes": [ { "id": "ltf", "title": "Leading the Future", "sub": "FEC C00916114; $75,788,224.10 receipts through June 30, 2026.", "pos": [ 1, 1, 1 ] }, { "id": "a16z", "title": "a16z Capital Management", "sub": "Two filed $25m contributions to LTF.", "pos": [ 2, 1, 1 ] }, { "id": "gregbrockman", "title": "Greg Brockman", "sub": "$12.5m individual contribution; not an OpenAI corporate payment.", "pos": [ 3, 1, 1 ] }, { "id": "annabrockman", "title": "Anna Brockman", "sub": "$12.5m individual contribution.", "pos": [ 1, 2, 1 ] }, { "id": "anthropic", "title": "Anthropic", "sub": "Announced restricted support for Public First Action.", "pos": [ 2, 2, 1 ] }, { "id": "publicfirst", "title": "Public First Action", "sub": "Separate from People First Marketing; Anthropic election-spending restriction.", "pos": [ 3, 2, 1 ] }, { "id": "buildai", "title": "Build American AI", "sub": "LTF affiliation and creator intermediary relationship reported by WIRED; no donor-to-creator allocation traced.", "pos": [ 1, 3, 1 ] }, { "id": "sm4", "title": "SM4", "sub": "Reported creator-pitch intermediary.", "pos": [ 2, 3, 1 ] } ], "edgeIds": [ "e144", "e145", "e146", "e148", "e149", "e150" ] }, { "id": "overview-rights", "title": "Digital-rights policy & funders", "description": "Follow the named relationships between the main participants in this part of the investigation.", "boundary": "Dates and relationship details are shown below the diagram. A historical connection is not necessarily a current one.", "topic": "The main branches", "nodes": [ { "id": "5rights", "title": "5Rights", "sub": "UK foundation advocating child-centred digital regulation. Donor records and a US lobbying contract are documented. Online-safety advocacy organisation founded by Beeban Kidron; California code sponsor.", "pos": [ 1, 1, 1 ] }, { "id": "omidyar", "title": "Omidyar Network", "sub": "Omidyar Network Fund Inc; charitable grants are not eBay corporate payments.", "pos": [ 2, 1, 1 ] }, { "id": "oak", "title": "Oak Foundation", "sub": "Common funder of 5Rights and WeProtect. Awards alone do not establish direction over recipient campaigns.", "pos": [ 3, 1, 1 ] }, { "id": "kidron", "title": "Beeban Kidron", "sub": "Founder of 5Rights and sponsor of the ControlAI-backed July 2026 amendment. Shared participation does not earmark 5Rights grants to ControlAI. ControlAI-listed supporter and Lords Amendment 84 sponsor. Other investigation maps her 5Rights role. 5Rights founder and listed sponsor of the 2026 AI emergency-shutdown amendment.", "pos": [ 1, 2, 1 ] }, { "id": "capitol-connection", "title": "Capitol Connection", "sub": "capconn LLC. Paid contractor to 5Rights for California AB2273 lobbying.", "pos": [ 2, 2, 1 ] }, { "id": "ab2273", "title": "California AB2273", "sub": "2022 Age-Appropriate Design Code campaign. Current enforcement and litigation are outside this investigation.", "pos": [ 3, 2, 1 ] }, { "id": "rothschild-foundation", "title": "Rothschild Foundation", "sub": "Specific UK charity 1138145; not a proxy for every institution with this name.", "pos": [ 1, 3, 1 ] }, { "id": "prudence", "title": "Prudence Trust", "sub": "Published multiyear core-work award to 5Rights.", "pos": [ 2, 3, 1 ] } ], "edgeIds": [ "e0", "e1", "e153", "e154", "e155", "e157", "e158", "e160", "e162", "e163", "w7-adv-e-chair" ] }, { "id": "overview-control", "title": "ControlAI & political access", "description": "Follow the named relationships between the main participants in this part of the investigation.", "boundary": "Dates and relationship details are shown below the diagram. A historical connection is not necessarily a current one.", "topic": "The main branches", "nodes": [ { "id": "controlai", "title": "ControlAI", "sub": "UK campaign traded by Secure Future Research. Existing major donors and grant conditions remain unresolved. AI advocacy organisation named in Hansard as the origin of emergency-shutdown proposals.", "pos": [ 1, 1, 1 ] }, { "id": "miotti", "title": "Andrea Miotti", "sub": "ControlAI founder/CEO; UK director and voting/appointment rights.", "pos": [ 2, 1, 1 ] }, { "id": "behrens", "title": "Eva Behrens", "sub": "ControlAI Special Projects Lead; UK director and voting/appointment rights.", "pos": [ 3, 1, 1 ] }, { "id": "conjecture", "title": "Conjecture", "sub": "2022 venture-backed alignment startup. Leahy described operational ending in March 2026, not corporate dissolution.", "pos": [ 1, 2, 1 ] }, { "id": "winga", "title": "Max Winga", "sub": "ControlAI Creator Outreach & Systems Lead; Frame past mentor.", "pos": [ 2, 2, 1 ] }, { "id": "sfr", "title": "Secure Future Research Ltd", "sub": "UK 15088415, limited by guarantee without share capital. Two named legal controllers; no exact 50/50 split inferred.", "pos": [ 3, 2, 1 ] }, { "id": "fabians", "title": "Fabian Society", "sub": "Named current/former leadership overlaps. No Society-to-ControlAI funding, commissioning or direction established.", "pos": [ 1, 3, 1 ] }, { "id": "sobel", "title": "Alex Sobel", "sub": "MP, historical Fabian executive; ControlAI-backed legislative activity.", "pos": [ 2, 3, 1 ] }, { "id": "casir", "title": "CASIR · draft", "sub": "Unincorporated, unfinished proposed fund. Its site claims affiliations without reciprocal confirmation; no verified donor commitment or executed grant.", "pos": [ 3, 3, 1 ] } ], "edgeIds": [ "e83", "e84", "e85", "e86", "e88", "e89", "e96", "e97", "e130", "e137", "e174" ] }, { "id": "overview-safety", "title": "Content & payment campaigns", "description": "Follow the named relationships between the main participants in this part of the investigation.", "boundary": "Dates and relationship details are shown below the diagram. A historical connection is not necessarily a current one.", "topic": "The main branches", "nodes": [ { "id": "ncose", "title": "NCOSE", "sub": "National Center on Sexual Exploitation; public co-signatory linking two distinct campaigns.", "pos": [ 1, 1, 1 ] }, { "id": "ai-moratorium-letter", "title": "June 2025 AI moratorium letter", "sub": "Shared letter/artifact. This map represents only selected signatories, not the full set of more than 140. A letter is not a directing organisation.", "pos": [ 2, 1, 1 ] }, { "id": "collective-shout", "title": "Collective Shout", "sub": "Australian campaigning organisation. Old corpus claim of Basic Religious Charity exemption was contradicted; no donor secrecy inference retained.", "pos": [ 3, 1, 1 ] }, { "id": "tankard-reist", "title": "Melinda Tankard Reist", "sub": "Collective Shout Movement Director; payment-letter signatory and trial stakeholder adviser.", "pos": [ 1, 2, 1 ] }, { "id": "rouse", "title": "Jon Rouse", "sub": "Payment-letter signatory and chair of the completed Australian trial's stakeholder advisory board.", "pos": [ 2, 2, 1 ] }, { "id": "payment-campaign", "title": "July 2025 payment campaign", "sub": "Open letter requested payment companies stop processing for targeted game platforms.", "pos": [ 3, 2, 1 ] }, { "id": "payment-processors", "title": "Payment-letter addressees", "sub": "Collection: PayPal, Mastercard, Visa, Paysafe and Discover leadership. Letter recipients, not a single legal entity or operational controller.", "pos": [ 1, 3, 1 ] }, { "id": "itch-io", "title": "itch.io", "sub": "Platform attributed processor scrutiny and adult-content deindexing to Collective Shout's campaign; private processor correspondence was not obtained.", "pos": [ 2, 3, 1 ] } ], "edgeIds": [ "e4", "e5", "e6", "e7", "e8", "e9", "e11", "e13" ] }, { "id": "overview-state", "title": "Regulators, alliances & trials", "description": "Follow the named relationships between the main participants in this part of the investigation.", "boundary": "Dates and relationship details are shown below the diagram. A historical connection is not necessarily a current one.", "topic": "The main branches", "nodes": [ { "id": "weprotect", "title": "WeProtect", "sub": "International alliance with government, industry and NGO board participation. UK and Dutch entities/accounts require careful separation. Participated in WEF digital-safety principles working group; separate from WEF.", "pos": [ 1, 1, 1 ] }, { "id": "home-office", "title": "UK Home Office", "sub": "Government records historical funding and a permanent WeProtect board seat.", "pos": [ 2, 1, 1 ] }, { "id": "european-commission", "title": "European Commission", "sub": "WeProtect board representation; proposal briefing and possible assistance requests recorded in July 2022 minutes.", "pos": [ 3, 1, 1 ] }, { "id": "wef", "title": "World Economic Forum", "sub": "Browne republication and NTI event links here; online-safety institutional links mapped separately. Convened Global Coalition for Digital Safety and related principles work. Digital Trust is a different WEF programme.", "pos": [ 1, 2, 1 ] }, { "id": "esafety", "title": "Australian eSafety", "sub": "Regulator with formal international policy relationships; age-assurance trial board is a different body.", "pos": [ 2, 2, 1 ] }, { "id": "inman-grant", "title": "Julie Inman Grant", "sub": "eSafety Commissioner; WeProtect board and WEF coalition involvement recorded in 2023–24 official report.", "pos": [ 3, 2, 1 ] }, { "id": "au-age-trial", "title": "Australian age-assurance trial", "sub": "Government-commissioned, contractor-run 2024–25 study; final report published 31 August 2025. New March–July minutes document objections and responses. April minutes clarify panel recruitment; implementation and operative review boundaries remain unresolved.", "pos": [ 1, 3, 1 ] }, { "id": "trial-board", "title": "Trial stakeholder advisory board", "sub": "Contractor trial stakeholder board; not a statutory regulator. Live roster does not establish an ongoing 2026 mandate. New March–July minutes document objections and responses. April minutes clarify panel recruitment; implementation and operative review boundaries remain unresolved.", "pos": [ 2, 3, 1 ] }, { "id": "avpa", "title": "Age Verification Providers Association", "sub": "Age-verification industry association. Its technical evidence is expressly credited in Ofcom’s January 2025 acceptance of email-based estimation; older Australian trial participation remains separately recorded.", "pos": [ 3, 3, 1 ] } ], "edgeIds": [ "e18", "e30", "e31", "e32", "e166", "e167", "e168", "e169", "e183", "n33", "w1-australia-edge-march-objections", "w1-australia-edge-unequal-draft-access" ] } ], "stories": [ { "slug": "the-big-picture", "title": "Who gets to operate the switches?", "deck": "Governments, funders, campaigners and companies exchange different kinds of power. The consequential moment is when someone turns another institution’s input into a decision.", "area": "The big picture", "people": [ "goodventures", "controlai", "oak", "ofcom", "mastercard", "stripe" ], "sections": [ { "title": "Money pays for capacity. Someone still chooses.", "text": [ "A donor can make research, campaigning or a legal challenge possible. An adviser can select the opportunities a donor sees. A campaign can supply a proposal. A regulator can adopt, narrow or reject it. A platform can then decide how to put a rule into practice. Following these decisions is more revealing than counting people who know one another.", "Good Ventures and its philanthropic adviser illustrate different allocation roles. Oak illustrates why a financial connection is not a complete political explanation: its grantees include both safety advocates and civil-liberties organisations resisting broad surveillance." ], "sources": [ { "title": "WeProtect Secretariat consultation · pp2–4", "url": "https://www.weprotect.org/wp-content/uploads/WPGA-consultation-response-EC-regulations-FINAL.pdf" }, { "title": "Oak annual report · PDF p38", "url": "https://oakfnd.org/wp-content/uploads/2026/06/Annual-report-2025-14_4.pdf" }, { "title": "Statewatch endorsement of EU CSA objection — September2023", "url": "https://statewatch.org/news/2023/september/statement-to-eu-countries-do-not-agree-to-mass-surveillance-proposal-warn-ngos/" }, { "title": "Apple v Home Secretary: public procedural judgment [2025] UKIPTrib 1", "url": "https://www.judiciary.uk/wp-content/uploads/2025/04/Apple-v-Secretary-of-State-for-the-Home-Department.pdf" }, { "title": "Liberty/Privacy International complaint announcement, 13 March 2025", "url": "https://www.libertyhumanrights.org.uk/issue/uk-governments-secret-apple-data-access-order-challenged-by-liberty-and-privacy-international/" }, { "title": "DFF August2026 grant announcement — June round", "url": "https://digitalfreedomfund.org/blog/dff-funds-16-new-legal-actions-for-digital-rights-2026-mid-year-update-on-new-grants/" }, { "title": "DFF one-year retrospective and reported March 2026 board adoption", "url": "https://digitalfreedomfund.org/blog/one-year-of-community-led-grant-decision-making-how-its-going-so-far/" }, { "title": "DFF Community Peer Group: current public process", "url": "https://digitalfreedomfund.org/community-peer-group/" } ] }, { "title": "The rules and the machinery belong to different hands.", "text": [ "Public authorities can impose legal duties. Companies operate payments, accounts and distribution. Advocacy organisations supply proposals and classifications. Their interests can align in one dispute and collide in another. These are overlapping coalitions with different resources, rather than permanent teams.", "Mastercard’s merchant-list rules permit onboarding after inquiry; Stripe’s published practice is generally disqualifying, with possible exceptions. The practical result depends on the recipient’s decision as well as the upstream rule." ], "sources": [ { "title": "eGumball G062863M: April2025 opinion and May6 modification", "url": "https://www.casemine.com/judgement/us/681cabf2d974c6edd357ad3b" }, { "title": "FilmOn.com v DoubleVerify, S244157: Supreme Court opinion", "url": "https://www.courts.ca.gov/opinions/archive/S244157.PDF" }, { "title": "Mastercard Security Rules and Procedures, Merchant Edition: Chapter11", "url": "https://www.mastercard.com/content/dam/mccom/shared/business/support/rules-pdfs/SPME-Manual.pdf" }, { "title": "Stripe: Terminated merchant files / MATCH acceptance practice", "url": "https://docs.stripe.com/disputes/match?locale=en-GB" } ] }, { "title": "Government can steer—and constrain—private power.", "text": [ "Ofcom’s decisions provide specific examples of selective uptake: a Full Fact contribution helped change a notice indicator; a Check My Ads request did not secure specific monetisation recommendations in one measure. In another domain, the CFPB required Block to provide notices and decide customer appeals within ten business days.", "Our interpretation is that power often accumulates across these interfaces. Proving a particular restriction still requires its own chain: who requested it, who had authority, who acted and what happened. A shared donor or compatible objective is a starting point for that work." ], "sources": [ { "title": "Response to consultation: Online Safety Act — Additional Safety Measures", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/check-my-ads-institute.pdf?v=409158" }, { "title": "Ofcom crisis response protocol — June 2026 statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/online-safety/protecting-children/illegal-harms/crisis-response-protocol-online-safety-additional-safety-measures.pdf?v=419248" }, { "title": "Full Fact's Submission to Ofcom's Additional Safety Measures Consultation", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/full-fact.pdf?v=409176" }, { "title": "Ofcom Part 5 guidance on highly effective age assurance and other duties", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/statement-age-assurance-and-childrens-access/guidance-on-highly-effective-age-assurance-and-other-part-5-duties.pdf?v=388810" }, { "title": "Ofcom Age Assurance and Children's Access statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/statement-age-assurance-and-childrens-access/statement-age-assurance-and-childrens-access.pdf?v=397036" }, { "title": "CFPB — Block consent order 2025-CFPB-0001", "url": "https://files.consumerfinance.gov/f/documents/cfpb_block-inc-consent-order_2025-01.pdf" }, { "title": "Block Q2 2026 Form 10-Q — scoped subsequent-matter disclosure", "url": "https://www.sec.gov/Archives/edgar/data/1512673/000162828026053368/xyz-20260630.htm" }, { "title": "California published proposed Block judgment", "url": "https://oag.ca.gov/system/files/attachments/press-docs/People%20v%20Block%20Judgment.pdf" }, { "title": "Oregon signed-party Block stipulated judgment packet", "url": "https://www.doj.state.or.us/wp-content/uploads/2026/07/2026-07-08-Cash-App-Multistate-Stipulated-General-Judgment-and-Money-Award-OR.pdf" } ] }, { "title": "The most revealing gap is often the return journey.", "text": [ "An adverse assessment can travel through a network quickly. What happens when it is corrected? Does the correction reach the buyer, the platform or the exported copy? Does someone reopen the decision? Does access return?", "A published appeal process, a software deletion mechanism and an actual restored account answer different questions. The missing evidence is often an ordinary case file or instruction trail that shows those events together." ], "sources": [ { "title": "Tech Coalition — Lantern transparency report 2025", "url": "https://technologycoalition.org/wp-content/uploads/lantern-transparency-report-2025.pdf" }, { "title": "Altitude pinned source — tcap-adapter", "url": "https://raw.githubusercontent.com/Jigsaw-Code/altitude/496951f7d7a83508e9fac5709d4db2e91c654685/signal-service/src/importers/tcap_api.py" }, { "title": "TCAP — current per-hash feedback documentation", "url": "https://docs.terrorismanalytics.org/docs/api-services/hash-feedback" }, { "title": "TCAP — current content-hash update documentation", "url": "https://docs.terrorismanalytics.org/docs/api-services/content-hash-list" }, { "title": "OHF — registered voting and appointment control", "url": "https://find-and-update.company-information.service.gov.uk/company/11656320/persons-with-significant-control" }, { "title": "CFPB — Block consent order 2025-CFPB-0001", "url": "https://files.consumerfinance.gov/f/documents/cfpb_block-inc-consent-order_2025-01.pdf" }, { "title": "Block Q2 2026 Form 10-Q — scoped subsequent-matter disclosure", "url": "https://www.sec.gov/Archives/edgar/data/1512673/000162828026053368/xyz-20260630.htm" }, { "title": "California published proposed Block judgment", "url": "https://oag.ca.gov/system/files/attachments/press-docs/People%20v%20Block%20Judgment.pdf" }, { "title": "Oregon signed-party Block stipulated judgment packet", "url": "https://www.doj.state.or.us/wp-content/uploads/2026/07/2026-07-08-Cash-App-Multistate-Stipulated-General-Judgment-and-Money-Award-OR.pdf" } ] } ] }, { "slug": "money-behind-ai-advocacy", "title": "Who funds the AI debate?", "deck": "Large grants, creator programmes and policy campaigns are visible. Their arrangements differ in how much control they give a sponsor.", "area": "AI & money", "people": [ "goodventures", "cais", "fli", "lightcone", "vitalik", "tegmark", "controlai", "frame", "miotti", "behrens", "winga" ], "sections": [ { "title": "A grant is a resource with a purpose.", "text": [ "FLI’s 2025 audited accounts report $45,197,838 in grants from Lightcone Foundation. That establishes scale. It does not tell us how much was spent on public persuasion or who approved a specific video. Asset transfers between related bodies also cannot be counted as independent donations each time they move." ], "sources": [ { "title": "FLI 2022 Form990", "url": "https://pdf.guidestar.org/PDF_Images/2022/471/052/2022-471052538-202510309349301606-.pdf" }, { "title": "FLI 2025 audited financial statements", "url": "https://www.lobbyregister.bundestag.de/media/ff/31/796351/Future-of-Life-Institute-2025-Audited-Financial-Statements-Final.pdf" } ] }, { "title": "Creator support comes in different forms.", "text": [ "People First’s published influencer terms include approval, revisions and conditional payment. Frame describes an independent creator track alongside work to institutional briefs, and denies funder control over content or selection. A creator acknowledged FLI, Frame and BlueDot support while asserting editorial independence.", "ControlAI and Rational Animations publicly acknowledged a partnership; its fee and agreement were not obtained. The sponsor behind Sabine Hossenfelder’s original account remains unidentified. These are different arrangements, not interchangeable evidence of script control." ], "sources": [ { "title": "Protect What Is Human creator recruitment", "url": "https://join.peoplefirst.cc/campaign/protectwhatshuman/" }, { "title": "People First Influencer Terms of Use", "url": "https://peoplefirst.cc/terms/" }, { "title": "ControlAI publication announcing Rational Animations collaboration", "url": "https://blog.controlai.org/p/watch-now-what-happens-if-ai-just" }, { "title": "Frame public FAQ content, shipped script index-CUMxxB1e.js", "url": "https://www.framefellowship.com/assets/index-CUMxxB1e.js" }, { "title": "Future Tense: The Dark Philosophy Driving AI Development", "url": "https://www.youtube.com/watch?v=Mky0HtSNE4M" }, { "title": "Rational Animations: What happens if AI just keeps getting smarter?", "url": "https://www.youtube.com/watch?v=0bnxF9YfyFI" } ] }, { "title": "Selection can influence the conversation.", "text": [ "Funding people who already share a concern can expand that concern’s reach without anyone rewriting a script. This is an interpretation of how selection works; identifying control over an individual output requires its actual agreement and approval history.", "ControlAI’s UK company articles put formal authority in its directors. Public filings help answer who can manage the organisation, while leaving major donor conditions unresolved." ], "sources": [ { "title": "Secure Future Research Ltd filed Articles of Association", "url": "https://find-and-update.company-information.service.gov.uk/company/15088415/filing-history/MzQxNDkzMjc4OWFkaXF6a2N4/document?download=0&format=pdf" }, { "title": "Control AI US Inc initial 2025 Form 990-EZ, IRS public XML", "url": "https://apps.irs.gov/pub/epostcard/990/xml/2026/2026_TEOS_XML_06A.zip" } ] }, { "title": "A campaign has a production chain.", "text": [ "The Global Call for AI Red Lines names CeSIA, The Future Society and CHAI as organisers, with ControlAI and FLI among partners providing feedback. Evan Hubinger is listed as a signatory. Those are specific roles in a shared campaign.", "The Call is different from A Narrow Path. A signature on one does not endorse every proposal in the other, and a proposed oversight institution is not an operating government body." ], "sources": [ { "title": "Global Call acknowledgements", "url": "https://red-lines.ai/acknowledgements" }, { "title": "ControlAI: International AI red lines", "url": "https://blog.controlai.org/p/red-lines" }, { "title": "Global Call individual signatories", "url": "https://red-lines.ai/individuals" } ] } ] }, { "slug": "one-donor-different-agendas", "title": "One donor. Conflicting agendas.", "deck": "Oak Foundation funds both online-safety advocacy and organisations challenging surveillance. Follow the grants, then follow each recipient’s choices.", "area": "Safety & rights", "people": [ "oak", "alan-parker", "5rights", "kidron", "weprotect", "w31-fa-statewatch", "w31-fa-dff", "w30-class-cma" ], "sections": [ { "title": "A common funding source is a real connection.", "text": [ "Oak’s reports identify multiyear support for 5Rights, WeProtect, Check My Ads, Statewatch and Digital Freedom Fund. These awards give organisations capacity and deserve scrutiny on their own terms: amount, duration, purpose and decision rights." ], "sources": [ { "title": "WeProtect Secretariat consultation · pp2–4", "url": "https://www.weprotect.org/wp-content/uploads/WPGA-consultation-response-EC-regulations-FINAL.pdf" }, { "title": "Oak annual report · PDF p38", "url": "https://oakfnd.org/wp-content/uploads/2026/06/Annual-report-2025-14_4.pdf" }, { "title": "Statewatch endorsement of EU CSA objection — September2023", "url": "https://statewatch.org/news/2023/september/statement-to-eu-countries-do-not-agree-to-mass-surveillance-proposal-warn-ngos/" }, { "title": "Apple v Home Secretary: public procedural judgment [2025] UKIPTrib 1", "url": "https://www.judiciary.uk/wp-content/uploads/2025/04/Apple-v-Secretary-of-State-for-the-Home-Department.pdf" }, { "title": "Liberty/Privacy International complaint announcement, 13 March 2025", "url": "https://www.libertyhumanrights.org.uk/issue/uk-governments-secret-apple-data-access-order-challenged-by-liberty-and-privacy-international/" } ] }, { "title": "The recipients disagree about the means.", "text": [ "Statewatch endorsed objections to broad scanning and possible identity requirements in the EU child-abuse proposal in September 2023. WeProtect’s Secretariat supported mandatory detection with safeguards in an earlier version of that debate. The positions diverge, although they concern different versions and dates.", "That is useful evidence against explaining every recipient’s action through a single donor programme. Shared funding does not erase disagreement or tell us which organisation will prevail." ], "sources": [ { "title": "WeProtect Secretariat consultation · pp2–4", "url": "https://www.weprotect.org/wp-content/uploads/WPGA-consultation-response-EC-regulations-FINAL.pdf" }, { "title": "Oak annual report · PDF p38", "url": "https://oakfnd.org/wp-content/uploads/2026/06/Annual-report-2025-14_4.pdf" }, { "title": "Statewatch endorsement of EU CSA objection — September2023", "url": "https://statewatch.org/news/2023/september/statement-to-eu-countries-do-not-agree-to-mass-surveillance-proposal-warn-ngos/" }, { "title": "Apple v Home Secretary: public procedural judgment [2025] UKIPTrib 1", "url": "https://www.judiciary.uk/wp-content/uploads/2025/04/Apple-v-Secretary-of-State-for-the-Home-Department.pdf" }, { "title": "Liberty/Privacy International complaint announcement, 13 March 2025", "url": "https://www.libertyhumanrights.org.uk/issue/uk-governments-secret-apple-data-access-order-challenged-by-liberty-and-privacy-international/" } ] }, { "title": "Money can sustain action that already exists.", "text": [ "Digital Freedom Fund reports a June 2026 grant supporting Liberty’s challenge concerning the alleged Apple technical-capability notice. Liberty filed in March 2025. The dates support a role in sustaining litigation; they do not support a claim that this grant originated the filing.", "DFF also describes external peers taking part in selection. A donor’s grant, peer selection and the execution of an award are separate decisions. Actual terms and approval records would reveal more about where each party’s discretion begins and ends." ], "sources": [ { "title": "WeProtect Secretariat consultation · pp2–4", "url": "https://www.weprotect.org/wp-content/uploads/WPGA-consultation-response-EC-regulations-FINAL.pdf" }, { "title": "Oak annual report · PDF p38", "url": "https://oakfnd.org/wp-content/uploads/2026/06/Annual-report-2025-14_4.pdf" }, { "title": "Statewatch endorsement of EU CSA objection — September2023", "url": "https://statewatch.org/news/2023/september/statement-to-eu-countries-do-not-agree-to-mass-surveillance-proposal-warn-ngos/" }, { "title": "Apple v Home Secretary: public procedural judgment [2025] UKIPTrib 1", "url": "https://www.judiciary.uk/wp-content/uploads/2025/04/Apple-v-Secretary-of-State-for-the-Home-Department.pdf" }, { "title": "Liberty/Privacy International complaint announcement, 13 March 2025", "url": "https://www.libertyhumanrights.org.uk/issue/uk-governments-secret-apple-data-access-order-challenged-by-liberty-and-privacy-international/" }, { "title": "DFF August2026 grant announcement — June round", "url": "https://digitalfreedomfund.org/blog/dff-funds-16-new-legal-actions-for-digital-rights-2026-mid-year-update-on-new-grants/" }, { "title": "DFF one-year retrospective and reported March 2026 board adoption", "url": "https://digitalfreedomfund.org/blog/one-year-of-community-led-grant-decision-making-how-its-going-so-far/" }, { "title": "DFF Community Peer Group: current public process", "url": "https://digitalfreedomfund.org/community-peer-group/" } ] } ] }, { "slug": "from-safety-campaign-to-law", "title": "How a safety proposal becomes a rule", "deck": "The UK and Australia offer something stronger than a list of shared contacts: documents showing policy ideas entering official consideration—and changing along the way.", "area": "Rules & government", "people": [ "ofcom", "avpa", "esafety", "digital-duty-of-care", "w10-duty-reset-australia", "w30-class-cma" ], "sections": [ { "title": "Policy travels in documents.", "text": [ "Carnegie-associated duty-of-care work, collaboration with Reset and citations in Australia’s statutory review connect a developed model to official consideration. Funding and commissioned work sustain parts of this policy infrastructure. Later funding, however, cannot be assigned authorship of an earlier idea." ], "sources": [ { "title": "Minister Wells: My feed, my way —8September2026", "url": "https://minister.infrastructure.gov.au/wells/media-release/my-feed-my-way" }, { "title": "Carnegie UK submission to Australian House Select Committee — January 2022", "url": "https://www1.essex.ac.uk/documents/research/carnegie/Submission-to-the-House-Select-Committee-on-Social-Media-and-Online-Safety.pdf" }, { "title": "Exposure draft: Online Safety Amendment (Digital Duty of Care) Bill2026 —8September2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/exposure-draft-online-safety-amendment-digital-duty-of-care-bill-2026-september2026.pdf" }, { "title": "Reset.Tech Australia: statutory-review submission70 —2024", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/osar-submission-70-reset-tech-australia.pdf" }, { "title": "Published updated Online Safety Act statutory review —February2025", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/report-of-the-statutory-review-of-the-online-safety-act-2021-february-2025.pdf" }, { "title": "Carnegie United Kingdom Trust:2021 accounts", "url": "https://carnegieuk.org/wp-content/uploads/2024/10/Annual-Report-and-Accounts-2021-2.pdf" }, { "title": "Carnegie United Kingdom Trust:2023 accounts", "url": "https://carnegieuk.org/wp-content/uploads/2024/10/Annual-report-and-accounts-31-December-2023-2.pdf" }, { "title": "Reset: Functioning or Failing? —May2024", "url": "https://au.reset.tech/uploads/Reset.Tech-Functioning-or-Failing-Report-V2C.pdf" }, { "title": "McKinnon: two Reset research commissions", "url": "https://mckinnon.co/democracy/social-media-mis-disinformation" } ] }, { "title": "The Australian draft narrows one requested power.", "text": [ "Reset sought vetted research access including third-sector organisations. The 8 September 2026 exposure draft instead specifies Australian-university employment and approval under prospective rules. That is a meaningful narrowing of eligibility.", "The proposed duty divides authority among a legislative framework, ministerial rules and eSafety. Its adult-consent provision limits what the duty requires; it does not create a universal adult veto. The text examined here does not establish universal identification." ], "sources": [ { "title": "Minister Wells: My feed, my way —8September2026", "url": "https://minister.infrastructure.gov.au/wells/media-release/my-feed-my-way" }, { "title": "Carnegie UK submission to Australian House Select Committee — January 2022", "url": "https://www1.essex.ac.uk/documents/research/carnegie/Submission-to-the-House-Select-Committee-on-Social-Media-and-Online-Safety.pdf" }, { "title": "Exposure draft: Online Safety Amendment (Digital Duty of Care) Bill2026 —8September2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/exposure-draft-online-safety-amendment-digital-duty-of-care-bill-2026-september2026.pdf" }, { "title": "Reset.Tech Australia: statutory-review submission70 —2024", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/osar-submission-70-reset-tech-australia.pdf" }, { "title": "Published updated Online Safety Act statutory review —February2025", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/report-of-the-statutory-review-of-the-online-safety-act-2021-february-2025.pdf" }, { "title": "Carnegie United Kingdom Trust:2021 accounts", "url": "https://carnegieuk.org/wp-content/uploads/2024/10/Annual-Report-and-Accounts-2021-2.pdf" }, { "title": "Carnegie United Kingdom Trust:2023 accounts", "url": "https://carnegieuk.org/wp-content/uploads/2024/10/Annual-report-and-accounts-31-December-2023-2.pdf" }, { "title": "Reset: Functioning or Failing? —May2024", "url": "https://au.reset.tech/uploads/Reset.Tech-Functioning-or-Failing-Report-V2C.pdf" }, { "title": "McKinnon: two Reset research commissions", "url": "https://mckinnon.co/democracy/social-media-mis-disinformation" } ] }, { "title": "Age assurance is a method question as well as a power question.", "text": [ "Ofcom’s acceptance of email-based age estimation followed technical evidence, including AVPA input. Accepting a method does not certify a vendor or award a procurement contract. Who supplies evidence, who chooses the threshold and who sells a service are distinct roles." ], "sources": [ { "title": "Response to consultation: Online Safety Act — Additional Safety Measures", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/check-my-ads-institute.pdf?v=409158" }, { "title": "Ofcom crisis response protocol — June 2026 statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/online-safety/protecting-children/illegal-harms/crisis-response-protocol-online-safety-additional-safety-measures.pdf?v=419248" }, { "title": "Full Fact's Submission to Ofcom's Additional Safety Measures Consultation", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/full-fact.pdf?v=409176" }, { "title": "Ofcom Part 5 guidance on highly effective age assurance and other duties", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/statement-age-assurance-and-childrens-access/guidance-on-highly-effective-age-assurance-and-other-part-5-duties.pdf?v=388810" }, { "title": "Ofcom Age Assurance and Children's Access statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/statement-age-assurance-and-childrens-access/statement-age-assurance-and-childrens-access.pdf?v=397036" } ] }, { "title": "A regulator can take one recommendation and leave another.", "text": [ "Ofcom attributes a crisis notice-indicator change to Full Fact’s submission and earlier meeting questions. It also considered Check My Ads’ crisis-revenue request without recommending specific monetisation practices in that measure.", "These decisions show influence and limits at the same time. They are a firmer basis for analysis than assuming either that consultation means control or that formal regulatory independence means nobody influenced the outcome." ], "sources": [ { "title": "Response to consultation: Online Safety Act — Additional Safety Measures", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/check-my-ads-institute.pdf?v=409158" }, { "title": "Ofcom crisis response protocol — June 2026 statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/online-safety/protecting-children/illegal-harms/crisis-response-protocol-online-safety-additional-safety-measures.pdf?v=419248" }, { "title": "Full Fact's Submission to Ofcom's Additional Safety Measures Consultation", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/full-fact.pdf?v=409176" }, { "title": "Ofcom Part 5 guidance on highly effective age assurance and other duties", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/statement-age-assurance-and-childrens-access/guidance-on-highly-effective-age-assurance-and-other-part-5-duties.pdf?v=388810" }, { "title": "Ofcom Age Assurance and Children's Access statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/statement-age-assurance-and-childrens-access/statement-age-assurance-and-childrens-access.pdf?v=397036" } ] } ] }, { "slug": "payments-and-blacklists", "title": "When a classification reaches your wallet", "deck": "A risk list, a payment-network rule and a processor’s decision can each contribute to a restriction. They are different steps, often controlled by different institutions.", "area": "Payments & platforms", "people": [ "mastercard", "stripe", "itch-io", "collective-shout", "w32-gdi-gdi", "w22-sig-lantern", "w22-sig-cashapp" ], "sections": [ { "title": "The recipient can tighten the rule.", "text": [ "Mastercard’s August 2026 MATCH rules require inquiry into listed merchants but permit onboarding. Stripe publishes a generally disqualifying practice with possible exceptions. A merchant can therefore face a restriction that depends on the processor’s own decision as well as the existence of a network record." ], "sources": [ { "title": "eGumball G062863M: April2025 opinion and May6 modification", "url": "https://www.casemine.com/judgement/us/681cabf2d974c6edd357ad3b" }, { "title": "FilmOn.com v DoubleVerify, S244157: Supreme Court opinion", "url": "https://www.courts.ca.gov/opinions/archive/S244157.PDF" }, { "title": "Mastercard Security Rules and Procedures, Merchant Edition: Chapter11", "url": "https://www.mastercard.com/content/dam/mccom/shared/business/support/rules-pdfs/SPME-Manual.pdf" }, { "title": "Stripe: Terminated merchant files / MATCH acceptance practice", "url": "https://docs.stripe.com/disputes/match?locale=en-GB" } ] }, { "title": "Information-sharing can prompt separate action.", "text": [ "A Lantern transparency case study reports Meta signals followed by investigations and action at Cash App and Xbox. The underlying case files were not inspected, so these remain operator-reported outcomes. The sequence does not establish that Lantern caused Steam removals.", "The useful question is what the receiving company did with the information, under which rules, and whether it would revisit the decision if the information changed." ], "sources": [ { "title": "Tech Coalition — Lantern transparency report 2025", "url": "https://technologycoalition.org/wp-content/uploads/lantern-transparency-report-2025.pdf" } ] }, { "title": "Historical distribution is not necessarily current distribution.", "text": [ "GDI’s risk information had a documented route into Oracle advertising tools, but the relevant Oracle service ended in September 2024. IAS later said it discontinued limited past GDI inputs. Drawing either relationship as an uninterrupted current route would mislead readers.", "A review process might change a GDI assessment without automatically restoring every advertising relationship. The customer’s receipt of the correction and subsequent decision remain separate facts to establish." ], "sources": [ { "title": "Clare Melford, co-founder and Chief Executive Officer, GDI: written evidence FON0071", "url": "https://committees.parliament.uk/writtenevidence/130306/html/" }, { "title": "IAS Enhancements to Context Control Avoidance", "url": "https://integralads.com/news/ias-enhancements-to-context-control-avoidance/" }, { "title": "Oracle Contextual Intelligence Segment Guide", "url": "https://www.oracle.com/a/ocom/docs/oracle-contextual-intelligence-segment-guide.pdf" }, { "title": "Oracle Advertising EOL Frequently Asked Questions", "url": "https://www.oracle.com/a/ocom/docs/oracle-advertising-end-of-life-faqs.pdf" } ] }, { "title": "There are constraints on the companies too.", "text": [ "The January 2025 CFPB order requires Block to provide specified restriction and appeal notices, decide appeals within ten business days and change status as appropriate. That narrows corporate discretion. It does not prove that every claimant recovered an account or received money." ], "sources": [ { "title": "CFPB — Block consent order 2025-CFPB-0001", "url": "https://files.consumerfinance.gov/f/documents/cfpb_block-inc-consent-order_2025-01.pdf" }, { "title": "Block Q2 2026 Form 10-Q — scoped subsequent-matter disclosure", "url": "https://www.sec.gov/Archives/edgar/data/1512673/000162828026053368/xyz-20260630.htm" }, { "title": "California published proposed Block judgment", "url": "https://oag.ca.gov/system/files/attachments/press-docs/People%20v%20Block%20Judgment.pdf" }, { "title": "Oregon signed-party Block stipulated judgment packet", "url": "https://www.doj.state.or.us/wp-content/uploads/2026/07/2026-07-08-Cash-App-Multistate-Stipulated-General-Judgment-and-Money-Award-OR.pdf" } ] } ] }, { "slug": "the-access-test", "title": "A positive audit. Nobody approved.", "deck": "An AliExpress audit reported readiness for public-data access while recording no approved applications. What does that result actually establish?", "area": "Rules & government", "people": [ "european-commission", "w32-gdi-gdi" ], "sections": [ { "title": "Separate the routes to access.", "text": [ "The Digital Services Act provides different mechanisms for project-specific vetted research access, access to public data and priority handling of trusted-flagger notices. These routes have different conditions and consequences. Funding for research does not itself confer access to a platform." ], "sources": [ { "title": "Directive2019/790 — incorporated research-organisation definition", "url": "https://eur-lex.europa.eu/eli/dir/2019/790/oj/eng" }, { "title": "Delegated Regulation2025/2050 — research-data access", "url": "https://eur-lex.europa.eu/eli/reg_del/2025/2050/oj/eng" }, { "title": "Digital Services Act 2022/2065 — Articles40 and22", "url": "https://eur-lex.europa.eu/eli/reg/2022/2065/oj/eng" } ] }, { "title": "Payment and approval belong to different parties.", "text": [ "Under AliExpress’s June 2025 commitments, the company nominates and pays a monitoring trustee, while the Commission approves and supervises the role. The payment relationship does not transfer the Commission’s approval power to the company." ], "sources": [ { "title": "C(2025)3982 final — AliExpress Article71 decision and annex", "url": "https://ec.europa.eu/newsroom/dae/redirection/document/124718" }, { "title": "Commission notice — Ankura monitoring-trustee approval", "url": "https://ec.europa.eu/newsroom/dae/redirection/document/124727" } ] }, { "title": "Read the result behind the opinion.", "text": [ "The FTI audit for the year ending June 2025 gave a positive opinion on the public-data access provision while reporting no approved applications. It said the applications examined were incomplete or invalid. This supports a narrower claim about assessed readiness, rather than proof that useful access was delivered.", "Legitimate rejection of deficient applications and an unnecessarily difficult access process are both possible explanations. Application files, responses, opportunities to remedy defects and actual delivery would distinguish them." ], "sources": [ { "title": "FTI AliExpress2024/25 Article37 audit — selected40(12) evidence", "url": "https://files.alicdn.com/tpsservice/118cffe16ea7541786b8e8ccf9365ce6.pdf" } ] }, { "title": "Trackers help, but do not decide the dispute.", "text": [ "The DSA40 Collaboratory collects voluntary application reports and supports researchers. Its September 2026 tracker contained inconsistent totals, and a public AliExpress entry lacked enough identifying detail to join it to the audit cohort.", "An applicant’s complaint is a lead to investigate. A platform’s positive audit is a claim to examine. Neither substitutes for the paired application and response." ], "sources": [ { "title": "DSA40 Collaboratory — entry and support mission", "url": "https://dsa40collaboratory.eu/" }, { "title": "Weizenbaum Policy Paper14 — Data Access, September2025", "url": "https://www.weizenbaum-institut.de/media/Publikationen/Weizenbaum_Policy_Paper/Weizenbaum_Policy_Paper_14.pdf" }, { "title": "DSA40 Tracker Insights —8September2026 snapshot", "url": "https://dsa40collaboratory.eu/tracker-insights/" } ] } ] }, { "slug": "who-builds-government-tools", "title": "Who gets paid to build the machinery?", "deck": "Public procurement connects policy ambitions with consultants, prototypes and evaluation. A budget announcement is only the beginning of that story.", "area": "Rules & government", "people": [ "ofcom" ], "sections": [ { "title": "A price, a budget and a payment are different facts.", "text": [ "The Counter Disinformation Data Project’s Oliver Wyman evaluation commission disclosed a £656,000 price excluding VAT, evaluation criteria and a Treasury-linked feasibility gate. The £8.4 million programme allocation was a separate budget figure. Adding these together as separate expenditure would double-count parts of the same programme.", "The acquired documents establish commissioning and conditions. They do not establish cash paid, the final evaluation or whether the funding gate was exercised." ], "sources": [ { "title": "CDDP FOI original bundle: selected business/contract/IP passages", "url": "https://assets.publishing.service.gov.uk/media/66803d6baec8650b10090192/FOI2023-18509_Document_bundle.pdf" }, { "title": "Official CDDP FOI release: inventory and withholding", "url": "https://www.gov.uk/government/publications/foi2024-18509-the-counter-disinformation-data-project/response-the-counter-disinformation-data-project" }, { "title": "Treasury Shared Outcomes Fund Round2 — CDDP entry", "url": "https://assets.publishing.service.gov.uk/media/6178283be90e07198108160d/Shared_Outcomes_Fund_Round_2_-_Pilot_Project_Summaries_-_FINAL.pdf" } ] }, { "title": "Advisers, suppliers and evaluators play different roles.", "text": [ "The Online Safety Data Initiative involved a CDEI-convened advisory group, suppliers including Faculty and PUBLIC, and PA Consulting programme evaluation. PUBLIC described a prototype sandbox for assessing classifiers. Its later work explicitly credited reuse of OSDI practices.", "These records connect people, programmes and methods. They do not establish that a prototype became a production system or identify every redacted subcontractor." ], "sources": [ { "title": "CDDP FOI original bundle: selected business/contract/IP passages", "url": "https://assets.publishing.service.gov.uk/media/66803d6baec8650b10090192/FOI2023-18509_Document_bundle.pdf" }, { "title": "CDEI OSDI expert-group account —2July2021", "url": "https://onlinesafetydata.blog.gov.uk/2021/07/02/cdei-convenes-expert-group-to-advise-on-online-safety-data-initiative-project/" }, { "title": "OSDI official programme description", "url": "https://onlinesafetydata.blog.gov.uk/about-us/" }, { "title": "OSDI appoints external evaluator PA Consulting", "url": "https://onlinesafetydata.blog.gov.uk/2021/09/23/working-with-an-independent-evaluator/" }, { "title": "PUBLIC reports participation in CDDP delivery", "url": "https://www.public.io/blog-post/better-collaboration-bigger-impact-how-data-sharing-accelerates-innovation" }, { "title": "PUBLIC G-Cloud14 service definition — OSDI sandbox case study", "url": "https://assets.applytosupply.digitalmarketplace.service.gov.uk/g-cloud-14/documents/706515/179204743065460-service-definition-document-2024-05-03-1318.pdf" }, { "title": "PUBLIC evaluation of Safety Tech Challenge Fund Round2", "url": "https://assets.publishing.service.gov.uk/media/6707a7a3e84ae1fd8592f125/safety_tech_challenge_fund_2_evaluation.pdf" } ] }, { "title": "Find the decision after the deliverable.", "text": [ "The most revealing next document would be an acceptance decision paired with the actual evaluation: what was tested, what failed, what changed and whether funding or deployment followed. A commissioned report and an effective government capability are different outcomes." ], "sources": [ { "title": "CDDP FOI original bundle: selected business/contract/IP passages", "url": "https://assets.publishing.service.gov.uk/media/66803d6baec8650b10090192/FOI2023-18509_Document_bundle.pdf" }, { "title": "Official CDDP FOI release: inventory and withholding", "url": "https://www.gov.uk/government/publications/foi2024-18509-the-counter-disinformation-data-project/response-the-counter-disinformation-data-project" }, { "title": "Treasury Shared Outcomes Fund Round2 — CDDP entry", "url": "https://assets.publishing.service.gov.uk/media/6178283be90e07198108160d/Shared_Outcomes_Fund_Round_2_-_Pilot_Project_Summaries_-_FINAL.pdf" }, { "title": "CDDP FOI original bundle: selected business/contract/IP passages", "url": "https://assets.publishing.service.gov.uk/media/66803d6baec8650b10090192/FOI2023-18509_Document_bundle.pdf" }, { "title": "CDEI OSDI expert-group account —2July2021", "url": "https://onlinesafetydata.blog.gov.uk/2021/07/02/cdei-convenes-expert-group-to-advise-on-online-safety-data-initiative-project/" }, { "title": "OSDI official programme description", "url": "https://onlinesafetydata.blog.gov.uk/about-us/" }, { "title": "OSDI appoints external evaluator PA Consulting", "url": "https://onlinesafetydata.blog.gov.uk/2021/09/23/working-with-an-independent-evaluator/" }, { "title": "PUBLIC reports participation in CDDP delivery", "url": "https://www.public.io/blog-post/better-collaboration-bigger-impact-how-data-sharing-accelerates-innovation" }, { "title": "PUBLIC G-Cloud14 service definition — OSDI sandbox case study", "url": "https://assets.applytosupply.digitalmarketplace.service.gov.uk/g-cloud-14/documents/706515/179204743065460-service-definition-document-2024-05-03-1318.pdf" }, { "title": "PUBLIC evaluation of Safety Tech Challenge Fund Round2", "url": "https://assets.publishing.service.gov.uk/media/6707a7a3e84ae1fd8592f125/safety_tech_challenge_fund_2_evaluation.pdf" } ] } ] } ], "questions": [ { "id": "w1-icmec-lead-1", "question": "Which funds, staff allocations and approvals supported ICMEC age-assurance drafting and advocacy in 2024–2026?", "why": "Unrestricted programme work vs earmarked drafting support vs commercially specified deliverable", "evidence": "Age-assurance project ledger, grant schedules, time allocations and approval record,2024–26", "status": "open", "people": [], "sources": [ { "title": "ICMEC consolidated financial statements 2024", "url": "https://cdn.icmec.org/wp-content/uploads/2026/03/ICMEC-2024-Consolidated-FS.pdf" }, { "title": "Device-Based Age Assurance technical whitepaper", "url": "https://cdn.icmec.org/wp-content/uploads/2024/10/Digital-Age-Assurance-Act-Technical-Whitepaper-FINAL-Feb-07-2025.pdf" }, { "title": "AB1043 Assembly Privacy and Consumer Protection analysis", "url": "https://apcp.assembly.ca.gov/system/files/2025-04/ab-1043-wicks-apcp-analysis.pdf" }, { "title": "ICMEC APFC roundtables", "url": "https://www.icmec.org/apfc-roundtables/" }, { "title": "ICMEC international impact report2024", "url": "https://cdn.icmec.org/wp-content/uploads/2025/05/ICMEC-Impact-Report-2024.pdf" } ], "answer": "" }, { "id": "w1-icmec-lead-2", "question": "What does the identified MPA memorandum identified in April21,2026 analysis; amendment instructions and submissions establish?", "why": "Who proposed surviving/deleted provisions and how approval occurred", "evidence": "MPA memorandum identified in April21,2026 analysis; amendment instructions and submissions", "status": "parked", "people": [], "sources": [ { "title": "AB1856 Assembly Privacy and Consumer Protection analysis", "url": "https://apcp.assembly.ca.gov/system/files/2026-04/ab-1856-wicks-apcp-analysis.pdf" } ], "answer": "" }, { "id": "w1-icmec-lead-3", "question": "What does the identified Deployed OS age-API admission terms and implementation specification establish?", "why": "Neutral statutory signal service vs discretionary admission control", "evidence": "Deployed OS age-API admission terms and implementation specification", "status": "parked", "people": [], "sources": [ { "title": "Device-Based Age Assurance technical whitepaper", "url": "https://cdn.icmec.org/wp-content/uploads/2024/10/Digital-Age-Assurance-Act-Technical-Whitepaper-FINAL-Feb-07-2025.pdf" }, { "title": "Microsoft Windows age-signals documentation", "url": "https://learn.microsoft.com/en-us/windows/apps/develop/security/age-signals/" } ], "answer": "" }, { "id": "w1-controlai-j-us-creditor", "question": "Who was owed Control AI US's $20,955 related-party balance at 31 December 2025, and what instrument created it?", "why": "Resolve the substantive gap after the bounded public-record pass.", "evidence": "2025 inter-organisation balance confirmation, accounts-payable ledger entry, supporting invoice/cost-allocation agreement or loan schedule; subsequent clearing entry", "status": "open", "people": [], "sources": [ { "title": "Control AI US Inc initial 2025 Form 990-EZ, IRS public XML", "url": "https://apps.irs.gov/pub/epostcard/990/xml/2026/2026_TEOS_XML_06A.zip" } ], "answer": "" }, { "id": "w1-controlai-j-uk-deferred", "question": "What components and conditions comprise the UK7596691 combined line?", "why": "Resolve the substantive gap after the bounded public-record pass.", "evidence": "Year-end accrual/deferred-income reconciliation, restricted-grant schedules, underlying agreements and full approved accounts", "status": "parked", "people": [], "sources": [ { "title": "Secure Future Research Ltd 2025 filleted unaudited micro-entity accounts", "url": "https://find-and-update.company-information.service.gov.uk/company/15088415/filing-history/MzU0MDI5NDI2MmFkaXF6a2N4/document?download=1&format=xhtml" } ], "answer": "" }, { "id": "w1-controlai-j-uk-exercise", "question": "Who actually approved and delegated cross-entity spending or creator mandates, and were there external approval rights?", "why": "Resolve the substantive gap after the bounded public-record pass.", "evidence": "Article27 decision minutes plus Article8 delegation/budget authorization, relevant grant/service agreement", "status": "parked", "people": [], "sources": [ { "title": "Secure Future Research Ltd filed Articles of Association", "url": "https://find-and-update.company-information.service.gov.uk/company/15088415/filing-history/MzQxNDkzMjc4OWFkaXF6a2N4/document?download=0&format=pdf" }, { "title": "Control AI US Inc initial 2025 Form 990-EZ, IRS public XML", "url": "https://apps.irs.gov/pub/epostcard/990/xml/2026/2026_TEOS_XML_06A.zip" } ], "answer": "" }, { "id": "w1-payments-lead-itch-original-chain", "question": "What instruction, account routing and choice produced itch's July 2025 payment restrictions?", "why": "Public post attributes the constraint to Stripe's banking partners; the actual instruction, exact merchant arrangement and implementation discussion are absent.", "evidence": "Identified July 2025 communication (medium unknown); referenced bank instruction; normally generated account/acquirer onboarding and change records; applicable executed terms.", "status": "parked", "people": [ "itch-io", "stripe", "stripe-bank-reference" ], "sources": [ { "title": "Retained itch.io reindexing post, new ancestry pointer", "url": "https://itch.io/t/5149036/reindexing-adult-nsfw-content" }, { "title": "Contemporaneous reproduction of itch's Stripe bank-restrictions account", "url": "https://www.gamingonlinux.com/2025/08/itch-io-to-reindex-free-adult-content-as-they-search-for-new-payment-processors/" } ], "answer": "" }, { "id": "w1-payments-lead-steam-original-chain", "question": "What did Mastercard, the unnamed processors and acquiring banks actually tell Valve?", "why": "Valve's direct-recipient-published account identifies the rule and rejected alternative; Mastercard's statement leaves message scope unresolved.", "evidence": "Identified Valve press replies; asserted processor notice and Valve response; asserted network/acquirer communications.", "status": "parked", "people": [ "steam", "mastercard", "w1-payments-steam-processors" ], "sources": [ { "title": "Valve gets pressured by payment processors: recipient-published Valve reply", "url": "https://www.gamingonlinux.com/2025/07/valve-gets-pressured-by-payment-processors-with-a-new-rule-for-game-devs-and-various-adult-games-removed/" }, { "title": "Mastercard statement article: recipient-published Valve response", "url": "https://www.gamingonlinux.com/2025/08/mastercard-release-a-statement-about-game-stores-payment-processors-and-adult-content/" } ], "answer": "" }, { "id": "w1-payments-lead-historical-policy", "question": "What Stripe policy and account terms governed itch in July 2025?", "why": "Archive index gives exact snapshots; current terms and bank lists cannot establish historical applicability.", "evidence": "Identified archived public policy; normally generated accepted account terms and notices.", "status": "parked", "people": [ "stripe", "itch-io" ], "sources": [ { "title": "Wayback CDX metadata for Stripe restricted-businesses July 2025 captures", "url": "https://web.archive.org/cdx/search/cdx?url=stripe.com%2Flegal%2Frestricted-businesses&output=json&filter=statuscode%3A200&from=20250701&to=20250802&fl=timestamp%2Coriginal%2Cdigest" }, { "title": "Identified pre-event Stripe restricted-businesses archive capture", "url": "https://web.archive.org/web/20250721113801/https://stripe.com/legal/restricted-businesses" }, { "title": "Current Stripe Services Agreement, background only", "url": "https://stripe.com/legal/ssa" } ], "answer": "" }, { "id": "w1-payments-lead-itch-id-selection", "question": "Did itch's July 2025 provider exploration yield a selected provider or deployed workflow?", "why": "A separately sourced proposal links geographic compliance to possible ID infrastructure, but identifies no provider or implemented relationship.", "evidence": "Conditional selection/contract/deployment records, or a later public status update.", "status": "parked", "people": [ "itch-io", "w1-payments-itch-digital-id-exploration" ], "sources": [ { "title": "Our update on the UK Online Safety Act", "url": "https://itch.io/t/5133739/our-update-on-the-uk-online-safety-act" } ], "answer": "" }, { "id": "w2-fli-holder-1", "question": "What does the 11April2022 founding deed and current statutes/amendments forCHE-450.633.132 establish?", "why": "Founder identity; appointment/removal; distribution voting; purpose change; dissolution destination", "evidence": "11April2022 founding deed and current statutes/amendments forCHE-450.633.132", "status": "parked", "people": [], "sources": [ { "title": "FLI 2025 audited financial statements", "url": "https://www.lobbyregister.bundestag.de/media/ff/31/796351/Future-of-Life-Institute-2025-Audited-Financial-Statements-Final.pdf" }, { "title": "FLI 2022 Form990", "url": "https://pdf.guidestar.org/PDF_Images/2022/471/052/2022-471052538-202510309349301606-.pdf" }, { "title": "Swiss founding notice reproduced by Moneyhouse", "url": "https://www.moneyhouse.ch/en/company/future-of-life-institute-europe-15209401511" }, { "title": "Swiss Gazette supervision notice HR02-1005510667", "url": "https://shab.lixt.ch/pdf/2022/7/1/9075c285-a959-4cf0-af6f-3626d4f095bd.pdf" }, { "title": "FLI current finance page", "url": "https://futureoflife.org/about-us/finances/" }, { "title": "FLI German lobby-register entry R000227", "url": "https://www.lobbyregister.bundestag.de/suche/R000227" }, { "title": "Geneva free public registry access guidance", "url": "https://www.ge.ch/consulter-registre-du-commerce-commander-documents/consulter-donnees-libre-acces" }, { "title": "Geneva physical dossier consultation", "url": "https://www.ge.ch/consulter-registre-du-commerce-commander-documents/consulter-dossier-physique-entreprise" } ], "answer": "" }, { "id": "w2-fli-holder-2", "question": "What does the Signed Europe2022–25 standalone statements with investment and related-party schedules; December30,2022 transfer acceptance establish?", "why": "Independent reserve, centrally appointable reserve, legally restricted assets and actual financial movements", "evidence": "Signed Europe2022–25 standalone statements with investment and related-party schedules; December30,2022 transfer acceptance", "status": "parked", "people": [], "sources": [ { "title": "FLI 2025 audited financial statements", "url": "https://www.lobbyregister.bundestag.de/media/ff/31/796351/Future-of-Life-Institute-2025-Audited-Financial-Statements-Final.pdf" }, { "title": "FLI 2022 Form990", "url": "https://pdf.guidestar.org/PDF_Images/2022/471/052/2022-471052538-202510309349301606-.pdf" }, { "title": "Swiss founding notice reproduced by Moneyhouse", "url": "https://www.moneyhouse.ch/en/company/future-of-life-institute-europe-15209401511" }, { "title": "Swiss Gazette supervision notice HR02-1005510667", "url": "https://shab.lixt.ch/pdf/2022/7/1/9075c285-a959-4cf0-af6f-3626d4f095bd.pdf" }, { "title": "FLI current finance page", "url": "https://futureoflife.org/about-us/finances/" }, { "title": "FLI German lobby-register entry R000227", "url": "https://www.lobbyregister.bundestag.de/suche/R000227" }, { "title": "Geneva free public registry access guidance", "url": "https://www.ge.ch/consulter-registre-du-commerce-commander-documents/consulter-donnees-libre-acces" }, { "title": "Geneva physical dossier consultation", "url": "https://www.ge.ch/consulter-registre-du-commerce-commander-documents/consulter-dossier-physique-entreprise" } ], "answer": "" }, { "id": "w1-creator-j-sabine", "question": "What does the Original offer email/thread including sender domain, attachments, fee, proposed wording/references and disclosure terms establish?", "why": "Separate publicly stated operating model from contractual conditions and exercised rights.", "evidence": "Original offer email/thread including sender domain, attachments, fee, proposed wording/references and disclosure terms", "status": "parked", "people": [], "sources": [ { "title": "Supplied Sabine transcript", "url": "https://www.youtube.com/watch?v=lPdmYMHrWKg" } ], "answer": "" }, { "id": "w1-creator-j-mateus", "question": "What does the Executed grant letter(s), milestones/reporting/renewal conditions, award amount and recipient entity, payment receipt, any draft/revision exchange for Mky0HtSNE4M establish?", "why": "Separate publicly stated operating model from contractual conditions and exercised rights.", "evidence": "Executed grant letter(s), milestones/reporting/renewal conditions, award amount and recipient entity, payment receipt, any draft/revision exchange for Mky0HtSNE4M", "status": "parked", "people": [], "sources": [ { "title": "Future Tense: The Dark Philosophy Driving AI Development", "url": "https://www.youtube.com/watch?v=Mky0HtSNE4M" }, { "title": "Frame Fellowship public homepage and output links", "url": "https://www.framefellowship.com/" } ], "answer": "" }, { "id": "w1-creator-j-frame", "question": "What does the Lead-funder grant agreements and acceptance, selection rubric/committee record, executed independent or amplifier placement agreement and legal grantor identity establish?", "why": "Separate publicly stated operating model from contractual conditions and exercised rights.", "evidence": "Lead-funder grant agreements and acceptance, selection rubric/committee record, executed independent or amplifier placement agreement and legal grantor identity", "status": "parked", "people": [], "sources": [ { "title": "Frame Fellowship public homepage and output links", "url": "https://www.framefellowship.com/" }, { "title": "Frame public FAQ content, shipped script index-CUMxxB1e.js", "url": "https://www.framefellowship.com/assets/index-CUMxxB1e.js" } ], "answer": "" }, { "id": "w1-creator-j-pf", "question": "What does the A creator-accepted Protect What Is Human SOW plus applicable dated terms, brief, revisions/approval, invoice/settlement and live post establish?", "why": "Separate publicly stated operating model from contractual conditions and exercised rights.", "evidence": "A creator-accepted Protect What Is Human SOW plus applicable dated terms, brief, revisions/approval, invoice/settlement and live post", "status": "parked", "people": [], "sources": [ { "title": "People First Influencer Terms of Use", "url": "https://peoplefirst.cc/terms/" }, { "title": "Protect What Is Human creator recruitment", "url": "https://join.peoplefirst.cc/campaign/protectwhatshuman/" } ], "answer": "" }, { "id": "w1-creator-j-rational", "question": "What does the Executed commissioning agreement and accepted brief for0bnxF9YfyFI, exact UK/US legal counterparty, revision approvals and invoice/payment establish?", "why": "Separate publicly stated operating model from contractual conditions and exercised rights.", "evidence": "Executed commissioning agreement and accepted brief for0bnxF9YfyFI, exact UK/US legal counterparty, revision approvals and invoice/payment", "status": "parked", "people": [], "sources": [ { "title": "Rational Animations: What happens if AI just keeps getting smarter?", "url": "https://www.youtube.com/watch?v=0bnxF9YfyFI" }, { "title": "ControlAI publication announcing Rational Animations collaboration", "url": "https://blog.controlai.org/p/watch-now-what-happens-if-ai-just" } ], "answer": "" }, { "id": "w1-australia-lead-revision-record", "question": "Which objections changed the approved protocol and what was actually implemented?", "why": "Minutes and final response table identify action; inconsistent document controls prevent clean revision chronology.", "evidence": "Identified follow-up email and control-register versions; normally generated approval and fieldwork records.", "status": "parked", "people": [ "au-age-trial", "trial-board", "w1-australia-protocol-revisions", "w1-australia-scout-insights" ], "sources": [ { "title": "March 2025 Stakeholder Advisory Board minutes", "url": "https://ageassurance.com.au/wp-content/uploads/2025/04/SAB-Minutes-13-March-2025-FINAL-2.pdf" }, { "title": "April 29, 2025 Stakeholder Advisory Board minutes", "url": "https://ageassurance.com.au/wp-content/uploads/2025/06/SAB-Minutes-29-April-2025_removed.pdf" }, { "title": "D-2.9 Evaluation Proposal, dated February 6, 2025", "url": "https://ageassurance.com.au/wp-content/uploads/2025/02/D-2.9-Evaluation-Proposal-Report.pdf" }, { "title": "A-1.3.1 Human Test Subjects Protocol, V1.3 filename", "url": "https://ageassurance.com.au/wp-content/uploads/2025/03/A-1.3.1-Human-Test-Subjects-Protocol-V1.3.docx" }, { "title": "Final Part B: statutory stakeholder liaison", "url": "https://ageassurance.com.au/wp-content/uploads/2025/08/AATT_Part_B_DIGITAL.pdf" } ], "answer": "" }, { "id": "w1-australia-lead-recusal-operation", "question": "How did the stakeholder-engagement-only role restriction apply to Corby's stated intention to review technical/vendor report sections, and what work was actually performed?", "why": "A declared exclusion and later review commitment identify a boundary; neither proves compliance nor breach.", "evidence": "Identified conflict declaration; normally generated task assignments, review permissions, signoffs and conflict/recusal minutes.", "status": "parked", "people": [ "corby", "au-age-trial", "w1-australia-trial-draft-review" ], "sources": [ { "title": "Retained trial conflict register", "url": "https://ageassurance.com.au/conflict-of-interest-register/" }, { "title": "July 8, 2025 Stakeholder Advisory Board minutes", "url": "https://ageassurance.com.au/wp-content/uploads/2025/09/SAB-Minutes-8-July-2025.pdf" } ], "answer": "" }, { "id": "w1-australia-lead-regulator-mou", "question": "What boundaries did the eSafety MOU impose and what regulator comments altered the report?", "why": "Agreement and draft input are identified, but their operative content is absent.", "evidence": "Identified signed MOU, mid-project impartiality report and draft-feedback communication.", "status": "parked", "people": [ "esafety", "au-age-trial", "w1-australia-esafety-mou" ], "sources": [ { "title": "April 29, 2025 Stakeholder Advisory Board minutes", "url": "https://ageassurance.com.au/wp-content/uploads/2025/06/SAB-Minutes-29-April-2025_removed.pdf" }, { "title": "Final Part B: statutory stakeholder liaison", "url": "https://ageassurance.com.au/wp-content/uploads/2025/08/AATT_Part_B_DIGITAL.pdf" } ], "answer": "" }, { "id": "w1-australia-lead-procurement", "question": "What were the exact contract, tax basis, variations and subcontracted amounts?", "why": "Original departmental brief confirms value; full contract/register inspection and payment reconciliation remain incomplete.", "evidence": "Identified CN4111198 and GST-labelled register; normally generated contract, variations, invoices and subcontracts.", "status": "parked", "people": [ "au-infrastructure", "accs" ], "sources": [ { "title": "Official departmental briefing confirming trial procurement value", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/foi26-145-documents-released.pdf" }, { "title": "FOI26-176 indexed GST-labelled contract register", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/foi-26-176_documents-released.pdf" }, { "title": "Original AusTender CN4111198 entry", "url": "https://www.tenders.gov.au/Cn/Show/77f527297b2c4c218e5f99476707aa98" } ], "answer": "" }, { "id": "reserve-5rights", "question": "Which records document the2022 5Rights–capconn terms and policy-output route?", "why": "Answered only for the narrowed route; remaining financial/approval questions are explicit successors.", "evidence": "Grant allocations, dated instructions, work-product approvals and invoices", "status": "answered", "people": [], "sources": [], "answer": "Original FARA materials, dated Kidron letters and California text uptake establish the bounded terms/output route. Exact grant-pool allocation, instructions, draft approvals, remittances and invoices remain not obtained." }, { "id": "reserve-weprotect", "question": "Which disclosures identify WeProtect grant/budget rights and the policy-input route?", "why": "Answered only for the narrowed route; remaining financial/approval questions are explicit successors.", "evidence": "Approval minutes, grant restrictions and correspondence", "status": "answered", "people": [], "sources": [], "answer": "Institutional grant/budget-right disclosures and credited/commissioned policy inputs establish the scoped resource/committee route. Signed project conditions, approval minutes, allocated expenditure and actual instruction remain not obtained." }, { "id": "reserve-cais", "question": "How does the $5.2m estate resolution reconcile to the $5,259,802 reserve and its later disappearance?", "why": "Retain a previously identified discriminating record outside the completed bounded wave.", "evidence": "Settlement instrument, ledger settlement and audit reconciliation", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "reserve-uk-status", "question": "What parliamentary action followed the planned 8 September 2026 Sobel introduction?", "why": "Retain a previously identified discriminating record outside the completed bounded wave.", "evidence": "Hansard, Votes and Proceedings, bill version and stage", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "reserve-opposing", "question": "Which named opposing-committee/vendor expenditure route was acquired?", "why": "Answered only for the narrowed route; remaining financial/approval questions are explicit successors.", "evidence": "Disbursements, invoices, contracts and intercommittee reconciliation", "status": "answered", "people": [], "sources": [], "answer": "One Think Big/Lantern Production Consultants March2026 route is acquired. Creative output, subvendors, invoices, approvals and complete committee purchase coverage remain not obtained." }, { "id": "reserve-fabian", "question": "Did a Fabian institution commission or approve a specific relevant policy intervention?", "why": "Retain a previously identified discriminating record outside the completed bounded wave.", "evidence": "Dated commission, minutes or formal submission; roster overlap is insufficient", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "closed-au-responses", "question": "Did the public record contain responses to trial objections?", "why": "Closes the missing-minute and unanswered-objection public-route question while preserving narrower implementation gaps.", "evidence": "", "status": "answered", "people": [], "sources": [], "answer": "Yes. Acquired March–July minutes and final methodology account record objections, responses and changed procedures; implementation and operative review boundaries remain unresolved." }, { "id": "w3-oak-l-oak-letters", "question": "Who approved the named5 Rights and WeProtect awards, and what rights did the signed letters reserve?", "why": "Generic process does not identify Alan Parker or prove recipient-specific termination/reporting/editorial rights.", "evidence": "Signed 2023 5 Rights and 2025 WeProtect grant letters, concept notes, approval/renewal record", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "w3-oak-l-subgrant", "question": "What budget deviations were actually approved or refused under the Dutch-to-UK agreement?", "why": "Disclosure establishes formal ability, not exercised control of a named output.", "evidence": "Executed 2021 subgrant agreement, annual schedules, amendment approvals and underspend adjustments", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "w3-oak-l-swiss", "question": "Which legal entity and hosted fund is Swiss Philanthropy in the430 k Safe Futures instrument?", "why": "A rerouting is explicit; its upstream funding source and exact legal vehicle are not.", "evidence": "2023 Together for Girls grant amendment and 2024-27 direct grant letter plus fund identifier", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "w3-oak-l-review-edits", "question": "Which contributor changes were accepted, and who had final publication approval?", "why": "Named Oak contribution is stronger than adjacency; no individual edit or veto evidence supplied.", "evidence": "Commissioning SOW, tracked drafts, feedback disposition and final signoff for 2024 review", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "w3-oak-l-survey-sow", "question": "Who paid and approved the specific 2023 European survey and associated advocacy products?", "why": "Commissioning and use are explicit; earmarked funding and exercise of manuscript rights remain unproved.", "evidence": "2023 survey SOW,invoices,funding cost code,questionnaire amendments and manuscript approvals", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "w3-opp-lead-lantern-invoices", "question": "Which workers, subvendors and placements produced the seven March anti-Bores expenditures, and what did Think Big approve?", "why": "Distinguishes an arms-length general contractor from an integrated payment conduit, and a filed generic service purpose from an exact commissioned output.", "evidence": "Executed March 2026 services contract/SOW, vendor invoices keyed to SE.4218-SE.4224 excluding SE.4217, subvendor ledger, placement orders and creative approval trail", "status": "open", "people": [], "sources": [ { "title": "Think Big April monthly amendment, March 2026", "url": "https://docquery.fec.gov/pdf/585/202605209870101585/202605209870101585.pdf" }, { "title": "CLC and Sophia Gonsalves-Brown complaint, American Mission and Think Big", "url": "https://campaignlegal.org/sites/default/files/2026-05/AISubvendorComplaintFinal%20%281%29.pdf" }, { "title": "FEC Advisory Opinion 1983-25", "url": "https://www.fec.gov/files/legal/aos/1983-25/1983-25.pdf" } ], "answer": "" }, { "id": "w3-icmec-l-westpac", "question": "What rights accompanied Westpac funding, and did they apply to Lighthouse, SaferAI or age-policy work?", "why": "Distinguishes capacity funding from programme-specific approval; resolves three-year website vs six-year report descriptions.", "evidence": "Original grant and amendments, programme budget/allocation schedules, reporting and renewal/termination provisions", "status": "open", "people": [], "sources": [ { "title": "Westpac 2024 Safer Children, Safer Communities report", "url": "https://www.westpac.com.au/content/dam/public/wbc/documents/pdf/aw/sustainability/wbc-2024-safer-children-safer-communities-report.pdf" }, { "title": "Westpac 2025 update, published April2026", "url": "https://www.westpac.com.au/content/dam/public/wbc/documents/pdf/aw/sustainability/wbc-2025-safer-children-safer-communities-report.pdf" }, { "title": "ICMEC Australia: origins and future outlook", "url": "https://icmec.org.au/about-icmec-australia/" } ], "answer": "" }, { "id": "w3-icmec-l-partners", "question": "Which organisations actually participated as unlisted SaferAI partners, and who approved a named output?", "why": "The rules establish a permissible unseen category; the records would identify actual participation and decisions.", "evidence": "Participation-category register, attendance record, publication clearance and contributor consent for one named policy paper", "status": "open", "people": [], "sources": [ { "title": "SaferAI for Children Coalition terms of reference", "url": "https://icmec.org.au/wp-content/uploads/2026/02/SaferAI-for-Children-Coalition-Terms-of-Reference_15012026-1.pdf" } ], "answer": "" }, { "id": "w3-icmec-l-lighthouse", "question": "Which contracted data and access rules govern Lighthouse and any resulting bank action?", "why": "Separates analytic capacity from identity adjudication or account denial; no sensitive raw records are requested.", "evidence": "Provider licence and provenance schedules, user access agreement, retention/correction rules and a de-identified alert-to-action audit trail", "status": "open", "people": [], "sources": [ { "title": "ICMEC Australia annual impact report2024-25", "url": "https://icmec.org.au/wp-content/uploads/2025/11/Impact_Report_ICMEC_Australia_online.pdf" }, { "title": "Westpac 2024 Safer Children, Safer Communities report", "url": "https://www.westpac.com.au/content/dam/public/wbc/documents/pdf/aw/sustainability/wbc-2024-safer-children-safer-communities-report.pdf" } ], "answer": "" }, { "id": "w3-icmec-l-virginia", "question": "What language and instructions followed Senator Craig's December2024 DLS draft request?", "why": "An explicit institutional response exists; the exact draft could distinguish agenda influence from provision authorship.", "evidence": "Draft request number, resulting bill draft and revisions, author instructions and published later bill counterpart", "status": "open", "people": [], "sources": [ { "title": "JCOTS 2024 annual report: ICMEC presentation and drafting request", "url": "https://dls.virginia.gov/commissions/jcots/materials/2024_annual_report.pdf" } ], "answer": "" }, { "id": "w3-5r-lead-1", "question": "Who originated and approved the August5 proposed amendments?", "why": "Separates a recorded amendment transmission from its authorship and accepted language.", "evidence": "CapConn email to Aaron Skaggs August5,2022 with attachment versions; surrounding July18–August11 draft exchanges", "status": "parked", "people": [], "sources": [ { "title": "CapConn Exhibit AB with 5Rights contract and activity/receipt appendices", "url": "https://efile.fara.gov/docs/7427-Exhibit-AB-20240627-1.pdf" }, { "title": "California State Archives: legislative resources", "url": "https://www.sos.ca.gov/archives/collections/legislative-resources" }, { "title": "California State Archives: reference services", "url": "https://www.sos.ca.gov/archives/services/reference-services" } ], "answer": "" }, { "id": "w3-5r-lead-2", "question": "Was client release or editorial permission exercised on these exact materials?", "why": "Distinguishes formal contract rights from exercised release or revision approval.", "evidence": "June9–10 background-sheet drafts and replies; June21 coalition-letter transmission/approval chain", "status": "parked", "people": [], "sources": [ { "title": "CapConn Exhibit AB with 5Rights contract and activity/receipt appendices", "url": "https://efile.fara.gov/docs/7427-Exhibit-AB-20240627-1.pdf" } ], "answer": "" }, { "id": "w3-5r-lead-3", "question": "Which funding pool and exact remittances paid the contract?", "why": "Identifies actual contemporaneous funding and receipt timing without projecting later grants backwards.", "evidence": "May–September2022 invoices, 5Rights ledger allocations and bank/remittance reconciliation for five reported ranges", "status": "parked", "people": [], "sources": [ { "title": "CapConn Exhibit AB with 5Rights contract and activity/receipt appendices", "url": "https://efile.fara.gov/docs/7427-Exhibit-AB-20240627-1.pdf" } ], "answer": "" }, { "id": "w4-fin-l-au-statements", "question": "What do AU component and subsequent standalone statements disclose about restricted grants and operating capacity?", "why": "6July2023 component audit explicitly referenced; current accessible standalone attachments not verified", "evidence": "After legitimate access-state change, inspect publicly lodged AIS attachments; specifically seek the AU audited component statements/report for period ended6July2023 named by the US audit. No ACNC retry while current admin-policy failure persists.", "status": "open", "people": [ "w3-icmec-australia" ], "sources": [ { "title": "ICMEC consolidated financial statements2023", "url": "https://cdn.icmec.org/wp-content/uploads/2025/02/ICMEC-2023-Consolidated-FS.pdf" }, { "title": "ACNC AIS2024 title/locator only, financial contents unavailable", "url": "https://www.acnc.gov.au/charity/charities/72001f82-e5d3-eb11-bacc-000d3a6aca25/documents/7af1f05b-5426-f011-8c4d-002248127014" } ], "answer": "" }, { "id": "w4-fin-l-grant-instruments", "question": "Which dates, permitted costs, reporting, renewal and sanction rights govern the AUD25m award?", "why": "Technical-capacity grant with recipient discretion versus explicit donor approvals/redirect rights; term change versus stale or differently scoped programme copy", "evidence": "Initial executed grant, FY24 extension and any legal-grantee transition instrument; do not infer novation existence merely from deconsolidation.", "status": "open", "people": [ "w3-icmec-westpac", "w3-icmec-australia", "icmec" ], "sources": [ { "title": "Westpac FY21 Modern Slavery Statement", "url": "https://www.cmsservices.westpac.com.au/content/dam/public/wbc/documents/pdf/aw/sustainability/WBC_2021_modern_slavery_statement.pdf" }, { "title": "Westpac2024 Modern Slavery Statement", "url": "https://www.westpac.com.au/content/dam/public/wbc/documents/pdf/aw/sustainability/wbc-2024-modern-slavery-statement.pdf" }, { "title": "Westpac 2024 Safer Children, Safer Communities report", "url": "https://www.westpac.com.au/content/dam/public/wbc/documents/pdf/aw/sustainability/wbc-2024-safer-children-safer-communities-report.pdf" } ], "answer": "" }, { "id": "w4-fin-l-software-allocation", "question": "Which grant or operating funds paid for platform development, licensing and maintenance?", "why": "Ordinary expected underlying business records; exact vendor allocations unverified", "evidence": "Project ledger, intangible-asset register, executed supplier/maintenance instruments and invoice coding; no attribution of USD20,699 fixed assets to Lighthouse.", "status": "open", "people": [ "w3-icmec-australia", "w3-icmec-lighthouse" ], "sources": [ { "title": "Westpac FY21 Modern Slavery Statement", "url": "https://www.cmsservices.westpac.com.au/content/dam/public/wbc/documents/pdf/aw/sustainability/WBC_2021_modern_slavery_statement.pdf" }, { "title": "ICMEC consolidated financial statements2023", "url": "https://cdn.icmec.org/wp-content/uploads/2025/02/ICMEC-2023-Consolidated-FS.pdf" } ], "answer": "" }, { "id": "w4-fin-l-membership-change", "question": "Which executed instruments and actual membership entries made the 6 July 2023 Australian control change effective?", "why": "Current constitutions and independent US filings establish rules or reported changes but no actual new member names, effective admission/resignation entries or donor reserved-rights instrument. Bounded independent publisher routes have reached a record frontier.", "evidence": "Member-register extract with effective dates; admission/resignation and constitutional amendment resolutions; adopted constitution and director appointment instruments effective 6 July 2023; any donor reserved-rights instrument", "status": "parked", "people": [ "icmec", "w3-icmec-australia" ], "sources": [ { "title": "ICMEC consolidated financial statements2023", "url": "https://cdn.icmec.org/wp-content/uploads/2025/02/ICMEC-2023-Consolidated-FS.pdf" }, { "title": "ICMEC Australia Constitution, July2023 headers", "url": "https://icmec.org.au/wp-content/uploads/2024/09/ICMEC-AU-Constitution-2023.pdf" }, { "title": "ICMEC Australia Constitution, March2025 headers", "url": "https://icmec.org.au/wp-content/uploads/2025/05/ICMEC-AU-Constitution-1.pdf" }, { "title": "ICMEC Australia public publisher media metadata: constitution", "url": "https://icmec.org.au/wp-json/wp/v2/media?search=constitution&per_page=30" }, { "title": "ICMEC Australia board page, current capture", "url": "https://icmec.org.au/icmec-board/" }, { "title": "ICMEC Australia actual linked governance page", "url": "https://icmec.org.au/icmec-australia-governance/" }, { "title": "ICMEC US 2023 Form 990, signed public disclosure copy", "url": "https://cdn.icmec.org/wp-content/uploads/2024/12/ICMEC_PD_990-2023-signed.pdf" }, { "title": "ICMEC Australia public metadata query: media-resolution", "url": "https://icmec.org.au/wp-json/wp/v2/media?search=resolution&per_page=50" }, { "title": "ICMEC Australia public metadata query: media-agm", "url": "https://icmec.org.au/wp-json/wp/v2/media?search=agm&per_page=50" }, { "title": "ICMEC Australia public metadata query: media-membership", "url": "https://icmec.org.au/wp-json/wp/v2/media?search=membership&per_page=50" }, { "title": "ICMEC Australia public metadata query: pages-agm", "url": "https://icmec.org.au/wp-json/wp/v2/pages?search=annual%20general&per_page=50" }, { "title": "ICMEC Australia public metadata query: posts-agm", "url": "https://icmec.org.au/wp-json/wp/v2/posts?search=annual%20general&per_page=50" } ], "answer": "" }, { "id": "w4-lh-lead-feed-governance", "question": "What provenance, permitted-use, retention and correction terms governed each Lighthouse feed and recipient access?", "why": "Public descriptions identify capabilities and a dated geographic restriction but not feed suppliers or enforceable licence schedules.", "evidence": "Versioned feed inventory, supplier licence schedules, recipient access agreements, privacy impact assessment and correction/withdrawal propagation procedure; exclude underlying sensitive data", "status": "open", "people": [ "w3-icmec-lighthouse", "w3-icmec-australia", "w4-lh-data-suppliers" ], "sources": [ { "title": "ICMEC Australia Symposium programme, 30 October 2024", "url": "https://26980386.fs1.hubspotusercontent-eu1.net/hubfs/26980386/ICMEC%20AU%20Symposium%202024%20-%20Program.pdf" }, { "title": "ICMEC Australia privacy-policy URL: original unavailable on this route", "url": "https://icmec.org.au/privacy-policy/" } ], "answer": "" }, { "id": "w4-lh-lead-alert-action", "question": "Can one de-identified institutional audit trail connect a Lighthouse result to a reviewed action or a rejected false match?", "why": "Operational metrics do not prove downstream decisions. Documented Huntsman closures must remain a separate operation until a join is recovered.", "evidence": "De-identified alert/query-to-decision audit trail, handling protocol, recipient acknowledgement and outcome evaluation", "status": "open", "people": [ "w3-icmec-lighthouse", "w4-lh-huntsman" ], "sources": [ { "title": "ICMEC Australia annual impact report2024-25", "url": "https://icmec.org.au/wp-content/uploads/2025/11/Impact_Report_ICMEC_Australia_online.pdf" }, { "title": "No Bank Left Behind, July 2026", "url": "https://icmec.org.au/wp-content/uploads/2026/08/ICMEC-Australia-Insights-Paper-No-Bank-Left-Behind-2.pdf" }, { "title": "AUSTRAC urges financial services to look out for child exploitation transactions", "url": "https://www.austrac.gov.au/news-and-media/media-release/austrac-urges-financial-services-look-out-child-exploitation-transactions" } ], "answer": "" }, { "id": "w4-lh-lead-support-instrument", "question": "What do the Lighthouse build and renewed support instruments allocate to ICMEC and Kablamo?", "why": "The vendor describes a continuing operating dependency without prices, deliverable acceptance or contract terms.", "evidence": "Build statement of work, acceptance record, Product Care agreement/renewals and handover schedule", "status": "open", "people": [ "w3-icmec-lighthouse", "w4-lh-kablamo", "w4-lh-product-owner" ], "sources": [ { "title": "Kablamo Lighthouse case study", "url": "https://www.kablamo.com.au/case-study/icmec-website-redesign" } ], "answer": "" }, { "id": "w4-pol-q-drafting", "question": "November2024 discussion-paper drafting history, contributor comments and publication clearance — what does the record show?", "why": "Who supplied the device-level sentence, who amended it and who approved publication.", "evidence": "November2024 discussion-paper drafting history, contributor comments and publication clearance", "status": "open", "people": [], "sources": [], "answer": "" }, { "id": "w4-pol-q-sponsor", "question": "Paper-specific budget, commissioning agreement and donor conditions if any — what does the record show?", "why": "General organisational support versus earmarked influence over this output.", "evidence": "Paper-specific budget, commissioning agreement and donor conditions if any", "status": "open", "people": [], "sources": [], "answer": "" }, { "id": "w4-pol-q-tas", "question": "Woolcott-review research file/source transmittal and response implementation log for section6.3.2 — what does the record show?", "why": "Independent literature discovery versus direct coalition submission; later acceptance/implementation of recommendations.", "evidence": "Woolcott-review research file/source transmittal and response implementation log for section6.3.2", "status": "open", "people": [], "sources": [], "answer": "" }, { "id": "w4-pol-q-esafety", "question": "Toolkit footnote34 source-selection and manuscript-review record — what does the record show?", "why": "How the2024 paper entered2026 guidance and whether any broader recommendations were accepted.", "evidence": "Toolkit footnote34 source-selection and manuscript-review record", "status": "open", "people": [], "sources": [], "answer": "" }, { "id": "w4-pol-q-encryption-clearance", "question": "Who introduced and cleared the homomorphic-encryption and E2EE-limit wording, and what source was intended for footnote29?", "why": "Bibliographic discrepancy plus missing technical support leave paragraph origin unresolved; FOI identifies actual draft and comment channels.", "evidence": "Attachment A and B May2025 drafts; final-draft redlines; comment-disposition log; footnote29 reference-manager entry and publication clearance", "status": "open", "people": [], "sources": [ { "title": "Designing for Safety: Preventing child sexual exploitation and abuse online", "url": "https://www.esafety.gov.au/sites/default/files/2026-02/Designing-for-Safety_Preventing-child-sexual-exploitation-and-abuse-online_CSEA-toolkit.pdf" }, { "title": "eSafety FOI25139 LOG118 released correspondence: toolkit consultation", "url": "https://www.esafety.gov.au/sites/default/files/2025-09/LOG_118_Document_Set.pdf?v=1765065600059" }, { "title": "Guiding principles for addressing technology-facilitated child sexual exploitation and abuse", "url": "https://eprints.lse.ac.uk/126219/3/van_der_Spuy_et_al_2024_Guiding_principles_for_addressing_technology-facilitated_CSEA.pdf" }, { "title": "LSE repository metadata for report126219", "url": "https://eprints.lse.ac.uk/126219/" } ], "answer": "" }, { "id": "w5-gov-l-component-audit", "question": "Who was the Australian component auditor and what did the statements/report through6July2023 disclose?", "why": "The US auditor expressly received the record; direct publisher metadata searches did not recover it.", "evidence": "Australian component financial statements and independent auditor report for period ended6July2023; audit engagement identity and transmission/cover record supplied to GRF", "status": "open", "people": [], "sources": [ { "title": "ICMEC consolidated financial statements2023", "url": "https://cdn.icmec.org/wp-content/uploads/2025/02/ICMEC-2023-Consolidated-FS.pdf" }, { "title": "ICMEC Australia annual impact report2024-25", "url": "https://icmec.org.au/wp-content/uploads/2025/11/Impact_Report_ICMEC_Australia_online.pdf" }, { "title": "ICMEC Australia publisher media search: audit", "url": "https://icmec.org.au/wp-json/wp/v2/media?search=audit&per_page=30" }, { "title": "ICMEC Australia publisher media search: financial", "url": "https://icmec.org.au/wp-json/wp/v2/media?search=financial&per_page=30" } ], "answer": "" }, { "id": "w5-gov-l-specific-approvals", "question": "How did Westpac-related funding and downstream programs pass through the board approval and conflict rules?", "why": "New charter identifies where approvals should be recorded; current board includes a Westpac executive. Only records of particular decisions can distinguish ordinary expertise from donor instruction or conditioned approvals.", "evidence": "Executed grant/extension and relevant board papers/minutes; delegated-authority schedule; interests disclosures, exceptions, recusals and votes; director appointment/representative agreement", "status": "open", "people": [], "sources": [ { "title": "ICMEC Australia Board Charter, adopted6December2023", "url": "https://icmec.org.au/wp-content/uploads/2024/09/ICMEC-AU-Board-Charter.pdf" }, { "title": "ICMEC Australia Conflict of Interest Policy", "url": "https://icmec.org.au/wp-content/uploads/2024/09/ICMEC-AU-Conflict-of-Interest.pdf" }, { "title": "ICMEC Australia board page, current capture", "url": "https://icmec.org.au/icmec-board/" }, { "title": "ICMEC Australia Constitution, March2025 headers", "url": "https://icmec.org.au/wp-content/uploads/2025/05/ICMEC-AU-Constitution-1.pdf" } ], "answer": "" }, { "id": "w5-enc-l-funding", "question": "What award and publication rights governed the Toolbox work package3 report?", "why": "Funding route is named but amounts and approval rights are absent; centre-level5Rights funding cannot supply missing grant terms. Award announcement, UKRI collaboration record and actual WP2 output acquired. Next target is the Cambridge/Bristol award and subaward terms, work-package cost schedule and publication/acceptance clauses. Parent centre value is not a project allocation.", "evidence": "Fourth Strategic Funding Call award agreement, work-package budget, deliverable and publication terms", "status": "parked", "people": [], "sources": [ { "title": "Guiding principles for addressing technology-facilitated child sexual exploitation and abuse", "url": "https://eprints.lse.ac.uk/126219/3/van_der_Spuy_et_al_2024_Guiding_principles_for_addressing_technology-facilitated_CSEA.pdf" }, { "title": "DFC REPHRAIN project description", "url": "https://www.digital-futures-for-children.net/our-work/rephrain" }, { "title": "UKRI Gateway to Research: REPHRAIN, EP/V011189/1", "url": "https://gtr.ukri.org/projects?ref=EP/V011189/1" }, { "title": "REPHRAIN: Launch of Strategic Funding Call 4 Projects", "url": "https://www.rephrain.ac.uk/launch-of-strategic-funding-call-4-projects/" }, { "title": "REPHRAIN Work Package 2, Project 1: Reliability Analyses", "url": "https://rephrain-f77d71.gitlab.io/wp2_p1.html" } ], "answer": "" }, { "id": "w5-enc-l-peerreview", "question": "Did eSafety-affiliated peer review alter recommendations in the2024 report or transmit it into toolkit drafting?", "why": "Named review participation is real access, but no comments, selection email or redline connects it to toolkit wording.", "evidence": "Peer-review invitation, comments, author response and source-selection correspondence", "status": "open", "people": [], "sources": [ { "title": "Guiding principles for addressing technology-facilitated child sexual exploitation and abuse", "url": "https://eprints.lse.ac.uk/126219/3/van_der_Spuy_et_al_2024_Guiding_principles_for_addressing_technology-facilitated_CSEA.pdf" }, { "title": "Designing for Safety: Preventing child sexual exploitation and abuse online", "url": "https://www.esafety.gov.au/sites/default/files/2026-02/Designing-for-Safety_Preventing-child-sexual-exploitation-and-abuse-online_CSEA-toolkit.pdf" } ], "answer": "" }, { "id": "w5-fund-l-allocation", "question": "Which project schedule and advisory record connect the 2021 award to the later financial-monitoring study?", "why": "A named award, funding acknowledgement and adviser roster exist. Contract schedules and comment dispositions can distinguish resourcing, consultation and editorial power. Public registration is now and explicitly post-analysis. Next target is the executed grant, scope/ethics amendments and CloudResearch engagement defining the approved expansion and cost baseline.", "evidence": "Executed award schedule, variations, research output list, advisory terms/minutes and reviewer dispositions", "status": "parked", "people": [], "sources": [ { "title": "Westpac 2021 Safer Children, Safer Communities grant recipients", "url": "https://www.westpac.com.au/about-westpac/sustainability/our-positions-and-perspectives/safer-children-safer-communities/scsc-2021-grant-recipients/" }, { "title": "Childlight Searchlight 2025: Study C funding, advisers and recommendations", "url": "https://www.childlight.org/uploads/publications/searchlight2025/Childlight_Searchlight_Report_2025.pdf" }, { "title": "Jesuit Social Services 2023 Australian men study", "url": "https://cdn.jss.org.au/wp-content/uploads/2023/11/20105310/Identifying-and-understanding-child-sexual-offending-behaviour-and-attitudes-among-Australian-men_web.pdf" }, { "title": "OSF registration j7xh2: The nature of online offending against children", "url": "https://osf.io/j7xh2/" }, { "title": "OSF j7xh2 response revision history", "url": "https://api.osf.io/v2/registrations/j7xh2/schema_responses/" }, { "title": "Whitten, Salter and Woodlock: study methodology, CrimRxiv release1", "url": "https://www.crimrxiv.com/pub/un6p337j/release/1" } ], "answer": "" }, { "id": "w6-ref-l-transmission", "question": "Which dated comment dispositions or tracked references connect the earlier encryption-policy inputs to the 2026 toolkit?", "why": "Actual2022–24 citations and submissions establish earlier consideration. The remaining mechanism is selection and revision of later guidance, including citation29.", "evidence": "Comment disposition, tracked draft and reference-manager export", "status": "open", "people": [], "sources": [ { "title": "eSafety end-to-end encryption position statement, October 2023", "url": "https://www.esafety.gov.au/sites/default/files/2023-10/End-to-end-encryption-position-statement-oct2023.pdf" }, { "title": "WeProtect Secretariat submission on draft Australian industry standards", "url": "https://www.esafety.gov.au/sites/default/files/2024-02/standards-submission-from-WeProtect-Global-Alliance.pdf" }, { "title": "Digital Rights Watch submission on draft Australian industry standards", "url": "https://www.esafety.gov.au/sites/default/files/2024-02/standards-submission-from-Digital-Rights-Watch_0.pdf" }, { "title": "Designing for Safety: Preventing child sexual exploitation and abuse online", "url": "https://www.esafety.gov.au/sites/default/files/2026-02/Designing-for-Safety_Preventing-child-sexual-exploitation-and-abuse-online_CSEA-toolkit.pdf" } ], "answer": "" }, { "id": "w6-draft-l-may-files", "question": "What text and edits did the May 2025 named attachments contain?", "why": "Exact filenames now identify the drafts; the email outline cannot establish their words or revision history.", "evidence": "OriginalDOCX attachments and any marked-up returns or version history", "status": "open", "people": [], "sources": [ { "title": "eSafety FOI25139 LOG118 released correspondence: toolkit consultation", "url": "https://www.esafety.gov.au/sites/default/files/2025-09/LOG_118_Document_Set.pdf?v=1765065600059" }, { "title": "eSafety disclosure log118: FOI25139 request scope and release summary", "url": "https://www.esafety.gov.au/about-us/corporate-documents/freedom-of-information/disclosure-log" } ], "answer": "" }, { "id": "w6-draft-l-decision", "question": "How did FOI25139 treat the named attachments, and which records were exempted?", "why": "Release summary says18partial/4whole exemptions but does not identify those records or distinguish omitted attachments from withholding.", "evidence": "FOI25139 decision notice, document schedule, attachment-treatment explanation and any already-released supplementary set", "status": "open", "people": [], "sources": [ { "title": "eSafety FOI25139 LOG118 released correspondence: toolkit consultation", "url": "https://www.esafety.gov.au/sites/default/files/2025-09/LOG_118_Document_Set.pdf?v=1765065600059" }, { "title": "eSafety disclosure log118: FOI25139 request scope and release summary", "url": "https://www.esafety.gov.au/about-us/corporate-documents/freedom-of-information/disclosure-log" } ], "answer": "" }, { "id": "w6-draft-l-csm-disposition", "question": "Which October 2025 draft did Common Sense Media review, and what happened to its requested changes?", "why": "Public comments name concrete requested edits and later review date; invitation is beyond the earlierFOI request window.", "evidence": "October2025 review invitation and attached draft, submission/receipt records, comment-disposition table, redlines and clearance records", "status": "open", "people": [], "sources": [ { "title": "Common Sense Media comments to Australian eSafety on Designing for Safety draft", "url": "https://www.commonsensemedia.org/sites/default/files/featured-content/files/aus-e-safety-comments-common-sense-media-october-2025.pdf" }, { "title": "Common Sense Media Public Filings and Letters: October 2025 eSafety comment entry", "url": "https://www.commonsensemedia.org/kids-action/public-filings" }, { "title": "eSafety disclosure log118: FOI25139 request scope and release summary", "url": "https://www.esafety.gov.au/about-us/corporate-documents/freedom-of-information/disclosure-log" } ], "answer": "" }, { "id": "w6-member-l-return-basis", "question": "What Australian tax period, control basis and inter-entity ledger support the 2023 Schedule R entries?", "why": "Any-time relationship inclusion is compatible with July deconsolidation, but the 100% entry uses a related-tax-year-end rule. The constitution defaults to calendar year unless changed. The type-L amount does not reveal actual service or cash direction.", "evidence": "2023 Schedule R supporting workpaper and Australian tax-year selector; any financial-year-change resolution; inter-entity type-L contract, valuation and general-ledger entries; Australian component statements/report through 6 July 2023", "status": "parked", "people": [], "sources": [ { "title": "ICMEC US 2023 Form 990, signed public disclosure copy", "url": "https://cdn.icmec.org/wp-content/uploads/2024/12/ICMEC_PD_990-2023-signed.pdf" }, { "title": "2023 Instructions for Schedule R (Form 990)", "url": "https://www.irs.gov/pub/irs-prior/i990sr--2023.pdf" }, { "title": "ICMEC consolidated financial statements2023", "url": "https://cdn.icmec.org/wp-content/uploads/2025/02/ICMEC-2023-Consolidated-FS.pdf" }, { "title": "ICMEC Australia Constitution, July2023 headers", "url": "https://icmec.org.au/wp-content/uploads/2024/09/ICMEC-AU-Constitution-2023.pdf" }, { "title": "ICMEC Australia Constitution, March2025 headers", "url": "https://icmec.org.au/wp-content/uploads/2025/05/ICMEC-AU-Constitution-1.pdf" } ], "answer": "" }, { "id": "w7-impl-l-processing-terms", "question": "What fields, retention instructions and UK method configuration do Reddit and Persona actually use?", "why": "Public customerFAQ identifies the controller and concrete data outputs. Actual processing schedule and rollout configuration would distinguish current promises, regional availability and implemented collection.", "evidence": "Executed processing schedule, configured output/retention schema, UK rollout specification and deletion-control evidence", "status": "open", "people": [], "sources": [ { "title": "Reddit Help: Why is Reddit asking for my age?", "url": "https://support.reddithelp.com/hc/en-us/articles/36429514849428-Why-is-Reddit-asking-for-my-age" }, { "title": "Persona customer-specific FAQ for Reddit users", "url": "https://help.withpersona.com/articles/7F6BaF9h8Fxf0XWkwQscXN/" }, { "title": "ICO monetary penalty notice to Reddit,23 February 2026", "url": "https://ico.org.uk/media2/hrlmvj14/reddit-mpn-20260223.pdf" } ], "answer": "" }, { "id": "w7-impl-l-regulator-decisions", "question": "Which specific ICO–Ofcom age-assurance decisions were coordinated around the July 2025 implementation period?", "why": "ICO confirms routine meetings, while broad multi-year FOI request was refused. A short, specific action log could reveal a decision link without assuming every meeting concernedReddit.", "evidence": "Topic-specific dated agenda/action log or joint decision record, bounded toJune–July 2025", "status": "open", "people": [], "sources": [ { "title": "ICO response IC-484385-D3Q2,24 March 2026", "url": "https://ico.org.uk/media2/xlvp0oq4/our-response-to-ic-484385-d3q2.pdf" }, { "title": "ICO monetary penalty notice to Reddit,23 February 2026", "url": "https://ico.org.uk/media2/hrlmvj14/reddit-mpn-20260223.pdf" } ], "answer": "" }, { "id": "w7-policy-l-email-evidence", "question": "Which submitted technical results and independent tests persuaded Ofcom that email-based age estimation can meet each HEAA criterion?", "why": "Statement paragraph 3.86 expressly relies on AVPA and VerifyMy evidence. Method acceptance is established, but the performance and provenance of the underlying tests require assessment of the actual records.", "evidence": "Complete consultation response, white paper, test dataset protocol, evaluation memorandum and relevant correspondence.", "status": "open", "people": [], "sources": [ { "title": "Ofcom Age Assurance and Children's Access statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/statement-age-assurance-and-childrens-access/statement-age-assurance-and-childrens-access.pdf?v=397036" }, { "title": "Ofcom Part 5 guidance on highly effective age assurance and other duties", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/statement-age-assurance-and-childrens-access/guidance-on-highly-effective-age-assurance-and-other-part-5-duties.pdf?v=388810" } ], "answer": "" }, { "id": "w7-policy-l-xgroovy-decision", "question": "What factual period, representations, legal provider identity and reasoning appear in the confirmation decision issued on 3 September 2026?", "why": "The official case update reports final penalties but promises the non-confidential decision later.", "evidence": "Non-confidential confirmation decision, including reasoning and penalty assessment.", "status": "parked", "people": [], "sources": [ { "title": "Ofcom investigation CW/01318/09/25: Xgroovy", "url": "https://www.ofcom.org.uk/online-safety/protecting-children/investigation-into-the-provider-of-xgroovy.coms-compliance-with-the-duty-to-prevent-children-from-encountering-pornographic-content-through-the-use-of-age-assurance" } ], "answer": "" }, { "id": "w7-adv-l-redlines", "question": "Which exact contributor drafts and government clearance records produced Amendment 124 and the accompanying age-assurance package?", "why": "Ministerial credit and sponsorship establish a contribution route; versioned drafts would distinguish specific accepted edits, government legal redrafting and broader campaigning.", "evidence": "May–July 2023 age-assurance amendment drafts, tracked changes, meeting readouts and policy/legal clearance records", "status": "parked", "people": [], "sources": [ { "title": "Hansard: Lords chamber, 6 July 2023", "url": "https://hansard.parliament.uk/Lords/2023-07-06/debates/600d4bb5-1775-4688-94c6-c1fcd7b54e27/LordsChamber" }, { "title": "Hansard: Online Safety Bill report stage and Schedule 4 decisions, 10 July 2023", "url": "https://hansard.parliament.uk/Lords/2023-07-10/debates/049FB9B5-C87F-4750-8523-63A07315339D/OnlineSafetyBill" }, { "title": "Online Safety Bill: Amendment 124, agreed", "url": "https://bills.parliament.uk/bills/3137/stages/17765/amendments/95907" }, { "title": "Online Safety Bill: Amendment 125, not moved", "url": "https://bills.parliament.uk/bills/3137/stages/17765/amendments/96131" } ], "answer": "" }, { "id": "w7-adv-l-project-funding", "question": "Were 5Rights' UK age-assurance submissions or parliamentary amendment work supported by earmarked funding or contributor clearance rights?", "why": "Existing general funding and personnel records do not allocate money, staff time or approval rights to these specific UK policy products.", "evidence": "2023–2025 UK age-assurance project ledger, grant schedules, staff allocations, commission terms and submission approval records", "status": "parked", "people": [], "sources": [ { "title": "5Rights: Part5 pornographic-content guidance consultation response, March 2024", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/272586-consultation-guidance-for-service-providers-publishing-pornographic-content/responses/5rights.pdf?v=404369" }, { "title": "5Rights: Protecting Children from Harms Online response, 17 July 2024", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/284469-consultation-protecting-children-from-harms-online/responses/5rights-foundation.pdf?v=385679" }, { "title": "5Rights: Additional Safety Measures response, October 2025", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/5rights-foundation.pdf?v=409144" }, { "title": "Parliamentary interests register", "url": "https://members.parliament.uk/member/4258/registeredinterests" } ], "answer": "" }, { "id": "w7-adv-l-certification-disposition", "question": "Did Ofcom later accept any part of the October 2025 IEEE/ACCS certification recommendation, or retain a provider-responsibility approach?", "why": "The submitted recommendation is concrete; its later disposition is not established by this advocacy lane and must not be silently turned into adoption.", "evidence": "Additional Safety Measures consultation statement, response-to-comments matrix and underlying age-assurance certification assessment", "status": "parked", "people": [], "sources": [ { "title": "5Rights: Additional Safety Measures response, October 2025", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/5rights-foundation.pdf?v=409144" } ], "answer": "" }, { "id": "w7-market-l-sla", "question": "What payment, appointment, liability and termination rights does the executed AVID–scheme agreement allocate?", "why": "Would identify revenue allocation, indemnity, renewal/termination and actual delegated rights beyond current public rules.", "evidence": "Executed AVID/scheme SLA, dated customer agreement and fee schedule", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "w7-market-l-ageaware", "question": "Which proposed AgeAware admission, exclusivity and fee rules became signed agreements or operating decisions?", "why": "Would convert proposed commercial/admission machinery into actual executed rights, or establish redesign/cessation. Test independent-assessment alternative before alleging self-certification.", "evidence": "Final approved AgeAware specification, procurement/exclusivity agreements, grant schedule, assessor appointment and impartiality assessment", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "w7-market-l-iso", "question": "Which committee contributions and dispositions shaped the final standard’s relevant requirements?", "why": "Public ISO page GET403; official BSI announcement used once. No paid standard acquisition. Cannot claim exact final clauses or votes from biographies.", "evidence": "Full normative ISO27566-1 and particular committee dispositions", "status": "parked", "people": [], "sources": [], "answer": "" }, { "id": "w7-adv-l-commission", "question": "What did the commissioned AgeAware rights review conclude, and what changed in response?", "why": "A participant account names the commissioned review and team. The terms, fee, declared interests, final report, clearance arrangements and response are not yet acquired. Those records can test the review’s independence and practical effects.", "evidence": "Terms of reference, signed commission, declarations, final report and response-to-recommendations record", "status": "parked", "people": [], "sources": [ { "title": "euCONSENT: Building a Safer Internet, the Journey to AgeAware", "url": "https://euconsent.eu/building-a-safer-internet-the-journey-to-ageaware/" } ], "answer": "" }, { "id": "w8-red-l-grant-terms", "question": "What conditions accompanied CeSIA's reported two-year Longview recommendation?", "why": "Recommendation letter, executed agreement and remittance schedule would identify legal grantor, actual receipts, restrictions and approval/reporting rights. Recommendation is evidenced; execution is not.", "evidence": "Recommendation letter; executed grant agreement if any; remittance and restricted-fund schedule", "status": "open", "people": [], "sources": [ { "title": "CeSIA diplomatic sprint proposal", "url": "https://manifund.org/projects/from-nobel-signatures-to-binding-red-lines-the-2026-diplomatic-sprint" }, { "title": "Longview organizational structure", "url": "https://www.longview.org/organisational-structure/" } ], "answer": "" }, { "id": "w8-red-l-campaign-spend", "question": "Who paid the launch and diplomatic campaign costs, and on what terms?", "why": "Named PR work and an explicit diplomatic budget locate likely finance records. Actual campaign cost centres, invoices and approval clauses remain not obtained.", "evidence": "Project ledger, purchase orders, vendor contracts, invoices, grant allocations", "status": "open", "people": [], "sources": [ { "title": "Global Call acknowledgements", "url": "https://red-lines.ai/acknowledgements" }, { "title": "CeSIA diplomatic sprint proposal", "url": "https://manifund.org/projects/from-nobel-signatures-to-binding-red-lines-the-2026-diplomatic-sprint" } ], "answer": "" }, { "id": "w8-red-l-text-authority", "question": "Who could approve or reject final campaign text and partner implementations?", "why": "Credits establish contribution, not final sign-off. Version history and partner terms could distinguish joint authorship, editorial control and broad coalition agreement.", "evidence": "Draft history, approval correspondence, partner memorandum and signatory text versions", "status": "open", "people": [], "sources": [ { "title": "Global Call acknowledgements", "url": "https://red-lines.ai/acknowledgements" }, { "title": "ControlAI: International AI red lines", "url": "https://blog.controlai.org/p/red-lines" } ], "answer": "" }, { "id": "w8-red-l-policy-transmission", "question": "Which campaign or subgroup document entered a government drafting process?", "why": "A transmitted briefing and tracked recipient edits would test conversion from access to adopted language; parliamentary advocacy and UNESCO hosting alone cannot.", "evidence": "Briefing as sent, recipient acknowledgment, tracked policy drafts, response record", "status": "open", "people": [], "sources": [ { "title": "UNESCO-network International AI Red Lines subgroup", "url": "https://www.unesco.org/ethics-ai/en/node/338" }, { "title": "Lords debate, 29 January 2026", "url": "https://hansard.parliament.uk/lords/2026-01-29/debates/4be73a9c-a273-4120-8c34-7d13cca24d61/LordsChamber" }, { "title": "TFS India AI Impact Summit account", "url": "https://thefuturesociety.org/indiaaisummit" }, { "title": "Government of India: New Delhi Declaration", "url": "https://www.pib.gov.in/PressReleasePage.aspx?PRID=2231208&lang=1®=3" } ], "answer": "" }, { "id": "w8-red-l-athens-support", "question": "What resources and conditions sat behind the Athens2025 support credits?", "why": "Named supporters narrow the record holders. Identity, amount, in-kind services and program approval rights are not disclosed by logos.", "evidence": "Sponsorship agreements, event budget, invoices and legal-entity identifiers", "status": "open", "people": [], "sources": [ { "title": "Athens Roundtable 2025 recap", "url": "https://thefuturesociety.org/wp-content/uploads/2023/04/The-Athens-Roundtable-2025-Recap-Report.pdf" } ], "answer": "" }, { "id": "w8-red-l-survey-curation", "question": "How did the original fifteen priorities become ten survey choices?", "why": "The published methodology records this reduction but not the mapping or exact questionnaire. These can reveal consequential merging, exclusions and wording before rankings.", "evidence": "Original workshop export, consolidation matrix, survey form and analysis workbook", "status": "open", "people": [], "sources": [ { "title": "Civil society priorities for international AI governance", "url": "https://thefuturesociety.org/cso-ai-governance-priorities/" } ], "answer": "" }, { "id": "w8-follow-l-g7-draft", "question": "Was CeSIA language accepted, softened or rejected in the French/G7 drafting process?", "why": "Published programme and partially overlapping official goals leave the actual transmission unestablished.", "evidence": "Dated submission, recipient acknowledgment, annotated draft and response; label award instrument separately", "status": "open", "people": [], "sources": [ { "title": "Sciences Po March2026 event programme", "url": "https://www.sciencespo.fr/fr/evenements/les-mardis-de-l-ia-i-from-principles-to-practice-implementing-secure-and-ethical-ai-frameworks-worldwide/" }, { "title": "CeSIA Sciences Po conference report", "url": "https://cesia.org/documents/8/Conference_Report_-_From_Principles_to_Practice.pdf" }, { "title": "G7 Digital and Technology Ministers declaration", "url": "https://www.entreprises.gouv.fr/files/files/Actualites/2026/g7/declaration-ministerielle-g7-numerique-technologies.pdf" }, { "title": "French G7 priorities", "url": "https://www.elysee.fr/G7evian/2026/02/05/conference-de-presse-sur-les-priorites-des-filieres-finances-et-numerique-du-g7-2026" }, { "title": "Macron G7 closing transcript", "url": "https://www.elysee.fr/en/emmanuel-macron/2026/06/17/g7-summit-in-evian-day-three" } ], "answer": "" }, { "id": "w8-follow-l-fathom-bill", "question": "What language and implementation powers did Fathom contribute to its advertised California AI-verification bill?", "why": "9September2026: SB813 sponsor and historical prospective operator interest acquired. Enrolled framework preserves voluntary use. Enactment, actual designation and commercial compulsion unresolved.", "evidence": "Governor action/chapter; GovOps implementing rules/designation; actual insurance/procurement/litigation conditions; Fathom application or decision", "status": "open", "people": [], "sources": [ { "title": "Fathom partnerships", "url": "https://fathom.org/partnerships" }, { "title": "McNerney announces Fathom-sponsored SB813", "url": "https://sd05.senate.ca.gov/news/mcnerney-introduces-bill-establish-safety-standards-artificial-intelligence-while-fostering" }, { "title": "SB813 Judiciary analysis: prospective sponsor interest", "url": "https://sjud.senate.ca.gov/system/files/2025-04/sb-813-mcnerney-sjud-analysis.pdf" }, { "title": "SB813 enrolled GovOps framework", "url": "https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB813" }, { "title": "SB813 official status", "url": "https://leginfo.legislature.ca.gov/faces/billStatusClient.xhtml?bill_id=202520260SB813" }, { "title": "SB813 final Senate Floor analysis", "url": "https://leginfo.legislature.ca.gov/faces/billAnalysisClient.xhtml?bill_id=202520260SB813" } ], "answer": "" }, { "id": "w8-follow-l-caif-fund", "question": "Who governs the new multi-agent safety fund involving CAIF, ARIA and industry partners?", "why": "9September2026: joint consideration and differentiated optional funding conditions read. USD10m ceiling, not receipts. Grantors, shares, conflicts protocol and disbursements unresolved.", "evidence": "Joint selection/conflicts and contribution agreements; contracting grantors/shares; executed awards/disbursements", "status": "open", "people": [], "sources": [ { "title": "Cooperative AI Foundation: backing and trustees", "url": "https://www.cooperativeai.com/foundation" }, { "title": "DeepMind multi-agent call announcement", "url": "https://deepmind.google/blog/investing-in-multi-agent-ai-safety-research/" }, { "title": "Multi-agent research competition: public call", "url": "https://schmidtsciences.smapply.io/prog/scaling_ai_safety_for_a_multi_agent_world/" }, { "title": "Call-specific multi-agent fund FAQ", "url": "https://docs.google.com/document/d/1LLRZvrKZD9rtKAWjWXPBVZpA5PTpWZatLBkPFT1IuzU/edit?tab=t.0" }, { "title": "Schmidt multi-agent research agenda", "url": "https://www.schmidtsciences.org/multi-agent-ai/" } ], "answer": "" }, { "id": "w8-follow-l-g7-minors", "question": "What does the separate2026 G7 minors instrument require, and who supplied its age-assurance language?", "why": "9September2026: three distinct G7 instruments read. Assurance endorsed, universal identity/premarket gate not adopted. UNICEF reports input; actual clause uptake and implementation plan unresolved.", "evidence": "Implementation plan and country/platform rule; UNICEF submission/receipt and negotiating drafts", "status": "open", "people": [], "sources": [ { "title": "Macron G7 closing transcript", "url": "https://www.elysee.fr/en/emmanuel-macron/2026/06/17/g7-summit-in-evian-day-three" }, { "title": "G7 ministers Common Set of Principles for minors", "url": "https://www.entreprises.gouv.fr/files/files/Actualites/2026/g7/principles-safer-and-more-secure-digital-space-for-minors.pdf" }, { "title": "G7 leaders call for safer digital space for minors", "url": "https://www.bundesregierung.de/resource/blob/975228/2442068/5fe0677f964699d24cd1c49eabe53784/2026-06-17-leader-s-call-on-a-safer-digital-space-for-minors-data.pdf?download=1" }, { "title": "G7 privacy-authorities age-assurance statement", "url": "https://www.priv.gc.ca/en/opc-news/speeches-and-statements/2026/s-d-g7-ageassurance_260626/" }, { "title": "UNICEF reports advance G7 evidence contribution", "url": "https://www.unicef.org/press-releases/note-correspondents-first-ever-g7-principles-protect-children-online" }, { "title": "5Rights calls for premarket certification after G7 outcome", "url": "https://5rightsfoundation.com/test-for-g7-leaders-will-be-holding-tech-firms-accountable/" } ], "answer": "" }, { "id": "w9-assure-l-practical-certification", "question": "Does an insurer, purchaser or court make optional AI verification practically necessary?", "why": "Proposed bill preserves optional route; workshop incentives and opposition predict another mechanism, not an exercised gate.", "evidence": "Executed terms, qualification schedule or refusal with applicable standard", "status": "open", "people": [], "sources": [ { "title": "SB813 enrolled GovOps framework", "url": "https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB813" }, { "title": "SB813 final Senate Floor analysis", "url": "https://leginfo.legislature.ca.gov/faces/billAnalysisClient.xhtml?bill_id=202520260SB813" }, { "title": "Ashby2026 verification/incentive workshop report", "url": "https://framerusercontent.com/assets/QrSLKhh1Ii8YzWNF9iLLM1hERI.pdf" } ], "answer": "" }, { "id": "w9-assure-l-g7-implementation", "question": "Which instrument implements G7 assurance commitments and enforces privacy limits?", "why": "Joint policy and guidance do not identify actual refusal/override. UNICEF calls a plan agreed but it was not recovered.", "evidence": "Published implementation plan, approved rule or platform notice; method, purpose limits and redress", "status": "open", "people": [], "sources": [ { "title": "G7 ministers Common Set of Principles for minors", "url": "https://www.entreprises.gouv.fr/files/files/Actualites/2026/g7/principles-safer-and-more-secure-digital-space-for-minors.pdf" }, { "title": "G7 leaders call for safer digital space for minors", "url": "https://www.bundesregierung.de/resource/blob/975228/2442068/5fe0677f964699d24cd1c49eabe53784/2026-06-17-leader-s-call-on-a-safer-digital-space-for-minors-data.pdf?download=1" }, { "title": "G7 privacy-authorities age-assurance statement", "url": "https://www.priv.gc.ca/en/opc-news/speeches-and-statements/2026/s-d-g7-ageassurance_260626/" }, { "title": "UNICEF reports advance G7 evidence contribution", "url": "https://www.unicef.org/press-releases/note-correspondents-first-ever-g7-principles-protect-children-online" } ], "answer": "" }, { "id": "w10-duty-l-fuller-drafts", "question": "Did selected recipients receive an earlier or materially fuller2026 duty-of-care draft?", "why": "Public availability does not settle selective access. The unexamined Butterfield clip is a user lead, not a verified quotation.2025 OPD withholding establishes a separate earlier secrecy dispute, not fuller corporate2026 versions.", "evidence": "Dated distribution register, exact attachments/version hashes, confidentiality terms, recipient acknowledgement; exact public clip before assessing its allegation", "status": "open", "people": [], "sources": [ { "title": "Department: public exposure-draft landing page —8September2026", "url": "https://www.infrastructure.gov.au/department/media/publications/exposure-draft-online-safety-amendment-digital-duty-care-bill-2026" }, { "title": "Minister Wells: My feed, my way —8September2026", "url": "https://minister.infrastructure.gov.au/wells/media-release/my-feed-my-way" }, { "title": "Exposure draft: Online Safety Amendment (Digital Duty of Care) Bill2026 —8September2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/exposure-draft-online-safety-amendment-digital-duty-of-care-bill-2026-september2026.pdf" }, { "title": "Senate outstanding orders: OPD169 duty-of-care preparation records", "url": "https://www.aph.gov.au/-/media/02_Parliamentary_Business/22_Chamber_Documents/221_Senate/Notice_Paper/Current_Notice_Paper/OPDs.pdf?hash=B52E85DFEBFCBD8E04405A2A98B1AD30CB3B330B&la=en" } ], "answer": "" }, { "id": "w10-duty-l-delegated-adult-rules", "question": "How would enacted rules and provider decisions treat lawful adult choice, adult-AI communication and researcher eligibility?", "why": "Exposure draft allocates prospective discretion and limits; future instruments, enactment and concrete implementation not established. No universal ID claim follows.", "evidence": "Enacted text/commencement; approved instruments and researcher rules; actual provider decision and review record, with privacy and adult-exception reasons", "status": "open", "people": [], "sources": [ { "title": "Exposure draft: Online Safety Amendment (Digital Duty of Care) Bill2026 —8September2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/exposure-draft-online-safety-amendment-digital-duty-of-care-bill-2026-september2026.pdf" }, { "title": "Online Safety Act2021: official compilation11December2024", "url": "https://www.legislation.gov.au/C2021A00076/2024-12-11/2024-12-11/text/original/epub/OEBPS/document_1/document_1.html" }, { "title": "Reset.Tech Australia: statutory-review submission70 —2024", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/osar-submission-70-reset-tech-australia.pdf" } ], "answer": "" }, { "id": "w11-fund-l-uk-terms", "question": "What control rights did Carnegie programme awards and network succession actually create?", "why": "Restricted purpose and named support establish capacity but not clause/publication veto. Carnegie has independent endowment/budget authority; legal Luminate/Reset and network host remain unresolved.", "evidence": "Signed grant schedules/reporting/clearance/clawback terms; trustee/programme decisions;2023 transition and fiscal-host/payroll agreement", "status": "open", "people": [], "sources": [ { "title": "Carnegie United Kingdom Trust:2021 accounts", "url": "https://carnegieuk.org/wp-content/uploads/2024/10/Annual-Report-and-Accounts-2021-2.pdf" }, { "title": "Carnegie United Kingdom Trust:2023 accounts", "url": "https://carnegieuk.org/wp-content/uploads/2024/10/Annual-report-and-accounts-31-December-2023-2.pdf" }, { "title": "Luminate funding process", "url": "https://luminategroup.com/funding-process/en" }, { "title": "Carnegie: programme transition to Online Safety Act Network —26October2023", "url": "https://carnegieuk.org/blog/online-safety-and-carnegie-uk/" } ], "answer": "" }, { "id": "w11-fund-l-alignment", "question": "What common reform position did the funded HRLC alignment work produce, and who approved it?", "why": "Named Reset-HRLC-DRW partnership is reported by the grantor, Rights-First acknowledges Mannifera support, and HRLC submission155 expressly endorses Reset model. A coalition-wide agreed position, approval rights and allocation of the specific alignment grant remain unresolved.", "evidence": "Executed2024 grant/payor and allocation records; coalition-wide agreed text and approval/decision record; attributed transmission of that specific common position", "status": "open", "people": [], "sources": [ { "title": "Mannifera: grants to Reset Australia and Human Rights Law Centre", "url": "https://mannifera.org.au/about-us/our-partners/improving-quality-public-debate/" }, { "title": "Mannifera membership guide: pooled route and governance", "url": "https://mannifera.org.au/wp-content/uploads/2024/04/Mannifera_Membership_Guide_web.pdf" }, { "title": "Australian Communities Foundation:2023/24 annual report", "url": "https://communityfoundation.org.au/wp-content/uploads/2024/12/Annual-Report-2023-24-Aus-Communities-Foundation.pdf" }, { "title": "Mannifera Executive Officer: dedicated disinformation round", "url": "https://mannifera.org.au/collaborating-to-tackle-disinformation-and-protect-democracy/" }, { "title": "Rights-First: Principles for Digital Platform Regulation", "url": "https://www.hrlc.org.au/app/uploads/2025/04/2410-Principles-for-Digital-Platforms-Regulation-Report.pdf" }, { "title": "HRLC submission155 —5 July2024", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/osar-submission-155-human-rights-law-centre.pdf" } ], "answer": "" }, { "id": "w11-fund-l-commission-terms", "question": "Did the McKinnon commissions constrain questions, findings or publication?", "why": "Named commissions and policy outputs are documented; substantive donor control and final-government-clause adoption are not.", "evidence": "2023/24 signed scopes, fees/payment, amendments, IP/publication clearance and acceptance correspondence", "status": "open", "people": [], "sources": [ { "title": "McKinnon: two Reset research commissions", "url": "https://mckinnon.co/democracy/social-media-mis-disinformation" }, { "title": "McKinnon privacy policy —September2025", "url": "https://mckinnon.co/privacy-policy" }, { "title": "Reset: Functioning or Failing? —May2024", "url": "https://au.reset.tech/uploads/Reset.Tech-Functioning-or-Failing-Report-V2C.pdf" }, { "title": "Reset: Achieving Digital Platform Public Transparency in Australia —June2024", "url": "https://au.reset.tech/Digital-platform-public-transparency.pdf" } ], "answer": "" }, { "id": "w11-fund-l-isf-reconciliation", "question": "How does the2024 annual support figure relate to the phase-two award and payments?", "why": "US$219,000 award and US$87,600 annual accrual support may overlap; no reconciliation or bank receipt acquired. Research purpose must remain separate from duty advocacy.", "evidence": "Executed award/amendments, award-account schedule, annual accrual-to-disbursement reconciliation and final report", "status": "open", "people": [], "sources": [ { "title": "Internet Society Foundation: phase-two children privacy project", "url": "https://www.isocfoundation.org/project/designing-defining-a-digital-world-in-childrens-best-interests-phase-2/" }, { "title": "Internet Society Foundation:2024 Form990", "url": "https://www.isocfoundation.org/wp-content/uploads/2025/12/2024-Filed-Form-990-Public-Disclosure-Copy-Internet-Society-Foundation.pdf" } ], "answer": "" }, { "id": "w12-gov-l-member-register", "question": "What current member and rights records reconcile the UK register after theMay2025 transition?", "why": "The2023 incorporation already records sole subscriber Scott and old PSC rights. New director-membership rules and Wood notification explain a plausible transition;April2026 change is details-only. No additive ownership mystery or proven current50/50. Public-record branch closed pending a genuinely new record.", "evidence": "Current member register, basis for PSC notices, any separate joint-voting/appointment instrument and relevant change record", "status": "parked", "people": [], "sources": [ { "title": "Reset Tech UK: replacement articles filed21May2025", "url": "https://find-and-update.company-information.service.gov.uk/company/14982650/filing-history/MzQ2NjkxNzExM2FkaXF6a2N4/document?download=0&format=pdf" }, { "title": "Reset Tech UK:7May2025 adoption resolution", "url": "https://find-and-update.company-information.service.gov.uk/company/14982650/filing-history/WUUyWDdNTUhhZGlxemtjeA/document?download=0&format=pdf" }, { "title": "Reset Tech UK: registered control categories", "url": "https://find-and-update.company-information.service.gov.uk/company/14982650/persons-with-significant-control" }, { "title": "Reset Tech UK: current officer register", "url": "https://find-and-update.company-information.service.gov.uk/company/14982650/officers" }, { "title": "Reset Tech UK:2023 incorporation and initial PSC", "url": "https://find-and-update.company-information.service.gov.uk/company/14982650/filing-history/MzM4NTM0MDY3NGFkaXF6a2N4/document?download=0&format=pdf" }, { "title": "Scott PSC04 filedApril2026: details-only change", "url": "https://find-and-update.company-information.service.gov.uk/company/14982650/filing-history/MzUxNTIxNzkwOGFkaXF6a2N4/document?download=0&format=pdf" }, { "title": "Official guarantee-company model articles", "url": "https://www.gov.uk/government/publications/model-articles-for-private-companies-limited-by-guarantee/model-articles-for-private-companies-limited-by-guarantee" } ], "answer": "" }, { "id": "w12-gov-l-committee-exercise", "question": "Did the historical Research Committee approve a Reset commission or its publication?", "why": "Accounts establish a general oversight institution, while separate branded records identify commissions. Their link is not established; legal counterparty, members and decision rights remain missing.", "evidence": "Research Committee terms, membership/appointment resolution, delegation thresholds and Reset engagement/clearance minutes", "status": "open", "people": [], "sources": [ { "title": "Susan McKinnon Research Centre:FY2024 report", "url": "https://a-ap.storyblok.com/f/3001038/x/abd4222701/fy24_smrc_financial-report.pdf" }, { "title": "Susan McKinnon Foundation trust:FY2024 financial report", "url": "https://a-ap.storyblok.com/f/3001038/x/5e9d8ac075/fy24_smf-financial-report.pdf" }, { "title": "McKinnon Funding and Financials: separate entities", "url": "https://mckinnon.co/about-us/funding-and-financials" }, { "title": "McKinnon: two Reset research commissions", "url": "https://mckinnon.co/democracy/social-media-mis-disinformation" } ], "answer": "" }, { "id": "w12-gov-l-network-delegation", "question": "Which entity and delegation governs the specific funded network output?", "why": "Privacy notice establishes shared operating services and constitution defines powers; neither identifies an OSN-specific budget, legal employer or exercised editorial authorization.", "evidence": "Operating/transition agreement, budget, employer/service allocation, board delegation and relevant conflict/approval minute", "status": "open", "people": [], "sources": [ { "title": "Online Safety Network privacy notice", "url": "https://www.onlinesafetynetwork.uk/data-and-privacy-policy/" }, { "title": "Reset Tech UK: replacement articles filed21May2025", "url": "https://find-and-update.company-information.service.gov.uk/company/14982650/filing-history/MzQ2NjkxNzExM2FkaXF6a2N4/document?download=0&format=pdf" }, { "title": "Reset Tech UK:2024 company accounts", "url": "https://find-and-update.company-information.service.gov.uk/company/14982650/filing-history/MzQ4MjM1MzE4MmFkaXF6a2N4/document?download=0&format=pdf" } ], "answer": "" }, { "id": "w13-disp-l-exact-uk-edit", "question": "Which specific submitted wording informed April2025 risk-age clarification?", "why": "Ofcom credits collective feedback including OSAN; inspected July response does not identify exact accepted minimum-age clause. Amendment credit is earned, exclusive authorship and decisive influence remain open.", "evidence": "Response-coding/decision log and drafting comparison for8.171; for search issue, final Register24.10-14 versus predecessor", "status": "open", "people": [], "sources": [ { "title": "OSAN children consultation response — July2024", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/284469-consultation-protecting-children-from-harms-online/responses/online-safety-act-network-2.pdf?v=385913" }, { "title": "Ofcom children risk-assessment statement —24 April2025", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/statement-protecting-children-from-harms-online/main-document/volume-3-assessing-the-risks-of-harms-to-children-online.pdf?v=396648" }, { "title": "SLSC25th Report: published Ofcom answer —15 May2025", "url": "https://publications.parliament.uk/pa/ld5901/ldselect/ldsecleg/123/12303.htm" }, { "title": "Ofcom further stakeholder responses —16 December2024", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/online-safety/information-for-industry/illegal-harms/annex-1--further-stakeholder-responses.pdf?v=388774" } ], "answer": "" }, { "id": "w13-disp-l-au-selection-reasons", "question": "What reasons and attributed input explain researcher and complaint-safeguard boundaries?", "why": "Government supports overarching duty but qualifies/rejects components. Draft university-only direct eligibility is narrower than NGO request; source-specific reason and clause authorship are absent.", "evidence": "Recommendation-to-clause instructions, attributed response matrix, university-eligibility decision reasons and future explanatory/data-access rules", "status": "open", "people": [], "sources": [ { "title": "HRLC submission155 —5 July2024", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/osar-submission-155-human-rights-law-centre.pdf" }, { "title": "Government response to Online Safety Act review —14 April2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/govt-response-to-the-independent-review-online-safety-act-2021.pdf" }, { "title": "Exposure draft: Online Safety Amendment (Digital Duty of Care) Bill2026 —8September2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/exposure-draft-online-safety-amendment-digital-duty-of-care-bill-2026-september2026.pdf" }, { "title": "Rights-First: Principles for Digital Platform Regulation", "url": "https://www.hrlc.org.au/app/uploads/2025/04/2410-Principles-for-Digital-Platforms-Regulation-Report.pdf" } ], "answer": "" }, { "id": "w14-tax-l-production-authority", "question": "Who selected, funded and approved the specific typology definitions?", "why": "Co-chair/drafting/foreword participation is evidenced; full credits, budget, commission and approval machinery remain uninspected. Missing finance is not proof of unpaid work.", "evidence": "Complete credits/acknowledgments, workstream terms, category-selection drafts/response log and disclosed commissioning or funding instrument", "status": "open", "people": [], "sources": [ { "title": "WEF2023 online-harms typology: scoped original text", "url": "https://www3.weforum.org/docs/WEF_Typology_of_Online_Harms_2023.pdf" }, { "title": "Resolver participant account of typology —24August2023", "url": "https://www.resolver.com/blog/online-harms-typology/" }, { "title": "SWGfL biography: John-Orr Hanna publications", "url": "https://swgfl.org.uk/about/" } ], "answer": "" }, { "id": "w14-tax-l-category-instructions", "question": "Which source-specific drafting decisions set25C-G categories and addition thresholds?", "why": "Official and NGO taxonomy use coexist with prior Australian law and broader Reset comparators. Draft differences do not identify the author, accepted clause or reason for each threshold. FOI count/penalty inconsistencies also need substantive version matching.", "evidence": "Category-by-category policy instructions, alternatives/disposition table, clearance and delivered-versus-published substantive version comparison", "status": "open", "people": [], "sources": [ { "title": "Australian review issues paper —April2024", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/online-safety-act-2021-review-issues-paper-26-april-2024.pdf" }, { "title": "Reset Digital Platform Regulation Green Paper —April2024", "url": "https://au.reset.tech/uploads/Digital-Platform-Regulation-Green-Paper.pdf" }, { "title": "Online Safety Act2021 as made: earlier harm tests", "url": "https://www.legislation.gov.au/C2021A00076/asmade/2021-07-23/text/original/epub/OEBPS/document_1/document_1.html" }, { "title": "Exposure draft: Online Safety Amendment (Digital Duty of Care) Bill2026 —8September2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/exposure-draft-online-safety-amendment-digital-duty-of-care-bill-2026-september2026.pdf" }, { "title": "FOI25-247 departmental briefings and review revision", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/foi-25-247-documents-released.pdf" }, { "title": "Government response to Online Safety Act review —14 April2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/govt-response-to-the-independent-review-online-safety-act-2021.pdf" } ], "answer": "" }, { "id": "w15-impl-l-paid-scope", "question": "Which executed scope and invoice generated DSIT transaction654609?", "why": "Original payment names supplier; report evidence is earlier and no invoice join exists. Contract, instructions and deliverables distinguish actual purchased work from capability.", "evidence": "Executed OrderForm/SOW, invoice, processing instructions, acceptance and selection records", "status": "open", "people": [], "sources": [ { "title": "DSIT February2025 spending: transaction654609", "url": "https://assets.publishing.service.gov.uk/media/69735cdea1311bdcfa0ed92b/dsit-spending-over-25000-february-2025.csv" }, { "title": "Resolver standard terms v3.0 —20August2021", "url": "https://www.resolver.com/legal/terms-of-service/" }, { "title": "ICO decision IC-374825-D9K4 — 11 November 2025", "url": "https://ico.org.uk/media2/ygomwbsa/ic-374825-d9k4.pdf" } ], "answer": "" }, { "id": "w15-impl-l-released-report", "question": "Where is the partly disclosed Crisp report and a matched referral outcome?", "why": "ICO confirms24October2025 partial release and ten June referrals. Underlying report and actual platform responses would test scope and results.", "evidence": "Released report, exact report date/version, matched escalation and platform response", "status": "open", "people": [], "sources": [ { "title": "ICO decision IC-374825-D9K4 — 11 November 2025", "url": "https://ico.org.uk/media2/ygomwbsa/ic-374825-d9k4.pdf" }, { "title": "Parliamentary testimony on NSOIT — 14 April 2026", "url": "https://committees.parliament.uk/oralevidence/17465/pdf/" } ], "answer": "" }, { "id": "w15-impl-l-adopted-policy", "question": "Which final instruction and independent operational audit governed NSOIT delivery?", "why": "Detailed public copy is draft with redacted approvals; scrutiny response did not commit to statutory/ISC change. Public testimony only partially corroborates operations.", "evidence": "Final policy/version/adoption, DPIA, completed assessment and operational audit", "status": "open", "people": [], "sources": [ { "title": "Disclosed DRAFT NSOIT compliance policy", "url": "https://bigbrotherwatch.org.uk/wp-content/uploads/2024/05/FOI2024-00576-NSOIT-Copy-compliance-policy_Redacted.pdf" }, { "title": "Current NSOIT privacy notice", "url": "https://www.gov.uk/government/publications/national-security-online-information-team-privacy-notice/national-security-online-information-team-privacy-notice" }, { "title": "Government response on NSOIT oversight — October2025", "url": "https://publications.parliament.uk/pa/cm5901/cmselect/cmsctech/1397/report.html" } ], "answer": "" }, { "id": "w15-impl-l-phone-implementation", "question": "Which elements of recommendation48 will create an actual identification obligation?", "why": "Qualified support precedes draft with no express phone-at-opening/preservation/tipoff mechanism located. A decision or instrument would distinguish deferral from implementation.", "evidence": "Disposition table, privacy/security assessment, drafting instructions and enacted instrument", "status": "open", "people": [], "sources": [ { "title": "Published updated Online Safety Act statutory review —February2025", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/report-of-the-statutory-review-of-the-online-safety-act-2021-february-2025.pdf" }, { "title": "Government response to Online Safety Act review —14 April2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/govt-response-to-the-independent-review-online-safety-act-2021.pdf" }, { "title": "Exposure draft: Online Safety Amendment (Digital Duty of Care) Bill2026 —8September2026", "url": "https://www.infrastructure.gov.au/sites/default/files/documents/exposure-draft-online-safety-amendment-digital-duty-of-care-bill-2026-september2026.pdf" } ], "answer": "" }, { "id": "w15-impl-l-current-supplier", "question": "What exact GSN scope,invoice and transition record explain the documented role andpayment?", "why": "Legal identity10433562,27January2026 officialanalysisrole and12September2025 payment nowestablished at their source scopes. Exactappointment,SOW,performancecriteria,transaction719926invoice andCrispsuccession/parallelwork remain unresolved.", "evidence": "ExecutedGSNcontract/SOW,invoice719926,performance/acceptancerecord anddatedtransitionorparallel-service arrangements", "status": "open", "people": [], "sources": [ { "title": "Current NSOIT privacy notice", "url": "https://www.gov.uk/government/publications/national-security-online-information-team-privacy-notice/national-security-online-information-team-privacy-notice" }, { "title": "DSIT February2025 spending: transaction654609", "url": "https://assets.publishing.service.gov.uk/media/69735cdea1311bdcfa0ed92b/dsit-spending-over-25000-february-2025.csv" }, { "title": "The Global Strategy Network Ltd10433562: official register", "url": "https://find-and-update.company-information.service.gov.uk/company/10433562" }, { "title": "Ministerial answer106529: GSN analysis and Faculty platform roles", "url": "https://questions-statements.parliament.uk/written-questions/detail/2026-01-19/106529" }, { "title": "DSIT September2025 spending: transaction719926", "url": "https://assets.publishing.service.gov.uk/media/6a6a1922d500fc17835859ab/dsit-spending-over-25000-sept-2025.csv" } ], "answer": "" }, { "id": "w15-impl-l-older-award", "question": "Can an accessible primary103528 award be matched to any payment or report?", "why": "Secondary listing supplies a locator only; primary route closed. No total contract/payment equivalence claimed.", "evidence": "Accessible primary award and exact supplier registration/scope with invoice reconciliation", "status": "parked", "people": [], "sources": [ { "title": "ContractsFinder103528 notice — original inaccessible", "url": "https://www.contractsfinder.service.gov.uk/Notice/47f6e657-8370-4133-98df-ee2f2194144e" }, { "title": "Secondary award listing103528 — discovery fields only", "url": "https://bidstats.uk/tenders/2023/W39/807534667" }, { "title": "DSIT February2025 spending: transaction654609", "url": "https://assets.publishing.service.gov.uk/media/69735cdea1311bdcfa0ed92b/dsit-spending-over-25000-february-2025.csv" } ], "answer": "" }, { "id": "w16-supp-l-acceptance-use", "question": "Which evaluation,acceptance andsource-access records show theplatform worked ascommissioned?", "why": "Accepted evaluation specification, price, Treasury-facing funding gate and visually redacted acceptance criteria are now read and integrated. Actual evaluation output, correctness/rights results, acceptance certificate, repository handover and exercised replacement remain not obtained.", "evidence": "Specific evaluationcommission/output,acceptancecriteria/tests,buyeracceptances,repositoryhandover andexerciseoflicensedthirdpartymaintenance", "status": "open", "people": [], "sources": [ { "title": "CDDP FOI original bundle: selected business/contract/IP passages", "url": "https://assets.publishing.service.gov.uk/media/66803d6baec8650b10090192/FOI2023-18509_Document_bundle.pdf" }, { "title": "Official CDDP FOI release: inventory and withholding", "url": "https://www.gov.uk/government/publications/foi2024-18509-the-counter-disinformation-data-project/response-the-counter-disinformation-data-project" }, { "title": "CDDP development/hosting/maintenance opportunity26087", "url": "https://production-cas-ui.crowncommercial.gov.uk/digital-outcomes/opportunities/opportunity-details/project/26087" } ], "answer": "" }, { "id": "w16-supp-l-variation-reconciliation", "question": "Which interveningterms reconcile102887 printedtotals andprove execution/payment?", "why": "Original135/137 containreal30000difference againstsimpleaddition;otherterms remainuninspected.102548visibleNo2signaturesareuncompleted. Separate contracts/caps do notformcombinedcashspend.", "evidence": "Completevariationchain,executedcounterparts,pricing/VATschedules,invoices andacceptances withcontractreferences", "status": "open", "people": [], "sources": [ { "title": "CDDP FOI original bundle: selected business/contract/IP passages", "url": "https://assets.publishing.service.gov.uk/media/66803d6baec8650b10090192/FOI2023-18509_Document_bundle.pdf" }, { "title": "Later Faculty call-off attachment — unverified indexed cover", "url": "https://www.contractsfinder.service.gov.uk/Notice/Attachment/1439ae7e-0b25-44f2-8280-1a9ed6c10620" } ], "answer": "" }, { "id": "w22-sig-l-correction", "question": "Does Lantern retraction reach prior recipients and reverse affected decisions?", "why": "Programme appeals/removal/TC notification exist. Accepted static pinned-client review establishes available deletion/opinion/checkpoint recovery paths; Lantern deployment, timely recipient application and decision reversal remain unresolved.", "evidence": "Lantern integration obligations and deployed-version evidence; recipient acknowledgements and anonymized case reopening/disposition", "status": "open", "people": [], "sources": [ { "title": "Tech Coalition — Lantern transparency report 2024", "url": "https://technologycoalition.org/wp-content/uploads/2025/06/Lantern-Transparency-Report-2024.pdf" }, { "title": "Tech Coalition — Lantern transparency report 2025", "url": "https://technologycoalition.org/wp-content/uploads/lantern-transparency-report-2025.pdf" }, { "title": "BSR — Lantern Human Rights Impact Assessment, November 2023", "url": "https://www.bsr.org/reports/BSR-Tech-Coalition-HRIA-Report.pdf" } ], "answer": "" }, { "id": "w22-sig-l-agreement", "question": "Which executed terms allocate Lantern admission, taxonomy, hosting, financial participation and funding rights?", "why": "Role descriptions omit contractual allocation, individual votes and amounts.", "evidence": "Executed multiparty agreement/amendments, Official Program Expectations, hosting instrument and programme budget", "status": "open", "people": [], "sources": [ { "title": "Tech Coalition — Lantern transparency report 2024", "url": "https://technologycoalition.org/wp-content/uploads/2025/06/Lantern-Transparency-Report-2024.pdf" }, { "title": "Tech Coalition — Lantern transparency report 2025", "url": "https://technologycoalition.org/wp-content/uploads/lantern-transparency-report-2025.pdf" }, { "title": "Tech Coalition — Lantern expands to financial institutions worldwide", "url": "https://technologycoalition.org/news/lantern-expands-to-financial-institutions-worldwide/" } ], "answer": "" }, { "id": "w22-sig-l-financial-assessment", "question": "What did the later financial-sector HRIA recommend and what was implemented?", "why": "2025 report describes review; public pilot summary is not complete financial HRIA or proof it is member-only.", "evidence": "Full financial HRIA, recommendation disposition and actual financial onboarding terms", "status": "open", "people": [], "sources": [ { "title": "Tech Coalition — Lantern transparency report 2025", "url": "https://technologycoalition.org/wp-content/uploads/lantern-transparency-report-2025.pdf" }, { "title": "Tech Coalition — Lantern financial pilot results", "url": "https://technologycoalition.org/news/lantern-financial-pilot-results/" }, { "title": "BSR — Lantern Human Rights Impact Assessment, November 2023", "url": "https://www.bsr.org/reports/BSR-Tech-Coalition-HRIA-Report.pdf" } ], "answer": "" }, { "id": "w22-sig-l-customer-execution", "question": "How did one customer map scores to Checkstep decisions and execute an appeal?", "why": "Webhook capabilities need configuration and execution records to establish consequences.", "evidence": "Executed SOW, settings/thresholds, handler implementation and redacted decision/appeal audit", "status": "open", "people": [], "sources": [ { "title": "Checkstep — Standard Integration", "url": "https://docs.checkstep.com/standard/" }, { "title": "Sightengine / Checkstep — GenAI moderation guide, current page", "url": "https://sightengine.com/the-ultimage-guide-to-genai-moderation" }, { "title": "Checkstep — Terms of Service", "url": "https://www.checkstep.com/terms-of-service" } ], "answer": "" }, { "id": "w22-sig-l-training-use", "question": "Which data actually entered model development under which accepted terms?", "why": "Two permissions differ in opt-out, controller and retention boundaries; neither proves execution.", "evidence": "Accepted terms, opt-out/feedback configuration, training lineage and correction/deletion handling", "status": "open", "people": [], "sources": [ { "title": "Checkstep — Terms of Service", "url": "https://www.checkstep.com/terms-of-service" }, { "title": "Sightengine — Terms of Service", "url": "https://sightengine.com/policies/terms" }, { "title": "Sightengine — Data Processing Addendum", "url": "https://sightengine.com/policies/dpa" } ], "answer": "" }, { "id": "w23-cash-l-applied-terms", "question": "Which accepted terms, reasons and balances governed one documented Cash App restriction?", "why": "Current terms identify powers, not their execution in a historical case or the source of an individual decision.", "evidence": "Dated accepted contract, account product, reason notice, source of restriction, appeal and fund-release chronology", "status": "open", "people": [], "sources": [ { "title": "Cash App US Terms — current page updated 26 August 2026", "url": "https://cash.app/us/en/legal/tos" }, { "title": "Cash App Acceptable Use Policy — undated current page", "url": "https://cash.app/us/en/legal/acceptable-use-policy" }, { "title": "Cash App — Your Account Is Closed, current help article", "url": "https://cash.app/help/us/en-us/8009-your-account-is-closed" }, { "title": "CFPB — Block consent order 2025-CFPB-0001", "url": "https://files.consumerfinance.gov/f/documents/cfpb_block-inc-consent-order_2025-01.pdf" } ], "answer": "" }, { "id": "w23-cash-l-implemented-appeals", "question": "How were the ordered restriction safeguards implemented and monitored?", "why": "108(c)–(d) supply real duties; the current help synopsis alone cannot prove compliance or noncompliance.", "evidence": "Implemented restriction/appeal procedure, 110–116 Compliance Plan and Board-approved annual report, request/response samples and written modifications", "status": "open", "people": [], "sources": [ { "title": "CFPB — Block consent order 2025-CFPB-0001", "url": "https://files.consumerfinance.gov/f/documents/cfpb_block-inc-consent-order_2025-01.pdf" }, { "title": "Cash App — Your Account Is Closed, current help article", "url": "https://cash.app/help/us/en-us/8009-your-account-is-closed" }, { "title": "Block — 16 January 2025 response to CFPB action", "url": "https://investors.block.xyz/investor-news/news-details/2025/Cash-Apps-Dedication-to-Customer-Service-and-Protection/default.aspx" }, { "title": "CFPB — Block enforcement case page", "url": "https://www.consumerfinance.gov/enforcement/actions/block-inc/" }, { "title": "CFPB — Block administrative docket", "url": "https://www.consumerfinance.gov/administrative-adjudication-proceedings/administrative-adjudication-docket/block-inc/" } ], "answer": "" }, { "id": "w23-cash-l-redress-execution", "question": "What was authorized and actually distributed under the CFPB Redress Plan?", "why": "Reserve, cap and company-reported civil-penalty payment do not establish consumer totals. Official AG referral now authenticates the domain link but not its administrator or checks.", "evidence": "119 non-objection and revisions, administrator appointment, 122 Redress Report and payment/receipt reconciliation", "status": "open", "people": [], "sources": [ { "title": "CFPB — Block consent order 2025-CFPB-0001", "url": "https://files.consumerfinance.gov/f/documents/cfpb_block-inc-consent-order_2025-01.pdf" }, { "title": "Block 2024 Form 10-K — scoped payment disclosure", "url": "https://www.sec.gov/Archives/edgar/data/1512673/000162828025007376/sq-20241231.htm" }, { "title": "Block 2025 Form 10-K — scoped litigation disclosure", "url": "https://www.sec.gov/Archives/edgar/data/1512673/000162828026012254/xyz-20251231.htm" }, { "title": "Cash App CFPB redress site — official AG referral established", "url": "https://www.cashappcfpbsettlement.com/" }, { "title": "Oregon AG announcement — redress-site referral only", "url": "https://www.doj.state.or.us/media-home/news-media-releases/ag-rayfield-announces-45m-multistate-settlement-in-cash-app-investigation/" } ], "answer": "" }, { "id": "w23-cash-l-consultant-execution", "question": "Who was engaged and what did the independent AML process produce?", "why": "Mandated role and reporting rights are documented; appointed firm and implemented CAP remain unidentified.", "evidence": "Executed engagement/selection notice, assessment, Board-approved CAP, state receipt, validation and extension records", "status": "open", "people": [], "sources": [ { "title": "Multistate Block consent order — Maine full copy", "url": "https://www.maine.gov/tools/whatsnew/attach.php?an=1&id=13145196" }, { "title": "Multistate Block BSA/AML consent order — CSBS body copy", "url": "https://www.csbs.org/node/521841" }, { "title": "Maine — Block order landing and transcription", "url": "https://www.maine.gov/pfr/consumercredit/enforcement/enforcement_item.shtml?id=13145196" } ], "answer": "" }, { "id": "w23-cash-l-state-payment", "question": "How do the party list, allocations and actual state payments reconcile?", "why": "Maine Appendix A lists 49 jurisdictions; Appendix B has 48 rows summing exactly to USD 79.075 m, includes Colorado and omits Wyoming. Difference is not an arithmetic shortfall or proof of missing money.", "evidence": "Payment instructions/receipts, corrected or explained Appendix B, counterpart/accession and allocation records", "status": "open", "people": [], "sources": [ { "title": "Multistate Block consent order — Maine full copy", "url": "https://www.maine.gov/tools/whatsnew/attach.php?an=1&id=13145196" }, { "title": "CSBS — 15 January 2025 multistate Block announcement", "url": "https://www.csbs.org/newsroom/state-regulators-issue-80-million-penalty-block-inc-cash-app-bsaaml-violations" }, { "title": "Block 2024 Form 10-K — scoped payment disclosure", "url": "https://www.sec.gov/Archives/edgar/data/1512673/000162828025007376/sq-20241231.htm" }, { "title": "Block 2025 Form 10-K — scoped litigation disclosure", "url": "https://www.sec.gov/Archives/edgar/data/1512673/000162828026012254/xyz-20251231.htm" } ], "answer": "" }, { "id": "w23-cash-l-july-disposition", "question": "What remedy, credit and reimbursement mechanisms are contained in the July2026 AG instruments?", "why": "The separately accepted r24 July instruments now answer this intake question; actual payment, judicial entry and later performance remain separate where unverified.", "evidence": "Acquired July agreed instruments; existing entry/reimbursement successor records and remaining redress-performance evidence.", "status": "answered", "people": [], "sources": [ { "title": "Block Q2 2026 Form 10-Q — scoped subsequent-matter disclosure", "url": "https://www.sec.gov/Archives/edgar/data/1512673/000162828026053368/xyz-20260630.htm" }, { "title": "Oregon AG announcement — redress-site referral only", "url": "https://www.doj.state.or.us/media-home/news-media-releases/ag-rayfield-announces-45m-multistate-settlement-in-cash-app-investigation/" }, { "title": "California published proposed Block judgment", "url": "https://oag.ca.gov/system/files/attachments/press-docs/People%20v%20Block%20Judgment.pdf" }, { "title": "Oregon signed-party Block stipulated judgment packet", "url": "https://www.doj.state.or.us/wp-content/uploads/2026/07/2026-07-08-Cash-App-Multistate-Stipulated-General-Judgment-and-Money-Award-OR.pdf" } ], "answer": "r24 integrates agreed redress, CFPB credits, conditional fallback approval and investigation-expense reimbursement. It does not prove extra consumer payouts or reimbursement paid. Existing successors retain entered-version and reimbursement-allocation questions; actual consumer redress, credit reconciliation and fallback performance remain unverified here." }, { "id": "w23-cash-l-doj-accrual", "question": "What later instrument, if any, resolves the DOJ matter behind the estimated loss?", "why": "Issuer Q2 2026 reports USD 526 m estimated loss; accrual is not a fine, payment or adjudicated outcome. Distinct from customer-remedy and Lantern cases.", "evidence": "Later primary disposition and accounting/payment reconciliation", "status": "parked", "people": [], "sources": [ { "title": "Block Q2 2026 Form 10-Q — scoped subsequent-matter disclosure", "url": "https://www.sec.gov/Archives/edgar/data/1512673/000162828026053368/xyz-20260630.htm" } ], "answer": "" }, { "id": "w24-ag-l-entered-july", "question": "Which final entered versions and timing records govern the July agreement?", "why": "Party assent is observed but judicial entry/case number not; literal June30 deadline precedes stipulated July8 effect. No breach or deliberate backdating inferred.", "evidence": "Entered judgment/docket entry, signed time amendment and implemented restriction procedure/annual report", "status": "open", "people": [], "sources": [ { "title": "California published proposed Block judgment", "url": "https://oag.ca.gov/system/files/attachments/press-docs/People%20v%20Block%20Judgment.pdf" }, { "title": "Oregon signed-party Block stipulated judgment packet", "url": "https://www.doj.state.or.us/wp-content/uploads/2026/07/2026-07-08-Cash-App-Multistate-Stipulated-General-Judgment-and-Money-Award-OR.pdf" } ], "answer": "" }, { "id": "w24-ag-l-reimbursement-allocation", "question": "Was the Cash App reimbursement paid, and how was it coded or allocated?", "why": "Agreed investigation-cost amount is inside the USD45m pool; no payment/expense allocation links it to older funds or NCII letter campaign.", "evidence": "Remittance/receipt, invoices, expense coding and fund/general-account allocation decision", "status": "open", "people": [], "sources": [ { "title": "Oregon signed-party Block stipulated judgment packet", "url": "https://www.doj.state.or.us/wp-content/uploads/2026/07/2026-07-08-Cash-App-Multistate-Stipulated-General-Judgment-and-Money-Award-OR.pdf" }, { "title": "NAAG FAQ — organizational resource and policy account", "url": "https://www.naag.org/about-naag/faqs/" }, { "title": "Rules and Regulations of the VW Settlement Fund", "url": "https://www.naag.org/wp-content/uploads/2020/08/VW-Settlement-Fund-Rules-and-Regulations-Final.pdf" }, { "title": "Financial Services and Consumer Protection Enforcement, Education, and Training Fund MOU", "url": "https://www.naag.org/wp-content/uploads/2020/08/FSF-MOU-Oct-2012.pdf" } ], "answer": "" }, { "id": "w24-ag-l-current-fund-rights", "question": "How do current fund rules and actual awards relate to the dated special-committee instruments?", "why": "2012/2016 committees and current consolidated roster differ; actual amendment, balance, grant or disbursement not acquired.", "evidence": "Charter/amendment/delegation, latest fund statements/audit notes, application, vote/recusal and payment records", "status": "open", "people": [], "sources": [ { "title": "Rules and Regulations of the VW Settlement Fund", "url": "https://www.naag.org/wp-content/uploads/2020/08/VW-Settlement-Fund-Rules-and-Regulations-Final.pdf" }, { "title": "Financial Services and Consumer Protection Enforcement, Education, and Training Fund MOU", "url": "https://www.naag.org/wp-content/uploads/2020/08/FSF-MOU-Oct-2012.pdf" }, { "title": "NAAG consolidated consumer-protection/charity fund committee — current page", "url": "https://www.naag.org/committee/consolidated-consumer-protection-and-charity-funds-committee/" }, { "title": "NAAG leadership and governing structure — current page", "url": "https://www.naag.org/about-naag/naag-leadership/" } ], "answer": "" }, { "id": "w24-ag-l-letter-origin", "question": "What drafting and jurisdiction sign-off records produced the August 2025 letters?", "why": "Clark authorship is officially attributed, while current May2026 procedure cannot prove 2025 circulation or external input.", "evidence": "Initiating draft/colleague letter, version history, 2025 procedure, sponsor instructions, signature audit and transmission log", "status": "open", "people": [], "sources": [ { "title": "47-AG combined payment/search letters — 22 August 2025", "url": "https://www.attorneygeneral.gov/wp-content/uploads/LTRS-TO-PAYMENT-AND-SEARCH-PLATFORMS_FINAL-47.pdf" }, { "title": "Pennsylvania AG — NCII letter authorship and coalition announcement", "url": "https://www.attorneygeneral.gov/taking-action/ag-sunday-leads-bipartisan-coalition-urging-search-engine-payment-app-companies-to-stop-the-spread-of-deepfake-videos-and-photos/" }, { "title": "NAAG Policy Letter Process — May 2026 version", "url": "https://www.naag.org/wp-content/uploads/2026/05/NAAG-Policy-Letter-Procedure_v5_2026.pdf" } ], "answer": "" }, { "id": "w24-ag-l-letter-outcome", "question": "Which recipient reply or decision is actually linked to the NCII requests?", "why": "Bounded response search found no company reply, meeting or implemented restriction tied to these letters. Idaho supplied a later threat, not outcome.", "evidence": "Reply, meeting/commitment log, policy-version change, merchant/acquirer instruction or resulting enforcement record", "status": "open", "people": [], "sources": [ { "title": "47-AG combined payment/search letters — 22 August 2025", "url": "https://www.attorneygeneral.gov/wp-content/uploads/LTRS-TO-PAYMENT-AND-SEARCH-PLATFORMS_FINAL-47.pdf" }, { "title": "Idaho AG Labrador — 19 September 2025 NCII warning", "url": "https://www.ag.idaho.gov/newsroom/labrador-letter-fighting-to-stop-big-techs-war-on-women-and-children/" } ], "answer": "" }, { "id": "w25-sf-l-notice-proof", "question": "Does the recipient-specific SF package satisfy the current statutory notice predicates?", "why": "Reported July notices and earlier general warning are distinct; confidential attachments, designated routing, receipt, responses and subsequent conduct determine the statutory bridge.", "evidence": "Complete notice, attachments, delivery/receipt proof, recipient response, court extension or adjudication", "status": "parked", "people": [ "w25-sf-july-notices", "w25-sf-notice-presumption", "w25-sf-city-attorney" ], "sources": [ { "title": "California Civil Code 1708.86 — current amended text", "url": "https://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?lawCode=CIV§ionNum=1708.86." }, { "title": "Chronicle — reported SF Apple/Google notices", "url": "https://www.sfchronicle.com/tech/article/apple-google-deepfake-nudify-apps-22350022.php" }, { "title": "WIRED — Apple and Google spokesperson responses", "url": "https://www.wired.com/story/san-francisco-demands-apple-and-google-delete-ai-nudify-apps-from-app-stores/" }, { "title": "SF City Attorney news — bounded publication recovery", "url": "https://www.sf.gov/cityattorney/news" }, { "title": "47-AG combined payment/search letters — 22 August 2025", "url": "https://www.attorneygeneral.gov/wp-content/uploads/LTRS-TO-PAYMENT-AND-SEARCH-PLATFORMS_FINAL-47.pdf" } ], "answer": "" }, { "id": "w25-sf-l-amendment-history", "question": "When did each enacted boundary enter, and what record attributes its selection?", "why": "March analysis documents proposed primary-purpose narrowing, but not obtained historical versions and later analysis prevent an adoption-date or lobbying-victory claim.", "evidence": "Permissibly supplied original dated versions, independently published committee analysis or drafting explanation", "status": "parked", "people": [ "w25-sf-ab621", "w25-sf-bauer-kahan", "w25-sf-civil-liberties-advocacy" ], "sources": [ { "title": "AB 621 — Assembly Privacy analysis, 18 March 2025", "url": "https://apcp.assembly.ca.gov/system/files/2025-03/ab-621-bauer-kahan.pdf" }, { "title": "AB 621 — Chapter 673, Statutes of 2025", "url": "https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202520260AB621" }, { "title": "AB 621 — official procedural history", "url": "https://leginfo.legislature.ca.gov/faces/billHistoryClient.xhtml?bill_id=202520260AB621" }, { "title": "Excluded historical-version attempt — chaptered bytes returned", "url": "https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202520260AB621&version=20250AB62199INT" } ], "answer": "" }, { "id": "w25-sf-l-briver-compliance", "question": "Were Briver’s required information, attestations and penalties delivered?", "why": "Required recurring records and payment are documented; no actual delivery, receipt, satisfaction or compliance adjudication was acquired. Court query stopped at a human-verification gateway.", "evidence": "Paragraph16 response, sworn attestations and domain evidence, payment/satisfaction or enforcement record", "status": "parked", "people": [ "w25-sf-briver", "w25-sf-briver-reporting", "w25-sf-city-attorney" ], "sources": [ { "title": "People v. Sol Ecom — signed Briver final judgment", "url": "https://kernochan.law.columbia.edu/sites/kernochan.law.columbia.edu/files/content/2025%20Symposium/People%20v.%20Sol%20Ecom%20Inc%20%28Briver%20Stipulated%20Final%20Judgment.PI%29.pdf" }, { "title": "SF Court Case Query — human-verification boundary", "url": "https://webapps.sftc.org/captcha/captcha.dll?referrer=https://webapps.sftc.org/ci/CaseInfo.dll?" } ], "answer": "" }, { "id": "w25-sf-l-provider-execution", "question": "What did named providers do beyond producing subpoena records?", "why": "City reports broader cooperation, while Tilak reports specific record productions. Neither establishes a named provider shutdown instruction or execution.", "evidence": "Original subpoena/production instruments, takedown direction, provider response and dated implementation record", "status": "parked", "people": [ "w25-sf-cloudflare", "w25-sf-namecheap", "w25-sf-namecom", "google", "w25-sf-case" ], "sources": [ { "title": "Karun Tilak declaration — alternative service and Exhibit 6", "url": "https://cdn.arstechnica.net/wp-content/uploads/2025/07/Chiu-v-Sol-Ecom-6-13-25-Declaration-of-Karun-Tilak.pdf" }, { "title": "SF City Attorney — ten-site outcome announcement", "url": "https://www.sf.gov/news-city-attorney-shuts-down-10-websites-that-create-nonconsensual-deepfake-pornography" } ], "answer": "" }, { "id": "w25-sf-l-drawnudes-assessment", "question": "What exactly did eSafety send to or conclude about the Drawnudes correspondent?", "why": "Interested-party claim is unverified. ClothOff messages show guidance and requests for information; unnamed later enforcement cannot establish service identity. The closest legal clock is22 December 2024, not a verified January approval regime.", "evidence": "Dated service-specific requirement, named legal basis/classification, response evidence and assessment/confirmation", "status": "parked", "people": [ "w25-sf-drawnudes", "w25-sf-operator-claim", "w25-sf-clothoff", "esafety", "w25-sf-dis-standard" ], "sources": [ { "title": "Karun Tilak declaration — alternative service and Exhibit 6", "url": "https://cdn.arstechnica.net/wp-content/uploads/2025/07/Chiu-v-Sol-Ecom-6-13-25-Declaration-of-Karun-Tilak.pdf" }, { "title": "eSafety FOI 25202 — ClothOff correspondence and circulars", "url": "https://www.esafety.gov.au/sites/default/files/2025-12/LOG_141_Document_Set.pdf" }, { "title": "eSafety disclosure Log 141 — correspondence release", "url": "https://www.esafety.gov.au/about-us/corporate-documents/freedom-of-information/disclosure-log" }, { "title": "DIS Class 1A/1B Industry Standard 2024 — F2024L00710", "url": "https://www.legislation.gov.au/F2024L00710/asmade/2024-06-21/text/original/pdf" }, { "title": "RES Class 1A/1B Industry Standard 2024 — companion clock", "url": "https://www.legislation.gov.au/F2024L00711/asmade/2024-06-21/text/original/pdf" }, { "title": "BOSE Amendment Determination 2024 — different May clock", "url": "https://www.legislation.gov.au/F2024L00590/asmade/2024-05-30/text/original/pdf" }, { "title": "Criminal Code Amendment (Deepfake Sexual Material) Act 2024", "url": "https://www.legislation.gov.au/C2024A00078/asmade/2024-09-02/text/original/pdf" }, { "title": "eSafety — unnamed UK provider formal warning", "url": "https://www.esafety.gov.au/newsroom/media-releases/esafety-moves-against-services-used-to-nudify-australian-school-children" }, { "title": "eSafety — reported withdrawal of three unnamed services", "url": "https://www.esafety.gov.au/newsroom/media-releases/esafety-action-prevents-services-nudifying-australian-school-children" }, { "title": "eSafety — unnamed Argentina-based provider direction", "url": "https://www.esafety.gov.au/newsroom/media-releases/esafety-takes-action-against-another-major-nudify-service-for-failing-to-protect-australian-children" } ], "answer": "" }, { "id": "w26-ver-l-related-party-instruments", "question": "What contracts, approvals and project ledgers explain the OHF–QuantSpark services and credit balances?", "why": "Filed control rights and two unaudited accounts establish a relationship, but not hourly rates, earmarked grant use, a classification instruction or settlement of balances. The GBP 80 invoice and GBP 1 balance differences remain unreconciled.", "evidence": "Executed service/loan agreements, related-party approvals, dated project cost ledgers and payment reconciliation", "status": "parked", "people": [ "w26-ver-ohf", "w26-ver-quantspark", "w26-ver-tat" ], "sources": [ { "title": "OHF — unaudited accounts to 31 July 2025", "url": "https://find-and-update.company-information.service.gov.uk/company/11656320/filing-history/MzUzNTQ3OTAwN2FkaXF6a2N4/document?download=0&format=pdf" }, { "title": "QuantSpark — unaudited accounts to 31 July 2025", "url": "https://find-and-update.company-information.service.gov.uk/company/10304694/filing-history/MzUzMjkyNjYyM2FkaXF6a2N4/document?format=pdf&download=0" }, { "title": "OHF — registered voting and appointment control", "url": "https://find-and-update.company-information.service.gov.uk/company/11656320/persons-with-significant-control" }, { "title": "QuantSpark — Adam Hadley registered control", "url": "https://find-and-update.company-information.service.gov.uk/company/10304694/persons-with-significant-control" } ], "answer": "" }, { "id": "w26-ver-l-canada-acceptance", "question": "Which executed terms, accepted milestones and payments apply to contributions 21415 and 22841?", "why": "Disclosure totals and a reported launch are not acceptance or disbursement. Current May 2026 programme terms cannot supply absent historical schedules, cost allocations or IP licences.", "evidence": "Signed agreements and amendments, incorporated schedules, acceptance decisions, progress reports and disbursement records", "status": "parked", "people": [ "w26-ver-public-safety-canada", "w26-ver-canada-centre", "w26-ver-ohf", "w26-ver-ca21415", "w26-ver-ca22841" ], "sources": [ { "title": "Canada — original TCAP contribution 21415", "url": "https://search.open.canada.ca/grants/record/ps-sp%2C214-2019-2020-Q4-0048%2Ccurrent" }, { "title": "Canada — amended TCAP total, agreement 21415", "url": "https://search.open.canada.ca/grants/record/ps-sp%2C088-2021-2022-Q3-00035%2Ccurrent" }, { "title": "Canada — TCAP Phase 2 contribution 22841", "url": "https://search.open.canada.ca/grants/record/ps-sp%2C088-2022-2023-Q4-00199%2Ccurrent" }, { "title": "CRF — current funding terms and conditions", "url": "https://www.publicsafety.gc.ca/cnt/cntrng-crm/crm-prvntn/fndng-prgrms/cmmnty-rslnc-fnd-trm-en.aspx?wbdisable=true" }, { "title": "Public Safety Canada — CAN033380 status account", "url": "https://commissioningerenceetrangere.ca/fileadmin/commission_ingerence_etrangere/Documents/Preuves_et_Presentations/Preuves/CAN033380.pdf" }, { "title": "Public Safety Canada — OHF Phase 2 entry", "url": "https://www.canada.ca/en/public-safety-canada/news/2023/05/government-of-canada-announces-more-than-25-million-for-projects-to-counter-radicalization-to-violence.html" } ], "answer": "" }, { "id": "w26-ver-l-us-funding-outcome", "question": "What became of the US funding process described in October 2023, and which agency held it?", "why": "The public Canadian status account identifies prospective funding but no awarding agency, value, instrument or payment. NSC coordination does not make the NSC the funder.", "evidence": "Dated funding decision, agency identification, agreement, award notice or termination of the reported process", "status": "parked", "people": [ "w26-ver-public-safety-canada", "w26-ver-us-nsc", "w26-ver-us-funding-process", "w26-ver-tat" ], "sources": [ { "title": "Public Safety Canada — CAN033380 status account", "url": "https://commissioningerenceetrangere.ca/fileadmin/commission_ingerence_etrangere/Documents/Preuves_et_Presentations/Preuves/CAN033380.pdf" } ], "answer": "" }, { "id": "w26-ver-l-uk-contract-execution", "question": "Was the pending Home Office pilot executed, and what acceptance and legal-party terms did it use?", "why": "The notice estimated a future signature and displayed unreconciled VAT values. A separate Ofcom listing remains an unverified lead in the accepted packet, not a demonstrated award or database-access right.", "evidence": "Executed contract, legal supplier identification, statement of work, pricing schedule, acceptance clauses and later performance or payment record", "status": "parked", "people": [ "w26-ver-uk-pilot", "w26-ver-tat", "w26-ver-ohf", "home-office" ], "sources": [ { "title": "Home Office — pending TCAP pilot UK5 notice", "url": "https://www.sell2wales.gov.wales/search/show/search_view.aspx?ID=SEP653567&catID=" }, { "title": "Home Office — voluntary infrastructure-provider guidance", "url": "https://www.gov.uk/government/publications/voluntary-guidance-for-internet-infrastructure-providers/voluntary-guidance-for-internet-infrastructure-providers-on-preventing-terrorism-online-accessible-version" } ], "answer": "" }, { "id": "w26-ver-l-paired-correction", "question": "Does an upstream classification correction reach a recipient and lead it to reopen a past restriction?", "why": "Current TCAP hash docs expose deletion and analyst feedback; GIFCT reports real internal corrections. Neither supplies a paired withdrawal, receipt, recipient review and restoration outcome. The old Altitude URL adapter is a distinct interface, not evidence of failure of current hash corrections.", "evidence": "Signal withdrawal record, delivery checkpoint or receipt, recipient correction policy, review decision and resulting content/account disposition", "status": "parked", "people": [ "w26-ver-tcap-hash-api", "w26-ver-tcap-hash-feedback", "w26-ver-hash-recipient", "w26-ver-gifct-hsdb", "w26-ver-altitude" ], "sources": [ { "title": "TCAP — current content-hash update documentation", "url": "https://docs.terrorismanalytics.org/docs/api-services/content-hash-list" }, { "title": "TCAP — current per-hash feedback documentation", "url": "https://docs.terrorismanalytics.org/docs/api-services/hash-feedback" }, { "title": "GIFCT — 2022 access and correction account", "url": "https://gifct.org/wp-content/uploads/2022/12/GIFCT-Transparency-Report-2022.pdf" }, { "title": "GIFCT — 2025 Annual and Transparency Report", "url": "https://gifct.org/wp-content/uploads/2026/05/2025-GIFCT-Annual-Report.pdf" }, { "title": "Altitude pinned source — tcap-adapter", "url": "https://raw.githubusercontent.com/Jigsaw-Code/altitude/496951f7d7a83508e9fac5709d4db2e91c654685/signal-service/src/importers/tcap_api.py" } ], "answer": "" }, { "id": "w26-ver-l-access-agreements", "question": "What adopted access and reuse agreements govern Archive users and GIFCT participants?", "why": "Public prerequisites, member criteria and the dated organizational Code establish gates, while exact access/reuse permissions, information-sharing correction clauses and adoption of current-linked draft IAC terms remain not obtained. No public author-wide challenge right is inferred from an authenticated member interface.", "evidence": "Current executed or publicly adopted access licence, information-sharing agreement, approved IAC terms and appointment/response minutes", "status": "parked", "people": [ "w26-ver-tcap-archive", "w26-ver-archive-audience", "w26-ver-gifct", "w26-ver-gifct-hsdb", "w26-ver-gifct-iac-draft" ], "sources": [ { "title": "TCAP — onboarding prerequisite for API access", "url": "https://docs.terrorismanalytics.org/docs/api-services/authentication" }, { "title": "TCAP — Archive policy destination without substantive text", "url": "https://terrorismanalytics.org/policies/archive" }, { "title": "GIFCT — membership dues and admission criteria", "url": "https://gifct.org/membership/" }, { "title": "GIFCT — Tech Solutions Code of Conduct v1.0 (2022)", "url": "https://gifct.org/wp-content/uploads/2022/12/GIFCT-Tech-Solution-Code-of-Ethics-GIFCT-For-Publishing-APPROVED-20221115-2.pdf" }, { "title": "GIFCT — current-linked draft IAC terms", "url": "https://gifct.org/wp-content/uploads/2026/01/IAC-Terms-of-Reference.pdf" } ], "answer": "" }, { "id": "w26-ver-l-altitude-deployment", "question": "Which named deployment used which version, source credentials, diagnostics and verdict receiver?", "why": "Pinned August 2024 code demonstrates defaults and local-log semantics. No customer configuration, actual verdict delivery, source diagnostic submission or tested correction/restoration chain was observed.", "evidence": "Versioned deployment/configuration, source-access agreement, receiver specification, acceptance test and correction-handling record", "status": "parked", "people": [ "w26-ver-altitude", "w26-ver-altitude-operator", "w26-ver-altitude-verdict-delivery", "w26-ver-tcap" ], "sources": [ { "title": "Altitude pinned source — readme", "url": "https://raw.githubusercontent.com/Jigsaw-Code/altitude/496951f7d7a83508e9fac5709d4db2e91c654685/README.md" }, { "title": "Altitude pinned source — import-config", "url": "https://raw.githubusercontent.com/Jigsaw-Code/altitude/496951f7d7a83508e9fac5709d4db2e91c654685/signal-service/src/api/importer.py" }, { "title": "Altitude pinned source — tasks", "url": "https://raw.githubusercontent.com/Jigsaw-Code/altitude/496951f7d7a83508e9fac5709d4db2e91c654685/signal-service/src/taskqueue/tasks.py" }, { "title": "Altitude pinned source — tcap-adapter", "url": "https://raw.githubusercontent.com/Jigsaw-Code/altitude/496951f7d7a83508e9fac5709d4db2e91c654685/signal-service/src/importers/tcap_api.py" }, { "title": "Altitude pinned source — base-importer", "url": "https://raw.githubusercontent.com/Jigsaw-Code/altitude/496951f7d7a83508e9fac5709d4db2e91c654685/signal-service/src/importers/importer.py" } ], "answer": "" }, { "id": "w27-roost-l-ccf-governing-instrument", "question": "Which adopted CCF deed and appointment records implement the proposed governance split?", "why": "Current self-description and the 2023 initial proposal cannot establish appointment quotas, vetoes, quorum, delegation or donor rights. Secondary identifier CC62350 and 27 May 2024 registration claim remain unverified against an official row; closed register/OData failures are not absence proof.", "evidence": "Adopted governing deed and amendments, current trustee appointment/removal record, delegations and reserved-matters schedule", "status": "parked", "people": [ "w27-roost-ccf", "w27-roost-call-leaders", "w27-roost-proposal2023" ], "sources": [ { "title": "2023 Summit — Next steps for a resilient Christchurch Call", "url": "https://www.christchurchcall.org/content/files/2024/06/Leaders-Summit-2023-Meeting-paper-Next-steps-for-a-resilient-Christchurch-Call.pdf" }, { "title": "NZ–France joint Foundation launch announcement", "url": "https://www.elysee.fr/en/emmanuel-macron/2024/05/14/joint-press-release-between-france-and-new-zealand-launch-of-the-christchurch-call-foundation" }, { "title": "CCF — current institutional self-description", "url": "https://www.christchurchcall.org/the-foundation" }, { "title": "CCF official registry entry attempt — content not acquired", "url": "https://register.charities.govt.nz/CharitiesRegister/ViewCharity?accountId=CC62350" }, { "title": "CCF official open-data query — no entity record acquired", "url": "https://www.odata.charities.govt.nz/Organisations" } ], "answer": "" }, { "id": "w27-roost-l-ccan-fundholder", "question": "Which member holds CCAN operational funds, on what terms and with what accounting?", "why": "Published terms establish an expected intermediary/MOU structure and Core decisions but not named custody, a signed memorandum, amounts, account signatories or actual compliance.", "evidence": "Executed funding memorandum, named legal fundholder, budget approvals, disbursement and reporting records", "status": "parked", "people": [ "w27-roost-ccan", "w27-roost-ccan-core", "w27-roost-ccan-fundholder", "w27-roost-ccf" ], "sources": [ { "title": "CCAN terms — internal decisions and fundholder", "url": "https://christchurchcall.network/wp-content/uploads/CCAN-TOR-2025.pdf" } ], "answer": "" }, { "id": "w27-roost-l-roost-resources", "question": "What restrictions, allocations and paid/in-kind reconciliation govern ROOST founding resources?", "why": "Three issuer totals are non-additive and unallocated. Founding labels and professional board affiliations do not establish donor instruction rights; published technical governance is not a grant contract.", "evidence": "Contribution agreements, cash/in-kind valuation schedules, receipts, donor reserved rights and board budget delegations", "status": "parked", "people": [ "w27-roost-roost", "w27-roost-resource-account", "w27-roost-roost-board", "w27-roost-tdc" ], "sources": [ { "title": "ROOST issuer launch — more than $27 million reported", "url": "https://www.prnewswire.com/news-releases/leading-technology-companies-and-foundations-back-new-initiative-to-provide-free-open-source-tools-for-a-safer-internet-in-the-ai-era-302371239.html" }, { "title": "ROOST president — resources and collaboration launch letter", "url": "https://roost.tools/blog/launching-roost-a-letter-of-gratitude/" }, { "title": "ROOST first 100 days — funding and in-kind total", "url": "https://roost.tools/blog/roost-s-first-100-days-building-planning/" }, { "title": "ROOST Project Roles — appointment, roadmap and merge powers", "url": "https://community.roost.tools/roles.html" }, { "title": "ROOST — current Board/TDC and professional affiliations", "url": "https://roost.tools/who-we-are/" } ], "answer": "" }, { "id": "w27-roost-l-roost-employer", "question": "What executed Global Impact–ROOST arrangement allocates employer, fiscal and approval responsibilities?", "why": "A current advertised vacancy establishes stated prospective employment, not universal payroll arrangements, fiscal sponsorship, grant ownership or actual hires.", "evidence": "Employer-services or fiscal agreement, authorized scope, legal parties, financial custody and appointment/approval delegations", "status": "parked", "people": [ "w27-roost-global-impact", "w27-roost-columbia-igp", "w27-roost-roost" ], "sources": [ { "title": "ROOST vacancy — Global Impact advertised employer", "url": "https://robust-open-online-safety-tools.breezy.hr/p/041580692d75-marketing-manager" }, { "title": "Columbia SIPA — ROOST incubation and intended independence", "url": "https://igp.sipa.columbia.edu/news/research-columbia-sipas-institute-global-politics-leads-launch-major-opensource-ai-safety" }, { "title": "ROOST — legal identity in website privacy statement", "url": "https://roost.tools/privacy-policy/" } ], "answer": "" }, { "id": "w27-roost-l-canada-project-acceptance", "question": "Did contribution26007 fund the CCF policy pack or any accepted ROOST output, and who approved it?", "why": "The Canadian recipient is CCF; ROOST appears in expected workstreams. A pre-award collaboration and later attributed policy pack establish real connections without a subaward, budget allocation, accepted milestone or government approval of definitions.", "evidence": "Executed contribution schedules, CCF–ROOST agreement/subaward if any, dated cost ledger, deliverable acceptance and payment record", "status": "parked", "people": [ "w27-roost-canada26007", "w27-roost-ccf", "w27-roost-ccf-elevate", "w27-roost-roost", "w27-roost-ccf-tvec-pack" ], "sources": [ { "title": "Canada contribution 26007 — CCF three workstreams", "url": "https://search.open.canada.ca/grants/record/ps-sp%2C088-2025-2026-Q3-00044%2Ccurrent" }, { "title": "CCF — Canadian investment and Elevate announcement", "url": "https://www.christchurchcall.org/canadian-investment-advances-christchurch-calls-work-to-tackle-violent-extremist-content-risk-in-emerging-technologies-and-to-support-small-and-medium-sized-platforms-to-counter-tvec/" }, { "title": "CCF partnership statement — authorized ROOST reprint", "url": "https://roost.tools/blog/christchurch-call-partners-with-roost-for-open-source-ai-tools-to-tackle-tvec/" }, { "title": "CCF contributed TVEC policy template", "url": "https://raw.githubusercontent.com/roostorg/model-community/22b3bd48562616f9379eb6fa63abe131346a6de6/resources/policy-packs/christchurch-call-foundation/tvec-policy/tvec-policy" }, { "title": "CCF policy evaluation — finite 170-item test", "url": "https://raw.githubusercontent.com/roostorg/model-community/22b3bd48562616f9379eb6fa63abe131346a6de6/resources/policy-packs/christchurch-call-foundation/tvec-policy/evaluation%20report.md" } ], "answer": "" }, { "id": "w27-roost-l-named-policy-deployment", "question": "Which exact production revisions, local policies and correction procedures governed the reported Osprey deployments?", "why": "Discord and Bluesky report production with local rules and effects. Neither account joins the CCF pack or Phoebe prototype to that deployment; no paired upstream correction, recipient update and reopened decision was acquired.", "evidence": "Versioned production configuration, adopted policy provenance, locally approved changes, effects/rollback specifications and permissible correction-to-outcome records", "status": "parked", "people": [ "w27-roost-osprey", "discord", "w27-roost-bluesky", "w27-roost-bluesky-effects", "w27-roost-ccf-tvec-pack", "w27-roost-phoebe" ], "sources": [ { "title": "Discord engineering — production Osprey architecture", "url": "https://discord.com/blog/osprey-open-sourcing-our-rule-engine" }, { "title": "Discord leader — Osprey donation and reintegration", "url": "https://discord.com/blog/how-roost-is-advancing-online-safety" }, { "title": "ROOST case study — Bluesky Osprey workflow", "url": "https://roost.tools/blog/how-bluesky-transformed-their-safety-operations-with-osprey/" }, { "title": "CCF contributed TVEC policy template", "url": "https://raw.githubusercontent.com/roostorg/model-community/22b3bd48562616f9379eb6fa63abe131346a6de6/resources/policy-packs/christchurch-call-foundation/tvec-policy/tvec-policy" }, { "title": "RMC integration guide — operator choices", "url": "https://raw.githubusercontent.com/roostorg/model-community/22b3bd48562616f9379eb6fa63abe131346a6de6/resources/RMC%20Guide%20to%20Using%20Open%20Safety%20Models.md" }, { "title": "Phoebe Osprey adapter — save and dry-run validation", "url": "https://raw.githubusercontent.com/haileyok/phoebe/dfb2bf206fc6cf757b5f10fb4c318ce9a3cce062/src/osprey/osprey.py" }, { "title": "Phoebe Ozone adapter — label methods unimplemented", "url": "https://raw.githubusercontent.com/haileyok/phoebe/dfb2bf206fc6cf757b5f10fb4c318ce9a3cce062/src/ozone/ozone.py" } ], "answer": "" }, { "id": "w27-roost-l-template-reconciliation", "question": "How were INCITE1, EDSA precedence and finite evaluation weaknesses resolved before any intended deployment?", "why": "Observed template ambiguity and a failed recruitment target warrant a versioned correction, but they are not proof of an actual bad moderation decision. The inspected August pin and finite history do not disclose every later revision.", "evidence": "Versioned policy clarification, category mapping, evaluation protocol/data provenance and release/acceptance record", "status": "parked", "people": [ "w27-roost-ccf-tvec-pack", "w27-roost-rmc", "w27-roost-maintainers" ], "sources": [ { "title": "CCF contributed TVEC policy template", "url": "https://raw.githubusercontent.com/roostorg/model-community/22b3bd48562616f9379eb6fa63abe131346a6de6/resources/policy-packs/christchurch-call-foundation/tvec-policy/tvec-policy" }, { "title": "CCF policy evaluation — finite 170-item test", "url": "https://raw.githubusercontent.com/roostorg/model-community/22b3bd48562616f9379eb6fa63abe131346a6de6/resources/policy-packs/christchurch-call-foundation/tvec-policy/evaluation%20report.md" }, { "title": "CCF policy relocation — repository history metadata", "url": "https://api.github.com/repos/roostorg/model-community/commits?path=resources%2Fpolicy-packs%2Fchristchurch-call-foundation%2Ftvec-policy&sha=22b3bd48562616f9379eb6fa63abe131346a6de6&per_page=5" }, { "title": "CCF policy earlier-path history", "url": "https://api.github.com/repos/roostorg/model-community/commits?path=gpt-oss-safeguard%2Fprojects%2Ftvec-policy&sha=22b3bd48562616f9379eb6fa63abe131346a6de6&per_page=5" } ], "answer": "" }, { "id": "w27-roost-l-pathways-engagement", "question": "Which accepted support or consulting agreement produced an implemented change, and when did Stability enroll?", "why": "Public conditions preserve implementation choice; a different Elevate terms destination is not obtained. The company says both July2024 and2025 for enrollment. No named support instruction, implemented policy change or ROOST/Lantern use follows.", "evidence": "Accepted registration/terms version, dated enrollment, scoped proposal and consulting agreement, delivered recommendation and participant adoption decision", "status": "parked", "people": [ "w27-roost-pathways", "w27-roost-pathways-terms", "w27-roost-tc-elevate", "w27-roost-stability-ai" ], "sources": [ { "title": "Tech Coalition Pathways — current eligibility and offerings", "url": "https://technologycoalition.org/programs/pathways/" }, { "title": "Pathways public terms — nine conditions", "url": "https://technologycoalition.org/programs/pathways/terms-and-conditions/" }, { "title": "Tech Coalition Elevate — tailored advisory service", "url": "https://technologycoalition.org/programs/elevate/" }, { "title": "Stability AI — Pathways and Coalition participation account", "url": "https://stability.ai/news-updates/stability-ai-joins-the-tech-coalition" }, { "title": "Published Brennan bundle — Stability AI 21 January 2025 email", "url": "https://www.brennancenter.org/sites/default/files/2025-02/tech-company-email-responses_redacted.pdf" } ], "answer": "" }, { "id": "w28-access-l-licence-contract", "question": "Which actual Microsoft–Coalition and applicant instruments govern software sublicensing?", "why": "Cloud2018terms, programme eligibility and reference-feed rights are distinct; missing contracts would locate reserved approvals and termination.", "evidence": "Master/applicant sublicense and reasoned grant/refusal record", "status": "parked", "people": [ "w28-access-photodna", "w28-access-photodna-cloud", "w28-access-tc-sublicensing", "w28-access-licensed-implementer" ], "sources": [ { "title": "PhotoDNA Cloud terms — May 2018 text", "url": "https://www.microsoft.com/en-us/photodna/termsofuse" }, { "title": "Coalition 2023 annual report — Microsoft licensing statement", "url": "https://technologycoalition.org/wp-content/uploads/Tech_Coalition_Annual-Report-2023.pdf" }, { "title": "Coalition 2024 annual report — non-member licensing and HMA", "url": "https://technologycoalition.org/wp-content/uploads/Tech-Coalition-Annual-Report-2024.pdf" }, { "title": "Coalition — expanded PhotoDNA licensing", "url": "https://technologycoalition.org/news/the-tech-coalition-empowers-industry-to-combat-online-child-sexual-abuse-with-expanded-photodna-licensing/" } ], "answer": "" }, { "id": "w28-access-l-reference-correction", "question": "Which public trace connects retraction/visibility changes to client correction and restored content?", "why": "NCMEC correction account and Snap status filtering are meaningful but do not establish universal recipient erasure/restoration.", "evidence": "Non-personal schema, audit and downstream acknowledgement", "status": "parked", "people": [ "w28-access-ncmec-reference", "w28-access-ncmec-client", "w28-access-stop-reference", "w28-access-stop-partner" ], "sources": [ { "title": "NCMEC Hash Sharing v2 — access and retraction API", "url": "https://hashsharing.ncmec.org/npo/v2/documentation/" }, { "title": "NCMEC Hash Sharing — client usage recommendations", "url": "https://hashsharing.ncmec.org/npo/v2/documentation/usage-recommendations.html" }, { "title": "Concentrix attestation — 2023 hash-list review", "url": "https://www.missingkids.org/content/dam/missingkids/pdfs/Concentrix-NCMEC-document.pdf" }, { "title": "StopNCII API v2 public reference", "url": "https://swgfl.org.uk/assets/documents/stopncii-api-reference-v2.pdf" }, { "title": "eSafety — Snap provider response on StopNCII", "url": "https://www.esafety.gov.au/sites/default/files/2026-02/BOSE-full-report-CSEA-sexual-extortion-periodic-notices-Feb2026Update_0.pdf" } ], "answer": "" }, { "id": "w28-access-l-partner-dsa", "question": "What signed current StopNCII agreement resolves purpose, fees, termination and screening review?", "why": "Posted public text has unfinished brackets, conflicting legal identifiers and inconsistent tier benefits; no participant-specific assent acquired.", "evidence": "Signed DSA, final terms/benefit schedule and reviewed decision", "status": "parked", "people": [ "w28-access-stop-partner-terms", "w28-access-stop-applicant-process", "w28-access-stop-approved", "w28-access-stop-programme" ], "sources": [ { "title": "StopNCII posted partner terms — unfinished clauses", "url": "https://stopncii.org/partners/join/register/terms-of-service/" }, { "title": "SWGfL privacy — applicant screening and challenge process", "url": "https://swgfl.org.uk/privacy/" }, { "title": "StopNCII — participation comparison", "url": "https://stopncii.org/partners/join/" }, { "title": "StopNCII Approved Participant route", "url": "https://stopncii.org/partners/join/approved-participants/" }, { "title": "StopNCII Programme Partner route", "url": "https://stopncii.org/partners/join/programme-partners/" } ], "answer": "" }, { "id": "w28-access-l-operator-grant-rights", "question": "What donor and internal-delegation instruments govern StopNCII money and policy decisions?", "why": "Pershing label supports purpose-specific income but not donor identity, full programme budget or donor approval powers.", "evidence": "Grant/partnership schedules, programme budget/delegation and approval minutes", "status": "parked", "people": [ "w28-access-swgfl", "w28-access-swgfl-trustees", "w28-access-pershing-fund", "w28-access-stopncii" ], "sources": [ { "title": "SWGfL filed FY2025 accounts — restricted funds", "url": "https://find-and-update.company-information.service.gov.uk/company/05589479/filing-history/MzQ5NjAyMTA2NGFkaXF6a2N4/document?format=pdf&download=0" }, { "title": "SWGfL articles adopted10October2024", "url": "https://find-and-update.company-information.service.gov.uk/company/05589479/filing-history/MzQ0MDQ5NTAzNmFkaXF6a2N4/document?format=pdf&download=0" } ], "answer": "" }, { "id": "w28-access-l-uksic-instruments", "question": "What do the actual UKSIC grant and consortium instruments assign for allocation, reporting, acceptance and renewal?", "why": "Childnet discloses a formal partnership, IWF describes financial-return/evaluator liaison, and Nominet reports a March 2025 programme end with continuing support. None supplies signed allocation, approval, termination or renewal terms.", "evidence": "Executed 2021/2022 grant and consortium partnership agreement; financial/reporting schedules; amendments or renewal instruments; named return/acceptance record", "status": "open", "people": [ "w28-access-nominet", "w28-access-uksic", "w28-access-uksic-consortium-agreement", "w28-access-childnet", "w28-access-swgfl", "iwf" ], "sources": [ { "title": "Childnet FY2025 annual report and accounts", "url": "https://find-and-update.company-information.service.gov.uk/company/03958416/filing-history/MzQ5NjAxODE2N2FkaXF6a2N4/document?format=pdf&download=0" }, { "title": "IWF Project Officer recruitment specification", "url": "https://www.iwf.org.uk/media/ns0pgbuk/recruitment-pack-project-officer.pdf" }, { "title": "Nominet 2021 UKSIC funding pledge", "url": "https://nominet.uk/news/uk-safer-internet-centre-receives-5-1-million-funding-lifeline-from-nominet/" }, { "title": "Nominet 2025 Impact Issue: UKSIC programme end and support", "url": "https://www.nominet.uk/wp-content/uploads/2025/08/The-2025-Impact-Issue-Final-Report.pdf" } ], "answer": "" }, { "id": "w28-access-l-uksic-reconciliation", "question": "What accounts for the GBP 132,208 gap between the FY 2025 recipient UKSIC-related rows and Nominet’s grant table?", "why": "SWGfL GBP 773,395 + IWF GBP 568,707 + Childnet GBP 437,106 = GBP 1,779,208, versus donor GBP 1,647,000, difference GBP 132,208. All year ends are 31 March 2025; common coverage/recognition bases remain unproved, especially Childnet’s complete fund composition. This is an accounting comparison, not missing or diverted cash.", "evidence": "FY2025 donor-to-recipient reconciliation, financial return, allocation schedule and restricted-fund income breakdown", "status": "open", "people": [ "w28-access-nominet", "w28-access-uksic", "w28-access-swgfl", "iwf", "w28-access-childnet", "w28-access-caf-account" ], "sources": [ { "title": "Nominet FY2025 annual report and accounts", "url": "https://nominet.uk/wp-content/uploads/2025/07/FY25-annual-report-and-accounts.pdf" }, { "title": "SWGfL filed FY2025 accounts — restricted funds", "url": "https://find-and-update.company-information.service.gov.uk/company/05589479/filing-history/MzQ5NjAyMTA2NGFkaXF6a2N4/document?format=pdf&download=0" }, { "title": "IWF FY2025 trustees report and accounts", "url": "https://www.iwf.org.uk/media/gxmk1ul2/iwf-trustees-annual-report-accounts-2024-2025.pdf" }, { "title": "Childnet FY2025 annual report and accounts", "url": "https://find-and-update.company-information.service.gov.uk/company/03958416/filing-history/MzQ5NjAxODE2N2FkaXF6a2N4/document?format=pdf&download=0" } ], "answer": "" }, { "id": "w28-access-l-sid-commission-rights", "question": "What approval rights, if any, accompanied Nominet’s disclosed input to the 2026 research design?", "why": "The published report states commissioning/funding, donor input, Childnet design lead and Opinium survey delivery. It does not reveal which questions changed or whether input was advisory, jointly agreed or backed by final approval.", "evidence": "2025/2026 research brief or commissioning agreement; design change/approval history; acceptance terms", "status": "open", "people": [ "w28-access-nominet", "w28-access-childnet", "w28-access-opinium", "w28-access-sid-research2026" ], "sources": [ { "title": "Safer Internet Day 2026: commissioning and study design", "url": "https://nominet.uk/wp-content/uploads/2026/01/Safer-Internet-Day-2026-Report.pdf" } ], "answer": "" }, { "id": "w29-enf-l-nominet-iwf-case", "question": "What exact IWF request protocol and redacted case trace connect request, registry action and correction?", "why": "2013 cooperation and later reported requests are real, but do not show a current IWF certification right, individual suspension or downstream restoration. Funding predates/follows independently.", "evidence": "Current cooperation agreement; redacted request/assessment/notice/outcome/reversal trace; current complaints instrument", "status": "parked", "people": [ "w28-access-nominet", "iwf", "w29-enf-cpp2026", "w29-enf-registrant" ], "sources": [ { "title": "Nominet 2013 registration-policy review terms of reference", "url": "https://media.nominet.uk/wp-content/uploads/2013/08/RegistrationPolicyReview-TermsofReference.pdf" }, { "title": "Nominet Criminal Practices Policy, 6 July 2026", "url": "https://nominet.uk/wp-content/uploads/2026/07/Criminal-Practices-Policy-06_07_2026.pdf" }, { "title": "Nominet criminality report for November 2022–October 2023", "url": "https://www.nominet.uk/wp-content/uploads/2024/09/Nominet-2023-Criminality-Report.pdf" }, { "title": "Nominet 2025 Impact Issue — criminal activity spread", "url": "https://www.nominet.uk/wp-content/uploads/2025/08/The-2025-Impact-Issue-Final-Report.pdf" } ], "answer": "" }, { "id": "w29-enf-l-police-registry-agreement", "question": "What operative agreement and case trace implement the Ashiko police-to-registry route?", "why": "Published2020 police policy describes an agreement and DCI authorisation, while July2026CPP has its own wording. Actual agreement/current reconciliation and case execution were not acquired.", "evidence": "Operative agreement/version schedule; authorisation and registry notification; redacted challenge/outcome record", "status": "parked", "people": [ "w29-enf-pipcu", "w29-enf-ashiko", "w29-enf-pipcu-dci", "w28-access-nominet", "w29-enf-registrar" ], "sources": [ { "title": "Operation Ashiko Suspension Policy, version1.2–20", "url": "https://www.cityoflondon.police.uk/SysSiteAssets/media/images/city-of-london/about-us/pipcu/pipcu---website-domain-suspension-policy--counterfeit-goods-.uk.pdf" }, { "title": "Nominet Criminal Practices Policy, 6 July 2026", "url": "https://nominet.uk/wp-content/uploads/2026/07/Criminal-Practices-Policy-06_07_2026.pdf" }, { "title": "IPO Intelligence Gateway intake change, effective 11 August 2025", "url": "https://www.gov.uk/government/news/new-intelligence-gateway-for-ip-crime-submissions" } ], "answer": "" }, { "id": "w29-enf-l-funding-instruments", "question": "What does the2026/27 signed Service Agreement fund and condition, and how does it reconcile with IPO additional dividends?", "why": "Official baseline commitment, paid dividend toDSIT and participant-attributed signing are distinct. Contract schedule, police receipt and programme allocation remain unknown.", "evidence": "Executed annual Service Agreement/schedules; funding assurance/payment reconciliation; permitted publication of previously non-public budget appendix", "status": "parked", "people": [ "w29-enf-ipo", "w15-impl-dsit", "w29-enf-city-police", "w29-enf-pipcu", "w29-enf-signing2026" ], "sources": [ { "title": "IPO FY2025/26 accounts — additional dividend and enforcement scope", "url": "https://www.gov.uk/government/publications/the-patent-office-annual-report-and-accounts-2025-to-2026/the-patent-office-annual-report-and-accounts-2025-to-2026" }, { "title": "IPO confirms PIPCU baseline funding and future industry co-funding pilot", "url": "https://www.gov.uk/government/news/ipo-confirms-continued-funding-for-specialist-ip-crime-unit" }, { "title": "Public post attributing 2026/27 Service Agreement signing and industry attendance", "url": "https://www.linkedin.com/posts/ipo-counter-infringement_following-our-confirmation-of-three-year-activity-7466146726517620738-7QmF" }, { "title": "City of London Police 2025 budget proposal for decision", "url": "https://democracy.cityoflondon.gov.uk/documents/s215023/FINAL_C%20-%20Copy.pdf" } ], "answer": "" }, { "id": "w29-enf-l-pilot-rules", "question": "Which co-funders, selection and conflict rules are adopted for the proposedApril2027industry pilot?", "why": "Announcement identifies joint IPO/PIPCU design but no contributor, payment, veto, case-right or accepted priority weight. Attendance at baseline signing does not answer it.", "evidence": "Approved pilot terms/criteria; contribution instruments; recusals; priority-setting and case-independence provisions", "status": "parked", "people": [ "w29-enf-ipo", "w29-enf-pipcu", "w29-enf-cofunding-pilot", "w29-enf-alliance", "w29-enf-sky-role" ], "sources": [ { "title": "IPO confirms PIPCU baseline funding and future industry co-funding pilot", "url": "https://www.gov.uk/government/news/ipo-confirms-continued-funding-for-specialist-ip-crime-unit" }, { "title": "Public post attributing 2026/27 Service Agreement signing and industry attendance", "url": "https://www.linkedin.com/posts/ipo-counter-infringement_following-our-confirmation-of-three-year-activity-7466146726517620738-7QmF" } ], "answer": "" }, { "id": "w29-enf-l-priority-current-rules", "question": "How do historical Steering Group priorities, current Gateway triage and the later IP Crime Group bodies connect in operative terms?", "why": "Records establish distinct stages and reported bodies, without a common charter, current weighting or private power over an individual case.", "evidence": "Current intake/triage guidance; delegation chart; group terms and decision/recusal records; requested-refused referral examples", "status": "parked", "people": [ "w29-enf-steering-group", "w29-enf-gateway", "w29-enf-pipcu", "w29-enf-ipcg-delivery", "w29-enf-ipcg-industry-advisory", "w29-enf-ipcg-executive" ], "sources": [ { "title": "PIPCU posted historical referral guide — remit, evidence and case acceptance", "url": "https://www.cityoflondon.police.uk/SysSiteAssets/media/images/city-of-london/about-us/pipcu/pipcu-referral-guide.pdf" }, { "title": "IPO Intelligence Gateway intake change, effective 11 August 2025", "url": "https://www.gov.uk/government/news/new-intelligence-gateway-for-ip-crime-submissions" }, { "title": "IPO FY2022/23 counter-infringement governance and KYBC forums", "url": "https://www.gov.uk/government/publications/ipo-annual-report-and-accounts-2022-to-2023/the-patent-office-annual-report-and-accounts-2022-to-2023-html" }, { "title": "IPO Counter-Infringement Strategy 2022–2027 — operational, governance and KYBC scopes", "url": "https://www.gov.uk/government/publications/ip-counter-infringement-strategy-2022-to-2027/intellectual-property-counter-infringement-strategy-2022-to-2027" } ], "answer": "" }, { "id": "w29-enf-l-rusi-commission", "question": "What funding and clearance instruments governed the March2021 RUSI research?", "why": "The five co-funders, funding-independent authorship statement, validation exclusions and later MPA citation are documented. Amounts, contracting identity and approval rights are not.", "evidence": "Grant/commission agreements, payment/allocations, review/clearance rights and design/validation record", "status": "parked", "people": [ "w29-enf-rusi", "w29-enf-rusi-report2021", "w29-enf-ipo", "w29-enf-alliance", "w1-icmec-mpa", "w29-enf-industry-trust", "w29-enf-premier-league" ], "sources": [ { "title": "RUSI Taking the Profit Out of IP Crime: Piracy and Organised Crime, March2021 corrected version", "url": "https://static.rusi.org/whr_ip_crime_web_version_0.pdf" }, { "title": "MPA FDF0068 and annexed rightsholder KYBC briefing", "url": "https://committees.parliament.uk/writtenevidence/108468/pdf/" } ], "answer": "" }, { "id": "w29-enf-l-kybc-government-disposition", "question": "Which specific KYBC options, objections and drafts were accepted or rejected after the named forums and roundtables?", "why": "The documentary path reaches consultation publication, parliamentary advocacy, report citation and government consideration. It does not yet reach an approved binding business-verification or service-cessation rule.", "evidence": "Options/disposition table; roundtable minutes; drafting/clearance record; accepted text, impact and challenge design", "status": "parked", "people": [ "w29-enf-kybc-forum", "w29-enf-music-response2021", "w29-enf-sector-vision2023", "w29-enf-alliance-kybc2020", "w29-enf-bpi-kybc2020", "w29-enf-cic-briefing2021" ], "sources": [ { "title": "Government response to Economics of Music Streaming, recommendation23", "url": "https://committees.parliament.uk/publications/7407/documents/77629/default/" }, { "title": "IPO Counter-Infringement Strategy 2022–2027 — operational, governance and KYBC scopes", "url": "https://www.gov.uk/government/publications/ip-counter-infringement-strategy-2022-to-2027/intellectual-property-counter-infringement-strategy-2022-to-2027" }, { "title": "IPO FY2022/23 counter-infringement governance and KYBC forums", "url": "https://www.gov.uk/government/publications/ipo-annual-report-and-accounts-2022-to-2023/the-patent-office-annual-report-and-accounts-2022-to-2023-html" }, { "title": "Creative Industries Sector Vision, June2023 — KYBC item", "url": "https://assets.publishing.service.gov.uk/media/64898de2b32b9e000ca96712/Creative_Industries_Sector_Vision__accessible_version_.pdf" }, { "title": "MPA FDF0068 and annexed rightsholder KYBC briefing", "url": "https://committees.parliament.uk/writtenevidence/108468/pdf/" } ], "answer": "" }, { "id": "w29-enf-l-advertising-terms-outcome", "question": "What do the agreed advertising working-group terms and eventual report establish, and how did the March2026 timetable develop?", "why": "February minutes record workplan agreement; March strategy gives expected report-back/conditional legislation and a separate OSA timetable. Neither establishes a completed2027output or current commencement status.", "evidence": "Agreed terms/workplan; delivered report/acceptance; dated consultation/commencement instruments; exact referral/action/correction rules", "status": "parked", "people": [ "w29-enf-oat", "w29-enf-ad-fraud-working-group", "w29-enf-iab-uk", "w29-enf-fraud-strategy2026", "ofcom", "w29-enf-oap2023" ], "sources": [ { "title": "Online Advertising Taskforce minutes, meeting26February2026", "url": "https://assets.publishing.service.gov.uk/media/69b92ee1635612b767a465e1/Online_Advertising_Taskforce__26_February_2026_minutes.pdf" }, { "title": "Fraud Strategy March2026 — advertising partnership and attributed statutory timetable", "url": "https://assets.publishing.service.gov.uk/media/69aea29d917847c0a4c8999e/E03551200_Fraud_Strategy_2026__English__ELAY.pdf" }, { "title": "Online Advertising Programme consultation response, July2023", "url": "https://www.gov.uk/government/consultations/online-advertising-programme-consultation/outcome/government-response-to-online-advertising-programme-consultation" } ], "answer": "" }, { "id": "w30-class-l-iwl-correction", "question": "What current IWL decision and update record shows an accepted challenge reaching advertising recipients?", "why": "Dated police removal duties and offered feeds do not establish current deployment or recipient restoration.", "evidence": "Current operating procedure or redacted inclusion/challenge/removal notice and recipient acknowledgment", "status": "parked", "people": [ "w29-enf-pipcu", "w30-class-iwl", "w30-class-ad-recipients" ], "sources": [ { "title": "Policing Fraud: Operation Creative case study released in FOIA-2021-094", "url": "https://assets.college.police.uk/s3fs-public/2021-10/FOIA-2021-094-Guidance-and-Reference-Materials.pdf" }, { "title": "Op Creative Portal contract notice 012517-2024", "url": "https://www.find-tender.service.gov.uk/Notice/012517-2024/PDF" }, { "title": "IAB UK Infringing Website List guide", "url": "https://www.iabuk.com/policy/infringing-website-list-iwl" } ], "answer": "" }, { "id": "w30-class-l-wipo-alert-agreements", "question": "What executed WIPO ALERT contributor/user terms govern updates and particular list permissions?", "why": "2019framework/currentFAQ are explicit but not executed contemporary agreements; actual PIPCU contribution remains unproved.", "evidence": "Current versioned LoU/user agreement and dated contributor admission/update record", "status": "parked", "people": [ "w30-class-wipo", "w30-class-wipo-alert", "w30-class-contributors" ], "sources": [ { "title": "WIPO/ACE/14/9 BRIP Database framework", "url": "https://www.wipo.int/edocs/mdocs/enforcement/en/wipo_ace_14/wipo_ace_14_9.pdf" }, { "title": "WIPO ALERT user FAQ: access, use and updates", "url": "https://www.wipo.int/en/web/wipo-alert/faqs" } ], "answer": "" }, { "id": "w30-class-l-advertiser-execution", "question": "Which named advertising buyer implemented a selected list and revisited it after correction?", "why": "Contract recommendations and marketplace offerings do not settle pre/post-purchase payment or restored placements.", "evidence": "Executed placement/configuration instrument and redacted correction/reconciliation record", "status": "parked", "people": [ "w30-class-ad-recipients", "w30-class-whitebullet", "w30-class-peer39" ], "sources": [ { "title": "Thinkbox supplementary parliamentary evidence ADV0038", "url": "https://committees.parliament.uk/writtenevidence/85971/html/" }, { "title": "White Bullet and Peer39 proposed Q3 launch", "url": "https://www.white-bullet.com/white-bullet-partners-with-peer39-to-help-defund-online-piracy" }, { "title": "Peer39 current marketplace: White Bullet offering", "url": "https://www.peer39.com/marketplace/" }, { "title": "LCCP 16.1.1: Responsible placement of digital adverts", "url": "https://www.gamblingcommission.gov.uk/licensees-and-businesses/lccp/condition/16-1-1-responsible-placement-of-digital-adverts" } ], "answer": "" }, { "id": "w30-class-l-pay-case", "question": "Which ALERT-PAY notice led to which provider decision and correction?", "why": "Pilot71%composite removals do not isolate merchant restriction, operator response, errors or review.", "evidence": "Versioned participation/notice/review terms and redacted completed case/outcome table", "status": "parked", "people": [ "w30-class-pay", "w30-class-payment-providers", "mastercard" ], "sources": [ { "title": "WIPO/ACE/18/40 presentation: ALERT-PAY", "url": "https://www.wipo.int/edocs/mdocs/enforcement/en/wipo_ace_18/wipo_ace_18_40_presentation.pdf" }, { "title": "WIPO/ACE/18/2/REV activities report: paragraph13", "url": "https://www.wipo.int/edocs/mdocs/govbody/en/wipo_ace_18/wipo_ace_18_2_rev.pdf" }, { "title": "Mastercard Anti-Piracy Policy", "url": "https://www.mastercard.com/us/en/anti-piracy-policy.html" } ], "answer": "" }, { "id": "w30-class-l-pay-resource", "question": "Who accepts and pays for ALERT-PAY deployment?", "why": "Approved priority and broad ER2.3 resources do not disclose a project earmark, supplier payment or go-live.", "evidence": "Project authorization, cost-centre/milestone/acceptance or dated go-live record", "status": "parked", "people": [ "w30-class-wipo", "w30-class-pay", "w30-class-wipo-er23" ], "sources": [ { "title": "WIPO approved Programme of Work and Budget 2026/27", "url": "https://www.wipo.int/documents/d/about-wipo/docs-en-budget-pdf-budget-2026-27.pdf" }, { "title": "WIPO Results, Budget and Performance", "url": "https://www.wipo.int/en/web/about-wipo/budget/index" }, { "title": "WIPO Performance Report 2024/25: ER2.3 and sector scope", "url": "https://www.wipo.int/documents/d/about-wipo/docs-en-budget-pdf-wpr-2024-25.pdf" }, { "title": "ACE18 recap: June update published August2026", "url": "https://www.wipo.int/en/web/respect-for-ip/w/ace/18-highlights-global-cooperation-and-innovation-in-ip-enforcement" } ], "answer": "" }, { "id": "w30-class-l-match-correction", "question": "How did a corrected MATCH entry affect a prior recipient's decision?", "why": "Response/removal/purge and privacy rights differ from recipient record retention and admission.", "evidence": "Redacted correction decision, recipient notice and separate admission outcome", "status": "parked", "people": [ "w30-class-match", "mc-acquirers", "stripe", "w30-class-merchants" ], "sources": [ { "title": "Mastercard Security Rules and Procedures, Merchant Edition: Chapter11", "url": "https://www.mastercard.com/content/dam/mccom/shared/business/support/rules-pdfs/SPME-Manual.pdf" }, { "title": "Mastercard MATCH Pro Privacy Notice", "url": "https://www.mastercard.com/global/en/match-privacy-notice.html" }, { "title": "Stripe: Terminated merchant files / MATCH acceptance practice", "url": "https://docs.stripe.com/disputes/match?locale=en-GB" } ], "answer": "" }, { "id": "w30-class-l-egumball-outcome", "question": "Was the ordered updated public brief filed and did eGumball later obtain a substantive remedy?", "why": "2023/2025procedural results are not delisting/restoration;28document public-brief order gives a precise reopening record.", "evidence": "Updated redacted brief inG062863 and later public disposition/removal confirmation; exact Paysafe entity reconciliation", "status": "parked", "people": [ "w30-class-egumball", "w30-class-merrick", "w30-class-egumball2025" ], "sources": [ { "title": "eGumball v Merrick Bank, G060853, 17May2023 opinion", "url": "https://www.casemine.com/judgement/us/6469db1542d466566c722fd1" }, { "title": "eGumball G062863M: April2025 opinion and May6 modification", "url": "https://www.casemine.com/judgement/us/681cabf2d974c6edd357ad3b" } ], "answer": "" }, { "id": "w30-class-l-filmon-remedy", "question": "What happened on remand to FilmOn's classification/correction claim?", "why": "Reversal and alleged2013demand establish a challenge, not label falsity, damages or restored advertising.", "evidence": "Later order/agreement and correction/recipient notice; underlying public appendix correspondence", "status": "parked", "people": [ "w30-class-filmon", "w30-class-dv-inc", "w30-class-historical-dv-reports" ], "sources": [ { "title": "FilmOn.com v DoubleVerify, S244157: Supreme Court opinion", "url": "https://www.courts.ca.gov/opinions/archive/S244157.PDF" }, { "title": "FilmOn opening brief: attributed correction account", "url": "https://supreme.courts.ca.gov/sites/default/files/supremecourt/default/documents/2-s244157-app-opening-brief-merits-011618.pdf" } ], "answer": "" }, { "id": "w30-class-l-dv-input-map", "question": "Which NCMEC files, columns and provider/domain mappings did DV use?", "why": "Incoming/outbound counts, repeated notices and threecategories make the exact transformation consequential; no misuse is inferred.", "evidence": "Dated source inventory, category version, mapping/filter and exclusion specification", "status": "parked", "people": [ "w30-class-dv", "w30-class-dv-ncmec-input", "w30-class-ncmec-outbound", "w30-class-hi-category" ], "sources": [ { "title": "DoubleVerify February2025 categories, defaults and NCMEC input account", "url": "https://doubleverify.com/blog/omnichannel/verify/taking-action-strengthening-protections-against-high-risk-ad-supported-content-and-attacking-the-csam-issue-together" }, { "title": "2021 Notifications Sent by NCMEC per ESP", "url": "https://www.missingkids.org/content/dam/missingkids/pdfs/2021-notifications-by-ncmec-per-esp.pdf" }, { "title": "2023 notifications to responsive providers", "url": "https://www.missingkids.org/content/dam/missingkids/pdfs/2023-notifications-by-ncmec-to-responsive-esp.pdf" }, { "title": "2023 notifications to unresponsive providers", "url": "https://www.missingkids.org/content/dam/missingkids/pdfs/2023-notifications-by-ncmec-to-unresponsive-esp.pdf" }, { "title": "2023 incoming CyberTipline reports by ESP", "url": "https://www.missingkids.org/content/dam/missingkids/pdfs/2023-reports-by-esp.pdf" } ], "answer": "" }, { "id": "w30-class-l-dv-correction", "question": "What governs category override and propagation from an IWF update to customer action?", "why": "Upstream daily maintenance and reported integration do not prove domain reclassification, platform update or ad restoration.", "evidence": "Applicable category/default/licence/update agreement or worked correction with recipient acknowledgment", "status": "parked", "people": [ "w30-class-iwf-url", "w30-class-hi-category", "w30-class-dv-platforms", "w30-class-advertiser-customers" ], "sources": [ { "title": "DoubleVerify publisher assessment and review statement", "url": "https://doubleverify.com/blog/omnichannel/dvmap/addressing-publisher-measurement-with-impartiality" }, { "title": "DoubleVerify April2025 default adoption and IWF partnership progress", "url": "https://doubleverify.com/company/newsroom/doubleverify-partners-with-iwf-to-drive-greater-ad-ecosystem-safety-and-strengthen-initiatives-against-child-sexual-abuse-material-online" }, { "title": "IWF annual report2025 dynamic-list maintenance and DV testimonial", "url": "https://www.iwf.org.uk/annual-data-insights-report-2025/organisation-insights/our-dynamic-lists/" } ], "answer": "" }, { "id": "w30-class-l-dv-membership-terms", "question": "What are the exact DV membership signatory, amount, period and licence/governance terms?", "why": "Over£25kband and general annual framework are not an invoice or assessment veto; FY2025accounts precede membership start.", "evidence": "Executed2025membership/fee/licence schedule, receipt and any council nomination record", "status": "parked", "people": [ "w30-class-dv", "iwf", "w30-class-iwf-council" ], "sources": [ { "title": "IWF DoubleVerify member card and contribution band", "url": "https://www.iwf.org.uk/membership/our-members/" }, { "title": "IWF annual fee framework", "url": "https://www.iwf.org.uk/membership/fees/" }, { "title": "IWF membership and Funding Council application route", "url": "https://www.iwf.org.uk/membership/" }, { "title": "IWF FY2025 accounts: membership governance and recognition scope", "url": "https://www.iwf.org.uk/media/gxmk1ul2/iwf-trustees-annual-report-accounts-2024-2025.pdf" } ], "answer": "" }, { "id": "w30-class-l-pressure-disposition", "question": "What did DV and TAG actually answer and which criteria changed?", "why": "Requests/claimed advocacy and later product reports do not isolate causation; TAG reports no BSCviolations.", "evidence": "Dated DV reply/product decision, TAG23-page response/findings and contribution correspondence", "status": "parked", "people": [ "w30-class-dv", "w30-class-tag", "w30-class-senate-inquiry", "w30-class-cma" ], "sources": [ { "title": "Senators letter to DoubleVerify", "url": "https://www.blackburn.senate.gov/services/files/F66FDFEE-380D-4C81-802A-2FD93626F8DB" }, { "title": "Senators letter to Trustworthy Accountability Group", "url": "https://www.blackburn.senate.gov/services/files/64135EE2-0EE5-46C2-9879-7FC7B454416B" }, { "title": "Check My Ads account of advocacy contribution to Senate inquiry", "url": "https://checkmyads.org/announcement/with-the-support-of-check-my-ads-institutes-advocacy-congress-launches-bipartisan-inquiry-into-adtech-firms-monetization-of-child-abuse/" }, { "title": "TAG Impact and Compliance Report2025: visible brand-safety section", "url": "https://2848641.fs1.hubspotusercontent-na1.net/hubfs/2848641/Impact%20and%20Compliance/TAG%20Impact&Compliance%20Report%202025.pdf" } ], "answer": "" }, { "id": "w30-class-l-counterpressure-outcome", "question": "What followed the preservation demand and separate Adalytics complaint?", "why": "CMAresponse asserts a position; court relief was merely requested against Adalytics and current status was not acquired.", "evidence": "Actual preservation confirmation, later instrument namingCMA, or case8:25-cv-01535disposition/settlement", "status": "parked", "people": [ "w30-class-cma", "w30-class-adalytics", "w30-class-dv-holdings", "w30-class-dv-complaint" ], "sources": [ { "title": "Paul Hastings/Bondi preservation letter for DV to Check My Ads", "url": "https://checkmyads.org/wp-content/uploads/2025/04/2025.04.14-Document-Preservation-Letter-to-Check-My-Ads-Institute_Redacted-1.pdf" }, { "title": "Check My Ads response to preservation letter", "url": "https://checkmyads.org/statement/check-my-ads-institute-statement-in-response-to-doubleverify-legal-letter/" }, { "title": "Filed DoubleVerify Holdings and DoubleVerify Inc complaint against Adalytics", "url": "https://www.adexchanger.com/wp-content/uploads/2025/05/DoubleVerify_Holdings_Inc_et_al_v_ADALYTICS_RESEARCH_LLC__mddce-25-01535__0001.0.pdf" } ], "answer": "" }, { "id": "w31-fa-l-grantconditions", "question": "What do the named CMA grant agreements permit donors to approve or reclaim?", "why": "Published awards and KR internal selection rights establish resources and entry decisions, but not CMA editorial control, tranche release or campaign allocation.", "evidence": "Executed award letters, approved applications/budgets, incorporated reporting/amendment/termination terms and recipient recognition", "status": "parked", "people": [ "w30-class-cma", "w31-fa-ford", "w31-fa-kr", "oak", "w31-fa-krboard" ], "sources": [ { "title": "Ford Foundation grant 144432 — historical approval", "url": "https://www.fordfoundation.org/work/our-grants/awarded-grants/grants-database/check-my-ads-institute-144432/" }, { "title": "KR Foundation 2023 annual report — donor-authored third-party republication", "url": "https://americansforpublictrust.org/wp-content/uploads/2025/09/2023-KR-Foundation-Annual-Report.pdf" }, { "title": "KR Foundation public 2025 grant list", "url": "https://krfnd.org/wp-content/uploads/2026/08/2025-List-of-grants.pdf" }, { "title": "Oak annual report · PDF p38", "url": "https://oakfnd.org/wp-content/uploads/2026/06/Annual-report-2025-14_4.pdf" }, { "title": "Grant application process and timeline", "url": "https://oakfnd.org/wp-content/uploads/2019/11/Oak-Foundation-Grant-application-process-and-timeline.pdf" } ], "answer": "" }, { "id": "w31-fa-l-donorrows", "question": "Can primary row facsimiles and award records resolve Democracy Fund amounts and the indexed Ford2025 entry?", "why": "Reader alignment and indexed discovery cannot establish exact payments, future commitment totals or award identity. Avoid summing future schedules and paid installments.", "evidence": "Readable original grant schedules, exact Ford2025 detail/ID and award-to-payment reconciliation", "status": "parked", "people": [ "w31-fa-democracy", "w31-fa-ford", "w30-class-cma" ], "sources": [ { "title": "Democracy Fund 2023 Form 990-PF — CMA paid and future schedules", "url": "https://democracyfund.org/wp-content/uploads/2025/02/2023-Democracy-Fund-Form-990-PF.pdf" }, { "title": "Democracy Fund 2024 Form 990-PF — CMA schedule scope", "url": "https://democracyfund.org/wp-content/uploads/2025/12/2024-Democracy-Fund-990-PF-FINAL.pdf" }, { "title": "Ford Foundation March 2025 indexed grant lead", "url": "https://www.fordfoundation.org/work/our-grants/awarded-grants/grants-database/?locations=918&order=asc&orderby=date&page=162" } ], "answer": "" }, { "id": "w31-fa-l-cmarest", "question": "What final decisions, if any, address CMA's non-crisis verification and sanctions proposals?", "why": "An explicit crisis-monetization rejection cannot be extrapolated to the rest of the31-page response; roadmap forecasts are not final adoption.", "evidence": "Final additional-safety decision, response matrix, code amendments and actual commencement/issue records", "status": "parked", "people": [ "w30-class-cma", "ofcom", "w31-fa-creatorchecks", "w31-fa-premonetization", "w31-fa-sanctions" ], "sources": [ { "title": "Response to consultation: Online Safety Act — Additional Safety Measures", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/check-my-ads-institute.pdf?v=409158" }, { "title": "Ofcom crisis response protocol — June 2026 statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/online-safety/protecting-children/illegal-harms/crisis-response-protocol-online-safety-additional-safety-measures.pdf?v=419248" }, { "title": "Ofcom roadmap — remaining additional measures forecast", "url": "https://www.ofcom.org.uk/online-safety/illegal-and-harmful-content/roadmap-to-regulation" }, { "title": "Ofcom regulatory index — issued/not-in-force labels", "url": "https://www.ofcom.org.uk/online-safety/illegal-and-harmful-content/online-safety-regulatory-documents" } ], "answer": "" }, { "id": "w31-fa-l-meetingrecord", "question": "What exact question and clearance record produced the ministerial-notice recommendation?", "why": "Ofcom gives two attributable inputs but no minute or drafting allocation. Later Full Fact text does not establish its role in the earlier meeting.", "evidence": "4August2025 agenda/minute; marked draft; initial planning document;17/22October2024 correspondence;21–26May2026 approval record", "status": "parked", "people": [ "ofcom", "w31-fa-fullfact", "w31-fa-dsitmeeting", "w31-fa-noticechange" ], "sources": [ { "title": "Ofcom crisis response protocol — June 2026 statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/online-safety/protecting-children/illegal-harms/crisis-response-protocol-online-safety-additional-safety-measures.pdf?v=419248" }, { "title": "Full Fact's Submission to Ofcom's Additional Safety Measures Consultation", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/consultations/category-1-10-weeks/consultation-online-safety---additional-safety-measures/responses/full-fact.pdf?v=409176" }, { "title": "Ofcom FOI02210914 — planning chronology and record existence", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/about-ofcom/foi/2026/june/crisis-response-protocol-criteria-and-internal-assessment-methodology.pdf?v=419823" } ], "answer": "" }, { "id": "w31-fa-l-crisisoperation", "question": "What is the issued crisis-code commencement, and what happens in a worked activation?", "why": "Statement, laid draft and SeptemberNCII commencement are distinct. Actual notice, criteria, deployment, records and correction results decide how these powers operate.", "evidence": "Issued crisis amendment and commencement; s175direction/notice; provider decision log, analysis, appeals and correction outcomes", "status": "parked", "people": [ "ofcom", "w31-fa-secretary", "w31-fa-providers", "w31-fa-protocol", "w31-fa-notice", "w31-fa-analysis" ], "sources": [ { "title": "Ofcom crisis response protocol — June 2026 statement", "url": "https://www.ofcom.org.uk/siteassets/resources/documents/online-safety/protecting-children/illegal-harms/crisis-response-protocol-online-safety-additional-safety-measures.pdf?v=419248" }, { "title": "Online Safety Act 2023 as enacted", "url": "https://www.legislation.gov.uk/ukpga/2023/50/pdfs/ukpga_20230050_en.pdf" }, { "title": "Draft amendments to children user-to-user code", "url": "https://www.gov.uk/government/publications/online-safety-act-draft-amendments-to-the-protection-of-children-code-of-practice-for-user-to-user-services/draft-amendments-to-the-protection-of-children-code-of-practice-for-user-to-user-services" }, { "title": "HCWS124 — ministerial laying statement", "url": "https://questions-statements.parliament.uk/written-statements/detail/2026-06-18/hcws124" }, { "title": "Ofcom regulatory index — issued/not-in-force labels", "url": "https://www.ofcom.org.uk/online-safety/illegal-and-harmful-content/online-safety-regulatory-documents" }, { "title": "Ofcom NCII enforcement programme — distinct commencement", "url": "https://www.ofcom.org.uk/online-safety/illegal-and-harmful-content/Enforcement-programme-into-measures-being-taken-by-services-to-prevent-users-from-encountering-or-sharing-non-consensual-intimate-imagery-NCII" } ], "answer": "" }, { "id": "w31-fa-l-divergentletters", "question": "What happened after the Statewatch-endorsed safeguards letter?", "why": "A signatory and requested remedy are established; recipient response, actual correction and legal allegations are not independently verified. Common date with DSIT meeting is not a link.", "evidence": "Acknowledgment, substantive response, case-specific appeal/restoration record or policy amendment", "status": "parked", "people": [ "w31-fa-statewatch", "w31-fa-swletter", "ofcom" ], "sources": [ { "title": "ORG letter to Ofcom and technology companies — Statewatch signatory", "url": "https://statewatch.org/wp-content/uploads/2026/04/letter_to_ofcom_tech_cos_palestine_action.pdf" }, { "title": "Statewatch published OSA letter — August2025 date and account", "url": "https://statewatch.org/news/2025/august/uk-authorities-must-clarify-whether-online-safety-act-will-lead-to-palestine-protest-censorship/" } ], "answer": "" }, { "id": "w31-fa-l-dffterms", "question": "What agreement and court record substantiate the Liberty grant and challenge?", "why": "r32 now includes13March2025 complaint and7April2025 procedural disclosure judgment. The complaint predates the identifiedJune2026grant; grant agreement/payment, original pleadings and merits remain open.", "evidence": "Grant agreement, recipient recognition/payment record, pleading and public judicial disposition", "status": "parked", "people": [ "oak", "w31-fa-dff", "w31-fa-liberty", "w31-fa-tcn" ], "sources": [ { "title": "DFF August2026 grant announcement — June round", "url": "https://digitalfreedomfund.org/blog/dff-funds-16-new-legal-actions-for-digital-rights-2026-mid-year-update-on-new-grants/" }, { "title": "DFF reported Liberty TCN grant and ongoing challenge", "url": "https://digitalfreedomfund.org/uk-government-forcing-apple-to-break-encryption/" }, { "title": "Oak annual report · PDF p38", "url": "https://oakfnd.org/wp-content/uploads/2026/06/Annual-report-2025-14_4.pdf" } ], "answer": "" }, { "id": "w32-gdi-l-dffterms", "question": "What legal force and safeguards attach to Peer Group selections?", "why": "Reported delegated selection is real enough to distinguish from staff-only decisions, but reserved institutional and contracting powers remain unspecified.", "evidence": "Signed Terms of Reference/conflict policy; March2026 adoption resolution; grant execution/payment policy", "status": "parked", "people": [ "w31-fa-dff", "w32-gdi-peer", "w32-gdi-dffboard" ], "sources": [ { "title": "DFF Community Peer Group: current public process", "url": "https://digitalfreedomfund.org/community-peer-group/" }, { "title": "DFF one-year retrospective and reported March 2026 board adoption", "url": "https://digitalfreedomfund.org/blog/one-year-of-community-led-grant-decision-making-how-its-going-so-far/" }, { "title": "DFF August2026 grant announcement — June round", "url": "https://digitalfreedomfund.org/blog/dff-funds-16-new-legal-actions-for-digital-rights-2026-mid-year-update-on-new-grants/" } ], "answer": "" }, { "id": "w32-gdi-l-pleadings", "question": "Which later public pleading or judgment settles the TCN claim and funding scope?", "why": "The acquired2025 procedural ruling confirms chronology and limited disclosure outcome, not a merits ruling or exact grant-funded work.", "evidence": "Original public complaint and later disposition; DFF grant agreement/work allocation", "status": "parked", "people": [ "w31-fa-liberty", "w32-gdi-pi", "w32-gdi-complaint" ], "sources": [ { "title": "Liberty/Privacy International complaint announcement, 13 March 2025", "url": "https://www.libertyhumanrights.org.uk/issue/uk-governments-secret-apple-data-access-order-challenged-by-liberty-and-privacy-international/" }, { "title": "Apple v Home Secretary: public procedural judgment [2025] UKIPTrib 1", "url": "https://www.judiciary.uk/wp-content/uploads/2025/04/Apple-v-Secretary-of-State-for-the-Home-Department.pdf" }, { "title": "DFF August2026 grant announcement — June round", "url": "https://digitalfreedomfund.org/blog/dff-funds-16-new-legal-actions-for-digital-rights-2026-mid-year-update-on-new-grants/" } ], "answer": "" }, { "id": "w32-gdi-l-gdiagreements", "question": "Which licence connects GDI classification changes to recipient obligations?", "why": "Historical appeal and Oracle/IAS distribution are described, but correction delivery, recipient acknowledgement and actual restoration remain separate missing steps.", "evidence": "Dated ratings-review policy/form; executed licence/update/deletion annex; resolved appeal and recipient instruction/acknowledgement; IAS versioned migration specification", "status": "parked", "people": [ "w32-gdi-gdi", "w32-gdi-ratingreview", "w32-gdi-buyers", "w32-gdi-iasproduct" ], "sources": [ { "title": "Oracle Contextual Intelligence Segment Guide", "url": "https://www.oracle.com/a/ocom/docs/oracle-contextual-intelligence-segment-guide.pdf" }, { "title": "Oracle Advertising EOL Frequently Asked Questions", "url": "https://www.oracle.com/a/ocom/docs/oracle-advertising-end-of-life-faqs.pdf" }, { "title": "Clare Melford, co-founder and Chief Executive Officer, GDI: written evidence FON0071", "url": "https://committees.parliament.uk/writtenevidence/130306/html/" }, { "title": "Privacy Policy", "url": "https://www.disinformationindex.org/privacy/" }, { "title": "IAS Enhancements to Context Control Avoidance", "url": "https://integralads.com/news/ias-enhancements-to-context-control-avoidance/" } ], "answer": "" }, { "id": "w32-gdi-l-fcdoassets", "question": "What did the FCDO GDI award permit the funded assets to become?", "why": "Outside-US funded activity, financial disclosure and commercial licensing coexist without an asset/reuse bridge. A geographic remit is not a demonstrated universal downstream-use prohibition.", "evidence": "Executed award/variations, deliverable and financial schedules, performance returns, acceptance/publication/IP and reuse terms", "status": "parked", "people": [ "w32-gdi-fcdo", "w32-gdi-cdmd", "w32-gdi-gdi" ], "sources": [ { "title": "Global Disinformation Index: written answer 149420", "url": "https://questions-statements.parliament.uk/written-questions/detail/2023-02-21/149420/" }, { "title": "Counter Disinformation and Media Development Programme Summary 2022 to 2023", "url": "https://assets.publishing.service.gov.uk/media/6773f1b34961c1185ea21b39/Counter_Disinformation_and_Media_Development_Programme_Summary_2022_to_2023.odt" }, { "title": "FCDO spending over GBP 25,000 for September 2022", "url": "https://www.gov.uk/csv-preview/671a4c7af7c956b7d065a47b/FCDO_spend_25k_September_2022.csv" }, { "title": "Clare Melford, co-founder and Chief Executive Officer, GDI: written evidence FON0071", "url": "https://committees.parliament.uk/writtenevidence/130306/html/" }, { "title": "Oracle Contextual Intelligence Segment Guide", "url": "https://www.oracle.com/a/ocom/docs/oracle-contextual-intelligence-segment-guide.pdf" } ], "answer": "" }, { "id": "w32-gdi-l-gdiidentity", "question": "Which agreement identifies the legal recipient and rights in GDI research assets?", "why": "German contribution, UK controller identity and GDI-branded output statements do not establish corporate ownership or transfer of funded background/results into commercial products.", "evidence": "Grant accession/payment record; corporate relationship; consortium background/results/exploitation and IP schedules; any named commercial licence", "status": "parked", "people": [ "w32-gdi-gdide", "w32-gdi-gdiuk", "w32-gdi-gdi", "w32-gdi-ai4trust" ], "sources": [ { "title": "AI4TRUST project 101070190 fact sheet", "url": "https://cordis.europa.eu/project/id/101070190" }, { "title": "Privacy Policy", "url": "https://www.disinformationindex.org/privacy/" }, { "title": "Very Large Online Platforms decide what counts as systemic risk research. That's self-regulation failing.", "url": "https://www.disinformationindex.org/blog/2026-04-20-very-large-online-platforms-decide-what-counts-as-systemic-risk-research-thats-self-regulation-failing/" }, { "title": "D2.1 Design of the methodological toolbox", "url": "https://ai4trust.eu/wp-content/uploads/2024/04/D2.1-Design-of-the-Methodological-Toolbox.pdf" }, { "title": "Clare Melford, co-founder and Chief Executive Officer, GDI: written evidence FON0071", "url": "https://committees.parliament.uk/writtenevidence/130306/html/" } ], "answer": "" }, { "id": "w32-gdi-l-decisionrules", "question": "Who approves AI4TRUST labels, aggregate thresholds and correction/override rules?", "why": "Sampling inputs and the reported model architecture are visible; category boundaries, approved thresholds, human override and distribution responsibilities remain not obtained. Component F1 cannot fill that gap.", "evidence": "GDI annotation codebook; final tool/warning-system release080166e529e5a48d; evaluation report080166e529e59930; approved thresholds and reviewer/recipient procedures", "status": "parked", "people": [ "w32-gdi-seeds", "w32-gdi-warning", "w32-gdi-validation", "w32-gdi-evaluation" ], "sources": [ { "title": "D2.1 Design of the methodological toolbox", "url": "https://ai4trust.eu/wp-content/uploads/2024/04/D2.1-Design-of-the-Methodological-Toolbox.pdf" }, { "title": "AI4TRUST periodic reporting, period 2", "url": "https://cordis.europa.eu/project/id/101070190/reporting" }, { "title": "AI4TRUST public results index: selected final deliverables", "url": "https://cordis.europa.eu/project/id/101070190/results" }, { "title": "Final release of disinformation detection tools and warning system: unavailable document", "url": "https://ec.europa.eu/research/participants/documents/downloadPublic?appId=PPGMS&documentIds=080166e529e5a48d" }, { "title": "Evaluation and impact assessment report: unavailable document", "url": "https://ec.europa.eu/research/participants/documents/downloadPublic?appId=PPGMS&documentIds=080166e529e59930" } ], "answer": "" }, { "id": "w32-gdi-l-researchaccess", "question": "What original GDI application/refusal and delivery record identifies its actual access channel?", "why": "r33 adds adopted general rules, AliExpress commitments, appointment and audit evidence. These do not establish GDI qualification, the paragraph governing its reported2024 refusal or delivered data; generic legal access is no longer an not obtained gap.", "evidence": "Dated GDI application, recipient/DSC response and cure exchange, applicant-matched reasoned request or actual fulfilment record", "status": "parked", "people": [ "w32-gdi-gdi", "w32-gdi-accessrequests", "w32-gdi-platforms" ], "sources": [ { "title": "Very Large Online Platforms decide what counts as systemic risk research. That's self-regulation failing.", "url": "https://www.disinformationindex.org/blog/2026-04-20-very-large-online-platforms-decide-what-counts-as-systemic-risk-research-thats-self-regulation-failing/" }, { "title": "Digital Services Act 2022/2065 — Articles40 and22", "url": "https://eur-lex.europa.eu/eli/reg/2022/2065/oj/eng" }, { "title": "Delegated Regulation2025/2050 — research-data access", "url": "https://eur-lex.europa.eu/eli/reg_del/2025/2050/oj/eng" }, { "title": "FTI AliExpress2024/25 Article37 audit — selected40(12) evidence", "url": "https://files.alicdn.com/tpsservice/118cffe16ea7541786b8e8ccf9365ce6.pdf" } ], "answer": "" }, { "id": "w33-con-l-trusteeoutput", "question": "What did Ankura actually report and what did the Commission do?", "why": "Approved appointment and required reporting do not establish submitted work, accepted findings or access performance.", "evidence": "Approved mandate/workplan; first interim report receipt/nonconfidential report; B.V.7 QA annex and Commission response", "status": "parked", "people": [ "w33-con-ankura", "w33-con-trustreport", "w33-con-ali" ], "sources": [ { "title": "C(2025)3982 final — AliExpress Article71 decision and annex", "url": "https://ec.europa.eu/newsroom/dae/redirection/document/124718" }, { "title": "Commission notice — Ankura monitoring-trustee approval", "url": "https://ec.europa.eu/newsroom/dae/redirection/document/124727" }, { "title": "Commission supervision index — AliExpress entry", "url": "https://digital-strategy.ec.europa.eu/en/policies/list-designated-vlops-and-vloses" } ], "answer": "" }, { "id": "w33-con-l-applicationpair", "question": "Can a tracked application be paired with its dated platform response and usable delivery?", "why": "FTI validity conclusions and organiser labels cannot be compared without matching identity, period, paragraph and submitted information.", "evidence": "Consented redacted application-response-cure/outcome pair and tracker coding record", "status": "parked", "people": [ "w33-con-tracker", "w33-con-aliopen", "w33-con-appcohort" ], "sources": [ { "title": "DSA40 Tracker Insights —8September2026 snapshot", "url": "https://dsa40collaboratory.eu/tracker-insights/" }, { "title": "Weizenbaum Policy Paper14 — Data Access, September2025", "url": "https://www.weizenbaum-institut.de/media/Publikationen/Weizenbaum_Policy_Paper/Weizenbaum_Policy_Paper_14.pdf" }, { "title": "FTI AliExpress2024/25 Article37 audit — selected40(12) evidence", "url": "https://files.alicdn.com/tpsservice/118cffe16ea7541786b8e8ccf9365ce6.pdf" } ], "answer": "" }, { "id": "w33-con-l-researchfunding", "question": "What project award and host funding instruments specify rights and money?", "why": "Mercator project credit differs from BMFTR/Berlin institute support; no common grant, amount, payment or approval right acquired.", "evidence": "Award agreement, recipient allocation/recognition and any reporting/editorial/data-access conditions", "status": "parked", "people": [ "w33-con-mercator", "w33-con-collab", "w33-con-weizenbaum" ], "sources": [ { "title": "Weizenbaum Policy Paper14 — Data Access, September2025", "url": "https://www.weizenbaum-institut.de/media/Publikationen/Weizenbaum_Policy_Paper/Weizenbaum_Policy_Paper_14.pdf" } ], "answer": "" }, { "id": "w33-con-l-evaluationresult", "question": "Which CDDP evaluation and Treasury decision show actual acceptance or continued funding?", "why": "Commissioned criteria and price now read; final output, agreed KPIs, acceptance certificate and exercised funding gate remain distinct gaps.", "evidence": "Initial KPI agreement, feasibility/interim/final report, HMT gateway minute, milestone acceptance and payment record", "status": "parked", "people": [ "w16-supp-cddp", "w33-con-evalcontract", "w33-con-fundgate" ], "sources": [ { "title": "CDDP FOI original bundle: selected business/contract/IP passages", "url": "https://assets.publishing.service.gov.uk/media/66803d6baec8650b10090192/FOI2023-18509_Document_bundle.pdf" }, { "title": "Official CDDP FOI release: inventory and withholding", "url": "https://www.gov.uk/government/publications/foi2024-18509-the-counter-disinformation-data-project/response-the-counter-disinformation-data-project" }, { "title": "Treasury Shared Outcomes Fund Round2 — CDDP entry", "url": "https://assets.publishing.service.gov.uk/media/6178283be90e07198108160d/Shared_Outcomes_Fund_Round_2_-_Pilot_Project_Summaries_-_FINAL.pdf" } ], "answer": "" }, { "id": "w33-con-l-osdioutput", "question": "What original OSDI assurance/product reports and advice dispositions were delivered?", "why": "PA programme assurance, PUBLIC prototype and later STCF method reuse are distinct. No original PA report, sandbox framework or acceptance certificate acquired.", "evidence": "PA inception/final report and management response; original prototype evaluation instruments; advisory minutes/shortlisting matrix", "status": "parked", "people": [ "w33-con-osdi", "pa-consulting", "w33-con-sandbox", "w33-con-expertgroup" ], "sources": [ { "title": "OSDI appoints external evaluator PA Consulting", "url": "https://onlinesafetydata.blog.gov.uk/2021/09/23/working-with-an-independent-evaluator/" }, { "title": "PUBLIC G-Cloud14 service definition — OSDI sandbox case study", "url": "https://assets.applytosupply.digitalmarketplace.service.gov.uk/g-cloud-14/documents/706515/179204743065460-service-definition-document-2024-05-03-1318.pdf" }, { "title": "PUBLIC evaluation of Safety Tech Challenge Fund Round2", "url": "https://assets.publishing.service.gov.uk/media/6707a7a3e84ae1fd8592f125/safety_tech_challenge_fund_2_evaluation.pdf" }, { "title": "CDEI OSDI expert-group account —2July2021", "url": "https://onlinesafetydata.blog.gov.uk/2021/07/02/cdei-convenes-expert-group-to-advise-on-online-safety-data-initiative-project/" } ], "answer": "" }, { "id": "w33-con-l-publiccontract", "question": "Which executed contract identifies PUBLIC's actual CDDP work?", "why": "Self-reported participation and redacted approved-subcontractor field permit several arrangements; OSDI consortium does not prove the CDDP subcontract.", "evidence": "Unredacted102548 approval, executed statement of work/subcontract and accepted allocated deliverable", "status": "parked", "people": [ "w33-con-public", "w16-supp-cddp", "w33-con-subapproval" ], "sources": [ { "title": "CDDP FOI original bundle: selected business/contract/IP passages", "url": "https://assets.publishing.service.gov.uk/media/66803d6baec8650b10090192/FOI2023-18509_Document_bundle.pdf" }, { "title": "PUBLIC reports participation in CDDP delivery", "url": "https://www.public.io/blog-post/better-collaboration-bigger-impact-how-data-sharing-accelerates-innovation" } ], "answer": "" } ], "hypotheses": [ { "id": "h-common-direction", "title": "Common direction across campaigns", "statement": "A common funding or governance centre sets consequential decisions across the separate AI, age-policy and payment campaigns. This remains an attributed investigation hypothesis; the acquired cross-sector record does not establish it.", "status": "open", "supports": [ "w2-fli-e-lightcone25", "w2-fli-ed-board-lightcone", "w1-icmec-e-paypal", "i05" ], "challenges": [ "w1-icmec-e-wicks-alternative", "w1-australia-edge-responses-reported", "w1-creator-e-output-frame" ], "assumptions": [ "Shared resources must carry decision rights and connect to the particular outputs; identity overlap alone is insufficient." ], "predictions": [ "A common instruction, approval schedule or matched restricted grant should connect otherwise separate campaign decisions." ], "next_test": "Compare named output instructions and project-specific allocations. New local budget rights, credited Oak input, coalition participation rules and legislative uptake do not establish a common direction across the separate campaigns.", "alternative": "Multiple autonomous initiatives use shared funders, talent and infrastructure; some local contracts convey specific approval rights." }, { "id": "h-aligned-selection", "title": "Aligned selection and voluntary collaboration", "statement": "Some apparent coordination arises through selecting aligned communicators and voluntary collaboration. The record supports this as an operating model in parts of the creator ecosystem, alongside separately specified production contracts.", "status": "supported", "supports": [ "n08", "w1-creator-e-output-fli", "w1-creator-e-output-frame", "w1-creator-e-rational-partnership" ], "challenges": [ "w1-creator-e-pf-terms", "e100" ], "assumptions": [ "Public independence statements may be incomplete; actual agreements can change this assessment." ], "predictions": [ "Grant letters permit editorial choice while selection criteria and renewal incentives constrain participation upstream." ], "next_test": "Compare a specific Future Tense grant letter and Frame placement agreement with the published output and revision history.", "alternative": "A named advertiser or donor exercises contract-specific approval beyond selection." }, { "id": "h-gateway-bargaining", "title": "Power through conditional access", "statement": "Providers of production money, payment access or technical interfaces can shape choices through eligibility and approval conditions. Exercised constraints are reported in the games-payment case; creator terms and proposed age interfaces show distinct capacities, not one universal command.", "status": "supported", "supports": [ "w1-payments-edge-steam-processor-objection", "w1-payments-edge-steam-retirement-choice", "w1-creator-e-pf-terms", "i11", "i13", "w3-oak-e-budget-permission", "w3-oak-e-underspend", "w3-icmec-e-coalition", "w9-assure-e-fund-google-route", "w9-assure-e-fund-google-conditions", "w9-assure-e-fund-schmidt-renewal", "w10-duty-e-draft-minister", "w10-duty-e-esafety-provider", "w10-duty-e-minister-research", "w30-class-e-match-admission", "w30-class-e-stripe-policy" ], "challenges": [ "n42", "w1-creator-e-output-frame" ], "assumptions": [ "Alternatives and switching costs determine practical leverage; no numerical bargaining power is inferred." ], "predictions": [ "The decision trail identifies eligibility terms, a rejected alternative and a consequence tied to access, even without a shared ideological organiser." ], "next_test": "Obtain Valve's dated processor/acquirer instruction chain and the exact governing account terms; distinguish network rule from downstream interpretation.", "alternative": "An upstream actor gave a campaign-specific directive; the refusal was not merely distributed risk interpretation.", "assessment_update": "The exposure draft specifies prospective ministerial and Commissioner decision points, reasonable-belief triggers and approval-rule dependencies. These are proposed authority, not exercised gates, payments or proof of universal identification." }, { "id": "h-contested-policy", "title": "Contested institutional decisions", "statement": "Policy and evaluation outcomes are negotiated among actors with different interests. The acquired California analyses and Australian minutes document revisions, objections and differentiated review access; influence is neither absent nor uniformly successful.", "status": "supported", "supports": [ "w1-icmec-e-wicks-scope", "w1-icmec-e-mpa", "w1-icmec-e-wicks-alternative", "w1-australia-edge-march-objections", "w1-australia-edge-responses-reported", "w3-5r-e-kidron-portantino", "w3-5r-e-prose-uptake", "w3-icmec-e-uptake", "w9-assure-e-sb813-version-change", "w9-assure-e-sb813-voluntary", "w9-assure-e-g7-privacy", "w9-assure-e-g7-premarket", "w10-duty-e-reset-review", "w10-duty-e-draft-adult", "w10-duty-e-research-narrower", "w31-fa-e-declinemonetization", "w31-fa-e-declineformalngo", "w31-fa-e-declinecentral" ], "challenges": [ "w1-australia-edge-unequal-draft-access", "w1-australia-edge-corby-review-commitment" ], "assumptions": [ "A minuted response need not establish adequate implementation or equal decision rights." ], "predictions": [ "Versioned drafts and signoffs show who won a disputed choice and whether safeguards constrained substantive work." ], "next_test": "Compare the operative Australian role restriction with assignment and review signoffs; obtain the MPA memorandum explicitly referenced by the California committee.", "alternative": "Nominal consultation leaves the substantive design controlled by an actor insulated from challenge.", "assessment_update": "Duty-care wave adds explicit recipient quotation of Reset, but also university-only eligibility narrower than its third-sector request and a limited private-adult exception. Input is documented; universal NGO access and final clause control are not." }, { "id": "w4-fin-h-recipient-direction", "statement": "A plausible Australian explanation is that Westpac financed technical capacity and the recipient chose programme and policy outputs under its own governance, subject to grant terms not yet recovered. The dated end of US control does not itself establish freedom from bank conditions.", "status": "open", "supports": [ "w3-icmec-e-grant", "w4-fin-e-control-ended", "w4-lh-e-owner" ], "challenges": [], "title": "Recipient direction within bank-supported capacity", "next_test": "Read actual membership/adoption resolutions, board approval/conflict records and signed grant conditions; compare them with published charter powers. Funding, self-governance and advisory access do not alone establish exercised donor direction.", "alternative": "The bank or another sponsor reserved and exercised a specific approval or policy condition; general support or vendor delivery decisions cannot establish that intervention.", "assumptions": [ "Loss of US sole membership/control concerns that relationship, not the entire funding or governance network.", "Technical delivery decisions and public policy approval are different responsibilities." ] } ], "decision_models": [ { "id": "g-valve-payment-choice", "title": "The choice Valve describes", "kind": "bargaining", "summary": "Qualitative reconstruction of Valve's attributed July–August 2025 account. Upstream communications and the full option set are not recovered; this is not a solved or scored game.", "players": [ { "node_id": "steam", "objective": "Retain payment methods while operating the catalogue, as Valve states.", "basis": "July17 and August1 company replies published by recipient GamingOnLinux.", "constraints": [ "Its proposed legality-based catalogue response was reportedly rejected." ], "outside_options": [ "Alternative providers or routes were not established in the acquired record." ], "moves": [ "Respond that legal content should remain permitted.", "Retire affected games to retain payment methods." ] }, { "node_id": "w1-payments-steam-processors", "objective": "Apply the restrictions attributed to them by Valve; internal objectives remain unobserved.", "basis": "Only Valve's account of unnamed processors/acquiring banks; no original instruction.", "constraints": [ "Reported citation of Mastercard Rule5.12.7." ], "moves": [ "Reject Valve's suggested response, according to Valve." ] } ], "sequence": [ "Valve reports receiving objections through processors/acquiring banks.", "Valve says it offered a legality-based response.", "It says processors rejected that response and cited a network brand-risk rule.", "Valve publicly attributes catalogue retirements to preserving payment methods." ], "scenarios": [ { "label": "Downstream risk interpretation", "assumptions": [ "Processors apply general account/network terms without a title-specific network command." ], "implications": [ "Account-level reasoning and appeal determine the practical decision." ], "signals": [ "Original bank notice cites its own interpretation or risk assessment." ] }, { "label": "Specific upstream instruction", "assumptions": [ "A network or other upstream party required the relevant restriction." ], "implications": [ "The decisive instruction predates and constrains the processors' choice." ], "signals": [ "Dated upstream notice identifies the scope and required action." ] } ], "depends_on": [ "edges:w1-payments-edge-steam-processor-objection", "edges:w1-payments-edge-mastercard-relay-attribution", "edges:w1-payments-edge-steam-retirement-choice", "edges:n42" ] } ], "network_layout": { "nodes": [ { "id": "5rights", "label": "5Rights", "group": "rights", "role": "Organisation", "summary": "UK foundation advocating child-centred digital regulation. Donor records and a US lobbying contract are documented. Online-safety advocacy organisation founded by Beeban Kidron; California code sponsor.", "x": 1908.5, "y": 1543.8, "degree": 17 }, { "id": "kidron", "label": "Beeban Kidron", "group": "rights", "role": "Person", "summary": "Founder of 5Rights and sponsor of the ControlAI-backed July 2026 amendment. Shared participation does not earmark 5Rights grants to ControlAI. ControlAI-listed supporter and Lords Amendment 84 sponsor. Other investigation maps her 5Rights role. 5Rights founder and listed sponsor of the 2026 AI emergency-shutdown amendment.", "x": 2090.8, "y": 1557.8, "degree": 12 }, { "id": "controlai", "label": "ControlAI", "group": "control", "role": "Organisation", "summary": "UK campaign traded by Secure Future Research. Existing major donors and grant conditions remain unresolved. AI advocacy organisation named in Hansard as the origin of emergency-shutdown proposals.", "x": 2137.3, "y": 1595.0, "degree": 28 }, { "id": "am84", "label": "Lords Amendment 84", "group": "control", "role": "Proposal", "summary": "Official status at Sep 5 cutoff: Not moved. Emergency shutdown proposal distinct from ASI-prohibition draft. Lords Amendment 84, July 2026; status Not moved. Proposal, not enacted authority.", "x": 2142.1, "y": 1585.1, "degree": 3 }, { "id": "ncose", "label": "NCOSE", "group": "safety", "role": "Organisation", "summary": "National Center on Sexual Exploitation; public co-signatory linking two distinct campaigns.", "x": 1936.2, "y": 1411.2, "degree": 5 }, { "id": "fli", "label": "Future of Life Institute", "group": "risk", "role": "Organisation", "summary": "EIN 47-1052538. Grantmaking, research and advocacy; affiliated asset transfers must not be added to persuasion spending. Named co-signatory of the June 2025 AI moratorium opposition letter.", "x": 2271.5, "y": 1399.6, "degree": 29 }, { "id": "collective-shout", "label": "Collective Shout", "group": "safety", "role": "Organisation", "summary": "Australian campaigning organisation. Old corpus claim of Basic Religious Charity exemption was contradicted; no donor secrecy inference retained.", "x": 1758.6, "y": 1442.7, "degree": 3 }, { "id": "payment-campaign", "label": "July 2025 payment campaign", "group": "safety", "role": "Campaign", "summary": "Open letter requested payment companies stop processing for targeted game platforms.", "x": 1804.1, "y": 1445.3, "degree": 7 }, { "id": "payment-processors", "label": "Payment-letter addressees", "group": "safety", "role": "Group of participants", "summary": "Collection: PayPal, Mastercard, Visa, Paysafe and Discover leadership. Letter recipients, not a single legal entity or operational controller.", "x": 1744.7, "y": 1438.5, "degree": 1 }, { "id": "itch-io", "label": "itch.io", "group": "safety", "role": "Company", "summary": "Platform attributed processor scrutiny and adult-content deindexing to Collective Shout's campaign; private processor correspondence was not obtained.", "x": 1732.8, "y": 1496.2, "degree": 8 }, { "id": "steam", "label": "Steam", "group": "safety", "role": "Platform", "summary": "Named target of the payment campaign; no separate internal decision record reviewed.", "x": 1723.5, "y": 1454.3, "degree": 3 }, { "id": "ai-moratorium-letter", "label": "June 2025 AI moratorium letter", "group": "safety", "role": "Campaign", "summary": "Shared letter/artifact. This map represents only selected signatories, not the full set of more than 140. A letter is not a directing organisation.", "x": 2101.1, "y": 1402.8, "degree": 2 }, { "id": "rouse", "label": "Jon Rouse", "group": "safety", "role": "Person", "summary": "Payment-letter signatory and chair of the completed Australian trial's stakeholder advisory board.", "x": 1857.6, "y": 1445.6, "degree": 2 }, { "id": "tankard-reist", "label": "Melinda Tankard Reist", "group": "safety", "role": "Person", "summary": "Collective Shout Movement Director; payment-letter signatory and trial stakeholder adviser.", "x": 1830.2, "y": 1447.4, "degree": 3 }, { "id": "au-age-trial", "label": "Australian age-assurance trial", "group": "state", "role": "Programme", "summary": "Government-commissioned, contractor-run 2024–25 study; final report published 31 August 2025. New March–July minutes document objections and responses. April minutes clarify panel recruitment; implementation and operative review boundaries remain unresolved.", "x": 2036.2, "y": 1473.7, "degree": 9 }, { "id": "au-infrastructure", "label": "Australian infrastructure department", "group": "state", "role": "Government", "summary": "Commissioned the trial and published its final report; historical departmental title abbreviated.", "x": 2086.9, "y": 1453.5, "degree": 3 }, { "id": "accs", "label": "ACCS · historical brand / scheme umbrella", "group": "industry", "role": "Company", "summary": "Earlier case records used ACCS for the trial consortium and certification brand. Exact legal persons now have separate nodes: AVID Certification Services Limited (14865982) and Age Check Certification Scheme Ltd (11493870). Historical claims keep their original counterpart scope; they are not silently reassigned.", "x": 2061.3, "y": 1471.1, "degree": 4 }, { "id": "trial-board", "label": "Trial stakeholder advisory board", "group": "state", "role": "Advisory body", "summary": "Contractor trial stakeholder board; not a statutory regulator. Live roster does not establish an ongoing 2026 mandate. New March–July minutes document objections and responses. April minutes clarify panel recruitment; implementation and operative review boundaries remain unresolved.", "x": 1922.7, "y": 1456.3, "degree": 5 }, { "id": "corby", "label": "Iain Corby", "group": "state", "role": "Person", "summary": "Trial stakeholder-engagement lead through SafetyTech Ltd; AVPA roles and euCONSENT role disclosed with mitigations.", "x": 2082.0, "y": 1537.7, "degree": 5 }, { "id": "avpa", "label": "Age Verification Providers Association", "group": "industry", "role": "Industry body", "summary": "Age-verification industry association. Its technical evidence is expressly credited in Ofcom’s January 2025 acceptance of email-based estimation; older Australian trial participation remains separately recorded.", "x": 2126.4, "y": 1501.6, "degree": 4 }, { "id": "euconsent", "label": "euCONSENT", "group": "credentials", "role": "Organisation", "summary": "Historical project umbrella spanning the earlier EC-funded consortium and later Belgian euCONSENT ASBL. The later ASBL published the Safe Online-supported AgeAware proposal. Proposed commercial rules and reported proof of concept are distinct from a current mandatory service.", "x": 2129.8, "y": 1585.2, "degree": 4 }, { "id": "innovate-uk", "label": "Innovate UK", "group": "state", "role": "Government", "summary": "Identified public funder of Project DefAI in the trial conflict register; amount not recovered.", "x": 2164.4, "y": 1420.3, "degree": 1 }, { "id": "defai", "label": "Project DefAI", "group": "state", "role": "Project", "summary": "Delivered by AVPA and ACCS, according to trial conflict register.", "x": 2120.3, "y": 1438.0, "degree": 4 }, { "id": "illuminate-tech", "label": "Illuminate Tech", "group": "state", "role": "Company", "summary": "Advises Project DefAI on ethics, data collection and outreach; trial delivery-team participant.", "x": 2114.9, "y": 1367.1, "degree": 3 }, { "id": "digi", "label": "DIGI", "group": "state", "role": "Industry body", "summary": "Industry body whose April 2025 objections appear in official FOI documents; objections are not adjudicated trial defects. New March–July minutes document objections and responses. April minutes clarify panel recruitment; implementation and operative review boundaries remain unresolved.", "x": 2070.2, "y": 1453.6, "degree": 1 }, { "id": "social-research-centre", "label": "Social Research Centre", "group": "state", "role": "Organisation", "summary": "Consumer research supplier; official August 2024 order records A$278,293.40 including GST.", "x": 2127.9, "y": 1433.6, "degree": 1 }, { "id": "esafety", "label": "Australian eSafety", "group": "state", "role": "Regulator", "summary": "Regulator with formal international policy relationships; age-assurance trial board is a different body.", "x": 1858.0, "y": 1579.1, "degree": 18 }, { "id": "inman-grant", "label": "Julie Inman Grant", "group": "state", "role": "Person", "summary": "eSafety Commissioner; WeProtect board and WEF coalition involvement recorded in 2023–24 official report.", "x": 1875.1, "y": 1567.1, "degree": 6 }, { "id": "weprotect", "label": "WeProtect", "group": "state", "role": "Organisation", "summary": "International alliance with government, industry and NGO board participation. UK and Dutch entities/accounts require careful separation. Participated in WEF digital-safety principles working group; separate from WEF.", "x": 1779.2, "y": 1594.7, "degree": 22 }, { "id": "wef", "label": "World Economic Forum", "group": "state", "role": "Organisation", "summary": "Browne republication and NTI event links here; online-safety institutional links mapped separately. Convened Global Coalition for Digital Safety and related principles work. Digital Trust is a different WEF programme.", "x": 1938.9, "y": 1536.1, "degree": 5 }, { "id": "wef-safety-principles", "label": "WEF digital-safety principles", "group": "state", "role": "Policy", "summary": "2023 joint policy-development work involving NGOs, regulated firms and regulators.", "x": 1888.8, "y": 1521.9, "degree": 6 }, { "id": "tech-principles-members", "label": "Microsoft / Meta / Amazon", "group": "state", "role": "Group of participants", "summary": "Named participants in the WEF digital-safety working group. Google is represented separately to preserve its other documented relationships.", "x": 1898.9, "y": 1504.4, "degree": 1 }, { "id": "ofcom", "label": "Ofcom", "group": "state", "role": "Regulator", "summary": "UK communications and online-safety regulator. Age-assurance guidance, consultation dispositions and enforcement are now mapped. Its earlier WEF working-group participation remains a separate relationship.", "x": 1965.1, "y": 1405.4, "degree": 37 }, { "id": "nij", "label": "US National Institute of Justice", "group": "safety", "role": "Government", "summary": "Awarded NCOSE a sex-trafficking demand-reduction research grant; no gaming-campaign earmark shown.", "x": 1896.8, "y": 1390.9, "degree": 1 }, { "id": "jensen-project", "label": "The Jensen Project", "group": "safety", "role": "Funder", "summary": "Identifies NCOSE as a 2021–2025 policy grantee; published purpose, no verified amount.", "x": 1896.3, "y": 1397.7, "degree": 1 }, { "id": "pentecost", "label": "Pentecost family foundation", "group": "safety", "role": "Funder", "summary": "Own foundation page says it supports NCOSE; amount and contractual conditions were not verified.", "x": 1902.9, "y": 1385.7, "degree": 1 }, { "id": "message-scanning-proposal", "label": "All-user message scanning proposal", "group": "safety", "role": "Policy", "summary": "Collective Shout proposed scanning all Australian users on relevant services admitting under-18s for specified lawful Class 1C / Class 2 material; not enacted or deployed evidence.", "x": 1696.0, "y": 1442.3, "degree": 1 }, { "id": "lightcone", "label": "Lightcone Foundation", "group": "risk", "role": "Organisation", "summary": "Nevada charity EIN 92-0636259. Not Lightcone Infrastructure/LessWrong or other similarly named entities.", "x": 2280.3, "y": 1351.0, "degree": 7 }, { "id": "vitalik", "label": "Vitalik Buterin", "group": "risk", "role": "Person", "summary": "2021 SHIB support confirmed. Total gift or liquidation value not established; FLI says gift unconditional and no decision role.", "x": 2304.1, "y": 1374.9, "degree": 2 }, { "id": "tegmark", "label": "Max Erik Tegmark", "group": "risk", "role": "Person", "summary": "FLI founder/chair in current September 2026 biography; overlapping 2024 Lightcone governance.", "x": 2291.4, "y": 1362.2, "degree": 3 }, { "id": "aguirre", "label": "Anthony Nicholas Aguirre", "group": "risk", "role": "Person", "summary": "FLI president/CEO; overlapping Lightcone officer record.", "x": 2297.6, "y": 1363.7, "degree": 3 }, { "id": "chita", "label": "Mihaela Chita-Tegmark", "group": "risk", "role": "Person", "summary": "FLI board secretary; filed name Mihaela Chita-Tegmark.", "x": 2296.5, "y": 1359.4, "degree": 3 }, { "id": "tallinn", "label": "Jaan Tallinn", "group": "risk", "role": "Person", "summary": "FLI and CAIS board bridge. Former DeepMind investor-director.", "x": 2332.2, "y": 1397.2, "degree": 5 }, { "id": "flieurope", "label": "Future of Life Institute Europe", "group": "risk", "role": "Organisation", "summary": "Related Swiss entity; recipient of 2022 asset transfer.", "x": 2305.5, "y": 1342.8, "degree": 10 }, { "id": "flfoundation", "label": "Future of Life Foundation", "group": "risk", "role": "Organisation", "summary": "Separate charity EIN 88-0926067; related-entity transactions.", "x": 2267.8, "y": 1347.7, "degree": 6 }, { "id": "baif", "label": "Beneficial AI Foundation", "group": "risk", "role": "Organisation", "summary": "Related charity EIN 88-4400777; employee-sharing and reimbursement disclosed.", "x": 2308.8, "y": 1382.0, "degree": 2 }, { "id": "cais", "label": "Center for AI Safety", "group": "risk", "role": "Organisation", "summary": "CAIS EIN 88-1751310. Separate from its Action Fund.", "x": 2402.0, "y": 1405.9, "degree": 16 }, { "id": "caisaction", "label": "CAIS Action Fund", "group": "risk", "role": "Organisation", "summary": "Separate 501(c)(4), EIN 93-2442608.", "x": 2546.4, "y": 1392.2, "degree": 1 }, { "id": "hendrycks", "label": "Dan Hendrycks", "group": "risk", "role": "Person", "summary": "CAIS executive/research director; industry advisory roles do not establish campaign financing.", "x": 2494.3, "y": 1402.9, "degree": 5 }, { "id": "goodventures", "label": "Good Ventures", "group": "risk", "role": "Organisation", "summary": "Historical CAIS award inventory ends with October 2023 exit grant. Not established current sponsor.", "x": 2388.3, "y": 1420.4, "degree": 10 }, { "id": "coefficient", "label": "Coefficient Giving", "group": "risk", "role": "Organisation", "summary": "Formerly Open Philanthropy; philanthropic adviser to Good Ventures.", "x": 2310.1, "y": 1438.1, "degree": 2 }, { "id": "tuna", "label": "Cari Tuna", "group": "risk", "role": "Person", "summary": "Good Ventures cofounder and board chair.", "x": 2436.6, "y": 1407.5, "degree": 1 }, { "id": "moskovitz", "label": "Dustin Moskovitz", "group": "risk", "role": "Person", "summary": "Good Ventures cofounder, secretary and treasurer; gifts are not Facebook/Asana corporate transfers.", "x": 2431.8, "y": 1401.0, "degree": 1 }, { "id": "makan", "label": "Divesh Makan", "group": "risk", "role": "Person", "summary": "Good Ventures board; ICONIQ partner. No ICONIQ-to-campaign payment established.", "x": 2440.8, "y": 1413.9, "degree": 1 }, { "id": "svcf", "label": "Silicon Valley Community Foundation", "group": "risk", "role": "Organisation", "summary": "2023 CAIS grant disclosed. Upstream donor/account unidentified.", "x": 2454.1, "y": 1389.8, "degree": 1 }, { "id": "cht", "label": "Center for Humane Technology", "group": "risk", "role": "Organisation", "summary": "Recipient of 2023 recommended FLI policy/messaging award.", "x": 2313.6, "y": 1374.0, "degree": 1 }, { "id": "dma", "label": "Digital Media Accelerator", "group": "risk", "role": "Programme", "summary": "FLI creator program; up to $100k/month planned across projects.", "x": 2317.1, "y": 1381.6, "degree": 2 }, { "id": "pwh", "label": "Protect What's Human", "group": "risk", "role": "Campaign", "summary": "FLI campaign launched February 2026; up to $8m announced envelope, not audited spend.", "x": 2321.4, "y": 1324.1, "degree": 3 }, { "id": "peoplefirst", "label": "People First Marketing, Inc.", "group": "risk", "role": "Organisation", "summary": "Creator recruitment agency. Separate from Public First Action.", "x": 2366.3, "y": 1269.1, "degree": 4 }, { "id": "pwhcreators", "label": "Commissioned campaign creators", "group": "risk", "role": "Audience", "summary": "Public offer and general editorial/disclosure terms; individual payment ledger unobserved.", "x": 2357.4, "y": 1287.4, "degree": 2 }, { "id": "frame", "label": "Frame Fellowship", "group": "risk", "role": "Programme", "summary": "Shared creator training and placement infrastructure; independent and amplifier tracks.", "x": 2385.9, "y": 1470.3, "degree": 15 }, { "id": "sara", "label": "Sara Landi Tortoli", "group": "risk", "role": "Person", "summary": "Announced CAIS placement for September-November 2026; completion not verified.", "x": 2436.8, "y": 1461.7, "degree": 1 }, { "id": "veronica", "label": "Veronica Hylak", "group": "risk", "role": "Person", "summary": "CAIS contract reported by Frame; executed contract and payment not read.", "x": 2453.6, "y": 1395.8, "degree": 1 }, { "id": "mikey", "label": "Mikey Posada", "group": "risk", "role": "Person", "summary": "CAIS contract reported by Frame; amount/date/terms unknown.", "x": 2460.0, "y": 1399.9, "degree": 1 }, { "id": "mateus", "label": "Mateus De Souza / Future Tense", "group": "risk", "role": "Person", "summary": "FLI funding reported by Frame; amount/date/terms unknown.", "x": 2333.8, "y": 1430.4, "degree": 3 }, { "id": "talos", "label": "Talos Network", "group": "risk", "role": "Organisation", "summary": "Membership and donor disclosure; gifts not earmarked to Eva Behrens on the retrieved record.", "x": 2230.1, "y": 1466.9, "degree": 4 }, { "id": "sff", "label": "Survival and Flourishing Fund", "group": "risk", "role": "Organisation", "summary": "Disclosed Talos supporter. No established grant to ControlAI.", "x": 2315.9, "y": 1469.1, "degree": 5 }, { "id": "sfr", "label": "Secure Future Research Ltd", "group": "control", "role": "Organisation", "summary": "UK 15088415, limited by guarantee without share capital. Two named legal controllers; no exact 50/50 split inferred.", "x": 2087.6, "y": 1583.5, "degree": 5 }, { "id": "miotti", "label": "Andrea Miotti", "group": "control", "role": "Person", "summary": "ControlAI founder/CEO; UK director and voting/appointment rights.", "x": 2110.6, "y": 1566.7, "degree": 9 }, { "id": "behrens", "label": "Eva Behrens", "group": "control", "role": "Person", "summary": "ControlAI Special Projects Lead; UK director and voting/appointment rights.", "x": 2116.7, "y": 1533.1, "degree": 10 }, { "id": "moes", "label": "Nicolas Moës", "group": "control", "role": "Person", "summary": "Former UK director/PSC; both ceased 7 November 2023.", "x": 2066.8, "y": 1614.0, "degree": 1 }, { "id": "leahy", "label": "Connor Leahy", "group": "control", "role": "Person", "summary": "Conjecture cofounder; moved to ControlAI US in March 2026.", "x": 2013.9, "y": 1589.4, "degree": 2 }, { "id": "alfour", "label": "Gabriel Alfour", "group": "control", "role": "Person", "summary": "Conjecture cofounder; ControlAI founding assistance and adviser.", "x": 2092.0, "y": 1583.2, "degree": 3 }, { "id": "winga", "label": "Max Winga", "group": "control", "role": "Person", "summary": "ControlAI Creator Outreach & Systems Lead; Frame past mentor.", "x": 2260.3, "y": 1535.4, "degree": 2 }, { "id": "conjecture", "label": "Conjecture", "group": "control", "role": "Organisation", "summary": "2022 venture-backed alignment startup. Leahy described operational ending in March 2026, not corporate dissolution.", "x": 2058.9, "y": 1568.9, "degree": 5 }, { "id": "controlaius", "label": "Control AI US Inc", "group": "control", "role": "Organisation", "summary": "Distinct US organization, described as 501(c)(4). Full board and UK-US transfer arrangements not established.", "x": 2016.2, "y": 1603.9, "degree": 5 }, { "id": "balde", "label": "Amir Benjamin Balde", "group": "control", "role": "Person", "summary": "Registered commissioned lobbyist for Control AI US in Germany; 2026 compensation unknown.", "x": 1973.2, "y": 1631.8, "degree": 1 }, { "id": "controlcreators", "label": "ControlAI creator partnerships", "group": "control", "role": "Programme", "summary": "Organization reports paid and unpaid collaborations; outreach role includes briefing/script drafting/review.", "x": 2123.7, "y": 1609.3, "degree": 3 }, { "id": "ukparliament", "label": "UK Parliament", "group": "control", "role": "Institution", "summary": "Recorded access, debates and proposed measures. MPs/peers are not interchangeable with executive government.", "x": 2103.6, "y": 1590.3, "degree": 6 }, { "id": "browne", "label": "Des Browne", "group": "control", "role": "Person", "summary": "Former UK Defence Secretary; ControlAI adviser. Retired from Lords 23 February 2026.", "x": 2044.1, "y": 1567.9, "degree": 5 }, { "id": "hunt", "label": "Lord Hunt of Kings Heath", "group": "control", "role": "Person", "summary": "Acknowledged ControlAI support in January 2026 debate; Amendment 84 sponsor.", "x": 2167.1, "y": 1604.3, "degree": 2 }, { "id": "sobel", "label": "Alex Sobel", "group": "control", "role": "Person", "summary": "MP, historical Fabian executive; ControlAI-backed legislative activity.", "x": 2105.1, "y": 1617.4, "degree": 4 }, { "id": "dodds", "label": "Anneliese Dodds", "group": "control", "role": "Person", "summary": "Fabian co-chair listed in observed current executive roster; ControlAI-listed campaign supporter.", "x": 2113.9, "y": 1630.4, "degree": 2 }, { "id": "russell", "label": "Sarah Russell", "group": "control", "role": "Person", "summary": "Fabian executive member; ControlAI-listed campaign supporter.", "x": 2128.9, "y": 1635.4, "degree": 2 }, { "id": "hedges", "label": "Mike Hedges", "group": "control", "role": "Person", "summary": "Welsh Fabians executive representative; ControlAI-listed supporter.", "x": 2125.8, "y": 1631.0, "degree": 2 }, { "id": "asato", "label": "Jess Asato", "group": "control", "role": "Person", "summary": "Fabian chair 2012-2014; ControlAI-listed supporter.", "x": 2107.6, "y": 1627.4, "degree": 2 }, { "id": "knight", "label": "Jim Knight", "group": "control", "role": "Person", "summary": "Fabian rural-Labour research adviser in 2017; ControlAI-listed supporter.", "x": 2120.3, "y": 1628.2, "degree": 2 }, { "id": "lister", "label": "Baroness Lister", "group": "control", "role": "Person", "summary": "Fabian June 2025 event speaker; ControlAI-listed supporter. Vice-presidency not verified.", "x": 2120.7, "y": 1635.6, "degree": 2 }, { "id": "fabians", "label": "Fabian Society", "group": "control", "role": "Organisation", "summary": "Named current/former leadership overlaps. No Society-to-ControlAI funding, commissioning or direction established.", "x": 2102.6, "y": 1641.4, "degree": 7 }, { "id": "nti", "label": "Nuclear Threat Initiative", "group": "control", "role": "Organisation", "summary": "Browne vice-chair in historical event report; co-hosted 2017 WEF roundtable.", "x": 1998.5, "y": 1551.8, "degree": 2 }, { "id": "cfg", "label": "Centre for Future Generations", "group": "control", "role": "Organisation", "summary": "Behrens former employer. Donor disclosures do not establish transfers to ControlAI.", "x": 2025.9, "y": 1528.7, "degree": 2 }, { "id": "cfgfunders", "label": "Dreamery / Tanka / Powoki", "group": "control", "role": "Funder", "summary": "CFG named supporters. Dreamery associated with cofounder Steven Schuurman; seed funding covers first five years baseline.", "x": 1955.1, "y": 1523.8, "degree": 1 }, { "id": "cdi", "label": "Center for Data Innovation / ITIF", "group": "control", "role": "Organisation", "summary": "2022 Behrens policy fellowship; CDI part of ITIF.", "x": 2082.8, "y": 1527.3, "degree": 1 }, { "id": "oecd", "label": "OECD", "group": "control", "role": "Institution", "summary": "Behrens prior fiscal-foresight work reported in employer biography.", "x": 2192.6, "y": 1501.5, "degree": 2 }, { "id": "bdi", "label": "Federation of German Industries", "group": "control", "role": "Organisation", "summary": "Behrens prior industrial-policy and Sino-European-relations work.", "x": 2098.2, "y": 1519.6, "degree": 1 }, { "id": "eidproposal", "label": "Persistent electronic-ID identifiers", "group": "control", "role": "Proposal", "summary": "Behrens 2022 authored EU eID design position; no demonstrated age-verification campaign involvement.", "x": 2085.7, "y": 1536.2, "degree": 1 }, { "id": "middlepowers", "label": "Middle-powers ASI proposal", "group": "control", "role": "Proposal", "summary": "November 2025 coauthored policy paper; state export controls/sanctions and specified self-defence redlines, not enacted authority.", "x": 2092.5, "y": 1560.1, "degree": 3 }, { "id": "asiuk", "label": "ControlAI proposed UK ASI bill", "group": "control", "role": "Proposal", "summary": "22-page download read September 2026. Proposes compute identity records, seizure/destruction powers, oversight. Not enacted.", "x": 2159.7, "y": 1625.0, "degree": 1 }, { "id": "nc12", "label": "Commons NC12", "group": "control", "role": "Proposal", "summary": "Official status at Sep 5 cutoff: Not called.", "x": 2124.5, "y": 1620.5, "degree": 2 }, { "id": "ltf", "label": "Leading the Future", "group": "industry", "role": "Organisation", "summary": "FEC C00916114; $75,788,224.10 receipts through June 30, 2026.", "x": 2517.0, "y": 916.6, "degree": 6 }, { "id": "a16z", "label": "a16z Capital Management", "group": "industry", "role": "Organisation", "summary": "Two filed $25m contributions to LTF.", "x": 2547.9, "y": 898.5, "degree": 1 }, { "id": "gregbrockman", "label": "Greg Brockman", "group": "industry", "role": "Person", "summary": "$12.5m individual contribution; not an OpenAI corporate payment.", "x": 2532.3, "y": 886.8, "degree": 1 }, { "id": "annabrockman", "label": "Anna Brockman", "group": "industry", "role": "Person", "summary": "$12.5m individual contribution.", "x": 2545.8, "y": 888.2, "degree": 1 }, { "id": "othercommittees", "label": "Other political committees", "group": "industry", "role": "Institution", "summary": "LTF transfer aggregate only; recipient-by-recipient allocation not mapped.", "x": 2537.6, "y": 894.0, "degree": 1 }, { "id": "anthropic", "label": "Anthropic", "group": "industry", "role": "Organisation", "summary": "Announced restricted support for Public First Action.", "x": 3242.1, "y": 989.3, "degree": 1 }, { "id": "publicfirst", "label": "Public First Action", "group": "industry", "role": "Organisation", "summary": "Separate from People First Marketing; Anthropic election-spending restriction.", "x": 3216.7, "y": 999.0, "degree": 1 }, { "id": "buildai", "label": "Build American AI", "group": "industry", "role": "Organisation", "summary": "LTF affiliation and creator intermediary relationship reported by WIRED; no donor-to-creator allocation traced.", "x": 2556.8, "y": 923.0, "degree": 2 }, { "id": "sm4", "label": "SM4", "group": "industry", "role": "Organisation", "summary": "Reported creator-pitch intermediary.", "x": 2595.6, "y": 913.5, "degree": 2 }, { "id": "pitchcreator", "label": "Creator prospect", "group": "industry", "role": "Audience", "summary": "Reported $5,000 offer; completed payment unobserved.", "x": 2625.7, "y": 895.1, "degree": 1 }, { "id": "conbackers", "label": "Conjecture's 2022 venture backers", "group": "industry", "role": "Funder", "summary": "Nat Friedman, Daniel Gross, Patrick and John Collison, Arthur Breitman, Andrej Karpathy, Sam Bankman-Fried. Amounts/current holdings unknown.", "x": 2013.3, "y": 1580.2, "degree": 1 }, { "id": "xai", "label": "xAI", "group": "industry", "role": "Organisation", "summary": "Hendrycks advisory connection; no campaign funding established.", "x": 2545.7, "y": 1399.1, "degree": 1 }, { "id": "scale", "label": "Scale AI", "group": "industry", "role": "Organisation", "summary": "Hendrycks advisory connection; no campaign funding established.", "x": 2543.2, "y": 1405.8, "degree": 1 }, { "id": "deepmind", "label": "DeepMind", "group": "industry", "role": "Organisation", "summary": "Historical Tallinn investor-director relationship.", "x": 2402.3, "y": 1383.9, "degree": 2 }, { "id": "omidyar", "label": "Omidyar Network", "group": "rights", "role": "Organisation", "summary": "Omidyar Network Fund Inc; charitable grants are not eBay corporate payments.", "x": 1857.7, "y": 1558.8, "degree": 4 }, { "id": "pierre-omidyar", "label": "Pierre Omidyar", "group": "rights", "role": "Person", "summary": "eBay founder and Omidyar Network co-founder.", "x": 1846.9, "y": 1588.2, "degree": 1 }, { "id": "oak", "label": "Oak Foundation", "group": "rights", "role": "Organisation", "summary": "Common funder of 5Rights and WeProtect. Awards alone do not establish direction over recipient campaigns.", "x": 1688.0, "y": 1557.3, "degree": 16 }, { "id": "alan-parker", "label": "Alan Parker", "group": "rights", "role": "Person", "summary": "Oak describes resources originating in the Duty Free Shoppers business he helped build.", "x": 1633.6, "y": 1563.7, "degree": 1 }, { "id": "prudence", "label": "Prudence Trust", "group": "rights", "role": "Organisation", "summary": "Published multiyear core-work award to 5Rights.", "x": 1931.9, "y": 1565.5, "degree": 1 }, { "id": "rothschild-foundation", "label": "Rothschild Foundation", "group": "rights", "role": "Organisation", "summary": "Specific UK charity 1138145; not a proxy for every institution with this name.", "x": 1908.0, "y": 1564.2, "degree": 1 }, { "id": "capitol-connection", "label": "Capitol Connection", "group": "rights", "role": "Organisation", "summary": "capconn LLC. Paid contractor to 5Rights for California AB2273 lobbying.", "x": 1855.3, "y": 1601.7, "degree": 7 }, { "id": "ab2273", "label": "California AB2273", "group": "rights", "role": "Policy", "summary": "2022 Age-Appropriate Design Code campaign. Current enforcement and litigation are outside this investigation.", "x": 1822.5, "y": 1531.8, "degree": 4 }, { "id": "reset", "label": "Reset", "group": "rights", "role": "Organisation", "summary": "Omidyar credits joint support for the California coalition in September 2022.", "x": 1729.0, "y": 1458.2, "degree": 4 }, { "id": "cross-street-films", "label": "Cross Street Films", "group": "rights", "role": "Organisation", "summary": "Parliamentary assistance declared in Kidron’s register; no allocation to the AI amendment established.", "x": 2144.3, "y": 1561.3, "degree": 1 }, { "id": "common-sense-media", "label": "Common Sense Media", "group": "rights", "role": "Organisation", "summary": "Kidron declares paid summit travel in March 2026; no amendment-specific support attributed.", "x": 2007.4, "y": 1589.8, "degree": 4 }, { "id": "home-office", "label": "UK Home Office", "group": "state", "role": "Organisation", "summary": "Government records historical funding and a permanent WeProtect board seat.", "x": 1913.7, "y": 1529.5, "degree": 7 }, { "id": "european-commission", "label": "European Commission", "group": "state", "role": "Organisation", "summary": "WeProtect board representation; proposal briefing and possible assistance requests recorded in July 2022 minutes.", "x": 1666.0, "y": 1762.1, "degree": 15 }, { "id": "eu-csa-proposal", "label": "EU child-abuse legislation", "group": "state", "role": "Policy", "summary": "Proposed legislation discussed by WeProtect’s board in July 2022. No present enactment inferred.", "x": 1736.8, "y": 1583.1, "degree": 4 }, { "id": "google", "label": "Google", "group": "state", "role": "Organisation", "summary": "WeProtect board participation; also represented in WEF digital-safety policy work.", "x": 1908.2, "y": 1582.2, "degree": 11 }, { "id": "snap", "label": "Snap", "group": "state", "role": "Organisation", "summary": "Representation on WeProtect’s published board; not an ownership share.", "x": 1841.2, "y": 1554.7, "degree": 2 }, { "id": "thorn", "label": "Thorn", "group": "state", "role": "Organisation", "summary": "Representation on WeProtect’s published board; not an ownership share.", "x": 1752.2, "y": 1538.0, "degree": 2 }, { "id": "casir", "label": "CASIR · draft", "group": "control", "role": "Organisation", "summary": "Unincorporated, unfinished proposed fund. Its site claims affiliations without reciprocal confirmation; no verified donor commitment or executed grant.", "x": 2118.8, "y": 1681.0, "degree": 5 }, { "id": "torchbearer", "label": "Torchbearer Community", "group": "control", "role": "Organisation", "summary": "Named by the CASIR draft site; affiliation is a site claim, not reciprocally authenticated.", "x": 2121.8, "y": 1727.8, "degree": 1 }, { "id": "pauseai-us", "label": "PauseAI US", "group": "control", "role": "Organisation", "summary": "Named by CASIR’s unfinished site; no completed funding relationship established.", "x": 2113.7, "y": 1727.6, "degree": 1 }, { "id": "alliance-secure-ai-action", "label": "Alliance for Secure AI Action", "group": "control", "role": "Organisation", "summary": "Named as a partner by CASIR’s unfinished site; not independently authenticated.", "x": 2131.3, "y": 1726.7, "degree": 1 }, { "id": "asi-message-grants", "label": "Message-based grant criteria", "group": "control", "role": "Policy", "summary": "CASIR draft: explicit ASI-extinction messaging, at least 65% member-category allocation and contemplated unspent-grant recovery. Proposed, not operational.", "x": 2138.8, "y": 1646.2, "degree": 2 }, { "id": "longview", "label": "Longview Philanthropy", "group": "risk", "role": "Organisation", "summary": "Personnel links plus indexed historical Good Ventures awards. No recovered Longview-to-CAIS or Australian age-assurance financing.", "x": 2359.6, "y": 1450.1, "degree": 6 }, { "id": "ogara", "label": "Aidan O’Gara", "group": "risk", "role": "Person", "summary": "Longview AI Program Officer in the prior source read, with CAIS affiliation on a 2023 paper.", "x": 2423.8, "y": 1427.3, "degree": 3 }, { "id": "ai-deception-paper", "label": "AI Deception paper", "group": "risk", "role": "Policy", "summary": "2023 paper discusses disclosure, watermarking and signatures for AI/human-origin content. Not equivalent to universal government identity.", "x": 2482.5, "y": 1418.0, "degree": 2 }, { "id": "singh", "label": "Akshyae Singh", "group": "risk", "role": "Person", "summary": "Frame executive director. Describes founding it with Austin Chen and earlier CAIS creator campaigns.", "x": 2416.7, "y": 1439.1, "degree": 2 }, { "id": "austin-chen", "label": "Austin Chen", "group": "risk", "role": "Person", "summary": "Manifund cofounder; Frame adviser and cofounding described by Singh. Specific roles do not establish control of every supported project.", "x": 2447.8, "y": 1470.4, "degree": 2 }, { "id": "ryan-kidd", "label": "Ryan Kidd", "group": "risk", "role": "Person", "summary": "Frame adviser and MATS co-executive director; named initial grantmaking.ai reviewer.", "x": 2460.2, "y": 1477.7, "degree": 2 }, { "id": "drew-spartz", "label": "Drew Spartz", "group": "risk", "role": "Person", "summary": "Frame adviser, creator and Nonlinear cofounder as listed by Frame.", "x": 2437.8, "y": 1467.7, "degree": 1 }, { "id": "manifund", "label": "Manifund", "group": "risk", "role": "Organisation", "summary": "Grantmaking/fiscal infrastructure. Its published account identifies fiscal sponsorship and payouts for grantmaking.ai.", "x": 2470.3, "y": 1473.8, "degree": 3 }, { "id": "grantmaking-ai", "label": "grantmaking.ai", "group": "risk", "role": "Programme", "summary": "Public funding round. Reviewers, organizers and fiscal processor are distinct roles. No ControlAI award inferred.", "x": 2506.4, "y": 1479.0, "degree": 2 }, { "id": "mozie", "label": "Mozie / Synapticks, Inc.", "group": "risk", "role": "Organisation", "summary": "Creator recruitment vendor. Its own case study describes a Frame engagement; contract and fee not recovered.", "x": 2444.7, "y": 1477.0, "degree": 1 }, { "id": "frame-selection", "label": "Frame selection committee", "group": "risk", "role": "Body", "summary": "Committee mentioned in Mozie's account. Membership and scoring rubric remain unidentified.", "x": 2432.0, "y": 1482.2, "degree": 1 }, { "id": "mox", "label": "Mox", "group": "risk", "role": "Organisation", "summary": "San Francisco AI-safety/EA workspace and incubator. Its fundraiser says Frame was developed in-house.", "x": 2405.3, "y": 1469.2, "degree": 4 }, { "id": "craig-eaif", "label": "Craig Falls / EA Infrastructure Fund", "group": "risk", "role": "Group of participants", "summary": "Grouped solely because Mox reports one combined $300,000 grant line; shares are not disaggregated.", "x": 2459.4, "y": 1463.8, "degree": 1 }, { "id": "frame-amplifier", "label": "Frame Amplifier track", "group": "risk", "role": "Programme", "summary": "Organizer's announced track to embed fellows in AI safety organizations. A programme offer is not proof every placement happened.", "x": 2438.4, "y": 1488.3, "degree": 1 }, { "id": "eclag", "label": "ECLAG", "group": "state", "role": "Coalition", "summary": "European Child Sexual Abuse Legislation Advocacy Group. Says it formed in April 2022 to align civil-society advocacy.", "x": 1744.6, "y": 1475.1, "degree": 7 }, { "id": "child-safety-on", "label": "Child Safety ON", "group": "state", "role": "Campaign", "summary": "EU child-safety legislation advocacy campaign; ECLAG is part of the coalition. Earlier branding: #ChildSafetyOnlineNow.", "x": 1808.9, "y": 1536.6, "degree": 3 }, { "id": "ecpat", "label": "ECPAT", "group": "state", "role": "Organisation", "summary": "Listed on ECLAG's current steering group. This source does not establish a campaign grant.", "x": 1708.4, "y": 1496.5, "degree": 1 }, { "id": "eurochild", "label": "Eurochild", "group": "state", "role": "Organisation", "summary": "Listed on ECLAG's current steering group. This source does not establish a campaign grant.", "x": 1703.3, "y": 1490.3, "degree": 1 }, { "id": "missing-children-europe", "label": "Missing Children Europe", "group": "state", "role": "Organisation", "summary": "Listed on ECLAG's current steering group. This source does not establish a campaign grant.", "x": 1713.0, "y": 1500.3, "degree": 1 }, { "id": "iwf", "label": "Internet Watch Foundation", "group": "state", "role": "Organisation", "summary": "Listed on ECLAG's current steering group. This source does not establish a campaign grant.", "x": 1751.3, "y": 1366.2, "degree": 12 }, { "id": "terre-des-hommes", "label": "Terre des Hommes", "group": "state", "role": "Organisation", "summary": "Listed on ECLAG's current steering group. This source does not establish a campaign grant.", "x": 1700.5, "y": 1496.2, "degree": 1 }, { "id": "wp-secretariat", "label": "WeProtect Secretariat", "group": "state", "role": "Body", "summary": "Its published EU submission expressly need not represent all Alliance members; its membership figures are dated 12 September 2022.", "x": 1707.6, "y": 1560.0, "degree": 3 }, { "id": "wp-ed", "label": "WeProtect executive director role", "group": "state", "role": "Decision-making role", "summary": "Financial authority in the 2023–25 Policy Plan. Role-level record; no historical occupant backfilled.", "x": 1782.6, "y": 1716.0, "degree": 2 }, { "id": "wp-coo", "label": "WeProtect operations role", "group": "state", "role": "Decision-making role", "summary": "COO expenditure/asset management responsibility described in the 2023–25 Policy Plan.", "x": 1784.3, "y": 1675.2, "degree": 3 }, { "id": "wp-frac", "label": "WeProtect finance/risk/audit committee", "group": "state", "role": "Body", "summary": "FRAC accountability function described in the Policy Plan.", "x": 1793.6, "y": 1732.0, "degree": 3 }, { "id": "wp-boards", "label": "WeProtect management / policy boards", "group": "state", "role": "Group of participants", "summary": "Two distinct boards grouped for the Policy Plan's common reporting endpoint; neither is the Secretariat.", "x": 1809.2, "y": 1775.1, "degree": 1 }, { "id": "wp-development", "label": "WeProtect Development Committee", "group": "state", "role": "Body", "summary": "Unpaid fundraising advisory/support body disclosed in its Policy Plan.", "x": 1750.8, "y": 1622.0, "degree": 1 }, { "id": "five-country-working-group", "label": "Five Country Working Group", "group": "state", "role": "Body", "summary": "Voluntary child-protection principles discussed with WeProtect in June 2024. Not an inferred covert actor.", "x": 1759.7, "y": 1623.2, "degree": 1 }, { "id": "pa-consulting", "label": "PA Consulting", "group": "state", "role": "Organisation", "summary": "Natalia Greene briefed WeProtect's board on the Global Threat Assessment in June 2023.", "x": 1553.5, "y": 1462.5, "degree": 3 }, { "id": "global-threat-assessment", "label": "Global Threat Assessment", "group": "state", "role": "Artifact", "summary": "WeProtect research output. Board comments establish a production interface, not dictated findings.", "x": 1661.8, "y": 1532.4, "degree": 2 }, { "id": "stripe", "label": "Stripe", "group": "safety", "role": "Organisation", "summary": "Named itch.io card processor. July 2025 bank restrictions are attributed through itch's public account.", "x": 1816.0, "y": 1536.4, "degree": 4 }, { "id": "paypal", "label": "PayPal", "group": "safety", "role": "Organisation", "summary": "Named itch.io payment/payout dependency. No recovered incident-specific PayPal order.", "x": 1757.9, "y": 1524.1, "degree": 7 }, { "id": "mastercard", "label": "Mastercard", "group": "safety", "role": "Organisation", "summary": "Card network: distinguish published standards, acquirer responsibility and a specific restriction decision.", "x": 1768.4, "y": 1508.8, "degree": 12 }, { "id": "stripe-bank-reference", "label": "Stripe banking partners · unidentified", "group": "safety", "role": "Unidentified party", "summary": "Real-world counterparties referenced in itch's relayed Stripe explanation; identities and original instruction are unresolved.", "x": 1774.7, "y": 1544.1, "degree": 1 }, { "id": "mc-acquirers", "label": "Mastercard acquirers · rule scope", "group": "safety", "role": "Group of participants", "summary": "Class of institutions bound by the rulebook. This does not name itch's acquirer.", "x": 1784.8, "y": 1516.6, "degree": 9 }, { "id": "itch-reindex", "label": "Free NSFW reindexing", "group": "safety", "role": "Event", "summary": "July 31, 2025 update required the No Payments setting for the described reindexing.", "x": 1678.3, "y": 1495.5, "degree": 1 }, { "id": "ftx-estate", "label": "FTX debtors / bankruptcy estate", "group": "risk", "role": "Organisation", "summary": "Recovery actor in a December 2023 court filing; distinct from original donors and former executives.", "x": 2443.8, "y": 1378.3, "degree": 1 }, { "id": "cais-reserve", "label": "CAIS potential-refund reserve", "group": "risk", "role": "Accounting record", "summary": "2023 carrying value $5,259,802. It is an accounting reserve, not a bank transfer.", "x": 2451.4, "y": 1380.5, "degree": 1 }, { "id": "cais-2024-close", "label": "CAIS 2024 liability close", "group": "risk", "role": "Accounting record", "summary": "Other liabilities close at $483,825, identified entirely as an operating lease liability; prior refund reserve no longer appears.", "x": 2448.3, "y": 1385.9, "degree": 1 }, { "id": "mastercard-statement", "label": "Mastercard gaming statement", "group": "safety", "role": "Artifact", "summary": "August 1, 2025 statement denying direct intervention. A published response, not a restriction instruction.", "x": 1729.6, "y": 1509.8, "degree": 1 }, { "id": "icmec", "label": "International Centre for Missing & Exploited Children (US)", "group": "state", "role": "Organisation", "summary": "US nonprofit; legislative sponsor and financial-coalition operator. Separate programme funding and policy work require their own evidence.", "x": 1704.1, "y": 1665.6, "degree": 18 }, { "id": "icmec-financial-coalition", "label": "ICMEC financial coalition", "group": "safety", "role": "Programme", "summary": "Financial Coalition Against Child Sexual Exploitation; programme donor listing supplies no policy earmark.", "x": 1763.1, "y": 1592.0, "degree": 5 }, { "id": "children-now", "label": "Children Now", "group": "rights", "role": "Organisation", "summary": "", "x": 1800.6, "y": 1638.7, "degree": 1 }, { "id": "visa", "label": "Visa", "group": "safety", "role": "Organisation", "summary": "", "x": 1826.3, "y": 1568.6, "degree": 3 }, { "id": "ab1043", "label": "California AB1043, Digital Age Assurance Act", "group": "rights", "role": "Legislation", "summary": "Chapter675 enacted October13,2025; age-signal framework operative January2027. Differences from earlier proposals remain in the evidence note.", "x": 1816.7, "y": 1608.9, "degree": 5 }, { "id": "ab1856", "label": "California AB1856", "group": "rights", "role": "Legislation", "summary": "Enrolled September1; presented Governor September4,2026 at4pm; no signature shown when checked September8. Pending status must be refreshed for later use.", "x": 1823.8, "y": 1415.4, "degree": 2 }, { "id": "wicks", "label": "Buffy Wicks", "group": "rights", "role": "Person", "summary": "", "x": 1800.2, "y": 1551.3, "degree": 4 }, { "id": "bob-cunningham", "label": "Bob Cunningham", "group": "state", "role": "Person", "summary": "ICMEC policy-engagement official quoted on legislative drafting.", "x": 1643.7, "y": 1696.5, "degree": 2 }, { "id": "icmec-age-whitepaper", "label": "ICMEC device-based age-assurance technical whitepaper", "group": "credentials", "role": "Proposal", "summary": "ICMEC-hosted technical proposal, with no named author/funder in the document; not an implementation record.", "x": 1638.5, "y": 1689.2, "degree": 2 }, { "id": "api-admission-role", "label": "Proposed OS-provider approval of website age-API integrations", "group": "credentials", "role": "Proposed function", "summary": "A proposed approval function; no unnamed secret holder is inferred.", "x": 1591.4, "y": 1706.1, "degree": 1 }, { "id": "microsoft", "label": "Microsoft", "group": "credentials", "role": "Organisation", "summary": "", "x": 1896.9, "y": 1568.9, "degree": 9 }, { "id": "meta", "label": "Meta", "group": "industry", "role": "Organisation", "summary": "", "x": 1905.4, "y": 1582.3, "degree": 6 }, { "id": "windows-age-signals", "label": "Windows age-signals APIs", "group": "credentials", "role": "Programme", "summary": "Official documentation describes planned APIs, not yet enabled at the September8 check.", "x": 1875.5, "y": 1566.4, "degree": 2 }, { "id": "microsoft-account", "label": "Microsoft-account session", "group": "credentials", "role": "Research resource", "summary": "", "x": 1863.1, "y": 1559.7, "degree": 1 }, { "id": "discord", "label": "Discord", "group": "credentials", "role": "Organisation", "summary": "", "x": 2133.9, "y": 1655.2, "degree": 7 }, { "id": "k-id", "label": "k-ID", "group": "credentials", "role": "Organisation", "summary": "", "x": 1920.7, "y": 1570.8, "degree": 3 }, { "id": "veratad", "label": "Veratad", "group": "credentials", "role": "Organisation", "summary": "", "x": 1928.7, "y": 1557.3, "degree": 1 }, { "id": "google-wallet", "label": "Google Wallet", "group": "credentials", "role": "Programme", "summary": "", "x": 1971.0, "y": 1596.1, "degree": 2 }, { "id": "discord-age-assurance", "label": "Discord age-assurance routes", "group": "credentials", "role": "Programme", "summary": "Distinct regional methods/tests; company documentation is not a live account audit.", "x": 2024.0, "y": 1610.3, "degree": 3 }, { "id": "discord-age-appeals", "label": "Discord age-appeals support records", "group": "custody", "role": "Research resource", "summary": "Separate historic ticketing custody path, not the normal verifier flow.", "x": 2214.1, "y": 1659.9, "degree": 2 }, { "id": "5ca", "label": "5CA", "group": "custody", "role": "Organisation", "summary": "", "x": 2270.0, "y": 1668.3, "degree": 1 }, { "id": "idscan", "label": "IDScan.net", "group": "custody", "role": "Organisation", "summary": "", "x": 784.7, "y": 741.0, "degree": 2 }, { "id": "veriscan", "label": "VeriScan", "group": "custody", "role": "Programme", "summary": "", "x": 804.9, "y": 744.6, "degree": 3 }, { "id": "planet13", "label": "Planet 13", "group": "custody", "role": "Organisation", "summary": "", "x": 834.1, "y": 764.5, "degree": 1 }, { "id": "dutchie", "label": "Dutchie", "group": "custody", "role": "Organisation", "summary": "", "x": 784.8, "y": 726.0, "degree": 1 }, { "id": "idscan-incident", "label": "IDScan 2026 security incident", "group": "custody", "role": "Event", "summary": "Vendor acknowledged possible account-data exposure; incident scope remains under investigation.", "x": 755.8, "y": 722.6, "degree": 1 }, { "id": "w1-icmec-icmec-singapore", "label": "ICMEC Limited (Singapore)", "group": "state", "role": "Organisation", "summary": "", "x": 1668.0, "y": 1691.1, "degree": 1 }, { "id": "w1-icmec-icmec-switzerland", "label": "ICMEC Switzerland", "group": "state", "role": "Organisation", "summary": "", "x": 1674.2, "y": 1686.8, "degree": 1 }, { "id": "w1-icmec-loans24", "label": "ICMEC FY2024 Schedule L operating-fund loans", "group": "state", "role": "Financial disclosure", "summary": "", "x": 532.3, "y": 1315.6, "degree": 3 }, { "id": "w1-icmec-humer", "label": "Franz Humer", "group": "state", "role": "Person", "summary": "", "x": 523.6, "y": 1306.3, "degree": 1 }, { "id": "w1-icmec-paul", "label": "Sally Paul (Form 990 spelling; supporters page: Sally Paull)", "group": "state", "role": "Person", "summary": "", "x": 559.9, "y": 1319.9, "degree": 1 }, { "id": "w1-icmec-li", "label": "Rick Li", "group": "state", "role": "Person", "summary": "", "x": 494.6, "y": 1311.0, "degree": 1 }, { "id": "w1-icmec-apfc-roundtables", "label": "ICMEC Asia-Pacific Financial Coalition country roundtables", "group": "state", "role": "Programme", "summary": "", "x": 1666.2, "y": 1521.2, "degree": 2 }, { "id": "w1-icmec-apfc-australia", "label": "APFC Australia Working Group", "group": "state", "role": "Working group", "summary": "", "x": 1589.0, "y": 1518.8, "degree": 2 }, { "id": "w1-icmec-auspaynet", "label": "Australian Payments Network, Fraud in Banking Forum", "group": "state", "role": "Organization programme", "summary": "", "x": 1533.7, "y": 1519.5, "degree": 1 }, { "id": "w1-icmec-model", "label": "ICMEC Digital Age Assurance model legislation", "group": "state", "role": "Proposal", "summary": "", "x": 1663.1, "y": 1682.4, "degree": 1 }, { "id": "w1-icmec-mpa", "label": "Motion Picture Association", "group": "state", "role": "Organisation", "summary": "", "x": 1849.4, "y": 1282.1, "degree": 4 }, { "id": "w1-controlai-uk-board", "label": "Secure Future Research directors / legal members", "group": "control", "role": "Governance body", "summary": "Articles define trustees as directors and make those directors the only legal members. Link existing directors by their separately supported appointment dates; no invented membership list.", "x": 2061.1, "y": 1608.9, "degree": 1 }, { "id": "w1-controlai-unnamed-related-org", "label": "Unnamed related organisation owed by Control AI US", "group": "control", "role": "Unidentified counterparty", "summary": "Do not merge with Secure Future Research, a donor, or any other named entity without a naming record. This is an unresolved reference, not proof of a new additional organisation.", "x": 1965.0, "y": 1625.5, "degree": 1 }, { "id": "w1-payments-steam-processors", "label": "Steam payment processors · unidentified", "group": "industry", "role": "Organisation", "summary": "Unresolved operational role in Valve's July/August 2025 statements. Not the existing payment-letter addressees node; no Stripe assignment.", "x": 1729.8, "y": 1483.2, "degree": 2 }, { "id": "w1-payments-steam-retirements-202507", "label": "Steam July 2025 game retirements", "group": "industry", "role": "Event", "summary": "Valve's reported removal-from-sale decision in response to payment-provider objections.", "x": 1663.2, "y": 1456.7, "degree": 1 }, { "id": "w1-payments-itch-payout-eligibility", "label": "itch collected-payout eligibility", "group": "industry", "role": "Programme", "summary": "Platform collection/payout arrangement with eligibility review and a documented notice template.", "x": 1674.9, "y": 1488.6, "degree": 2 }, { "id": "w1-payments-itch-uk-restrictions-202507", "label": "itch UK geographic restrictions · July 2025", "group": "industry", "role": "Event", "summary": "Separate geographic compliance response described by itch; do not merge with global payment deindexing.", "x": 1670.2, "y": 1495.6, "degree": 1 }, { "id": "w1-payments-itch-digital-id-exploration", "label": "itch proposed digital-ID partnerships · July 2025", "group": "industry", "role": "Programme", "summary": "Exploration of unnamed third-party providers, not a deployed identity system or a named commercial relationship.", "x": 1675.0, "y": 1502.4, "degree": 1 }, { "id": "icmec-2024-accounts", "label": "ICMEC 2024 consolidated accounts", "group": "state", "role": "Research resource", "summary": "US and Singapore perimeter, year endedDecember31,2024; issuedJanuary29,2026. Historical accounts, not a current balance sheet.", "x": 1674.1, "y": 1696.4, "degree": 3 }, { "id": "icmec-financing-dependence", "label": "ICMEC fundraising dependence", "group": "state", "role": "Research resource", "summary": "Audit going-concern note links continuation to unrestricted fundraising; no particular donor directive identified.", "x": 1682.7, "y": 1694.7, "degree": 1 }, { "id": "w2-fli-fihn", "label": "Anna Béatrice Ramsay Fihn", "group": "risk", "role": "Person", "summary": "See dated sourced relations; no additional authority inferred.", "x": 2342.4, "y": 1313.1, "degree": 1 }, { "id": "w2-fli-echo", "label": "ECHO SA", "group": "risk", "role": "Organisation", "summary": "See dated sourced relations; no additional authority inferred.", "x": 2336.6, "y": 1308.9, "degree": 1 }, { "id": "w2-fli-supervisor", "label": "Swiss Federal Department of Home Affairs, foundation supervision", "group": "risk", "role": "Public authority", "summary": "See dated sourced relations; no additional authority inferred.", "x": 2348.5, "y": 1317.4, "degree": 1 }, { "id": "w2-fli-founder-role", "label": "FLI Europe founder named in unrecovered deed", "group": "risk", "role": "Unresolved documentary role", "summary": "See dated sourced relations; no additional authority inferred.", "x": 2328.8, "y": 1307.0, "degree": 1 }, { "id": "w1-creator-futuretense-output", "label": "The Dark Philosophy Driving AI Development", "group": "risk", "role": "Published output", "summary": "One specific output with creator-side funding disclosure; no estimate of paid reach or contract value", "x": 2342.1, "y": 1424.8, "degree": 4 }, { "id": "w1-creator-bluedot", "label": "BlueDot Impact", "group": "risk", "role": "Organisation", "summary": "Named by creator; exact grant-making entity not established in this pass", "x": 2384.1, "y": 1408.4, "degree": 1 }, { "id": "w1-creator-ltff", "label": "Long Term Future Fund", "group": "risk", "role": "Fund", "summary": "Named as a lead funder in Frame public FAQ; donor-side award not inspected", "x": 2436.9, "y": 1474.5, "degree": 1 }, { "id": "w1-creator-aitof", "label": "AI Safety Tactical Opportunities Fund", "group": "risk", "role": "Fund", "summary": "Named as a lead funder in Frame public FAQ; administrator and grant agreement not established here", "x": 2443.3, "y": 1483.3, "degree": 1 }, { "id": "rational", "label": "Rational Animations", "group": "risk", "role": "Creator", "summary": "See dated sourced relations; no additional authority inferred.", "x": 2168.6, "y": 1622.4, "degree": 1 }, { "id": "w1-creator-peoplefirst-terms", "label": "People First general influencer terms, capture 8 September 2026", "group": "risk", "role": "Terms document", "summary": "General proposed/applicable terms, not a recovered signed SOW or historic terms version", "x": 2403.8, "y": 1240.0, "degree": 1 }, { "id": "w1-creator-dots", "label": "Dots payout provider", "group": "risk", "role": "Service", "summary": "Named as third-party payout provider in People First current terms; exact operating legal entity not verified", "x": 2395.9, "y": 1235.5, "degree": 1 }, { "id": "w1-australia-scout-insights", "label": "Scout Insights", "group": "state", "role": "Organisation", "summary": "Named mystery-shopping research provider in April29 trial minutes; operational role reported by participant Melissa Stevenson.", "x": 2076.4, "y": 1461.6, "degree": 1 }, { "id": "w1-australia-esafety-mou", "label": "Trial–eSafety MOU · original not obtained", "group": "state", "role": "Policy", "summary": "Reported signed boundary-setting instrument, identified in April minutes and final B.18.", "x": 1848.5, "y": 1604.2, "degree": 1 }, { "id": "w1-australia-trial-draft-review", "label": "Trial draft technical/vendor report review", "group": "state", "role": "Event", "summary": "Multistage trial report review. eSafety input concerns the Draft Preliminary Report; July final-report access and Corby review intention are separate dated stages.", "x": 1973.9, "y": 1559.1, "degree": 2 }, { "id": "w1-australia-protocol-revisions", "label": "Trial methodology and safeguard revisions", "group": "state", "role": "Event", "summary": "Reported response to stakeholder feedback. Complete dated revision/approval trail remains missing.", "x": 2060.1, "y": 1450.1, "degree": 1 }, { "id": "fli-action-research", "label": "Future of Life Action and Research", "group": "risk", "role": "Organisation", "summary": "Separate related organisation named in FLI cost-sharing balances.", "x": 2273.4, "y": 1346.9, "degree": 2 }, { "id": "fli-ed-audit-role", "label": "FLI executive director role in 2025 audit", "group": "risk", "role": "Decision-making role", "summary": "Named in 2025 standalone audit Note 6. Role bearer not identified by that passage.", "x": 2275.1, "y": 1308.9, "degree": 7 }, { "id": "fli-asia", "label": "Future of Life Asia", "group": "risk", "role": "Organisation", "summary": "Named as an organisation with overlapping executive board service in FLI 2025 audit Note 6; no resource transfer inferred.", "x": 2300.5, "y": 1273.9, "degree": 1 }, { "id": "wise-ancestors", "label": "Wise Ancestors", "group": "risk", "role": "Organisation", "summary": "Named as an organisation with overlapping executive board service in FLI 2025 audit Note 6; no resource transfer inferred.", "x": 2285.9, "y": 1270.0, "degree": 1 }, { "id": "windfall-trust", "label": "Windfall Trust", "group": "risk", "role": "Organisation", "summary": "Named as an organisation with overlapping executive board service in FLI 2025 audit Note 6; no resource transfer inferred.", "x": 2293.4, "y": 1270.7, "degree": 1 }, { "id": "w3-5r-portantino", "label": "Anthony Portantino", "group": "rights", "role": "Person", "summary": "California senator and Senate Appropriations chair in the2022 records; no current role asserted", "x": 2141.0, "y": 1569.7, "degree": 1 }, { "id": "w3-5r-stern", "label": "Henry Stern", "group": "rights", "role": "Person", "summary": "California senator addressed in Kidron August3,2022 letter; no current role asserted", "x": 2139.6, "y": 1552.7, "degree": 1 }, { "id": "w3-5r-mckinley", "label": "Shannon McKinley", "group": "rights", "role": "Person", "summary": "Wicks chief of staff in2022 FARA activity log; recipient of draft background-sheet exchanges", "x": 1849.8, "y": 1640.6, "degree": 1 }, { "id": "w3-5r-skaggs", "label": "Aaron Skaggs", "group": "rights", "role": "Person", "summary": "Portantino office recipient in August5,2022 log; exact title not stated", "x": 1842.8, "y": 1641.4, "degree": 1 }, { "id": "w3-5r-dang", "label": "Eric Dang", "group": "rights", "role": "Person", "summary": "Senate President pro Tempore policy-unit consultant in2022 log", "x": 1854.3, "y": 1635.7, "degree": 1 }, { "id": "w3-5r-aurre", "label": "Christine Aurre", "group": "rights", "role": "Person", "summary": "Governor Newsom deputy legislative secretary in August2022 log; earlier June entry has a different then-role and must not be projected backwards", "x": 1856.2, "y": 1645.0, "degree": 1 }, { "id": "w3-5r-judiciary", "label": "California Senate Judiciary Committee", "group": "rights", "role": "Government body", "summary": "Official2022 committee analysis and FARA recipient entries", "x": 1836.2, "y": 1639.9, "degree": 1 }, { "id": "w3-5r-coalition-letter", "label": "AB2273 coalition letter · 10June2022", "group": "rights", "role": "Document", "summary": "Original FARA material19; document node represents advocacy prose, not a separate organisation", "x": 2731.3, "y": 313.0, "degree": 1 }, { "id": "w3-5r-concurrence-analysis", "label": "AB2273 Assembly concurrence analysis · 30Aug2022", "group": "rights", "role": "Document", "summary": "Official analysis356279; document node represents published legislative analysis", "x": 2747.2, "y": 290.1, "degree": 1 }, { "id": "w3-oak-senior-decision", "label": "Oak senior funding-decision roles", "group": "rights", "role": "Decision-making role", "summary": "Unspecified senior foundation members receive programme-officer recommendations under the published generic process. Do not substitute Alan Parker for this role.", "x": 1645.6, "y": 1568.5, "degree": 1 }, { "id": "w3-oak-wp-nl", "label": "Stichting The WePROTECT Global Alliance", "group": "state", "role": "Organisation", "summary": "Dutch legal foundation, KvK76585379. Financial/legal actor distinct from the UK company and umbrella Alliance.", "x": 1601.5, "y": 1577.5, "degree": 7 }, { "id": "w3-oak-wp-uk", "label": "WePROTECT Global Alliance Limited", "group": "state", "role": "Organisation", "summary": "UK company named in the Dutch accounts; controlled entity and subgrant recipient. UK registry number not in this lane.", "x": 1548.8, "y": 1588.9, "degree": 4 }, { "id": "w3-oak-safe-futures", "label": "Safe Futures Hub", "group": "state", "role": "Programme", "summary": "Joint initiative co-led by SVRI, Together for Girls and WeProtect; acknowledged Oak support. Not a separate incorporated entity established by this evidence.", "x": 1675.1, "y": 1595.9, "degree": 7 }, { "id": "w3-oak-svri", "label": "Sexual Violence Research Initiative (SVRI)", "group": "state", "role": "Organisation", "summary": "Safe Futures Hub co-lead and institution of the 2025 guidance author.", "x": 1623.6, "y": 1608.3, "degree": 1 }, { "id": "w3-oak-together", "label": "Together for Girls, Inc.", "group": "state", "role": "Organisation", "summary": "Safe Futures Hub co-lead and original intermediary for a grant subsequently rerouted.", "x": 1618.0, "y": 1595.7, "degree": 2 }, { "id": "w3-oak-swiss-philanthropy", "label": "Swiss Philanthropy — named grant counterparty", "group": "state", "role": "Unidentified counterparty", "summary": "Named this way in WeProtect 2023 accounts. Precise legal vehicle, hosted fund and upstream donor unresolved. Do not automatically merge with Swiss Philanthropy Foundation or Oak.", "x": 1546.8, "y": 1582.0, "degree": 1 }, { "id": "w3-oak-helena-duch", "label": "Helena Duch", "group": "rights", "role": "Person", "summary": "Oak Solutions sub-programme head; Safe Futures advisory member and credited 2024 report contributor. No evidence of unilateral editorial veto or approval of the named Oak awards.", "x": 1653.0, "y": 1584.0, "degree": 3 }, { "id": "w3-oak-evidence-review", "label": "Building Safe Futures — 2024 evidence review", "group": "state", "role": "Publication", "summary": "Commissioned by Safe Futures Hub with Oak support; named authors and stakeholder draft-feedback process. Focus is low/middle-income countries and humanitarian settings.", "x": 1634.1, "y": 1599.8, "degree": 2 }, { "id": "w3-oak-guidance", "label": "Practice-based knowledge — 2025 guidance", "group": "state", "role": "Publication", "summary": "Safe Futures Hub publication credited to Arti Mohan with institutional and named external review; practitioner guidance, not binding regulation.", "x": 1628.9, "y": 1613.4, "degree": 1 }, { "id": "w3-oak-economist", "label": "Economist Impact", "group": "state", "role": "Organisation", "summary": "Conducted WeProtect's 2023 European survey. Commissioned research role does not establish ownership of findings by the donor.", "x": 1700.4, "y": 1636.1, "degree": 1 }, { "id": "w3-oak-eu-survey", "label": "WeProtect European online-harms survey 2023", "group": "state", "role": "Publication", "summary": "Retrospective survey of2,000 eighteen-year-olds in France,Germany,Netherlands,Poland; fieldedFeb–Mar 2023; used in EU policy advocacy.", "x": 1730.1, "y": 1611.9, "degree": 3 }, { "id": "w3-oak-policy-board", "label": "WeProtect Global Policy Board", "group": "state", "role": "Body", "summary": "Specific Policy Board component of the previously grouped wp-boards. December 2023 minutes record budget approval. Not identical to the Dutch Management Board.", "x": 1764.8, "y": 1627.9, "degree": 1 }, { "id": "w3-opp-thinkbig", "label": "Think Big", "group": "industry", "role": "Organisation", "summary": "Federal super PAC C00923417. Recipient of LTF committee contributions; filer of the March 2026 Lantern spending records. Separate from Build American AI and Public First Action.", "x": 2463.7, "y": 930.7, "degree": 3 }, { "id": "w3-opp-lantern", "label": "Lantern Production Consultants LLC", "group": "industry", "role": "Organisation", "summary": "Named payee in Think Big Schedule B and E. This pass identifies the filed vendor; it does not identify beneficial owners, individual workers or ultimate subvendors.", "x": 2417.3, "y": 931.8, "degree": 3 }, { "id": "w3-opp-bores", "label": "Alexander Bores — NY-12 candidate", "group": "state", "role": "Person", "summary": "Target of seven March 2026 independent expenditures marked opposition in Think Big filings; federal House NY-12 primary 2026. No recipient of those expenditures.", "x": 2372.0, "y": 901.5, "degree": 1 }, { "id": "w3-opp-march-operation", "label": "Think Big March 2026 Bores opposition operation", "group": "industry", "role": "Campaign", "summary": "Seven filed digital, print, texting and cable expenditure rows, reported dissemination March 1-30, 2026. No specific creative or platform invoice was matched.", "x": 2382.5, "y": 921.8, "degree": 2 }, { "id": "w3-icmec-australia", "label": "ICMEC Australia", "group": "custody", "role": "Organisation", "summary": "Australian legal company ABN86645682828, ACN645682828; US audit dates loss of sole membership/control to 6 July 2023. Published July2023-labelled/March2025 constitutions describe membership and appointment rules, including a US-specific renewal exception; actual members and exact adoption history unresolved. Global Centre For Child Protection is a same-ABN business name from 3 July 2026.", "x": 1623.5, "y": 1796.8, "degree": 17 }, { "id": "w3-icmec-westpac", "label": "Westpac", "group": "custody", "role": "Organisation", "summary": "Bank that financed the Australian data-sharing programme; recipient-specific agreement remains not obtained.", "x": 1651.9, "y": 1861.9, "degree": 5 }, { "id": "w3-icmec-lighthouse", "label": "Lighthouse", "group": "custody", "role": "Project", "summary": "ICMEC Australia investigative data application; data access and downstream account decisions are separate questions.", "x": 1549.4, "y": 1838.3, "degree": 5 }, { "id": "w3-icmec-saferai", "label": "SaferAI for Children", "group": "custody", "role": "Coalition", "summary": "ICMEC Australia-led AI and child-safety coalition, with published membership and participation rules.", "x": 1627.6, "y": 1781.3, "degree": 3 }, { "id": "w3-icmec-partner-category", "label": "Unlisted partner category", "group": "custody", "role": "Decision-making role", "summary": "A permitted participation category in the published rules, not an identified person or inferred secret membership instance.", "x": 1586.3, "y": 1805.6, "degree": 1 }, { "id": "w3-icmec-jodie-arthur", "label": "Jodie Arthur", "group": "custody", "role": "Person", "summary": "2024-25 report lists Westpac financial-crime executive and ICMEC Australia non-executive director roles.", "x": 1608.7, "y": 1843.7, "degree": 3 }, { "id": "w3-icmec-gannon", "label": "Colm Gannon", "group": "credentials", "role": "Person", "summary": "ICMEC Australia CEO and author of device-level age-assurance advocacy in May2026.", "x": 1562.2, "y": 1831.8, "degree": 2 }, { "id": "w3-icmec-device-advocacy", "label": "Australian device-age advocacy", "group": "credentials", "role": "Publication", "summary": "May2026 advocacy for device-centred age signals, described as reducing repeat verification and platform data exposure.", "x": 1519.6, "y": 1858.0, "degree": 1 }, { "id": "w3-icmec-jcots", "label": "Virginia JCOTS", "group": "state", "role": "Organisation", "summary": "Joint Commission on Technology and Science;2024 report records presentation and subsequent draft request.", "x": 1599.9, "y": 1716.8, "degree": 1 }, { "id": "w3-icmec-craig", "label": "Senator Craig", "group": "state", "role": "Person", "summary": "Senator named in JCOTS December3,2024 record as requesting a device-age bill draft.", "x": 1586.4, "y": 1694.3, "degree": 1 }, { "id": "w3-icmec-draft-request", "label": "Virginia device-age draft request", "group": "state", "role": "Decision", "summary": "Drafting request submitted to Division of Legislative Services, according to December3,2024 JCOTS proceedings.", "x": 1635.7, "y": 1681.6, "degree": 2 }, { "id": "w4-fin-deconsolidation", "label": "Australia deconsolidation · 6July2023", "group": "custody", "role": "Event", "summary": "Accounting removal of USD13,519,626 net assets from ICMEC consolidated accounts; not a cash-transfer event. Contains cash, investments, receivables and other balances.", "x": 1588.7, "y": 1829.5, "degree": 1 }, { "id": "w4-pol-paper", "label": "SaferAI November 2024 discussion paper", "group": "credentials", "role": "Publication", "summary": "Published AI-and-child-protection recommendations; collective-writing attribution, no named final approver.", "x": 1684.1, "y": 1745.3, "degree": 5 }, { "id": "w4-pol-device-proposal", "label": "SaferAI device/encrypted-environment detection proposal", "group": "credentials", "role": "Proposal", "summary": "Paper p8 proposes AI detection/reporting within encrypted environments and blocking at device level. Not an age or identity signal design or implementation.", "x": 1646.3, "y": 1771.2, "degree": 1 }, { "id": "w4-pol-tas-review", "label": "Woolcott review to Tasmania's Premier (2025)", "group": "credentials", "role": "Publication", "summary": "Independent review Part A, November2025; explicitly cites and draws on SaferAI paper in section6.3.2.", "x": 1598.5, "y": 1752.6, "degree": 3 }, { "id": "w4-pol-woolcott", "label": "Peter Woolcott AO", "group": "credentials", "role": "Person", "summary": "Named independent reviewer; no assertion he drafted every cited paragraph.", "x": 1550.2, "y": 1763.9, "degree": 1 }, { "id": "w4-pol-premier-office", "label": "Premier of Tasmania (report addressee)", "group": "credentials", "role": "Office", "summary": "Report cover addresses Jeremy Rockliff as Premier. No separate transmittal receipt or acceptance of each recommendation acquired.", "x": 1547.6, "y": 1757.4, "degree": 1 }, { "id": "w4-pol-esafety-toolkit", "label": "eSafety CSEA Safety by Design toolkit (March2026)", "group": "credentials", "role": "Guidance", "summary": "Government regulator guidance cites the paper for AI investigative triage; not a legal instrument.", "x": 1816.1, "y": 1684.5, "degree": 5 }, { "id": "w4-pol-jago", "label": "Mikaela Jago", "group": "credentials", "role": "Person", "summary": "Named in original paper/call PDF metadata; 2024-25 institutional report lists Senior Capacity and Prevention Coordinator. Production metadata does not prove prose authorship or approval.", "x": 1581.1, "y": 1820.1, "degree": 1 }, { "id": "w4-lh-kablamo", "label": "Kablamo", "group": "custody", "role": "Organisation", "summary": "Lighthouse delivery and support vendor as described by its own case study; contracting legal entity unverified.", "x": 1573.3, "y": 1825.6, "degree": 2 }, { "id": "w4-lh-product-owner", "label": "Lighthouse product owner — unnamed delivery role", "group": "custody", "role": "Decision-making role", "summary": "An ICMEC delivery role described by the vendor, not an identified person or a demonstrated financial/policy decision-maker.", "x": 1501.6, "y": 1862.6, "degree": 1 }, { "id": "w4-lh-data-suppliers", "label": "Lighthouse data suppliers — unnamed", "group": "custody", "role": "Decision-making role", "summary": "Provider category reported by ICMEC; no named feed supplier, licence counterparty or dataset ownership inferred.", "x": 1507.0, "y": 1868.6, "degree": 1 }, { "id": "w4-lh-austrac", "label": "AUSTRAC", "group": "custody", "role": "Organisation", "summary": "Australian financial intelligence agency; named joint participant in Operation Huntsman.", "x": 1138.6, "y": 2582.4, "degree": 1 }, { "id": "w4-lh-accce", "label": "Australian Centre to Counter Child Exploitation (ACCCE)", "group": "custody", "role": "Organisation", "summary": "Named joint participant in Operation Huntsman. The AUSTRAC page expands the acronym with an additional word Sexual; the programme uses the standard name shown here.", "x": 1128.1, "y": 2587.1, "degree": 1 }, { "id": "w4-lh-huntsman", "label": "Operation Huntsman", "group": "custody", "role": "Programme", "summary": "Separate joint ACCCE-AUSTRAC initiative, reported by AUSTRAC as operating since June 2022. No recovered Lighthouse join.", "x": 1148.8, "y": 2562.0, "degree": 3 }, { "id": "w4-lh-huntsman-closures", "label": "Huntsman bank-account closures — reported aggregate", "group": "custody", "role": "Event", "summary": "More than 3,000 Australian accounts linked to sextortion payments, as reported on 15 September 2025; no individual account identifiers acquired.", "x": 1167.2, "y": 2535.9, "degree": 1 }, { "id": "w4-pol-esafety-encryption-guidance", "label": "eSafety encrypted-service recommendations", "group": "credentials", "role": "Policy proposal", "summary": "Toolkit asks industry to consider client-side scanning, homomorphic encryption, risk-based encryption deployment and conditional restrictions. Published guidance, not proof of technical efficacy, legal compulsion or deployment.", "x": 1813.7, "y": 1719.7, "degree": 1 }, { "id": "w5-gov-au-members", "label": "ICMEC Australia company members", "group": "custody", "role": "Governance body", "summary": "Legal voting membership; actual names and change dates not recovered. Distinct from portal users, donors or SaferAI coalition members.", "x": 1528.1, "y": 1879.8, "degree": 3 }, { "id": "w5-gov-au-board", "label": "ICMEC Australia board", "group": "custody", "role": "Governance body", "summary": "Published approval body for strategy, budgets, major partnerships and material programs; specific exercised decisions unresolved.", "x": 1569.6, "y": 1850.0, "degree": 5 }, { "id": "w5-gov-au-secretary", "label": "ICMEC Australia company secretary", "group": "custody", "role": "Decision-making role", "summary": "Named role holding membership/minute and interests registers; individual incumbent not established.", "x": 1533.4, "y": 1884.2, "degree": 2 }, { "id": "w5-gov-constitution", "label": "ICMEC Australia published constitutions", "group": "custody", "role": "Publication", "summary": "July2023-labeled and March2025 texts preserve a US-specific membership-renewal exception. Actual adoption resolutions and member register not obtained.", "x": 1662.6, "y": 1675.8, "degree": 1 }, { "id": "w5-gov-grf", "label": "GRF CPAs & Advisors", "group": "custody", "role": "Organisation", "summary": "US consolidating auditor; report header identifies GRF and its receipt of Australian component auditor reports. Australian auditor identity unresolved.", "x": 1587.6, "y": 1825.0, "degree": 1 }, { "id": "w5-enc-paper", "label": "Guiding principles for technology-facilitated CSEA (December 2024)", "group": "credentials", "role": "Publication", "summary": "61-page DFC technical report; cited by eSafety footnotes26 and29. Scoped reading, not whole-report technical audit.", "x": 1933.3, "y": 1710.2, "degree": 12 }, { "id": "w5-enc-dfc", "label": "Digital Futures for Children centre", "group": "state", "role": "Organisation", "summary": "Joint LSE and 5Rights research centre; report acknowledges centre funding from 5Rights.", "x": 1922.1, "y": 1629.4, "degree": 2 }, { "id": "w5-enc-ukri", "label": "UK Research and Innovation (UKRI)", "group": "state", "role": "Organisation", "summary": "", "x": 2150.7, "y": 1658.5, "degree": 2 }, { "id": "w5-enc-rephrain", "label": "REPHRAIN research centre", "group": "state", "role": "Programme", "summary": "", "x": 2040.3, "y": 1718.0, "degree": 3 }, { "id": "w5-enc-toolbox", "label": "Toolbox for Measuring Online CSEA", "group": "state", "role": "Programme", "summary": "Report identifies itself as work package3, funded through REPHRAIN fourth Strategic Funding Call.", "x": 2007.2, "y": 1760.2, "degree": 4 }, { "id": "w5-enc-nicholas", "label": "Mariesa Nicholas", "group": "credentials", "role": "Person", "summary": "Report p5 names her as independent peer reviewer with Australian e-Safety Commissioner Office affiliation; no toolkit authorship inferred.", "x": 1896.9, "y": 1647.5, "degree": 2 }, { "id": "w5-enc-orben", "label": "Amy Orben", "group": "credentials", "role": "Person", "summary": "Report p8 identifies her as the Toolbox project lead at University of Cambridge.", "x": 2027.8, "y": 1804.0, "degree": 2 }, { "id": "w5-enc-workshop", "label": "December 2024 Safety by Design CSEA workshop", "group": "state", "role": "Event", "summary": "Public FOI correspondence identifies workshop and subsequent toolkit drafting. No paragraph is backdated to the workshop.", "x": 1954.8, "y": 1615.6, "degree": 2 }, { "id": "w5-enc-consultation", "label": "May 2025 eSafety toolkit industry consultation", "group": "state", "role": "Event", "summary": "Published FOI25139 Doc19 preserves 5May invitation and 16May X response; attachment contents not acquired.", "x": 1896.2, "y": 1624.9, "degree": 6 }, { "id": "w5-enc-x", "label": "X Corp.", "group": "state", "role": "Organisation", "summary": "Corporate recipient identified by released correspondence and @x.com domains; private names remain redacted.", "x": 1872.3, "y": 1611.4, "degree": 3 }, { "id": "w5-enc-spuy", "label": "Anri van der Spuy", "group": "credentials", "role": "Person", "summary": "Named report author on December2024 cover; no allocation of particular sentences.", "x": 1945.1, "y": 1750.7, "degree": 1 }, { "id": "w5-enc-witting", "label": "Sabine Witting", "group": "credentials", "role": "Person", "summary": "Named report author on December2024 cover; no allocation of particular sentences.", "x": 2014.7, "y": 1677.1, "degree": 2 }, { "id": "w5-enc-burton", "label": "Patrick Burton", "group": "credentials", "role": "Person", "summary": "Named report author on December2024 cover; no allocation of particular sentences.", "x": 1939.0, "y": 1753.7, "degree": 1 }, { "id": "w5-enc-day", "label": "Emma Day", "group": "credentials", "role": "Person", "summary": "Named report author on December2024 cover; no allocation of particular sentences.", "x": 2017.5, "y": 1679.5, "degree": 2 }, { "id": "w5-enc-livingstone", "label": "Sonia Livingstone", "group": "credentials", "role": "Person", "summary": "Named report author on December2024 cover; no allocation of particular sentences.", "x": 2012.0, "y": 1674.6, "degree": 2 }, { "id": "w5-enc-sylwander", "label": "Kim R. Sylwander", "group": "credentials", "role": "Person", "summary": "Named report author on December2024 cover; no allocation of particular sentences.", "x": 1931.8, "y": 1751.8, "degree": 1 }, { "id": "w5-fund-unsw", "label": "UNSW Sydney", "group": "custody", "role": "Organisation", "summary": "", "x": 1602.5, "y": 1955.3, "degree": 1 }, { "id": "w5-fund-project2021", "label": "UNSW online-offending survey project", "group": "custody", "role": "Research", "summary": "Named in Westpac 2021 research-grant recipients. Related 2023 Australian report and 2025 multinational analysis are distinct publications.", "x": 1622.9, "y": 1916.0, "degree": 2 }, { "id": "w5-fund-study2025", "label": "Financial-behaviour study · 2025", "group": "credentials", "role": "Publication", "summary": "Study C in Childlight Searchlight2025. Recommendations are researchers' proposals, not observed bank rules.", "x": 1734.0, "y": 1881.8, "degree": 7 }, { "id": "w5-fund-childlight", "label": "Childlight", "group": "safety", "role": "Organisation", "summary": "", "x": 1722.7, "y": 1939.6, "degree": 2 }, { "id": "w5-fund-hdf", "label": "Human Dignity Foundation", "group": "rights", "role": "Organisation", "summary": "", "x": 1714.6, "y": 1980.5, "degree": 1 }, { "id": "w5-fund-advisory", "label": "Study C advisory committee", "group": "credentials", "role": "Committee", "summary": "Report-listed advisers; no minutes, attendance or individual edit record obtained.", "x": 1669.5, "y": 1849.8, "degree": 3 }, { "id": "w5-fund-monitoring", "label": "Broader financial monitoring proposal", "group": "custody", "role": "Proposal", "summary": "Research recommendation to expand transaction detection and cross-sector information sharing; no validated operational classifier or implementation shown.", "x": 1728.2, "y": 1924.3, "degree": 1 }, { "id": "w5-fund-jss", "label": "Jesuit Social Services", "group": "safety", "role": "Organisation", "summary": "", "x": 1718.6, "y": 1922.9, "degree": 1 }, { "id": "w5-fund-salter", "label": "Michael Salter", "group": "credentials", "role": "Person", "summary": "Named author of Childlight2025 Study C and independent peer reviewer of DFC2024 report. Roles in separate documents do not establish control over either institution.", "x": 1838.9, "y": 1801.4, "degree": 2 }, { "id": "w6-ref-levy", "label": "Ian Levy", "group": "state", "role": "Person", "summary": "Listed as UK National Cyber Security Centre technical director on the July 2022 paper; historical role, not a current-employment claim.", "x": 1939.1, "y": 1808.7, "degree": 2 }, { "id": "w6-ref-robinson", "label": "Crispin Robinson", "group": "state", "role": "Person", "summary": "Listed as GCHQ technical director for cryptanalysis on the July 2022 paper; historical role.", "x": 1948.0, "y": 1807.1, "degree": 2 }, { "id": "w6-ref-ncsc", "label": "UK National Cyber Security Centre", "group": "state", "role": "Government body", "summary": "Institution identified on the paper cover; no new funding or command relationship inferred.", "x": 1951.5, "y": 1852.6, "degree": 1 }, { "id": "w6-ref-gchq", "label": "GCHQ", "group": "state", "role": "Government body", "summary": "Institution identified on the paper cover; authors expressly disclaim government policy and requirements.", "x": 1965.7, "y": 1849.8, "degree": 1 }, { "id": "w6-ref-paper", "label": "Thoughts on Child Safety on Commodity Platforms (2022)", "group": "credentials", "role": "Research paper", "summary": "Technical proposal later cited by EU and Australian authorities; contains explicit scope limitations.", "x": 1921.6, "y": 1748.0, "degree": 5 }, { "id": "w6-ref-ec-defence", "label": "European Commission public defence (August 2022)", "group": "state", "role": "Policy communication", "summary": "Archived argument for the then-proposed child-abuse regulation, including a citation to Levy and Robinson.", "x": 1794.9, "y": 1760.6, "degree": 2 }, { "id": "w6-ref-position", "label": "eSafety encryption position (October 2023)", "group": "state", "role": "Policy statement", "summary": "Agency discussion of encrypted-service detection options with technical limitations, predating the two2024 reports.", "x": 1892.3, "y": 1666.6, "degree": 2 }, { "id": "w6-ref-consultation", "label": "Australian draft industry standards consultation (2023–24)", "group": "state", "role": "Consultation", "summary": "Common destination of contrasting WeProtect Secretariat and Digital Rights Watch submissions.", "x": 1643.1, "y": 1510.5, "degree": 2 }, { "id": "w6-ref-drw", "label": "Digital Rights Watch", "group": "rights", "role": "Civil society organisation", "summary": "Australian consultation participant opposing proposed encrypted-content detection requirements.", "x": 1600.3, "y": 1454.1, "degree": 2 }, { "id": "w6-draft-may-a", "label": "May 2025 CSEA toolkit Attachment A: core content", "group": "credentials", "role": "Draft document", "summary": "Exact indexed filename: Attachment A - DRAFT CSEA SbD Toolkit - Industry feedback - May 2025.docx. Contents not acquired.", "x": 1896.2, "y": 1659.5, "degree": 1 }, { "id": "w6-draft-may-b", "label": "May 2025 CSEA toolkit Attachment B: practical tools", "group": "credentials", "role": "Draft document", "summary": "Exact indexed filename: Attachment B - DRAFT CSEA SbD Toolkit - Tools only - Industry feedback - May 2025.docx. Contents not acquired.", "x": 1886.8, "y": 1658.7, "degree": 1 }, { "id": "w6-draft-csm-comments", "label": "Common Sense Media comments to eSafety ·31Oct2025", "group": "credentials", "role": "Publication", "summary": "Published3-page response to requested draft review, hosted by Common Sense Media.", "x": 1939.8, "y": 1592.0, "degree": 3 }, { "id": "w6-draft-october-draft", "label": "Designing for Safety draft reviewed October 2025", "group": "credentials", "role": "Draft document", "summary": "Title identified by31October 2025 comments: Designing for Safety: Preventing Grievous Harm Before It Happens. Exact draft bytes, version label and redlines not acquired.", "x": 1982.3, "y": 1602.6, "degree": 2 }, { "id": "w6-award-survey", "label": "Multinational survey under UNSW ethics HC220317", "group": "custody", "role": "Research", "summary": "Survey fielded November–December 2022; later analyses include Childlight2025 Study C. Survey dataset, 2021 award and later analyses are separate objects; per-output financial allocation remains unknown.", "x": 1714.9, "y": 1951.6, "degree": 3 }, { "id": "w6-award-protocol", "label": "OSF j7xh2 post-analysis registration", "group": "custody", "role": "Record", "summary": "Registered 10 November 2023, approved response revision submitted6 December 2023. Both versions explicitly disclose registration after analysis. Not evidence of a pre-data analysis plan or donor control.", "x": 1707.9, "y": 1993.7, "degree": 1 }, { "id": "w6-award-cloudresearch", "label": "CloudResearch / PrimePanels recruitment service", "group": "custody", "role": "Organisation", "summary": "Authors name CloudResearch as recruitment/administering provider using PrimePanels. No contract, fee, invoice or participant-panel payment schedule acquired.", "x": 1698.5, "y": 1991.8, "degree": 1 }, { "id": "w6-award-wp2", "label": "Toolbox WP2 reliability analyses, October 2024", "group": "custody", "role": "Publication", "summary": "A dated, credited analytical output linked by UKRI; separate from December2024 DFC WorkPackage3 report. Its rationale describes code sharing and later toolbox updating, without identifying grant payment or editorial rights.", "x": 2035.1, "y": 1801.8, "degree": 2 }, { "id": "w7-impl-reddit", "label": "Reddit", "group": "industry", "role": "Platform", "summary": "UK age-access implementation and separate ICO enforcement examined; provider descriptions are not live technical audits.", "x": 2020.2, "y": 1250.0, "degree": 3 }, { "id": "w7-impl-persona", "label": "Persona Identities, Inc.", "group": "custody", "role": "Verification provider", "summary": "Named Reddit processor. Exact contract/fees and independently verified deletion remain unknown.", "x": 2052.6, "y": 1214.1, "degree": 2 }, { "id": "w7-impl-ico", "label": "Information Commissioner’s Office", "group": "state", "role": "Regulator", "summary": "UK data-protection regulator. Its Reddit decision concerns UK GDPR, distinct from Ofcom OSA duties.", "x": 1962.8, "y": 1311.1, "degree": 3 }, { "id": "w7-impl-reddit-age", "label": "Reddit UK mature-content age gate", "group": "custody", "role": "Access condition", "summary": "ICO notice records 14 July 2025 Persona implementation; current general help describes additional region-dependent options.", "x": 2045.9, "y": 1210.8, "degree": 2 }, { "id": "w7-impl-government", "label": "UK government", "group": "state", "role": "Government", "summary": "Cabinet Office issuedMarch 2026 national digital ID consultation; September ministerial answer confirms cancellation. Department/date scopes remain explicit.", "x": 2760.4, "y": 540.5, "degree": 1 }, { "id": "w7-impl-national-id", "label": "National digital-ID scheme · cancelled July 2026", "group": "state", "role": "Cancelled programme", "summary": "March 2026 consultation described possible online-age use. Cancellation confirmed by parliamentary sources; not an active universal online-ID requirement.", "x": 2740.1, "y": 563.2, "degree": 2 }, { "id": "w7-impl-age-use", "label": "Proposed national credential for online age proof", "group": "credentials", "role": "Historical policy proposal", "summary": "March 2026 proposed use explicitly non-exclusive; parent national scheme later cancelled.", "x": 2724.0, "y": 581.1, "degree": 1 }, { "id": "w7-impl-discord-settings", "label": "Discord UK age-gated settings and content", "group": "custody", "role": "Access condition", "summary": "Company documents checks for particular content/settings; declining does not close the whole account.", "x": 2188.3, "y": 1666.9, "degree": 1 }, { "id": "w7-policy-osa", "label": "UK Online Safety Act 2023", "group": "state", "role": "Statute", "summary": "Statutory duties and regulator powers; current sections 12 and 81 read alongside targeted enacted provisions.", "x": 2000.4, "y": 1390.7, "degree": 2 }, { "id": "w7-policy-heaa", "label": "Ofcom highly effective age assurance framework", "group": "state", "role": "Regulatory guidance", "summary": "Four technical criteria plus implementation requirements; method examples do not approve named vendors.", "x": 1991.2, "y": 1377.8, "degree": 4 }, { "id": "w7-policy-child-code", "label": "Issued UK Protection of Children user-to-user Code", "group": "state", "role": "Code of practice", "summary": "Issued 4 July 2025 and in force 25 July 2025. Age-assurance measures have distinct content, risk and functionality scopes.", "x": 1998.0, "y": 1380.8, "degree": 2 }, { "id": "w7-policy-email-estimation", "label": "Email-based age estimation as a HEAA-capable method", "group": "state", "role": "Technical method", "summary": "Added after AVPA and VerifyMy evidence. This is conditional method acceptance, not provider accreditation or purchase.", "x": 2084.9, "y": 1462.8, "degree": 3 }, { "id": "w7-policy-xgroovy", "label": "Provider of Xgroovy.com", "group": "state", "role": "Service provider", "summary": "Ofcom case CW/01318/09/25 identifies the provider by service. The exact legal company remains unresolved.", "x": 1987.8, "y": 1368.1, "degree": 2 }, { "id": "w7-market-avid", "label": "AVID Certification Services Limited", "group": "state", "role": "Company", "summary": "Exact legal company14865982; supplies ACCS audit/certification/test purchasing under an SLA and holds scheme-company director appointment/removal rights.", "x": 2335.4, "y": 1573.2, "degree": 6 }, { "id": "w7-market-scheme", "label": "Age Check Certification Scheme Ltd", "group": "state", "role": "Company", "summary": "Exact legal company11493870, described as nonprofit limited by guarantee. Distinct from AVID service company. Historical accs node remains an umbrella.", "x": 2269.2, "y": 1545.3, "degree": 4 }, { "id": "w7-market-allen", "label": "Tony Iain Allen / Tony Allen", "group": "state", "role": "Person", "summary": "AVID PSC and ACCS executive identified in company register/BSI respectively. Holds disclosed majority stake band; BSI lists ISO27566-1 editorial and national-panel roles.", "x": 2282.6, "y": 1580.5, "degree": 4 }, { "id": "w7-market-bate", "label": "Nicholas John Bate", "group": "state", "role": "Person", "summary": "Current AVID PSC with disclosed more-than25%-to50% share/voting band. Historical ceased Nick Bate entry is not treated as another person or unexplained exit.", "x": 2394.7, "y": 1586.0, "degree": 1 }, { "id": "w7-market-iso27566", "label": "ISO/IEC27566-1:2025 age-assurance framework", "group": "state", "role": "Standard", "summary": "International age-assurance framework, released after the July 2025 UK implementation deadline. Normative full text not acquired. Role/commercial-offering links do not imply mandatory regulator endorsement.", "x": 2340.2, "y": 1589.6, "degree": 3 }, { "id": "w7-market-bsi-panel", "label": "BSI IST/33/5/5 age-assurance panel", "group": "state", "role": "Standards body panel", "summary": "National standards panel with Tony Allen listed as chair in BSI's October 2025 event biography; coordinated UK input described in February 2026 announcement.", "x": 2338.9, "y": 1597.6, "degree": 2 }, { "id": "w7-market-ageaware", "label": "AgeAware interoperability proposal and proof of concept", "group": "state", "role": "Programme", "summary": "2024 consultation and January 2025 reported build under euCONSENT ASBL. Commercial/admission provisions remain proposed unless separately implemented; current AVPA describes development through2025.", "x": 2166.9, "y": 1565.8, "degree": 8 }, { "id": "w7-market-safeonline", "label": "Safe Online funding programme", "group": "state", "role": "Funding programme", "summary": "Named sponsor of euCONSENT Phase II/AgeAware: reported project-level purpose is established. Executed award value, subproject/vendor allocation, ultimate donor allocation and governing award terms remain not obtained.", "x": 2184.6, "y": 1596.7, "degree": 1 }, { "id": "w7-market-verifymy", "label": "VerifyMy Limited", "group": "state", "role": "Company", "summary": "Company12050874 identified in official product-page footer. Offers email-based age estimation; credited participant in AgeAware design. No customer-specific deployment inferred.", "x": 2179.1, "y": 1515.0, "degree": 3 }, { "id": "w7-market-yoti", "label": "Yoti (credited provider brand)", "group": "state", "role": "Provider brand", "summary": "Named as AgeAware detailed-design participant in January 2025 project account. This lane has not allocated that contribution to a specific legal subsidiary.", "x": 2219.1, "y": 1579.8, "degree": 1 }, { "id": "w7-market-agechecked", "label": "AgeChecked (credited provider brand)", "group": "state", "role": "Provider brand", "summary": "Named as AgeAware detailed-design participant in January 2025 project account. Legal operator/shareholding and compensation not here.", "x": 2220.9, "y": 1572.8, "degree": 1 }, { "id": "w7-adv-parkinson", "label": "Lord Parkinson of Whitley Bay", "group": "rights", "role": "Person", "summary": "Minister who presented the July 2023 government age-assurance amendments and credited specific parliamentary collaborators.", "x": 2135.5, "y": 1559.0, "degree": 2 }, { "id": "w7-adv-amend124", "label": "Online Safety Bill Amendment 124 · July 2023", "group": "rights", "role": "Legislative amendment", "summary": "Government age-assurance principles amendment, co-sponsored by Kidron and agreed 10 July 2023. One contribution route into Ofcom's statutory framework.", "x": 2122.4, "y": 1578.6, "degree": 2 }, { "id": "w7-adv-amend125", "label": "Kidron privacy Amendment 125 · July 2023", "group": "rights", "role": "Legislative amendment", "summary": "Proposed addition to Amendment 124 concerning privacy, purpose limits and deletion/storage; minister declined it and it was not moved 10 July 2023.", "x": 2134.1, "y": 1565.6, "degree": 2 }, { "id": "w7-adv-privacy-criterion-proposal", "label": "5Rights privacy criterion proposal · March 2024", "group": "rights", "role": "Policy proposal", "summary": "Part5 consultation request for privacy-preserving age assurance as an express criterion; Ofcom declined that criterion architecture in January 2025.", "x": 1941.4, "y": 1474.5, "degree": 2 }, { "id": "w7-adv-token-proposal", "label": "5Rights age-token proposal · July 2024", "group": "rights", "role": "Policy proposal", "summary": "Part of a broader proposed ecosystem for age-appropriate services; Ofcom later added conditional token guidance after responses from 5Rights, Yoti and ACT.", "x": 1940.5, "y": 1474.4, "degree": 2 }, { "id": "w7-adv-certification-proposal", "label": "5Rights standards/certification proposal · October 2025", "group": "rights", "role": "Policy proposal", "summary": "Additional-measures response promotes risk-based IEEE 2089.1 confidence levels and asks Ofcom to consider certification, citing ACCS requirements. Adoption or payment not established.", "x": 1989.4, "y": 1509.1, "degree": 2 }, { "id": "w7-adv-ageaware-rights-review", "label": "Commissioned AgeAware rights review · reported January 2025", "group": "rights", "role": "Commissioned review", "summary": "euCONSENT reported commissioning a child/human-rights review through team academics, with Emma Day and Sabine Witting of Tech Legality as intended co-leads. The review, contract, fees and completion were not acquired.", "x": 2079.8, "y": 1633.1, "degree": 5 }, { "id": "w8-red-call", "label": "Global Call for AI Red Lines", "group": "redlines", "role": "Campaign", "summary": "Launched 22 September 2025. General demand for internationally enforceable AI boundaries; endorsement does not establish agreement with every partner implementation.", "x": 2232.0, "y": 1514.1, "degree": 22 }, { "id": "w8-red-tfs", "label": "The Future Society", "group": "redlines", "role": "Organisation", "summary": "Campaign co-organizer; named operational staff and policy-forum work. Organizational financing is distinct from campaign expenditure.", "x": 2296.1, "y": 1457.9, "degree": 12 }, { "id": "w8-red-cesia", "label": "CeSIA", "group": "redlines", "role": "Organisation", "summary": "Centre pour la Securite de l'Intelligence Artificielle; co-organizer and author of a 2026 diplomatic funding proposal.", "x": 2321.6, "y": 1492.8, "degree": 11 }, { "id": "w8-red-chai", "label": "Center for Human-Compatible AI (CHAI)", "group": "redlines", "role": "Organisation", "summary": "UC Berkeley research centre named as campaign co-organizer.", "x": 2278.8, "y": 1510.4, "degree": 1 }, { "id": "w8-red-segerie", "label": "Charbel-Raphael Segerie", "group": "redlines", "role": "Person", "summary": "Credited initial drafter; CeSIA executive director in the diplomatic proposal.", "x": 2285.1, "y": 1523.5, "degree": 1 }, { "id": "w8-red-iliadis", "label": "Niki Iliadis", "group": "redlines", "role": "Person", "summary": "Credited TFS project management and policy outreach.", "x": 2273.5, "y": 1500.3, "degree": 1 }, { "id": "w8-red-zoumpalova", "label": "Tereza Zoumpalova", "group": "redlines", "role": "Person", "summary": "Credited TFS campaign management and signatory work; named UNESCO-network subgroup representative.", "x": 2285.9, "y": 1517.2, "degree": 2 }, { "id": "w8-red-tatevossian", "label": "Anoush Rima Tatevossian", "group": "redlines", "role": "Person", "summary": "Credited UN-engagement advice and policy/civil-society outreach.", "x": 2271.0, "y": 1506.5, "degree": 1 }, { "id": "w8-red-grimonpont", "label": "Arthur Grimonpont", "group": "redlines", "role": "Person", "summary": "Credited approach to Maria Ressa and press-conference coordination.", "x": 2280.8, "y": 1532.4, "degree": 2 }, { "id": "w8-red-shimi", "label": "Adam Shimi", "group": "redlines", "role": "Person", "summary": "Credited drafting feedback on the campaign statement; no final approval right established.", "x": 2275.6, "y": 1525.3, "degree": 1 }, { "id": "w8-red-hubinger", "label": "Evan Hubinger", "group": "redlines", "role": "Person", "summary": "Listed campaign signatory. This record does not establish campaign management or endorsement of A Narrow Path.", "x": 2273.5, "y": 1515.9, "degree": 1 }, { "id": "w8-red-ressa", "label": "Maria Ressa", "group": "redlines", "role": "Person", "summary": "Named campaign advocate in UN-hosted prepared speech dated 22 September 2025.", "x": 2297.8, "y": 1537.1, "degree": 3 }, { "id": "w8-red-bengio", "label": "Yoshua Bengio", "group": "redlines", "role": "Person", "summary": "International red-lines advocacy at the 24 September 2025 Security Council meeting; scoped automatic-transcript evidence.", "x": 2064.8, "y": 214.4, "degree": 1 }, { "id": "w8-red-apollo-pr", "label": "Apollo Strategic Communications", "group": "redlines", "role": "Organisation", "summary": "Campaign acknowledgements credit press-conference preparation/moderation. Distinct from Apollo Research and Apollo.io; fee and contract not obtained.", "x": 2266.3, "y": 1519.4, "degree": 1 }, { "id": "w8-red-narrow-path", "label": "A Narrow Path", "group": "redlines", "role": "Proposal", "summary": "ControlAI-promoted AI governance proposal. Distinct from the Global Call; proposed institutions are not established bodies.", "x": 2177.2, "y": 1617.1, "degree": 1 }, { "id": "w8-red-sprint", "label": "CeSIA 2026 diplomatic sprint", "group": "redlines", "role": "Proposal", "summary": "Fundraising proposal seeks lab-threshold harmonization and state-led international agreements. Goals and proposed partnerships are not completed actions.", "x": 2288.3, "y": 1503.3, "degree": 2 }, { "id": "w8-red-effektiv", "label": "Effektiv Spenden", "group": "risk", "role": "Organisation", "summary": "Reports organizational support to CeSIA and The Future Society following Longview recommendations.", "x": 2337.9, "y": 1472.9, "degree": 2 }, { "id": "w8-red-effisciences", "label": "EffiSciences", "group": "risk", "role": "Organisation", "summary": "SFF lists as receiving charity for the 2024 CeSIA recommendation; fiscal-sponsorship terms not obtained.", "x": 2373.3, "y": 1505.2, "degree": 1 }, { "id": "w8-red-unesco", "label": "UNESCO", "group": "state", "role": "Institution", "summary": "", "x": 2261.8, "y": 1530.7, "degree": 2 }, { "id": "w8-red-unesco-network", "label": "UNESCO civil-society and academic AI network", "group": "state", "role": "Programme", "summary": "Umbrella network launched June 2025; funding and chronology distinct from the September campaign.", "x": 2243.5, "y": 1587.9, "degree": 4 }, { "id": "w8-red-unesco-subgroup", "label": "International AI Red Lines subgroup", "group": "redlines", "role": "Programme", "summary": "TFS and Ada co-lead; CeSIA member. Independent CSO work does not commit UNESCO or Member States.", "x": 2299.7, "y": 1524.0, "degree": 5 }, { "id": "w8-red-ada", "label": "Ada Lovelace Institute", "group": "redlines", "role": "Organisation", "summary": "Listed as co-lead of the UNESCO-network red-lines subgroup.", "x": 2354.0, "y": 1535.0, "degree": 1 }, { "id": "w8-red-mcgovern", "label": "Patrick J. McGovern Foundation", "group": "risk", "role": "Organisation", "summary": "UNESCO attributes initial umbrella-network support to this foundation; no campaign-specific grant established.", "x": 2294.4, "y": 1606.8, "degree": 1 }, { "id": "w8-red-clement-jones", "label": "Lord Tim Clement-Jones", "group": "control", "role": "Person", "summary": "Self-reported signatory and parliamentary advocate of the Global Call.", "x": 2271.0, "y": 1535.0, "degree": 1 }, { "id": "w8-red-athens2025", "label": "Athens Roundtable — London, 4 December 2025", "group": "redlines", "role": "Event", "summary": "TFS-organized policy forum; report links agenda to Global Call and records a 5Rights policy-model presentation.", "x": 2259.9, "y": 1470.9, "degree": 13 }, { "id": "w8-red-barrington-leach", "label": "Leanda Barrington-Leach", "group": "rights", "role": "Person", "summary": "5Rights executive director identified in the event recap; advocated the UK Children's Code as an AI safeguard model.", "x": 2088.5, "y": 1505.8, "degree": 3 }, { "id": "w8-red-childrens-code", "label": "UK Age Appropriate Design Code", "group": "rights", "role": "Policy", "summary": "Policy model explicitly invoked at the AI forum. This source does not establish a new identity-check mandate.", "x": 2134.8, "y": 1496.7, "degree": 1 }, { "id": "w8-red-cooperative-ai", "label": "Cooperative AI Foundation", "group": "risk", "role": "Organisation", "summary": "CAIF, UK charity1201294/company13485176. Reports hosting Athens speakers' reception and advocating multi-agent monitoring. Organizational backing distinct from event allocation.", "x": 2373.5, "y": 1432.4, "degree": 6 }, { "id": "w8-red-arnold-porter", "label": "Arnold & Porter", "group": "redlines", "role": "Organisation", "summary": "Credited event supporter; exact contracting entity, support form and amount unspecified.", "x": 2305.4, "y": 1467.2, "degree": 1 }, { "id": "w8-red-fathom", "label": "Fathom (AI-governance nonprofit)", "group": "redlines", "role": "Organisation", "summary": "Resolved to fathom.org, self-disclosed FEIN93-4840479. Current TFS verified-redlines partnership; event-series financial support separately reported. No verified2025amount.", "x": 2341.8, "y": 1413.4, "degree": 7 }, { "id": "w8-red-deliberaide", "label": "deliberAIde platform", "group": "redlines", "role": "Organisation", "summary": "Brand resolved to deliberaide.com; attributed TFS testimonial describes event/report workflow. Legal operator and event-specific terms remain unresolved.", "x": 2303.0, "y": 1459.7, "degree": 2 }, { "id": "w8-red-india-workshop", "label": "AI Safety Connect India red-lines workshop", "group": "redlines", "role": "Event", "summary": "February 2026 TFS/CeSIA workshop. Not the government summit declaration or a negotiated agreement.", "x": 2334.4, "y": 1468.6, "degree": 2 }, { "id": "w8-red-unga80", "label": "UNGA 80th anniversary session — 22 September 2025", "group": "state", "role": "Event", "summary": "", "x": 2349.4, "y": 1552.0, "degree": 1 }, { "id": "w8-red-unsc10005", "label": "UN Security Council meeting 10005", "group": "state", "role": "Event", "summary": "", "x": 2066.2, "y": 200.0, "degree": 1 }, { "id": "w8-red-renaissance", "label": "Renaissance Numerique", "group": "redlines", "role": "Organisation", "summary": "TFS survey methodology credits jointly consolidating fifteen workshop priorities into ten survey options.", "x": 2340.4, "y": 1497.1, "degree": 1 }, { "id": "w8-red-survey", "label": "February 2025 civil-society priority survey", "group": "redlines", "role": "Research", "summary": "44 anonymous self-identified representatives ranked curated options; not a representative global consensus measurement.", "x": 2291.2, "y": 1491.2, "degree": 3 }, { "id": "w8-follow-macroscopic", "label": "Macroscopic Ventures", "group": "risk", "role": "Organisation", "summary": "CAIF names this philanthropic backer; legal grantor and receipt schedule remain not obtained.", "x": 2428.2, "y": 1430.2, "degree": 2 }, { "id": "w8-follow-clifton", "label": "Jesse Clifton", "group": "risk", "role": "Person", "summary": "CAIF trustee listed as Macroscopic grantmaking officer. Director appointment6June2025 separately recorded; funder appointment rights not established.", "x": 2424.1, "y": 1421.5, "degree": 2 }, { "id": "w8-follow-athens-series", "label": "Athens Roundtable event series", "group": "redlines", "role": "Programme", "summary": "Series-level financial support must not be allocated to a particular annual event without evidence.", "x": 2310.3, "y": 1441.9, "degree": 2 }, { "id": "w8-follow-athens-dialogue", "label": "Athens2025 private red-lines dialogue", "group": "redlines", "role": "Event", "summary": "Participant accounts describe policy contributions; acceptance or official adoption requires separate evidence.", "x": 2326.2, "y": 1450.7, "degree": 2 }, { "id": "w8-follow-france-g7", "label": "French G7 presidency2026", "group": "state", "role": "Institution", "summary": "", "x": 2374.1, "y": 1496.8, "degree": 2 }, { "id": "w8-follow-mirlesse", "label": "Alexandre Mirlesse", "group": "state", "role": "Person", "summary": "French foreign-ministry AI/tech diplomat in March2026 programme; scheduled moderator, not inferred attendee or G7 drafting official.", "x": 2421.0, "y": 1509.6, "degree": 1 }, { "id": "w8-follow-sciencespo-march", "label": "Sciences Po/CeSIA conference — 10March2026", "group": "redlines", "role": "Event", "summary": "Institutional event programme records co-organization and scheduled diplomatic participants.", "x": 2383.0, "y": 1502.2, "degree": 3 }, { "id": "w8-follow-g7-ministers", "label": "G7 Digital and Technology Ministers", "group": "state", "role": "Institution", "summary": "", "x": 3308.7, "y": 1478.2, "degree": 5 }, { "id": "w8-follow-risk-comparability", "label": "G7 risk-framework comparability discussions", "group": "state", "role": "Policy", "summary": "Public commitment to future discussions in ministerial declaration; not an adopted common threshold or slowdown trigger.", "x": 3343.4, "y": 1480.6, "degree": 1 }, { "id": "w8-follow-g7-open-ai", "label": "G7 AI openness and adoption programme", "group": "state", "role": "Policy", "summary": "Official ministerial declaration also promotes openness, innovation and broader AI adoption.", "x": 3339.5, "y": 1488.2, "degree": 1 }, { "id": "w9-assure-multi-agent-fund", "label": "Scaling AI Safety for a Multi-Agent World", "group": "risk", "role": "Programme", "summary": "Advertised research competition ceiling up to USD10m, not paid cash or separately established legal entity. Awards expected autumn2026; no partner allocations acquired.", "x": 2462.5, "y": 1387.8, "degree": 9 }, { "id": "w9-assure-schmidt-sciences", "label": "Schmidt Sciences", "group": "risk", "role": "Organisation", "summary": "FAQ identifies Schmidt Sciences LLC, subsidiary of Eric and Wendy Schmidt Fund for Strategic Innovation. No parent EIN assigned. Administrative front door is not sole grantor proof.", "x": 2518.3, "y": 1379.6, "degree": 3 }, { "id": "w9-assure-aria", "label": "ARIA", "group": "state", "role": "Organisation", "summary": "Named joint supporting party. Separate contribution, award and allocation not acquired.", "x": 2514.0, "y": 1369.3, "degree": 1 }, { "id": "w9-assure-google-option", "label": "Google.org optional funding route", "group": "industry", "role": "Funding route", "summary": "Opt-in route carries openness/Google AI Principles conditions. Legal grantor and executed terms unknown; opting out preserves other-partner consideration.", "x": 2518.3, "y": 1373.6, "degree": 2 }, { "id": "w9-assure-agent-infrastructure", "label": "Agent infrastructure and admission-control research", "group": "risk", "role": "Research agenda", "summary": "Identity/authentication, admission and revocation are research topics. Call excludes one privileged controller of all agents; no compulsory operational identity requirement.", "x": 2509.2, "y": 1364.7, "degree": 1 }, { "id": "w9-assure-sb813", "label": "California SB813 — enrolled, enactment unestablished", "group": "state", "role": "Legislation", "summary": "Passed30August/enrolled1September2026; no signature/chapter shown9September. Would establish GovOps IVO qualification but preserves operation without this route. Earlier certification liability benefits absent.", "x": 2414.4, "y": 1357.0, "degree": 5 }, { "id": "w9-assure-govops", "label": "California Government Operations Agency", "group": "state", "role": "Public authority", "summary": "Enrolled SB813 would delegate verifier qualification and suspension procedures if enacted; no exercised designation acquired.", "x": 2463.7, "y": 1325.8, "degree": 2 }, { "id": "w9-assure-ivo-role", "label": "SB813 independent-verifier role", "group": "state", "role": "Proposed role", "summary": "No appointed operator. Proposed independence, reporting and non-outcome-contingent payment rules apply to the role.", "x": 2490.3, "y": 1306.1, "degree": 2 }, { "id": "w9-assure-assessed-role", "label": "SB813 assessed developer/deployer/operator", "group": "industry", "role": "Decision-making role", "summary": "Generic role, not a named company. Verification optional under this proposed framework; other laws/contracts remain separate.", "x": 2458.8, "y": 1322.2, "degree": 2 }, { "id": "w9-assure-old-mro", "label": "SB813 earlier MRO role —2025", "group": "state", "role": "Superseded policy", "summary": "Historical sponsor sought recognition under an earlier model-review/certification scheme. No later appointment or surviving liability privilege inferred.", "x": 2388.7, "y": 1378.7, "degree": 2 }, { "id": "w9-assure-bsa", "label": "Business Software Alliance", "group": "industry", "role": "Organisation", "summary": "Opposition and prediction of practical litigation pressure reported in final Floor analysis; not an observed compulsory-audit outcome.", "x": 2460.3, "y": 1335.3, "degree": 1 }, { "id": "w9-assure-g7-principles", "label": "G7 minors principles — ministers", "group": "state", "role": "Policy statement", "summary": "Age assurance encouraged consistently with law/minimal intrusion. No specified vendor, real-name system, universal access age or mandatory certification.", "x": 3263.2, "y": 1474.1, "degree": 3 }, { "id": "w9-assure-g7-leaders", "label": "G7 heads of state and government", "group": "state", "role": "Institution", "summary": "Authors of separate17June call; distinct from digital ministers and privacy authorities.", "x": 3328.8, "y": 1470.0, "degree": 2 }, { "id": "w9-assure-g7-call", "label": "G7 leaders minors call —17June2026", "group": "state", "role": "Policy statement", "summary": "Requests conversational-AI assurance/default protections. Under18 defines beneficiaries, not a universal minimum access age; no enacted global gate.", "x": 3302.4, "y": 1469.9, "degree": 2 }, { "id": "w9-assure-g7-dpas", "label": "G7 data-protection and privacy authorities", "group": "state", "role": "Institution", "summary": "Separate roundtable, not a unitary G7 regulator; guidance does not establish compliant implementation.", "x": 3174.3, "y": 1472.7, "degree": 1 }, { "id": "w9-assure-g7-privacy", "label": "G7 privacy limits on age assurance", "group": "state", "role": "Policy statement", "summary": "Opposes indiscriminate use and identification/tracking/profiling/monitoring; purpose/data limitations. Notes ISO27566 without required certification.", "x": 3220.7, "y": 1473.2, "degree": 2 }, { "id": "w9-assure-5rights-premarket", "label": "5Rights premarket certification demand", "group": "rights", "role": "Policy proposal", "summary": "17June demand goes beyond inspected G7 adopted texts; no operator specified.", "x": 1913.0, "y": 1569.9, "degree": 1 }, { "id": "w9-assure-unicef", "label": "UNICEF", "group": "rights", "role": "Institution", "summary": "Reports advance evidence contribution; exact package and clause uptake not obtained.", "x": 3344.5, "y": 1472.2, "degree": 1 }, { "id": "w9-assure-paris-simulation", "label": "Paris verification simulation —2026", "group": "state", "role": "Publication", "summary": "Organizer-reported findings about revocation, appeals and conflicting incentives; not actual control or capture.", "x": 2386.1, "y": 1394.4, "degree": 1 }, { "id": "w9-assure-ashby-proposals", "label": "Ashby2026 verification/incentive proposals", "group": "state", "role": "Publication", "summary": "Optional verification could affect insurance/liability; child-rating and high-risk fraud-identity concepts remain proposals.", "x": 2379.2, "y": 1388.7, "degree": 1 }, { "id": "w10-duty-carnegie", "label": "Carnegie UK", "group": "dutycare", "role": "Organisation", "summary": "Institutional author of Australian committee submission. Statutory review cites its earlier model work; no current funding or clause-clearance rights inferred.", "x": 1695.9, "y": 1323.7, "degree": 4 }, { "id": "w10-duty-woods", "label": "Lorna Woods", "group": "dutycare", "role": "Person", "summary": "Model originator credited by statutory review, network principal adviser in2024 and joint article author. No inferred personal grant or control right.", "x": 1764.9, "y": 1368.5, "degree": 5 }, { "id": "w10-duty-perrin", "label": "William Perrin", "group": "dutycare", "role": "Person", "summary": "Statutory review credits duty-of-care conception with Woods; no financing or individual drafting approval established in this wave.", "x": 1669.4, "y": 1343.0, "degree": 2 }, { "id": "w10-duty-network", "label": "Online Safety Network", "group": "dutycare", "role": "Organisation", "summary": "Same network previously named Online Safety Act Network; renamed1September2026.2024 government-hosted submission and advisory role at their dates.", "x": 1765.0, "y": 1381.3, "degree": 8 }, { "id": "w10-duty-reset-australia", "label": "Reset.Tech Australia", "group": "dutycare", "role": "Organisation", "summary": "Author of submission70 and Farthing affiliation in joint article. Distinct scoped Australian node; not silently merged with existing ambiguous global Reset node. No new donor-direction edge.", "x": 1574.7, "y": 1393.2, "degree": 16 }, { "id": "w10-duty-farthing", "label": "Rys Farthing", "group": "dutycare", "role": "Person", "summary": "Joint article author identified with Reset Australia; coauthorship does not prove organization-wide control or commissioning.", "x": 1628.2, "y": 1368.8, "degree": 2 }, { "id": "w10-duty-joint-output", "label": "A singular Australian duty: Woods/Farthing article", "group": "dutycare", "role": "Publication", "summary": "22April2024 joint public argument. Demonstrates a shared authored output, not formal inter-organizational agreement or adopted clause.", "x": 1698.3, "y": 1359.3, "degree": 2 }, { "id": "w10-duty-review2025", "label": "Published updated statutory review —February2025", "group": "dutycare", "role": "Publication", "summary": "Section5.5 credits Woods/Perrin, cites Carnegie model work and directly quotes Reset submission70 while recommending an overarching duty. Recipient-side acknowledgement; no exclusive drafting attribution. Initially deliveredOctober2024, then revised beforeFebruary2025 publication; available text is not assumed identical to initial delivery. FOI reports departmental acceptance brief clearance, not signed ministerial acceptance.", "x": 1689.9, "y": 1385.7, "degree": 21 }, { "id": "w10-duty-draft2026", "label": "Australia digital duty of care — exposure draft", "group": "dutycare", "role": "Proposed legislation", "summary": "Published8September2026. Creates proposed ministerial/Commissioner discretion and provider duties if enacted. Operative rules and certain simplified outlines remain incomplete.", "x": 1755.1, "y": 1472.6, "degree": 13 }, { "id": "w10-duty-minister-role", "label": "Australian communications minister — statutory role", "group": "dutycare", "role": "Decision-making role", "summary": "Role that would make specified legislative instruments under this draft; distinct from claims about current individual decisions.", "x": 1746.2, "y": 1471.0, "degree": 5 }, { "id": "w10-duty-provider-role", "label": "Covered online-service providers", "group": "dutycare", "role": "Decision-making role", "summary": "Generic covered class and day-to-day controllers. Proportionate safe-environment duties proposed; no named implementation or universal identity collection established.", "x": 1779.9, "y": 1508.1, "degree": 4 }, { "id": "w10-duty-private-adult-limit", "label": "Private consenting-adult limit on the duty", "group": "dutycare", "role": "Proposed safeguard", "summary": "Section26(7): duty does not require action concerning lawful communications privately solely between consenting adults. Not blanket immunity; public viewing and adult-AI application differ or remain unresolved.", "x": 1748.4, "y": 1450.2, "degree": 1 }, { "id": "w10-duty-privacy-factor", "label": "Privacy in reasonably practicable measures", "group": "dutycare", "role": "Proposed safeguard", "summary": "Sections26-26A proportionality considers privacy with severity, likelihood, knowledge, availability and cost. Not a blanket encryption exemption or proven compliant implementation.", "x": 1718.1, "y": 1459.7, "degree": 1 }, { "id": "w10-duty-merits-review", "label": "Review of specified enforcement decisions", "group": "dutycare", "role": "Proposed safeguard", "summary": "Draft provides ART review of specified decisions including duty directions; existing internal review may apply. No automatic stay or universal merits appeal established.", "x": 1805.7, "y": 1525.1, "degree": 2 }, { "id": "w10-duty-research-role", "label": "Approved Australian-university researchers", "group": "dutycare", "role": "Proposed role", "summary": "Section205B university employment plus approval under future legislative rules, university ethics and prescribed public-interest research. No automatic NGO or company data right.", "x": 1713.0, "y": 1464.6, "degree": 5 }, { "id": "w10-duty-third-sector-request", "label": "Reset third-sector research-access request", "group": "dutycare", "role": "Policy proposal", "summary": "2024 submission sought vetted access including third-sector organizations. Independent NGO eligibility is broader than draft205B university requirement; collaboration remains possible.", "x": 1637.1, "y": 1429.6, "degree": 2 }, { "id": "w11-fund-luminate", "label": "Luminate — funding label in Carnegie accounts", "group": "funding", "role": "Organisation", "summary": "Accounts record Luminate partnership income; narrative says Luminate/Reset. Contracting legal entity unresolved; not two separate grants or automatic identity with Luminate Foundation Inc.", "x": 1608.0, "y": 1209.0, "degree": 2 }, { "id": "w11-fund-minderoo", "label": "Minderoo Foundation", "group": "funding", "role": "Organisation", "summary": "Named restricted project funder and Frontier Technology initiative parent in advice announcement. Exact grant instrument/individual decision rights not obtained.", "x": 1620.8, "y": 1203.7, "degree": 2 }, { "id": "w11-fund-carnegie-programme", "label": "Carnegie harm-regulation / online-safety programme", "group": "funding", "role": "Programme", "summary": "Reported restricted project income sustains duty-model work and later OSA completion. Earlier model work predates the2021 income inspected; no claim this money first invented it.", "x": 1643.8, "y": 1245.0, "degree": 5 }, { "id": "w11-fund-un-advice", "label": "Carnegie advice to UN minority-issues rapporteur", "group": "funding", "role": "Publication", "summary": "Identifiable international advice financially supported by Minderoo initiative per22November2021 announcement. No extra award amount or adopted UN recommendations established.", "x": 1614.2, "y": 1205.4, "degree": 2 }, { "id": "w11-fund-carnegie-board", "label": "Carnegie trustees — programme budget authority", "group": "funding", "role": "Governance role", "summary": "Trustees agree strategy/budgets, management executes; own endowment supplies capacity. No individual trustee vote, donor veto or grant-clearance record acquired.", "x": 1667.6, "y": 1310.8, "degree": 2 }, { "id": "w11-fund-mannifera", "label": "Mannifera", "group": "funding", "role": "Funding program", "summary": "Pooled funding programme with ACF sub-fund and member collective decisions. Named2024 grants; exact contracting payor, individual votes and recipient terms unresolved.", "x": 1546.0, "y": 1366.4, "degree": 7 }, { "id": "w11-fund-acf", "label": "Australian Communities Foundation", "group": "funding", "role": "Organisation", "summary": "Mannifera funds route and administrator described by guide and own report. No automatic assignment as legal payor of specific Reset/HRLC contracts.", "x": 1503.0, "y": 1336.9, "degree": 2 }, { "id": "w11-fund-hrlc", "label": "Human Rights Law Centre", "group": "funding", "role": "Organisation", "summary": "Mannifera records2024 grant for expert alignment and advocacy. A later grantor account expressly names HRLC and Digital Rights Watch as partners in Reset transparency advocacy; exact joint output, endorsement and adopted clause remain unresolved.", "x": 1608.8, "y": 1397.0, "degree": 7 }, { "id": "w11-fund-alignment-work", "label": "HRLC expert alignment and advocate training", "group": "funding", "role": "Funded purpose", "summary": "Grantor describes aligning legal/tech experts and training advocates for2024 misinformation, online-safety and privacy reforms. Purpose rather than verified completed common text.", "x": 1563.9, "y": 1386.5, "degree": 1 }, { "id": "w11-fund-reset-advocacy", "label": "Reset transparency-advocacy grant purpose", "group": "funding", "role": "Funded purpose", "summary": "2024 Mannifera grant supports transparency research/advocacy into policy and social-media inquiry, including OSA review. No specific2026 clause earmark.", "x": 1524.3, "y": 1380.0, "degree": 1 }, { "id": "w11-fund-mckinnon", "label": "McKinnon — named research commissioner", "group": "funding", "role": "Organisation", "summary": "Names two2023/24 Reset commissions. Present privacy policy identifies a group entity; historic contractual party and publication-clearance rights unresolved.", "x": 1418.4, "y": 1333.0, "degree": 6 }, { "id": "w11-fund-functioning", "label": "Functioning or Failing? —May2024", "group": "funding", "role": "Publication", "summary": "Commissioned misinformation-code research recommending consideration of a duty, EU-style assessments and researcher access. No bill commission or governmental adoption established.", "x": 1487.5, "y": 1362.9, "degree": 2 }, { "id": "w11-fund-transparency", "label": "Achieving Digital Platform Public Transparency —June2024", "group": "funding", "role": "Publication", "summary": "Commissioned comparison proposes Australian transparency/risk/audit/research mechanisms using EU DSA/UK OSA models. Not adopted legislation or proven donor-required conclusions.", "x": 1489.5, "y": 1359.3, "degree": 2 }, { "id": "w11-fund-isf", "label": "Internet Society Foundation", "group": "funding", "role": "Organisation", "summary": "2024 return EIN82-3285688. Distinct from Internet Society and unspecified global Reset. Accrual return and project award not cash evidence.", "x": 1522.2, "y": 1387.2, "degree": 2 }, { "id": "w11-fund-privacy-project", "label": "Children privacy expectations — phase-two research", "group": "funding", "role": "Funded project", "summary": "US$219,000 award record for8Jan2024-7July2025 privacy research in education technology/targeting. Separate purpose from duty-of-care advocacy; implementation expectations not completed findings.", "x": 1521.4, "y": 1393.1, "degree": 2 }, { "id": "w12-gov-reset-uk", "label": "Reset Tech UK Limited", "group": "governance", "role": "Company", "summary": "Company14982650, limited by guarantee without share capital. Listed among privacy-notice entities; no proof it alone operates/contracts for OSN. Whole-company accounts are not network budget.", "x": 1598.4, "y": 1441.2, "degree": 5 }, { "id": "w12-gov-uk-director-members", "label": "Reset UK director-member decision body", "group": "governance", "role": "Governance role", "summary": "2025 articles make directors the only members. Board management/appointments and revocable delegation; one vote each on written resolutions. No named donor reserve found.", "x": 1509.7, "y": 1446.8, "degree": 4 }, { "id": "w12-gov-scott", "label": "Dale Benjamin Scott", "group": "governance", "role": "Person", "summary": "Listed UK director/PSC;2023 sole-subscriber origin qualifies register categories. May2025 adoption agreement does not establish a surviving individual veto.", "x": 1538.0, "y": 1445.3, "degree": 2 }, { "id": "w12-gov-wood", "label": "Poppy Wood", "group": "governance", "role": "Person", "summary": "Listed UK director; PSC voting band notified7May2025. Constitution makes directors members; current register and basis remain to reconcile without summing bands.", "x": 1448.6, "y": 1446.2, "degree": 2 }, { "id": "w12-gov-uk-constitutional-limits", "label": "Reset UK entrenched purposes and conflict rules", "group": "governance", "role": "Constitutional rule", "summary": "Certain amendments require member unanimity. Conflicts require declaration but permit specified participation, including connected Group Company matters, subject to exceptions and exclusions. Not proof rules exercised.", "x": 1459.5, "y": 1451.7, "degree": 1 }, { "id": "w12-gov-reset-action-ggmbh", "label": "Reset Tech Action gGmbH — named debtor", "group": "governance", "role": "Organisation", "summary": "2024 UK accounts name this entity as owingGBP1,156. Do not silently identify its legal registration with similarly spelled Action GmbH in constitutional list; no parent control inferred.", "x": 1547.9, "y": 1442.0, "degree": 1 }, { "id": "w12-gov-mann-members", "label": "Mannifera collective decision process", "group": "governance", "role": "Governance role", "summary": "Historical guide: EO assesses, member panel shortlists, Taskforce endorses, members reach consensus at Muster.2023 timetable not proven execution of special2024 round.", "x": 1496.8, "y": 1340.9, "degree": 2 }, { "id": "w12-gov-myer", "label": "The Myer Foundation", "group": "governance", "role": "Organisation", "summary": "FY24 report discloses earlier three-year Mannifera membership award and representative/board role. Not each member paid the current minimum or an HRLC earmark.", "x": 1493.4, "y": 1345.8, "degree": 2 }, { "id": "w12-gov-rights-first", "label": "Rights-First —4October2024", "group": "governance", "role": "Publication", "summary": "David Mejia-Canales authored framework acknowledging Mannifera. Human-rights constraints accompany duty/regulator proposals. Author disclaimer prevents attribution to every supporter; exact grant allocation unknown.", "x": 1597.0, "y": 1396.9, "degree": 4 }, { "id": "w12-gov-mejia", "label": "David Mejia-Canales", "group": "governance", "role": "Person", "summary": "Rights-First author; HRLC reports parliamentary representation. No governmental adoption or donor approval inferred.", "x": 1556.5, "y": 1381.3, "degree": 2 }, { "id": "w12-gov-hrlc-engagement", "label": "HRLC parliamentary engagement — reportedOctober2024", "group": "governance", "role": "Reported event", "summary": "HRLC account reports Parliament House/government engagement. Displayed8October2024 differs fromURL2025 string; official minutes and exact adopted wording not obtained.", "x": 1568.4, "y": 1375.8, "degree": 2 }, { "id": "w12-gov-mckinnon-trust", "label": "Trustee for Susan McKinnon Foundation", "group": "governance", "role": "Trust", "summary": "ABN28 739100711 discretionary investment trust/private ancillary fund. Exact corporate trustee name unresolved; group finance not Reset-specific money.", "x": 1370.7, "y": 1311.1, "degree": 2 }, { "id": "w12-gov-research-centre", "label": "Former Susan McKinnon Research Centre Ltd", "group": "governance", "role": "Company", "summary": "Historical entity ABN/ACN40 656129127 /656129127; now OH-RULE FAMILY PHILANTHROPY LTD from14April2026. Same identifier, no assignment or wind-up inferred. Committee route not yet tied to Reset contracts.", "x": 1355.7, "y": 1305.9, "degree": 3 }, { "id": "w12-gov-charitable-foundation", "label": "Susan McKinnon Charitable Foundation Ltd", "group": "governance", "role": "Company", "summary": "ABN12 653756597, distinct from trust and Research Centre. Current group listing/existence during2023-24 does not identify the Reset contract signatory.", "x": 1366.6, "y": 1318.5, "degree": 1 }, { "id": "w12-gov-research-committee", "label": "Historical McKinnon Research Committee", "group": "governance", "role": "Governance role", "summary": "Research Centre FY2024 report places research-fund application and publishing/dissemination under committee oversight. Members, appointment rules, vote thresholds and Reset-specific exercise unresolved.", "x": 1310.5, "y": 1286.9, "degree": 1 }, { "id": "w12-gov-march-roundtable", "label": "Reset transparency roundtable —March2024", "group": "governance", "role": "Event", "summary": "Twenty academic/civil-society experts; Chatham House Rule. April briefing reports process. Attendance/contribution expressly not endorsement; roster not expanded here.", "x": 1502.2, "y": 1368.1, "degree": 2 }, { "id": "w12-gov-april-briefing", "label": "Regulating for Transparency —April2024", "group": "governance", "role": "Publication", "summary": "Briefing translates March discussion into recommendations and credits McKinnon support. Reset owns errors/omissions; no new independent commission/payment inferred.", "x": 1444.1, "y": 1346.2, "degree": 2 }, { "id": "w13-disp-osan-feb", "label": "OSAN February2024 illegal-harms response", "group": "disposition", "role": "Publication", "summary": "Dated response with civil-inference and search-risk proposals. Maeve Walsh is contact; not proof she wrote every passage.", "x": 1838.8, "y": 1373.2, "degree": 3 }, { "id": "w13-disp-civil-inference", "label": "Proposed contextual civil-inference approach", "group": "disposition", "role": "Policy proposal", "summary": "OSAN argues wider context and civil balance of probabilities. Cyberflashing example contrasts suicide-offence difficulties; not a general test that unbeneficial speech is illegal.", "x": 1866.6, "y": 1379.5, "degree": 3 }, { "id": "w13-disp-search-risk", "label": "Search risk and obligations mismatch concern", "group": "disposition", "role": "Policy proposal", "summary": "OSAN p97 raises direct and indirect routes and risk/measure fit. Ofcom reports a risk-register amendment while retaining statutory search-content limits.", "x": 1908.7, "y": 1378.8, "degree": 2 }, { "id": "w13-disp-osan-july", "label": "OSAN July2024 children response", "group": "disposition", "role": "Publication", "summary": "Names Woods and Maeve Walsh. Issue8 criticises age gates without service design change, seeks evidence/metrics and supports not selecting one tool; exact accepted minimum-age clause unestablished.", "x": 1757.0, "y": 1342.4, "degree": 3 }, { "id": "w13-disp-maeve-walsh", "label": "Maeve Walsh", "group": "disposition", "role": "Person", "summary": "Named co-author of OSAN July2024 children response and contact on February response. No formal Ofcom appointment established in this bounded pass; unrelated Siobhan Walsh is not this person.", "x": 1742.2, "y": 1307.5, "degree": 1 }, { "id": "w13-disp-child-risk-ages", "label": "Children risk-assessment age assumptions", "group": "disposition", "role": "Guidance", "summary": "April2025 clarification: terms-of-service minimums/self-declaration insufficient evidence. Without HEAA enforcement, conservatively assess likely younger users. Not itself a universal age gate or ID-document mandate.", "x": 1867.0, "y": 1388.6, "degree": 2 }, { "id": "w13-disp-hrlc155", "label": "HRLC July2024 review submission155", "group": "disposition", "role": "Publication", "summary": "Expressly recommends Reset model and uses WEF2023 typology; requests research and civil-society access. Submitted advocacy is distinct from government adoption.", "x": 1717.7, "y": 1434.4, "degree": 4 }, { "id": "w13-disp-reset-five-pillars", "label": "Reset Five Pillars regulatory model", "group": "disposition", "role": "Policy proposal", "summary": "April2024 Green Paper proposes duty, assessment, mitigation, transparency and accountability across broader harms. Cites EU/UK/Canadian comparators and frames proposals as contribution to debate. HRLC expressly endorses it; no evidence HRLC mediated reviewer citation of separate Reset submission70.", "x": 1641.4, "y": 1412.8, "degree": 3 }, { "id": "w13-disp-wef-typology", "label": "WEF2023 Typology of Online Harms", "group": "disposition", "role": "Publication", "summary": "August2023 WEF coalition typology: contextual common language, not severity rating or regulatory compliance. Foreword/introduction scoped original text read; complete matrix/credits remain uninspected. Separately cited byApril2024 official issues paper, adopted byJuly HRLC advocacy and used in current eSafety education; none alone is statutory incorporation.", "x": 1855.7, "y": 1479.9, "degree": 12 }, { "id": "w13-disp-au-commitment", "label": "Government duty-of-care commitment —November2024", "group": "disposition", "role": "Event", "summary": "Rowland13November speech grounds direction in received review and international alignment;14November release follows. Does not identify an NGO/funder as decision-maker.", "x": 1654.2, "y": 1362.5, "degree": 1 }, { "id": "w13-disp-au-briefing", "label": "Department briefing cleared26November2024", "group": "disposition", "role": "Government record", "summary": "FOI Document2 media table summarises HRLC and Reset public advocacy. Post-announcement official awareness, not proof of minister reading before decision.", "x": 1564.1, "y": 1396.7, "degree": 2 }, { "id": "w13-disp-au-response", "label": "Government review response —April2026", "group": "disposition", "role": "Government record", "summary": "Published recommendation-by-recommendation support, qualification and non-support. Support of policy architecture is distinct from enacted law or adoption of every NGO proposal.", "x": 1674.4, "y": 1391.4, "degree": 12 }, { "id": "w13-disp-rights-safeguards", "label": "Rights-First reasons, remedies and review", "group": "disposition", "role": "Policy proposal", "summary": "Report seeks reasons for content/account restrictions, accessible remedies and review of significant regulatory decisions. Safeguards accompany its proposed regulatory powers.", "x": 1672.3, "y": 1432.7, "degree": 2 }, { "id": "w13-disp-hrlc-civil-research", "label": "HRLC civil-society research-access request", "group": "disposition", "role": "Policy proposal", "summary": "Submission155 p11 expressly includes vetted researchers and civil society organizations. September draft205B requires Australian-university employment plus prospective approvals.", "x": 1688.6, "y": 1440.7, "degree": 2 }, { "id": "w14-tax-hildreth", "label": "Adam Hildreth", "group": "taxonomy", "role": "Person", "summary": "Resolver August2023 account describes Crisp founder and co-chair of typology workstream. Foreword signatory; exact2023 legal company/contract and commercial-control rights not established here.", "x": 1909.6, "y": 1394.5, "degree": 3 }, { "id": "w14-tax-hanna", "label": "John-Orr Hanna", "group": "taxonomy", "role": "Person", "summary": "Current SWGfL biography reports drafting-committee participation in2023 typology. Retrospective institutional account, not appointment instrument or final editorial authority.", "x": 1874.6, "y": 1461.9, "degree": 1 }, { "id": "w14-tax-dobrygowski", "label": "Daniel Dobrygowski", "group": "taxonomy", "role": "Person", "summary": "Named WEF-affiliated foreword signatory on indexed originalp3. Foreword credit does not establish main-text authorship or a veto.", "x": 1855.5, "y": 1459.0, "degree": 1 }, { "id": "w14-tax-bantourakis", "label": "Minos Bantourakis", "group": "taxonomy", "role": "Person", "summary": "Named WEF-affiliated foreword signatory and launch-article coauthor. No final approval or government drafting right inferred.", "x": 1878.6, "y": 1469.0, "degree": 1 }, { "id": "w14-tax-core4cs", "label": "CO:RE4Cs research — attributed antecedent", "group": "taxonomy", "role": "Publication", "summary": "Resolver links2021 Livingstone et al research as typology antecedent. Earlier original blocked/unread; no category-by-category copying or complete authorship/acknowledgment audit.", "x": 1864.0, "y": 1465.0, "degree": 1 }, { "id": "w14-tax-issues2024", "label": "Australian review issues paper —April2024", "group": "taxonomy", "role": "Publication", "summary": "Government discussion uses WEF typology atfn17,88,93 before HRLC July submission. Descriptive framing, not law or proof officials were independent of all advocates.", "x": 1871.8, "y": 1475.2, "degree": 1 }, { "id": "w14-tax-esafety-harms-guidance", "label": "eSafety harm categories — educational guidance", "group": "taxonomy", "role": "Guidance", "summary": "Current Safety by Design page uses WEF typology for overlapping categories. Last-update display does not date initial adoption; industry education, not binding duty definitions.", "x": 1857.4, "y": 1529.5, "degree": 2 }, { "id": "w14-tax-review-categories", "label": "Review recommendation5 harm families", "group": "taxonomy", "role": "Policy proposal", "summary": "Published updated reviewp56 proposes harm families and defers broader financial/societal/misinformation consolidation to other regimes or future review. Government supportsrec5 in principle.", "x": 1657.8, "y": 1375.2, "degree": 2 }, { "id": "w14-tax-deferred-consolidation", "label": "Broader harm-regime consolidation deferred", "group": "taxonomy", "role": "Policy proposal", "summary": "Reviewer considers broader harms compatible with duty model but within other existing/proposed regimes; possible future expanded commission. Deliberate scope choice, not rejection of WEF taxonomy or claim harms unreal.", "x": 1660.3, "y": 1358.3, "degree": 1 }, { "id": "w14-tax-draft-categories", "label": "Draft initial categories and ministerial additions", "group": "taxonomy", "role": "Policy proposal", "summary": "Sections25B-D separate serious-harm, child andunder16 feature obligations. Addition thresholds differ for serious harm and child harm;25G feature additions lack express serious-harm threshold. Not enacted taxonomy incorporation.", "x": 1735.3, "y": 1441.5, "degree": 3 }, { "id": "w14-tax-draft-risk-scope", "label": "Draft foreseeable, combined and non-user risks", "group": "taxonomy", "role": "Policy proposal", "summary": "Sections25E/26A extend assessment beyond named labels to reasonably foreseeable risks and combined/non-user harms. Absence of a general misinformation label does not settle all future scope.", "x": 1725.5, "y": 1450.6, "degree": 1 }, { "id": "w14-tax-act2021", "label": "As-made2021 Australian harm tests", "group": "taxonomy", "role": "Law", "summary": "Earlier Act already contains serious harm, targeted child/adult abuse, contextual merit/consent and classification tests. A competing antecedent to similarities; not proof each later clause copied it.", "x": 1709.9, "y": 1414.0, "degree": 1 }, { "id": "w14-tax-review-version-brief", "label": "FOI review-version record — inconsistencies retained", "group": "taxonomy", "role": "Government record", "summary": "PostOctober delivery revision brief,8November2024 departmental clearance. AttachmentC says68 recommendations, later response67; daily-action penalty and removal-notice penalty descriptions differ. No reconstructed total/deletion or signed minister acceptance.", "x": 1647.2, "y": 1366.3, "degree": 1 }, { "id": "w15-impl-crisp-brand", "label": "Crisp commercial brand", "group": "implementation", "role": "Commercial brand", "summary": "Current policy identifies GroupLimited trading as Crisp;2024 announcement aligns brand under Resolver. Do not substitute this brand for each historical contracting entity.", "x": 1950.2, "y": 1310.3, "degree": 5 }, { "id": "w15-impl-crisp-group", "label": "Crisp Thinking Group Limited", "group": "implementation", "role": "Company", "summary": "Company05636614; current policy controller trading as Crisp. Register identifies changing Kroll corporate controllers. Not automatically every historic contract signer.", "x": 1898.0, "y": 1258.4, "degree": 4 }, { "id": "w15-impl-crisp-uk", "label": "Crisp Thinking(UK) Limited", "group": "implementation", "role": "Company", "summary": "Company06442193. Registry identifies Group05636614 share control. DSIT supplier name matched across sources; payment CSV itself lacks company number.", "x": 1795.4, "y": 1261.0, "degree": 2 }, { "id": "w15-impl-kroll-international", "label": "Kroll International(UK) Limited", "group": "implementation", "role": "Company", "summary": "Company09897805; historical registered CrispGroup controller, notified25May2022 and ceased29December2022. Not generic Kroll ultimate parent.", "x": 1925.4, "y": 1221.1, "degree": 1 }, { "id": "w15-impl-kroll-advisory", "label": "Kroll Advisory Ltd.", "group": "implementation", "role": "Company", "summary": "Company05568550; registered current CrispGroup controller notified29December2022. Registry75%+ shares/votes does not establish exact percentage or ultimate ownership.", "x": 1932.7, "y": 1225.3, "degree": 1 }, { "id": "w15-impl-resolver-brand", "label": "Resolver / Trust&Safety commercial brand", "group": "implementation", "role": "Commercial brand", "summary": "Crisp brand alignment announcedMarch2024; later compliance partnership reported. Brand is distinct from ResolverInc and exact supplier entities.", "x": 2052.7, "y": 1318.7, "degree": 3 }, { "id": "w15-impl-resolver-inc", "label": "Resolver Inc.", "group": "implementation", "role": "Company", "summary": "Named provider with applicable affiliates in currently served standard terms v3.0. Exact government engagement/signatory and incorporation jurisdiction unestablished.", "x": 2096.9, "y": 1286.6, "degree": 1 }, { "id": "w15-impl-customer-instructions", "label": "Customer-specified moderation instructions", "group": "implementation", "role": "Operational policy", "summary": "Supplier policy describes customer instructions for automatic removal, expert escalation and challenge; customer-channel processor role differs from crisis-index controller. Actual customer configuration unobserved.", "x": 1983.0, "y": 1275.4, "degree": 1 }, { "id": "w15-impl-oscar", "label": "OSCAR compliance workflow tool", "group": "implementation", "role": "Product", "summary": "Illuminate tool translating requirements into compliance workflows; supplier partnership announcement, not a mandated product or awarded procurement.", "x": 2157.8, "y": 1341.4, "degree": 1 }, { "id": "w15-impl-dsit", "label": "UK Department for Science,Innovation and Technology", "group": "implementation", "role": "Government department", "summary": "Publisher of2025 supplier payment under Security&OnlineHarms/NSOIT&TechProgramme. Operational commission and outputs require their own evidence.", "x": 1690.1, "y": 1268.0, "degree": 11 }, { "id": "w15-impl-licensing45", "label": "Review45: consider licensing major services", "group": "implementation", "role": "Policy proposal", "summary": "Published review147 proposes possible condition of operation, with enforcement/newsbargaining/ombuds/cost uses and jurisdiction/design difficulties. Government explicitly did not support it.", "x": 1651.8, "y": 1380.4, "degree": 2 }, { "id": "w15-impl-identification48", "label": "Review48: account identification and preservation", "group": "implementation", "role": "Policy proposal", "summary": "Review proposes phone-number collection at account opening, preventing notice tip-offs and compelled preservation. Response supports in principle. Express implementation of these elements not located in inspected duty-draft amendments; not an enacted universal-ID finding.", "x": 1710.2, "y": 1420.7, "degree": 3 }, { "id": "w15-impl-nsoit", "label": "National Security Online Information Team", "group": "implementation", "role": "Government unit", "summary": "DSIT operational unit: public-source monitoring and platform terms referrals within ministerial remit. Draft detailed instructions, current notice and actual testimony have distinct evidence states.", "x": 1757.9, "y": 1226.2, "degree": 8 }, { "id": "w15-impl-dsit-ministers", "label": "DSIT ministers — scope-setting role", "group": "implementation", "role": "Government role", "summary": "Public answers say ministers agree and review high-risk remit. A role over task selection, not evidence a named minister approved each referral.", "x": 1741.3, "y": 1188.3, "degree": 1 }, { "id": "w15-impl-nsoit-protocol", "label": "NSOIT delivery and escalation instructions — DRAFT", "group": "implementation", "role": "Draft policy", "summary": "Requires agreed sources/search terms/task scope and legal/privacy assurance; sustained monitoring needs ministerial agreement. Sensitive-origin exclusions and source limits apply in draft. Exact escalation approvals partly redacted; adoption/applicability unestablished.", "x": 1727.9, "y": 1170.8, "degree": 3 }, { "id": "w15-impl-nsoit-partners", "label": "NSOIT external delivery partners — role", "group": "implementation", "role": "Contractor role", "summary": "Draft describes processors following DSIT documented instructions; actual executed order/SOW and relation toFebruary2025 payment remain missing.", "x": 1704.3, "y": 1128.7, "degree": 1 }, { "id": "w15-impl-platform-review", "label": "Platforms reviewing NSOIT referrals", "group": "implementation", "role": "Operational role", "summary": "Department describes trusted channels, warnings and review under each platform own terms. Platforms retain stated action choice; actual referral outcomes and pressure not resolved.", "x": 1733.8, "y": 1184.6, "degree": 2 }, { "id": "w15-impl-crisp-report2024", "label": "Crisp report requested for week beginning24June2024", "group": "implementation", "role": "Document", "summary": "ICO independently confirms one held report and partial disclosure24October2025. Underlying report unread, exact date/contractor entity unspecified. Department later calls electionweek; do not replace request period or join to2025 invoice.", "x": 1878.8, "y": 1270.4, "degree": 3 }, { "id": "w15-impl-gsn", "label": "The Global Strategy Network Ltd", "group": "implementation", "role": "Company", "summary": "Official register identifies active company10433562, incorporated18October2016. Current notice names analysis contractor;27January2026 answer distinguishes GSN analysis fromFaculty platform development. OneSeptember2025 payment is recorded. Exact appointment,contract scope andCrisp succession/parallel provision remain unresolved.", "x": 1693.3, "y": 1207.4, "degree": 5 }, { "id": "w15-impl-science-committee", "label": "Commons Science,Innovation and Technology Committee", "group": "implementation", "role": "Parliamentary committee", "summary": "Recommended statutory footing and Intelligence and Security Committee oversight for NSOIT in2025. Government reply offers existing channels without commitment to requested change.", "x": 1712.3, "y": 1230.8, "degree": 2 }, { "id": "w16-supp-barrett", "label": "Giles Henry Donne Barrett", "group": "platform", "role": "Person", "summary": "Registered GSNPSC, notified24June2021:more than25% through50% shares/votes. Exact holding and family relationship not inferred.", "x": 1669.0, "y": 1166.5, "degree": 1 }, { "id": "w16-supp-king", "label": "Matthew Thomas King", "group": "platform", "role": "Person", "summary": "Registered GSNPSC, notified28October2020:more than25% through50% shares/votes. Bands are not exact half-shares or a separate operating-decision mandate.", "x": 1661.1, "y": 1169.4, "degree": 1 }, { "id": "w16-supp-faculty", "label": "Faculty Science Limited", "group": "platform", "role": "Company", "summary": "Supplier registration08873131 in disclosed order102548. January2026 answer namesFacultyScience asCDDP developer. Later inaccessible indexed registration mismatch not merged or promoted.", "x": 1584.4, "y": 1335.1, "degree": 7 }, { "id": "w16-supp-cddp", "label": "Counter-Disinformation Data Platform", "group": "platform", "role": "Government technical platform", "summary": "Official answer identifiesFaculty development; tender allocates technical priorities to government. Business case describes cross-government information sharing. Actual effectiveness,query choices andreferral outcomes unestablished.", "x": 1554.5, "y": 1349.6, "degree": 6 }, { "id": "w16-supp-procurement26087", "label": "CDDP procurement26087 — requirements", "group": "platform", "role": "Procurement record", "summary": "September2023 opportunity specifies DSIT-directed development,hosting,maintenance. IncumbentFaculty,indicative budget andplanned assessment do not establish award,actual delivery or replacement.", "x": 1539.2, "y": 1338.3, "degree": 2 }, { "id": "w16-supp-faculty102548", "label": "Faculty102548 variationNo2 — disclosed copy", "group": "platform", "role": "Contract variation", "summary": "Dated16August2023/effective17March2023;expiry31March2024;capGBP510000excludingVAT;printed totalGBP2373873. Visible signature fields uncompleted without black redaction; execution unestablished. AppendixC variesIPallocation.", "x": 1628.6, "y": 1297.7, "degree": 3 }, { "id": "w16-supp-licensed-ip", "label": "CDDP retained licensed IP", "group": "platform", "role": "Contractual resource", "summary": "AppendixC changes arrangements: newIP/specially-written software shall be supplier-owned,not assigned;buyer licence perpetual. Permitted third-party service rights andsource access obligations are terms,not proof delivery or successful switching.", "x": 1629.4, "y": 1290.6, "degree": 4 }, { "id": "w16-supp-authorised-maintainers", "label": "Buyer-authorised CDDP contractors", "group": "platform", "role": "Contractual role", "summary": "Variation permits third parties to maintain,host,develop/adapt forbuyer without directFaculty agreement. Other commercial exploitation/competing products restricted;actual appointment or repository access unestablished.", "x": 1644.9, "y": 1261.5, "degree": 2 }, { "id": "w16-supp-faculty102887", "label": "Faculty102887 variation trail — unreconciled totals", "group": "platform", "role": "Contract variation", "summary": "No1 dated13December2022 prints cap1514650 after1264650;No2 dated1March2023/effective16January adds400000excludingVAT butprints1944650. GBP30000discrepancy is real in visible records;intervening terms unknown,notfraud or a corrected total.", "x": 1519.4, "y": 1370.5, "degree": 2 }, { "id": "w16-supp-dcms", "label": "UK Department for Digital,Culture,Media and Sport — historical buyer", "group": "platform", "role": "Government department", "summary": "Named buyer in2021business case,2022orders andMarch2023Faculty variation. Distinct historical party fromDSIT named inAugust2023variation; no blanket contract novation inferred.", "x": 1479.7, "y": 1407.7, "degree": 7 }, { "id": "w16-supp-newswhip", "label": "NewsWhip Media Limited", "group": "platform", "role": "Company", "summary": "Irish registration485016 in disclosed contract103186 forrealtime social/predictive media monitoring intelligence. Charge unread; no establishedCDDP ingestion,deployment or exercised extension.", "x": 1422.0, "y": 1416.3, "degree": 1 }, { "id": "w22-sig-tc", "label": "Tech Coalition", "group": "signals", "role": "Organisation", "summary": "Programme administrator. Member-supported resources, admission/compliance; legal funding allocations and full participant agreement not obtained.", "x": 1952.5, "y": 1715.4, "degree": 12 }, { "id": "w22-sig-lantern", "label": "Lantern — child-safety signals", "group": "signals", "role": "Programme", "summary": "Tech Coalition OCSEA investigation programme; distinct from political vendor w 3-opp-lantern. Participants review and decide under own policies; not a universal restriction list.", "x": 1872.8, "y": 1673.4, "degree": 12 }, { "id": "w22-sig-participants", "label": "Lantern participating companies", "group": "signals", "role": "Decision-making role", "summary": "Collective participant role, not a legal person or proof identical upload rights. Financial institutions currently receive only.", "x": 1882.5, "y": 1709.8, "degree": 4 }, { "id": "w22-sig-taxonomy", "label": "Lantern signal taxonomy", "group": "signals", "role": "Policy", "summary": "Investigation labels may cover platform-policy violations without legal determination. Supplemental tags cannot stand alone. Steam tag is not membership or sanctions evidence.", "x": 1905.2, "y": 1708.5, "degree": 3 }, { "id": "w22-sig-pilot", "label": "Lantern financial pilot", "group": "signals", "role": "Programme", "summary": "August 2024–first half 2025 pilot; six named participants,1935 signals and 108 investigations reported, some ongoing. Full pilot report offered to members/participants, not acquired.", "x": 1801.6, "y": 1614.5, "degree": 3 }, { "id": "w22-sig-financial-eligibility", "label": "Worldwide financial eligibility — receive only", "group": "signals", "role": "Policy", "summary": "14 May 2026 announcement extends eligibility to individually evaluated institutions; not universal bank enrolment or financial uploads.", "x": 1867.1, "y": 1710.6, "degree": 1 }, { "id": "w22-sig-noncontinuation", "label": "Reported participant non-continuation", "group": "signals", "role": "Event", "summary": "2024 report says unnamed participant unable to meet compliance would not continue in 2025. Identity and executed revocation date unknown.", "x": 1957.4, "y": 1751.3, "degree": 1 }, { "id": "w22-sig-block", "label": "Block", "group": "signals", "role": "Organisation", "summary": "Company publication identifies Tech Coalition membership and Cash App participation. Case outcomes remain attributed; no adjacent donor allocation inferred.", "x": 1813.7, "y": 1759.2, "degree": 27 }, { "id": "w22-sig-cashapp", "label": "Cash App", "group": "signals", "role": "Service", "summary": "Block financial service participating in Lantern; receiving-role pilot and reported 18-account case separately evidenced.", "x": 1787.9, "y": 1675.4, "degree": 5 }, { "id": "w22-sig-case2025", "label": "Reported Meta → financial/Xbox case", "group": "signals", "role": "Event", "summary": "Operator report: police request→Meta investigation→signals→Block 18 Cash App closures and Microsoft Xbox action. No underlying case file read; exact dates unknown.", "x": 1891.2, "y": 1617.5, "degree": 3 }, { "id": "w22-sig-bsr", "label": "BSR", "group": "signals", "role": "Organisation", "summary": "Author of commissioned 2023 HRIA. Assessed risks and commitments are not an audit of later outcomes.", "x": 1971.9, "y": 1755.1, "degree": 3 }, { "id": "w22-sig-hria2023", "label": "Lantern human-rights assessment (2023)", "group": "signals", "role": "Publication", "summary": "November 2023 report from April–August work; shared-input/fragmented-remedy risks. Predates later financial integration. Scoped reader only.", "x": 1963.5, "y": 1755.9, "degree": 2 }, { "id": "w22-sig-checkstep", "label": "Checkstep", "group": "signals", "role": "Organisation", "summary": "Moderation workflow provider; public specification/terms describe customer-specific decisions and training permission. No particular executed contract/deployment observed.", "x": 2132.0, "y": 2503.6, "degree": 6 }, { "id": "w22-sig-sightengine", "label": "Sightengine (Kozelo SAS)", "group": "signals", "role": "Organisation", "summary": "Terms identify service owned/operated by Kozelo SAS, French SAS; classifiers in coauthored workflow. No corporate-control register investigation.", "x": 2123.4, "y": 2487.0, "degree": 4 }, { "id": "w22-sig-customer-workflow", "label": "Customer moderation workflow", "group": "signals", "role": "Programme", "summary": "Documented integration: scores feed configured policy and decisions for a customer. Capability, not shared cross-platform restriction list or observed deployment.", "x": 2145.1, "y": 2517.5, "degree": 2 }, { "id": "w22-sig-customer-role", "label": "Customer platform / subscriber role", "group": "signals", "role": "Decision-making role", "summary": "Abstract contractual/workflow role, not identified customer. Chooses settings and implements webhook handler; actual accepted terms unknown.", "x": 2137.0, "y": 2499.1, "degree": 6 }, { "id": "w22-sig-checkstep-training", "label": "Checkstep customer-data training permission", "group": "signals", "role": "Policy", "summary": "Terms 5.5 permit service ML use with email opt-out. Permission, customer acceptance and actual use remain separate.", "x": 2137.4, "y": 2522.8, "degree": 2 }, { "id": "w22-sig-feedback", "label": "Voluntary Sightengine feedback data", "group": "signals", "role": "Research resource", "summary": "Voluntarily submitted raw media, optionally labelled. Separate from ordinary scanning and general suggestions/corrections; special licence/controller/retention terms.", "x": 2108.6, "y": 2496.4, "degree": 2 }, { "id": "w23-cash-terms", "label": "Current Cash App US terms", "group": "cashapp-authority", "role": "Policy", "summary": "Current agreement with Block: displayed effective 7 February 2024, updated 26 August 2026. Gray annotations non-operative; no inference these bytes governed the earlier Lantern case.", "x": 1732.7, "y": 1729.6, "degree": 2 }, { "id": "w23-cash-aup", "label": "Cash App acceptable-use policy", "group": "cashapp-authority", "role": "Policy", "summary": "Incorporated current policy; criminal activity, adult-content sales and risky-business restrictions are distinct. Adoption date and specific external instruction unknown.", "x": 1661.5, "y": 1692.5, "degree": 2 }, { "id": "w23-cash-external-policies", "label": "Referenced bank/card-network policies", "group": "cashapp-authority", "role": "Policy input", "summary": "Generic external-rule category referenced by Cash App AUP; no identified policy version, bank instruction, ownership chain or individual closure direction acquired.", "x": 1612.8, "y": 1653.8, "degree": 1 }, { "id": "w23-cash-help", "label": "Cash App closure-help procedure", "group": "cashapp-authority", "role": "Publication", "summary": "Undated provider guidance: contact Support, eligibility screen, then appropriate-team review; up to 10 days, not expressly business days. Not observed performance or complete legal duties.", "x": 1766.5, "y": 1704.8, "degree": 1 }, { "id": "w23-cash-customers", "label": "Cash App US customer accounts", "group": "cashapp-authority", "role": "Decision-making role", "summary": "General class under current terms and 2025 order, not the particular 18 accounts in Lantern report. Account access, balance withdrawal and transfer-error remedies are distinct.", "x": 1793.9, "y": 1721.4, "degree": 4 }, { "id": "w23-cash-cfpb", "label": "Consumer Financial Protection Bureau", "group": "cashapp-authority", "role": "Organisation", "summary": "US agency issuing and enforcing 2025-CFPB-0001. Order, findings, response and implementation evidence remain separate.", "x": 1805.2, "y": 1798.6, "degree": 4 }, { "id": "w23-cash-cfpb-order", "label": "CFPB Block consent order — 16 January 2025", "group": "cashapp-authority", "role": "Policy", "summary": "Final order, with Block consenting without factual/legal admissions except jurisdiction. Imposes anti-fraud controls, restriction remedies, redress and monitoring; conditional duration/modification provisions.", "x": 1783.6, "y": 1800.8, "degree": 5 }, { "id": "w23-cash-restriction-procedure", "label": "Required restriction notice and appeal", "group": "cashapp-authority", "role": "Policy", "summary": "Order 108(c)–(d):24 h written notice with specified exception and 10 business-day decision on appeal/change status. Not guaranteed restoration or the transfer-error document-access rule.", "x": 1794.0, "y": 1792.9, "degree": 2 }, { "id": "w23-cash-error-procedure", "label": "Required transfer-error investigation", "group": "cashapp-authority", "role": "Policy", "summary": "Defined Notice of Error process includes investigation, provisional credit, findings and evidence access. Not a generic right to all evidence behind every account closure.", "x": 1753.2, "y": 1831.0, "degree": 1 }, { "id": "w23-cash-compliance-plan", "label": "Required CFPB Compliance Plan", "group": "cashapp-authority", "role": "Required record", "summary": "Create and implement within 120 days; furnish to CFPB on request. Actual plan/implementation not acquired; not the separate prior non-objection Redress Plan.", "x": 1807.5, "y": 1808.5, "degree": 3 }, { "id": "w23-cash-redress-plan", "label": "Required CFPB Redress Plan and allocation", "group": "cashapp-authority", "role": "Required record", "summary": "Director review/non-objection;75 million reserve obligation and 120 million consumer-redress cap. Distribution report not obtained; no actual consumer total established.", "x": 1776.4, "y": 1816.8, "degree": 3 }, { "id": "w23-cash-cfpb-enforcement", "label": "CFPB Enforcement Director role", "group": "cashapp-authority", "role": "Decision-making role", "summary": "Order-defined Director/delegate with redress non-objection/revision and specified monitoring/modification duties; not an identified individual adjudicating every appeal.", "x": 1761.3, "y": 1856.8, "degree": 1 }, { "id": "w23-cash-block-board", "label": "Block Board compliance responsibility", "group": "cashapp-authority", "role": "Governing body", "summary": "Board has ultimate responsibility under both orders; specified review/approval/reporting duties. No inference particular directors exercised every required action.", "x": 1804.5, "y": 1840.0, "degree": 3 }, { "id": "w23-cash-working-groups", "label": "Required Cash App risk/support working groups", "group": "cashapp-authority", "role": "Mandated committee", "summary": "CFPB109 committee includes risk, compliance, business, customer-operations, data-science roles; customer-support, fraud/security, EFTA groups. Actual meeting/resource records not obtained.", "x": 1819.9, "y": 1796.8, "degree": 1 }, { "id": "w23-cash-state-regulators", "label": "Participating state money-transmission regulators", "group": "cashapp-authority", "role": "Collective role", "summary": "Authorities in multistate order. Maine Appendix A lists 49 jurisdictions includingDC;48 Appendix B allocations. Not the eight examination states, seven enforcement leads or five-member executive committee.", "x": 1818.0, "y": 1825.0, "degree": 3 }, { "id": "w23-cash-state-order", "label": "Multistate Block BSA/AML order — January 2025", "group": "cashapp-authority", "role": "Policy", "summary": "Effective 15 January 2025, fuller Maine copy includes counterparts/appendices. Underlying ROE not attached; no admission of violations, implemented controls or payment receipt inferred.", "x": 1822.3, "y": 1867.2, "degree": 1 }, { "id": "w23-cash-state-exec", "label": "State order Executive Committee", "group": "cashapp-authority", "role": "Committee", "summary": "Initial Arkansas, California, Massachusetts, Texas, Washington representatives; receives reports, questions consultant and accesses workpapers. Not express approver of every CAP or replacement consultant.", "x": 1811.0, "y": 1852.5, "degree": 3 }, { "id": "w23-cash-company-committee", "label": "Block Compliance Management Committee", "group": "cashapp-authority", "role": "Mandated committee", "summary": "Company-appointed committee under state order, broad relevant-division representation; reporting/quarterly meetings. Existence of mandate does not prove actual meetings or dissolution.", "x": 1828.0, "y": 1797.6, "degree": 1 }, { "id": "w23-cash-consultant", "label": "Mandated independent AML consultant", "group": "cashapp-authority", "role": "Mandated role", "summary": "Unidentified qualified independent third-party role: assessment, priority, validation, workpapers. No particular firm appointed or contracted is established.", "x": 1808.5, "y": 1822.0, "degree": 4 }, { "id": "w23-cash-state-cap", "label": "Required AML Corrective Action Plan", "group": "cashapp-authority", "role": "Required record", "summary": "Board-approved plan after consultant identifies needed corrections; priorities incorporated, state committee receives. Actual CAP, acceptance, extension or validation not obtained.", "x": 1802.7, "y": 1859.0, "degree": 3 }, { "id": "w24-ag-july-resolution", "label": "July 2026 Block / state-AG agreed resolution", "group": "state-coordination", "role": "Agreed instrument", "summary": "Published state-specific texts: Oregon party signatures observed; California expressly proposed; court entry unverified. Stipulated effective 8 July 2026. State pool is USD 45m, with NAAG allocation inside it; no payment receipt.", "x": 1783.8, "y": 1806.8, "degree": 6 }, { "id": "w24-ag-july-ags", "label": "46 participating July Block-settlement AGs", "group": "state-coordination", "role": "Collective role", "summary": "Announced coalition co-led by Oregon and Texas. Separate state instruments incorporate substantive terms, not one document bearing all 46 signatures. Distinct from January 2025 money-transmission regulators.", "x": 1800.4, "y": 1793.5, "degree": 2 }, { "id": "w24-ag-july-exec", "label": "July settlement Multistate AG Executive Committee", "group": "state-coordination", "role": "Committee", "summary": "Oregon, Texas, Colorado, Connecticut, Florida, Indiana, Vermont and Washington. Specific fallback-plan approval/reporting rights; internal vote rule unknown. Distinct from NAAG and January state-order committees.", "x": 1724.6, "y": 1874.5, "degree": 1 }, { "id": "w24-ag-july-redress", "label": "July agreed redress credit and 2030 fallback", "group": "state-coordination", "role": "Conditional plan", "summary": "Federal-order redress payments satisfy this agreement. Conditional fallback plan is proposed by 18 March 2030 for eight-state committee approval before payment; not a second additive USD 75–120m award.", "x": 1751.7, "y": 1842.0, "degree": 3 }, { "id": "w24-ag-july-compliance", "label": "July required company compliance committee/report", "group": "state-coordination", "role": "Mandated role", "summary": "Agreed committee with quarterly Board reporting; annual report is committee-approved and available to state on request. No observed appointments, reports or monitor implementation.", "x": 1814.6, "y": 1810.1, "degree": 2 }, { "id": "w24-ag-naag", "label": "National Association of Attorneys General", "group": "state-coordination", "role": "Organisation", "summary": "Member resource/coordination association and specified July investigation-expense reimbursement payee. No receipt or allocation to another fund/campaign established. Distinct from Mission Foundation and state enforcement committees.", "x": 1733.2, "y": 1718.6, "degree": 12 }, { "id": "w24-ag-naag-exec", "label": "NAAG Executive Committee", "group": "state-coordination", "role": "Governing body", "summary": "Current page describes 13 AGs governing funds/property/management and employing executive director; dated older fund instruments assign additional specified approval rights. No universal authority over state cases.", "x": 1685.0, "y": 1772.0, "degree": 5 }, { "id": "w24-ag-naag-president", "label": "NAAG president role", "group": "state-coordination", "role": "Decision-making role", "summary": "Role appoints historical fund special committees subject to executive approval. No inference current incumbent made earlier decisions or named individual voted a grant.", "x": 1642.4, "y": 1816.3, "degree": 2 }, { "id": "w24-ag-naag-director", "label": "NAAG executive director role", "group": "state-coordination", "role": "Decision-making role", "summary": "Employed for day-to-day operations; specified historical fund applications on NAAG behalf may be signed by this role. No identity inferred from current roster for earlier events.", "x": 1657.5, "y": 1804.7, "degree": 1 }, { "id": "w24-ag-member-offices", "label": "NAAG member AG offices", "group": "state-coordination", "role": "Collective role", "summary": "Members receive described collaboration/services and may seek purpose-limited fund grants. Not every member is a signatory, grantee or recipient of a particular investigation grant.", "x": 1700.4, "y": 1750.1, "degree": 5 }, { "id": "w24-ag-consolidated", "label": "Current consolidated consumer/charity fund committee", "group": "state-coordination", "role": "Committee", "summary": "Seven-person current roster and enforcement/training remit, observed 9 September 2026. No amendment transfers every 2012/2016 special-committee rule or Cash App allocation to this body.", "x": 1709.9, "y": 1745.6, "degree": 1 }, { "id": "w24-ag-vw-fund", "label": "VW Settlement Fund — 2016 rules", "group": "state-coordination", "role": "Fund", "summary": "Purpose-limited segregated fund. Rules describe USD 20m settlement provision to NAAG, principal grants, conditional repayment and 1% fees. Current balance/application and actual receipts unknown.", "x": 1695.4, "y": 1762.7, "degree": 4 }, { "id": "w24-ag-vw-committee", "label": "VW fund three-AG special committee", "group": "state-coordination", "role": "Committee", "summary": "Under 2016 rules: presidential appointment with Executive Committee approval, no more than two same-party members, two votes for grants, applicant-office recusal/replacement. Present authority/amendments unknown.", "x": 1667.8, "y": 1794.7, "degree": 3 }, { "id": "w24-ag-vw-allocation", "label": "2016 VW partial-settlement allocation as described in rules", "group": "state-coordination", "role": "Historical instrument", "summary": "2016 rules recite a partial settlement providing USD 20m to NAAG for specified consumer-protection purposes and investigation reimbursement. Underlying payment records not acquired.", "x": 1713.3, "y": 1752.0, "degree": 1 }, { "id": "w24-ag-fsf", "label": "Financial Services and Consumer Protection Fund", "group": "state-coordination", "role": "Fund", "summary": "MOU effective 31 October 2012: purpose-limited segregated resources for investigation, experts and training. Five-judgment USD 15m wording preserved without inferring aggregate USD 75m received.", "x": 1689.4, "y": 1758.6, "degree": 3 }, { "id": "w24-ag-fsf-committee", "label": "Financial Services Fund five-AG special committee", "group": "state-coordination", "role": "Committee", "summary": "2012 MOU: annual terms, no more than three same-party members; three votes for grants at least USD 5,000, two cross-party votes below. Present authority/amendments unknown.", "x": 1660.9, "y": 1791.1, "degree": 3 }, { "id": "w24-ag-foundation", "label": "NAAG Mission Foundation Inc.", "group": "state-coordination", "role": "Organisation", "summary": "Separate IRS-return identity; 2024 Type I supporting-organization reporting names NAAG as sole supported body. Foundation support categories are not association-wide revenue or a complete control map.", "x": 1722.0, "y": 1753.7, "degree": 1 }, { "id": "w24-ag-procedure", "label": "NAAG Policy Letter Process — May 2026", "group": "state-coordination", "role": "Published procedure", "summary": "Four bipartisan sponsors, at least 10 business days, jurisdictional sign-on audit and recirculation on edits; 36-state threshold for NAAG letterhead/distribution. Not automatically the August 2025 procedure.", "x": 1699.6, "y": 1741.6, "degree": 3 }, { "id": "w24-ag-payment-letter", "label": "22 August 2025 NCII payment-platform letter", "group": "state-coordination", "role": "Advocacy instrument", "summary": "Asks six payment firms for controls and proactive seller-authorization removal using private rules. No compulsory deadline/order or established recipient response; NCII scope, not blanket pornography prohibition.", "x": 1783.7, "y": 1594.8, "degree": 11 }, { "id": "w24-ag-search-letter", "label": "22 August 2025 NCII search-platform letter", "group": "state-coordination", "role": "Advocacy instrument", "summary": "Separate body asks for search restrictions, warnings and redirects; Take It Down Act discussion belongs here. Not a proved company change or judicial disposition.", "x": 1716.5, "y": 1655.3, "degree": 1 }, { "id": "w24-ag-letter-coalition", "label": "47-AG NCII letter coalition", "group": "state-coordination", "role": "Collective role", "summary": "Dated coalition described by official participant account and combined signed letters. Collective endorsement does not mean every signatory authored the text or adopted Idaho subsequent threat framing.", "x": 1742.0, "y": 1631.0, "degree": 2 }, { "id": "w24-ag-clark", "label": "Charity Clark — Vermont AG, letter attribution", "group": "state-coordination", "role": "Person", "summary": "Pennsylvania official announcement attributes letter authorship to Vermont AG Charity Clark; original drafts/staff authorship not obtained. No claimed fund-control or reimbursement link.", "x": 1743.2, "y": 1606.3, "degree": 1 }, { "id": "w24-ag-labrador", "label": "Raul Labrador — Idaho AG, September 2025 statement", "group": "state-coordination", "role": "Person", "summary": "Signatory whose later official newsletter warns of legal consequences. His statement is not an order, company response or proof of every AG position.", "x": 1756.2, "y": 1615.0, "degree": 1 }, { "id": "w24-ag-amex", "label": "American Express Company", "group": "state-coordination", "role": "Organisation", "summary": "Payment-letter addressee through its chief legal officer. Requested action does not establish receipt, agreement or implementation.", "x": 1747.0, "y": 1612.6, "degree": 1 }, { "id": "w24-ag-apple", "label": "Apple Inc.", "group": "state-coordination", "role": "Organisation", "summary": "Payment letter addresses company counsel concerning Apple Pay. No policy change or reply tied to this letter is established.", "x": 1741.5, "y": 1583.0, "degree": 4 }, { "id": "w25-sf-city-attorney", "label": "San Francisco City Attorney’s Office", "group": "sf-enforcement", "role": "Public authority", "summary": "Prosecuting office in the California action; named recipient of Briver judgment payments and compliance information. Office assertions and court duties remain separate.", "x": 1864.6, "y": 1451.1, "degree": 13 }, { "id": "w25-sf-chiu", "label": "David Chiu — San Francisco City Attorney", "group": "sf-enforcement", "role": "Person", "summary": "Named sponsor of AB 621 in March 2025 committee analysis; public official bringing the cited litigation. No private role or motive inferred.", "x": 1884.8, "y": 1388.0, "degree": 2 }, { "id": "w25-sf-bauer-kahan", "label": "Rebecca Bauer-Kahan — AB 621 author", "group": "sf-enforcement", "role": "Person", "summary": "Bill author identified in the official March analysis; offered primary-purpose clarification. Exact subsequent adoption amendment not acquired.", "x": 1915.2, "y": 1293.3, "degree": 1 }, { "id": "w25-sf-civil-liberties-advocacy", "label": "California Civil Liberties Advocacy", "group": "sf-enforcement", "role": "Organisation", "summary": "Listed as opposing AB 621 unless amended in the18 March analysis; objections concern expression, federal law and service-provider burdens. No amendment victory attributed.", "x": 1924.2, "y": 1297.8, "degree": 1 }, { "id": "w25-sf-ab621", "label": "AB 621 / Civil Code 1708.86", "group": "sf-enforcement", "role": "Legislation", "summary": "Chapter 673 approved/filed 13 October 2025, effective 1 January 2026. Third-party liability and a conditional notice presumption have separate predicates and statutory exceptions.", "x": 1905.7, "y": 1342.7, "degree": 7 }, { "id": "w25-sf-notice-presumption", "label": "Service-specific evidence → conditional presumption", "group": "sf-enforcement", "role": "Legal mechanism", "summary": "CIV 1708.86(c)(2): sufficient evidence, qualifying sender, named service and enabling relationship, designated process if displayed, receipt and30 days; court-extension and statutory protections remain. No automatic liability judgment.", "x": 1931.4, "y": 1294.6, "degree": 1 }, { "id": "w25-sf-facilitator-class", "label": "Enabling-service providers — statutory class", "group": "sf-enforcement", "role": "Regulated class", "summary": "A role defined by actual enabling conduct and the legal predicates; not a finding that every app store, bank, host or AI service is liable.", "x": 1922.8, "y": 1291.2, "degree": 1 }, { "id": "w25-sf-prosecutor-role", "label": "Public prosecutors under 1708.86", "group": "sf-enforcement", "role": "Public authority class", "summary": "Includes AGs, city attorneys, county counsel, district attorneys and other city/county prosecutors. Courts determine remedies; no automatic assessed amount from a letter.", "x": 1934.3, "y": 1300.4, "degree": 1 }, { "id": "w25-sf-july-notices", "label": "SF Apple/Google notices — July 2026 reporting", "group": "sf-enforcement", "role": "Reported instrument", "summary": "Reported July 16 notices and confidential attachments were not acquired. Reported28-day response request is distinct from the statutory30-day receipt condition.", "x": 1828.3, "y": 1542.3, "degree": 6 }, { "id": "w25-sf-case", "label": "People v. Sol Ecom et al. — CGC-24-617237", "group": "sf-enforcement", "role": "Court case", "summary": "Multi-defendant California action. Briver’s settlement is company-specific; the Tilak declaration supports alternative service on Sol Ecom. No all-defendant judgment inferred.", "x": 1849.1, "y": 1398.9, "degree": 2 }, { "id": "w25-sf-superior-court", "label": "San Francisco Superior Court", "group": "sf-enforcement", "role": "Court", "summary": "Court issuing the signed, filed Briver judgment; continuing jurisdiction and noticed-motion conditions are documented. Separate docket query was not performed.", "x": 1668.3, "y": 1278.4, "degree": 2 }, { "id": "w25-sf-briver", "label": "Briver LLC", "group": "sf-enforcement", "role": "Organisation", "summary": "Stipulating defendant in the signed judgment. Defined associated-person scope is in the instrument; no merger with other defendants or service providers. Payment/compliance unverified.", "x": 1789.7, "y": 1379.0, "degree": 6 }, { "id": "w25-sf-sol-ecom", "label": "Sol Ecom, Inc. — captioned respondent", "group": "sf-enforcement", "role": "Organisation", "summary": "Captioned respondent and subject of Tilak’s alternative-service motion; no Briver settlement obligations assigned to it.", "x": 1838.2, "y": 1359.9, "degree": 1 }, { "id": "w25-sf-briver-judgment", "label": "Briver final judgment — May 2025", "group": "sf-enforcement", "role": "Court order", "summary": "Signed29 May; filed/served30 May 2025. Briver-specific negotiated obligations without proof or admissions; effective date is court-signature date. No separate register-of-actions acquired.", "x": 1715.8, "y": 1313.5, "degree": 6 }, { "id": "w25-sf-briver-reporting", "label": "Briver compliance reports and network disclosure", "group": "sf-enforcement", "role": "Required deliverable", "summary": "Paragraph16 information,14-day initial and annual attestations with domain evidence for five years. No actual response or attestation acquired.", "x": 1675.3, "y": 1274.1, "degree": 1 }, { "id": "w25-sf-announced-outcome", "label": "Ten sites offline or inaccessible in California — reported", "group": "sf-enforcement", "role": "Reported outcome", "summary": "City’s2 June 2025 account combines offline and California-inaccessibility outcomes and unnamed registrar/registry cooperation. Not an independent global shutdown audit or provider-specific instruction.", "x": 1847.2, "y": 1425.1, "degree": 1 }, { "id": "w25-sf-cloudflare", "label": "Cloudflare, Inc.", "group": "sf-enforcement", "role": "Organisation", "summary": "Named subpoena recipient in the filed declaration; responsive records are reported by counsel. No provider shutdown command or execution established.", "x": 1866.4, "y": 1422.3, "degree": 1 }, { "id": "w25-sf-namecheap", "label": "Namecheap, Inc.", "group": "sf-enforcement", "role": "Organisation", "summary": "Named registrar/subpoena recipient; counsel reports contact records for the two services. No termination inferred from disclosure.", "x": 1876.4, "y": 1425.7, "degree": 1 }, { "id": "w25-sf-namecom", "label": "Name.com", "group": "sf-enforcement", "role": "Organisation", "summary": "Named former registrar for ClothOff and subpoena recipient. Exact corporate suffix/identifier not established in inspected scope.", "x": 1856.9, "y": 1424.0, "degree": 1 }, { "id": "w25-sf-drawnudes", "label": "drawnudes.io — described service", "group": "sf-enforcement", "role": "Online service", "summary": "Service associated by city counsel with the Exhibit6 correspondent. No target-site visit, legal-operator identity resolution or eSafety clearance established.", "x": 1843.2, "y": 1493.5, "degree": 1 }, { "id": "w25-sf-clothoff", "label": "ClothOff — service correspondence", "group": "sf-enforcement", "role": "Online service", "summary": "Named in official eSafety correspondence and separately in the city declaration. No legal-operator merge with Drawnudes, Briver or Sol Ecom is inferred.", "x": 1837.3, "y": 1603.4, "degree": 1 }, { "id": "w25-sf-operator-claim", "label": "Drawnudes correspondent’s eSafety claim", "group": "sf-enforcement", "role": "Attributed claim", "summary": "25 December 2024 email claims requirements, implementation and confirmation tied to January 2025 legislation;4 January follow-up repeats a regulator list. No Australian notice or approval attached in this exhibit.", "x": 1858.1, "y": 1512.4, "degree": 3 }, { "id": "w25-sf-dis-standard", "label": "DIS Standard2024 — commenced 22 December", "group": "sf-enforcement", "role": "Legislation", "summary": "F2024L00710 made 19 June, registered 21 June 2024; explicit22 December commencement corroborated by contemporaneous agency correspondence. Mandatory, category-specific duties; no named-service approval.", "x": 1937.6, "y": 1642.5, "degree": 6 }, { "id": "w25-sf-high-impact-genai", "label": "Covered high-impact generative-AI DIS", "group": "sf-enforcement", "role": "Regulated class", "summary": "Statutory category with scope, classification and control qualifications. Not every AI service, all adult content, or a finding about a particular operator.", "x": 1941.8, "y": 1676.8, "degree": 2 }, { "id": "w26-ver-tat", "label": "Tech Against Terrorism", "group": "private-verification", "role": "Organisation", "summary": "Counterterrorism-support initiative implemented by Online Harms Foundation. Its public products, corporate implementer and external funders are distinct roles; no item-level donor command inferred.", "x": 2026.1, "y": 1699.0, "degree": 15 }, { "id": "w26-ver-ohf", "label": "Online Harms Foundation", "group": "private-verification", "role": "Organisation", "summary": "Company 11656320, limited by guarantee without share capital; formerly QuantSpark Foundation until 11 May 2021. Implementer and disclosed contribution recipient, distinct from QuantSpark Limited.", "x": 2084.6, "y": 1745.2, "degree": 8 }, { "id": "w26-ver-quantspark", "label": "QuantSpark Limited", "group": "private-verification", "role": "Organisation", "summary": "Company 10304694, private company limited by shares. Filed controller of OHF and its related supplier/creditor; not the foundation named in Canadian awards.", "x": 2101.4, "y": 1771.4, "degree": 6 }, { "id": "w26-ver-hadley", "label": "Adam Hadley", "group": "private-verification", "role": "Person", "summary": "Filed QuantSpark director/controller; company account identifies founder/CEO and shared TAT leadership. Corporate rights are not proof of an individual content decision.", "x": 2071.1, "y": 1743.9, "degree": 2 }, { "id": "w26-ver-tcap", "label": "Terrorist Content Analytics Platform", "group": "private-verification", "role": "Technical system", "summary": "TAT service for collection, classification, alerts and related archive interfaces. Historic reports, current API documentation and the 2024 Altitude adapter have different dates and scopes.", "x": 1994.7, "y": 1660.6, "degree": 13 }, { "id": "w26-ver-tcap-policy", "label": "TCAP inclusion and verification rules", "group": "private-verification", "role": "Policy", "summary": "Selected legal designation inputs combined with TAT source/content assessment, imitation/splinter rules and delisting choices. Full verification policy not acquired.", "x": 2026.3, "y": 1689.5, "degree": 2 }, { "id": "w26-ver-public-safety-canada", "label": "Public Safety Canada", "group": "private-verification", "role": "Public authority", "summary": "Department disclosing TCAP contributions and their recipient/workstream scope. Departmental endorsement, contractual acceptance and payment are separate evidence states.", "x": 2139.4, "y": 1763.2, "degree": 8 }, { "id": "w26-ver-canada-centre", "label": "Canada Centre — Public Safety Canada", "group": "private-verification", "role": "Public authority unit", "summary": "Named departmental centre in the 2023 status account, coordinating TAT engagement and hosting the launch event. This does not itself identify the grant acceptance official.", "x": 2068.3, "y": 1754.5, "degree": 3 }, { "id": "w26-ver-us-nsc", "label": "US National Security Council", "group": "private-verification", "role": "Public authority", "summary": "Canadian departmental account reports coordination with the Canada Centre on engagement with TAT as of October 2023. No US award or classification authority assigned to this body.", "x": 2098.6, "y": 1796.1, "degree": 1 }, { "id": "w26-ver-us-funding-process", "label": "US funding for TAT — process reported in October 2023", "group": "private-verification", "role": "Reported funding process", "summary": "Canadian status account says US funding was in process; agency, amount, agreement, exact approval stage and payment are unknown. This is not an awarded grant or an NSC payment.", "x": 2090.0, "y": 1737.6, "degree": 2 }, { "id": "w26-ver-ca21415", "label": "Canada TCAP contribution — agreement 21415", "group": "private-verification", "role": "Funding instrument", "summary": "Original CAD 1,155,622.44; amended total CAD 1,256,688.09. Both name QuantSpark Foundation; totals are not additive. Signed instrument/payment absent; one-day start-date discrepancy retained.", "x": 2125.8, "y": 1778.4, "degree": 4 }, { "id": "w26-ver-ca22841", "label": "Canada TCAP Phase 2 — agreement 22841", "group": "private-verification", "role": "Funding instrument", "summary": "CAD 1,945,904 disclosed contribution to OHF for archive/classification and moderation-workflow workstreams; stated period 1 April 2023–31 March 2026. No project acceptance or disbursement acquired.", "x": 2072.0, "y": 1718.2, "degree": 3 }, { "id": "w26-ver-crf-terms", "label": "Community Resilience Fund — current terms", "group": "private-verification", "role": "Funding rules", "summary": "Current May 2026 programme architecture: budget/report acceptance, conditional payments, possible IP licence and bounded redistribution approval. Not automatically the terms incorporated in historical TCAP agreements.", "x": 2174.7, "y": 1799.6, "degree": 2 }, { "id": "w26-ver-uk-pilot", "label": "Home Office TCAP pilot — pending September 2026", "group": "private-verification", "role": "Proposed contract", "summary": "UK5 notice SEP653567, pending with estimated 14 September signature and six-month term. Displayed GBP 192,000 ex VAT / 240,000 inc VAT remain unreconciled; no execution/payment inferred.", "x": 1974.9, "y": 1614.8, "degree": 2 }, { "id": "w26-ver-uk-voluntary-guide", "label": "UK voluntary infrastructure guidance — 2024", "group": "private-verification", "role": "Policy guidance", "summary": "Explicitly encourages TCAP and related measures where applicable while creating no legal obligation. Separate statutory/police duties are not automatically triggered by any TAT alert.", "x": 1957.4, "y": 1596.0, "degree": 2 }, { "id": "w26-ver-gifct", "label": "Global Internet Forum to Counter Terrorism", "group": "private-verification", "role": "Organisation", "summary": "GIFCT describes itself as a member-funded nonprofit. Company-led governance, admission conditions and shared-tool access are documented; primary incorporation/IRS identity and donor-specific commands were not acquired.", "x": 1988.1, "y": 1710.5, "degree": 12 }, { "id": "w26-ver-gifct-board", "label": "GIFCT Operating Board", "group": "private-verification", "role": "Governing body", "summary": "Current governance body supplies strategic oversight; Meta chairs in 2026. Founding-company and elected at-large roles are distinct; roster does not prove item-level intervention.", "x": 1969.7, "y": 1660.5, "degree": 5 }, { "id": "w26-ver-gifct-iac", "label": "GIFCT Independent Advisory Committee", "group": "private-verification", "role": "Advisory body", "summary": "Current body advises on priorities/performance. Current-linked terms remain labelled draft; advisory participation is not board voting, database access or a content veto.", "x": 2001.4, "y": 1689.6, "degree": 2 }, { "id": "w26-ver-gifct-iac-draft", "label": "GIFCT IAC terms — current-linked draft", "group": "private-verification", "role": "Draft governance instrument", "summary": "Published draft describes Board-endorsed appointments, nonvoting chair participation, support and formal responses. Signed/adopted status and exercised appointments/conflicts were not acquired.", "x": 2011.6, "y": 1686.0, "degree": 2 }, { "id": "w26-ver-gifct-members", "label": "GIFCT member technology companies", "group": "private-verification", "role": "Participant class", "summary": "Membership requires substantive criteria and tiered annual dues. Shared-database access additionally requires the information-sharing agreement; integrated or integrating is not proof of all operational use.", "x": 1987.3, "y": 1698.4, "degree": 6 }, { "id": "w26-ver-gifct-hsdb", "label": "GIFCT Hash-Sharing Database", "group": "private-verification", "role": "Technical system", "summary": "Qualified hashes/labels supplied by members and GIFCT; recipient platforms retain moderation decisions. Reported corrections and withdrawal visibility do not establish downstream restoration.", "x": 1998.2, "y": 1696.9, "degree": 8 }, { "id": "w26-ver-gifct-taxonomy", "label": "GIFCT HSDB taxonomy — December 2022 v1.0", "group": "private-verification", "role": "Classification policy", "summary": "Distinct legal-list, qualifying incident and behavioral admission routes; TCAP URL hashes must also satisfy GIFCT parameters. Not automatic import of every upstream label.", "x": 2018.2, "y": 1735.5, "degree": 1 }, { "id": "w26-ver-gifct-code", "label": "GIFCT Tech Solutions Code — 2022", "group": "private-verification", "role": "Access governance policy", "summary": "Leadership/Board process can restrict organizational technology access or membership. Organization appeal differs from platform-user content appeal; exercise not proved here.", "x": 2007.4, "y": 1751.5, "degree": 1 }, { "id": "w26-ver-gifct-hria", "label": "BSR assessment for GIFCT — 2021", "group": "private-verification", "role": "Commissioned assessment", "summary": "GIFCT funded the assessment while BSR editorial control. Scoped future organizational risk/recommendations, not an audit of every member outcome; original PDF absent after 403.", "x": 1990.6, "y": 1754.4, "degree": 3 }, { "id": "w26-ver-gifct-hr-policy", "label": "GIFCT human-rights policy — May 2026", "group": "private-verification", "role": "Policy", "summary": "Assessment/action/tracking/reporting commitments and responsibility matrix; not completed measurements or an individual remedy record.", "x": 2000.4, "y": 1754.8, "degree": 1 }, { "id": "w26-ver-gifct-map", "label": "GIFCT Membership Advisory Program", "group": "private-verification", "role": "Programme", "summary": "MAP launch in 2025 is reported; transition of previously BSR-delivered human-rights services into MAP was announced in May 2026, without a verified completion date.", "x": 2015.3, "y": 1750.8, "degree": 1 }, { "id": "w26-ver-tcap-archive", "label": "TCAP Archive and Research Portal", "group": "private-verification", "role": "Technical system", "summary": "TAT reports building a reusable media archive and advertises platform/researcher access. Current download/reuse licence and detailed access criteria were not acquired.", "x": 2050.5, "y": 1739.5, "degree": 4 }, { "id": "w26-ver-tcap-hash-api", "label": "TCAP content-hash API — current public docs", "group": "private-verification", "role": "Technical interface", "summary": "Onboarded users can receive fingerprints and incremental metadata/deletion updates. Published interface, not tested service; separate from Altitude’s older URL-alert importer.", "x": 2033.0, "y": 1720.0, "degree": 4 }, { "id": "w26-ver-tcap-hash-feedback", "label": "TCAP per-hash feedback and analyst response", "group": "private-verification", "role": "Review mechanism", "summary": "Authenticated users can report individual-hash errors and read their own analyst decisions. Three-to-five-working-day reply is an aim; examples are illustrative, not actual cases.", "x": 2046.6, "y": 1732.6, "degree": 2 }, { "id": "w26-ver-trusted-flagger", "label": "TAT Trusted Flagger Platform", "group": "private-verification", "role": "Technical system", "summary": "Advertised permission-controlled partner submissions, analyst/AI vetting, metadata referrals and archive expansion. No DSA appointment, live partner configuration or automatic removal authority inferred.", "x": 2032.6, "y": 1705.4, "degree": 2 }, { "id": "w26-ver-altitude", "label": "Altitude — inspected public implementation", "group": "private-verification", "role": "Software", "summary": "TAT announced stewardship with Jigsaw collaboration; code pinned to 8 August 2024 commit. Trust-based prioritisation, configurable inputs and optional diagnostics are static findings, not deployed outcomes.", "x": 1972.7, "y": 1657.9, "degree": 10 }, { "id": "w26-ver-altitude-operator", "label": "Platform operating Altitude — deployment role", "group": "private-verification", "role": "Operator role", "summary": "Integrator supplies source credentials/authentication and determines how reviewed verdicts affect its own service. No named deployment or actual settings acquired.", "x": 1983.6, "y": 1689.1, "degree": 3 }, { "id": "w26-ver-altitude-verdict-delivery", "label": "Altitude verdict delivery and local-log fallback", "group": "private-verification", "role": "Implementation mechanism", "summary": "Inspected code publishes reviewed decisions to a configured receiver or writes a local log. Internal ACCEPTED is not proof of live removal; later action requires the platform’s implementation.", "x": 1975.5, "y": 1692.2, "degree": 3 }, { "id": "w26-ver-tcap-launch2023", "label": "TCAP 2.0 Canadian launch — May 2023", "group": "private-verification", "role": "Reported event", "summary": "Departmental account records joint TAT launch at Canada Centre event 1–5 May 2023. Launch participation does not establish grant milestone acceptance or completed Phase 2.", "x": 2057.7, "y": 1739.5, "degree": 2 }, { "id": "w26-ver-archive-audience", "label": "Archive access applicants — platforms and researchers", "group": "private-verification", "role": "Access role", "summary": "Advertised audience for Archive access. This role does not establish an admitted organization, a download licence or actual access by an Altitude deployment.", "x": 2076.9, "y": 1781.2, "degree": 1 }, { "id": "w26-ver-hash-recipient", "label": "Onboarded TCAP hash-list recipient", "group": "private-verification", "role": "Recipient role", "summary": "Authenticated recipient role in current public API documentation; may maintain a checkpointed local cache. No actual customer, synchronization result or account restoration is identified.", "x": 2054.7, "y": 1762.8, "degree": 1 }, { "id": "w27-roost-call", "label": "Christchurch Call", "group": "christchurch-roost", "role": "Campaign", "summary": "Voluntary multi-sector Call. Community programme direction is distinct from Foundation corporate powers and domestic law.", "x": 2452.1, "y": 1845.2, "degree": 3 }, { "id": "w27-roost-call-leaders", "label": "Christchurch Call Leaders", "group": "christchurch-roost", "role": "Group of participants", "summary": "Published multi-sector direction/entry role; electorate, quorum, veto and exercised decisions remain not obtained.", "x": 2492.8, "y": 1865.2, "degree": 2 }, { "id": "w27-roost-ccf", "label": "Christchurch Call Foundation", "group": "christchurch-roost", "role": "Organisation", "summary": "Named Canadian contribution recipient and independent support organization. Current roles and donor pages do not substitute for its not obtained governing deed or trustee appointment instruments.", "x": 2232.6, "y": 1765.4, "degree": 13 }, { "id": "w27-roost-support-secretariat", "label": "Call support functions / secretariat", "group": "christchurch-roost", "role": "Institutional function", "summary": "Support function announced for transfer to the Foundation from 1 July 2024 and retrospectively reported transitioned; staff, assets and IP novation instruments absent.", "x": 2320.4, "y": 1795.2, "degree": 2 }, { "id": "w27-roost-proposal2023", "label": "Resilient Call — 2023 institutional proposal", "group": "christchurch-roost", "role": "Proposal", "summary": "Proposed independent legal capacity and split between financial/corporate oversight and Leaders work-programme direction. Not the adopted constitution.", "x": 2273.8, "y": 1795.1, "degree": 1 }, { "id": "w27-roost-ccan", "label": "Christchurch Call Advisory Network", "group": "christchurch-roost", "role": "Organisation", "summary": "Advisory network with its own published internal decision rules, funding relationship and member fundholder; advice is not a general Foundation veto.", "x": 2307.0, "y": 1806.6, "degree": 5 }, { "id": "w27-roost-ccan-core", "label": "CCAN Core Committee", "group": "christchurch-roost", "role": "Governance body", "summary": "At most 25; admissions, internal decisions, co-chair election and fund-use rules under December 2025-labelled terms.", "x": 2350.8, "y": 1830.2, "degree": 2 }, { "id": "w27-roost-ccan-fundholder", "label": "CCAN member fundholder — identity unacquired", "group": "christchurch-roost", "role": "Decision-making role", "summary": "Terms describe operational funds held by a member organization under an agreed memorandum. This role is not a named legal entity or proven signed instrument.", "x": 2346.0, "y": 1835.7, "degree": 2 }, { "id": "w27-roost-mfat", "label": "New Zealand Ministry of Foreign Affairs and Trade", "group": "christchurch-roost", "role": "Government", "summary": "Department reporting Call support transition and its own expenditure. MFAT costs are not the whole programme budget.", "x": 2393.8, "y": 1821.8, "degree": 2 }, { "id": "w27-roost-packard", "label": "David and Lucile Packard Foundation", "group": "christchurch-roost", "role": "Organisation", "summary": "Donor grant database publishes a 2024 CCF general-support entry; receipt and governing grant terms not acquired.", "x": 2278.9, "y": 1790.8, "degree": 1 }, { "id": "w27-roost-carnegie-ny", "label": "Carnegie Corporation of New York", "group": "christchurch-roost", "role": "Organisation", "summary": "Grantmaking corporation in New York, distinct from the existing Carnegie UK node.", "x": 2268.8, "y": 1799.9, "degree": 2 }, { "id": "w27-roost-pivotal-leaders-fund", "label": "Pivotal Leaders Fund", "group": "christchurch-roost", "role": "Funding programme", "summary": "CCF names support through this fund; legal payer, amount, terms and receipt are not established by the inspected page.", "x": 2281.5, "y": 1784.9, "degree": 1 }, { "id": "w27-roost-canada26007", "label": "Canada–CCF contribution 26007", "group": "christchurch-roost", "role": "Funding instrument", "summary": "CAD 1,369,064 disclosed contribution to CCF, 23 October 2025–31 March 2028, with three expected workstreams including a ROOST partnership. No ROOST subaward or acceptance record.", "x": 2206.4, "y": 1784.0, "degree": 3 }, { "id": "w27-roost-ccf-elevate", "label": "CCF Elevate programme", "group": "christchurch-roost", "role": "Programme", "summary": "CCF programme associates the Canadian contribution with new-sector work, ROOST tools and community uptake. Distinct from Tech Coalition Elevate.", "x": 2245.6, "y": 1795.2, "degree": 2 }, { "id": "w27-roost-roost", "label": "ROOST — Robust Open Online Safety Tools Inc", "group": "christchurch-roost", "role": "Organisation", "summary": "Website legal entity and reported EIN 33-3305434; tax status self-reported. Open-tool organization with published repository governance; incorporation and donor reserved rights not acquired.", "x": 2091.0, "y": 1690.9, "degree": 13 }, { "id": "w27-roost-columbia-igp", "label": "Columbia SIPA Institute of Global Politics", "group": "christchurch-roost", "role": "Organisation", "summary": "Institution reports ROOST incubation and prospective independent nonprofit operation; no university ownership or transfer agreement established.", "x": 2133.9, "y": 1711.8, "degree": 1 }, { "id": "w27-roost-global-impact", "label": "Global Impact — advertised ROOST employer", "group": "christchurch-roost", "role": "Organisation", "summary": "Current ROOST-linked vacancy says US individuals would be employed by Global Impact on behalf of ROOST. No filled-vacancy, fiscal-sponsor or grant-custody finding.", "x": 2136.8, "y": 1717.9, "degree": 1 }, { "id": "w27-roost-roost-board", "label": "ROOST Board", "group": "christchurch-roost", "role": "Governance body", "summary": "Current Roles policy assigns TDC appointment. Listed members are individuals with professional affiliations; no employer-reserved seats or donor veto established.", "x": 2149.4, "y": 1698.4, "degree": 3 }, { "id": "w27-roost-tdc", "label": "ROOST Technical Design Committee", "group": "christchurch-roost", "role": "Governance body", "summary": "Current overarching technical/roadmap and tie-break role. November 2025 initial recruitment statement is a separate historical account.", "x": 2145.1, "y": 1709.5, "degree": 3 }, { "id": "w27-roost-maintainers", "label": "ROOST project maintainers", "group": "christchurch-roost", "role": "Decision-making role", "summary": "Project development, releases and merge approvals under branch protections; repository authority is not downstream platform enforcement power.", "x": 2159.9, "y": 1721.3, "degree": 4 }, { "id": "w27-roost-community", "label": "ROOST contributor / community role", "group": "christchurch-roost", "role": "Decision-making role", "summary": "May propose changes; ordinary membership carries no explicit permissions. Partner access teams and individual code contribution are distinct.", "x": 2210.3, "y": 1742.8, "degree": 1 }, { "id": "w27-roost-resource-account", "label": "ROOST reported founding resources", "group": "christchurch-roost", "role": "Reported finance", "summary": "Issuer accounts report more than $27m, $28m and later $28.5m funding/in-kind. These are non-additive descriptions, not a reconciled cash balance or donor allocation schedule.", "x": 2129.5, "y": 1720.2, "degree": 1 }, { "id": "w27-roost-rmc", "label": "ROOST Model Community (RMC)", "group": "christchurch-roost", "role": "Repository", "summary": "Open policy/model integration resources at inspected commit 22b3bd4, dated 20 August 2026. Repository content is distinct from adopted platform policy or deployment.", "x": 2151.1, "y": 1730.5, "degree": 6 }, { "id": "w27-roost-ccf-tvec-pack", "label": "CCF TVEC policy pack", "group": "christchurch-roost", "role": "Policy template", "summary": "CCF-contributed policy, contextual annexes and finite evaluation in RMC. Contains removal/escalation guidance, contextual exceptions and an unresolved INCITE1 reference; not an executed classifier or legal mandate.", "x": 2181.7, "y": 1748.8, "degree": 6 }, { "id": "w27-roost-rmc-licence", "label": "RMC CC BY 4.0 licence", "group": "christchurch-roost", "role": "Licence", "summary": "Permission to share/adapt subject to attribution and other conditions; no endorsement, platform data access, mandatory updates or proof of adoption.", "x": 2161.9, "y": 1751.2, "degree": 2 }, { "id": "w27-roost-tool-operator", "label": "Tool adopter / platform operator", "group": "christchurch-roost", "role": "Decision-making role", "summary": "Chooses local policy, configuration, endpoint and response. This abstract role is not a discovered customer or a grant recipient.", "x": 2120.7, "y": 1728.9, "degree": 15 }, { "id": "w27-roost-osprey", "label": "Osprey", "group": "christchurch-roost", "role": "Software", "summary": "Discord-origin open rules engine donated to ROOST; reported production use and platform-specific rules/effects are separate from the reusable CCF policy pack.", "x": 2174.4, "y": 1687.0, "degree": 6 }, { "id": "w27-roost-bluesky", "label": "Bluesky", "group": "christchurch-roost", "role": "Organisation", "summary": "Named operator in ROOST-published Osprey case study. Reported architecture/effects are not independently observed telemetry.", "x": 2229.7, "y": 1699.6, "degree": 3 }, { "id": "w27-roost-bluesky-effects", "label": "Bluesky-owned Osprey effects receiver", "group": "christchurch-roost", "role": "Implementation", "summary": "Reported Kafka-to-receiver-to-Ozone path applies platform decisions; distinguishes engine outputs from local enforcement.", "x": 2226.9, "y": 1705.3, "degree": 2 }, { "id": "w27-roost-phoebe", "label": "Phoebe investigation prototype", "group": "christchurch-roost", "role": "Software", "summary": "Pinned 7 February 2026 agent prototype for Osprey investigation/rule drafting. Not the same demonstrated stack as Discord or Bluesky production, and Ozone label methods are unimplemented.", "x": 2182.9, "y": 1750.9, "degree": 6 }, { "id": "w27-roost-phoebe-rules", "label": "Phoebe local rule drafting / dry-run path", "group": "christchurch-roost", "role": "Implementation", "summary": "Static code writes local .sml files and separately runs push-rules with --dry-run. Save does not enforce prior validation or publish remote production changes.", "x": 2228.0, "y": 1776.3, "degree": 2 }, { "id": "w27-roost-pathways", "label": "Tech Coalition Pathways", "group": "christchurch-roost", "role": "Programme", "summary": "No-cost, selective capacity-building programme for eligible non-member company employees. Conditions govern access and materials while implementation is voluntary.", "x": 2022.4, "y": 1708.8, "degree": 6 }, { "id": "w27-roost-pathways-terms", "label": "Pathways non-member terms", "group": "christchurch-roost", "role": "Terms", "summary": "Nine published conditions on registration, requests, limited site IP and mentoring confidentiality, with user implementation discretion; no acceptance event acquired.", "x": 2078.7, "y": 1731.3, "degree": 5 }, { "id": "w27-roost-tc-elevate", "label": "Tech Coalition Elevate", "group": "christchurch-roost", "role": "Programme", "summary": "Tailored fee-quoted advisory/implementation service, separate from free Pathways and from CCF Elevate. Its distinct linked service terms were not acquired.", "x": 1993.9, "y": 1735.1, "degree": 2 }, { "id": "w27-roost-stability-ai", "label": "Stability AI", "group": "christchurch-roost", "role": "Organisation", "summary": "Company reports Pathways participation and later Coalition membership. Its two statements disagree on enrollment year; no ROOST or Lantern use inferred.", "x": 1991.9, "y": 1731.0, "degree": 2 }, { "id": "w27-roost-clint-smith", "label": "Clint Smith", "group": "christchurch-roost", "role": "Person", "summary": "Current ROOST co-chair and listed Discord legal chief; also the named participant author of an Osprey account. Personal office and corporate appointment rights are distinct.", "x": 2173.9, "y": 1680.9, "degree": 2 }, { "id": "w28-access-photodna", "label": "Microsoft PhotoDNA", "group": "reference-access-resources", "role": "Research resource", "summary": "Signature-comparison technology. Free provision/source-code licensing is not an open-source licence or automatic reference-data access.", "x": 1941.1, "y": 1574.3, "degree": 4 }, { "id": "w28-access-photodna-cloud", "label": "PhotoDNA Cloud terms and service", "group": "reference-access-resources", "role": "Policy", "summary": "Live-posted May2018 service-specific instrument; separate from installed software and TC sublicensing. Incorporated versions and current commercial offer unresolved.", "x": 1747.8, "y": 1486.2, "degree": 3 }, { "id": "w28-access-tc-sublicensing", "label": "Tech Coalition PhotoDNA sublicense route", "group": "reference-access-resources", "role": "Programme", "summary": "Reported delegated member/non-member software licensing, expanded January2025; master/applicant agreements and actual selection decisions not obtained.", "x": 1955.6, "y": 1644.5, "degree": 5 }, { "id": "w28-access-licensed-implementer", "label": "Licensed PhotoDNA implementer — role", "group": "reference-access-resources", "role": "Audience", "summary": "Abstract recipient role, not a discovered applicant or guaranteed entitlement to all reference lists.", "x": 1969.5, "y": 1673.6, "degree": 1 }, { "id": "w28-access-cloud-customer", "label": "PhotoDNA Cloud customer — role", "group": "reference-access-resources", "role": "Audience", "summary": "Abstract customer of the dated hosted-service terms, not every installed PhotoDNA licensee.", "x": 1700.3, "y": 1478.9, "degree": 1 }, { "id": "w28-access-ncmec", "label": "National Center for Missing & Exploited Children", "group": "reference-access-resources", "role": "Organisation", "summary": "NCMEC administers differentiated reference-sharing environments and reports review/correction work; software and data agreements remain distinct.", "x": 1609.0, "y": 1402.6, "degree": 8 }, { "id": "w28-access-ncmec-reference", "label": "NCMEC reference-sharing environments", "group": "reference-access-resources", "role": "Research resource", "summary": "Separate credentials/permissions and submitter-controlled visibility. Labels do not legally classify every entry in every environment.", "x": 1520.3, "y": 1363.2, "degree": 5 }, { "id": "w28-access-ncmec-submitter", "label": "Reference submitter — member role", "group": "reference-access-resources", "role": "Audience", "summary": "Controls visibility and retraction of its submitted entries through documented functions; not an identified person or universal database authority.", "x": 1469.3, "y": 1336.8, "degree": 2 }, { "id": "w28-access-ncmec-client", "label": "Reference recipient / local dataset — role", "group": "reference-access-resources", "role": "Audience", "summary": "Selects local subset, polling and downstream actions; no specific client cache or restored account observed.", "x": 1465.8, "y": 1343.4, "degree": 2 }, { "id": "w28-access-concentrix", "label": "Concentrix — audit provider", "group": "reference-access-resources", "role": "Organisation", "summary": "Named attestation author; review contract reported July2023 and donated service valued by NCMEC. No broader ownership inquiry.", "x": 1563.5, "y": 1381.8, "degree": 2 }, { "id": "w28-access-ncmec-audit2023", "label": "2023 NCMEC list audit / reported correction", "group": "reference-access-resources", "role": "Event", "summary": "Finite review of538922 unique files.59 differentiated exceptions; auditor reports being told NCMEC removed all59, not direct deletion or downstream restoration evidence.", "x": 1555.8, "y": 1390.0, "degree": 2 }, { "id": "w28-access-swgfl", "label": "South West Grid for Learning Trust / SWGfL", "group": "reference-access-resources", "role": "Organisation", "summary": "Charitable guarantee company05589479; charities1120354/SC051351. Legal operator distinct from programmes and platform participation; no share capital.", "x": 1914.9, "y": 1450.0, "degree": 15 }, { "id": "w28-access-swgfl-trustees", "label": "SWGfL trustees / corporate members", "group": "reference-access-resources", "role": "Organisation", "summary": "Trustees are the only corporate members; management, appointments and recorded/revocable delegation under2024articles. No donor-reserved seat proved.", "x": 1941.7, "y": 1492.8, "degree": 3 }, { "id": "w28-access-stopncii", "label": "StopNCII", "group": "reference-access-resources", "role": "Programme", "summary": "SWGfL programme with public-user and institutional contribution routes. Hash signals, partner access, withdrawal and platform policy decisions are distinct.", "x": 1935.7, "y": 1503.9, "degree": 9 }, { "id": "w28-access-rph", "label": "Revenge Porn Helpline", "group": "reference-access-resources", "role": "Programme", "summary": "SWGfL support service; not a separately established corporation or the same budget as StopNCII.", "x": 1944.4, "y": 1428.8, "degree": 1 }, { "id": "w28-access-stop-user", "label": "StopNCII public submitter — role", "group": "reference-access-resources", "role": "Audience", "summary": "Depicted adult with material and case credentials under public-user FAQ; this is not the only database-input route or an observed identity verification.", "x": 1936.5, "y": 1509.5, "degree": 3 }, { "id": "w28-access-stop-reference", "label": "StopNCII reference/status dataset", "group": "reference-access-resources", "role": "Research resource", "summary": "Documented multiple cases, provider/NGO/public-user inputs and feedback. Withdrawal/status markers are not automatic restoration.", "x": 1903.0, "y": 1512.0, "degree": 8 }, { "id": "w28-access-stop-partner", "label": "StopNCII recipient platform — role", "group": "reference-access-resources", "role": "Audience", "summary": "Abstract organizational recipient; own policy judgement and access conditions. Not a customer list or a verified vote holder.", "x": 1943.3, "y": 1476.6, "degree": 10 }, { "id": "w28-access-stop-approved", "label": "Approved Participant access route", "group": "reference-access-resources", "role": "Programme", "summary": "Lower annual fee schedule and depicted onboarding; advertised benefits conflict with generic join comparison.", "x": 1944.0, "y": 1450.4, "degree": 3 }, { "id": "w28-access-stop-programme", "label": "Programme Partner access route", "group": "reference-access-resources", "role": "Programme", "summary": "Higher annual fee schedule; advertised network/policy engagement, no defined vote/veto or decision exercise.", "x": 1951.1, "y": 1448.9, "degree": 3 }, { "id": "w28-access-stop-partner-terms", "label": "Posted StopNCII partner terms / missing DSA", "group": "reference-access-resources", "role": "Policy", "summary": "Observed public text with conflicting identifiers, unfinished brackets and purpose definitions. Separate data-sharing agreement and executed participant assent remain not obtained.", "x": 1951.7, "y": 1454.4, "degree": 5 }, { "id": "w28-access-stop-applicant-process", "label": "Partner business screening / challenge process", "group": "reference-access-resources", "role": "Programme", "summary": "Applicant/business-official data, not depicted users. Human adverse-result consideration and narrow challenge commitments, not universal refusal appeal.", "x": 2000.3, "y": 1422.6, "degree": 3 }, { "id": "w28-access-deloitte", "label": "Deloitte LLP — applicant verifier", "group": "reference-access-resources", "role": "Organisation", "summary": "Named business-verification/process actor in SWGfL notice, represented as separate controller; no vendor contract acquired.", "x": 2050.0, "y": 1402.6, "degree": 1 }, { "id": "w28-access-neterium", "label": "Neterium — applicant screening service", "group": "reference-access-resources", "role": "Organisation", "summary": "Named sanctions/PEP/watchlist/adverse-media screen; vendor legal contract/individual decisions not acquired.", "x": 2049.4, "y": 1409.0, "degree": 1 }, { "id": "w28-access-pershing-fund", "label": "Pershing Grant — legal donor unresolved", "group": "reference-access-resources", "role": "Research resource", "summary": "Accounts fund label earmarked for StopNCII. No automatic match to Pershing Square Foundation or other legal donor.", "x": 1946.2, "y": 1469.2, "degree": 2 }, { "id": "w28-access-karuna-nain", "label": "Karuna Nain", "group": "reference-access-resources", "role": "Person", "summary": "SWGfL trustee/director; operator biography describes former Meta safety-policy role and launch collaboration. No Meta-nominated seat established.", "x": 1937.8, "y": 1540.3, "degree": 2 }, { "id": "w28-access-fsc", "label": "Free Speech Coalition", "group": "reference-access-resources", "role": "Organisation", "summary": "Industry association advocates StopNCII adoption, including payment-partner credibility. No bank instruction established.", "x": 1974.0, "y": 1493.1, "degree": 1 }, { "id": "w28-access-nominet", "label": "Nominet UK", "group": "reference-access-resources", "role": "Organisation", "summary": "Company 03203859, limited by guarantee. Registry and DNS business plus accumulated reserves support public-benefit commitments; membership is not shareholder profit entitlement. FY 2025 is the inspected financial snapshot; FY 2026 content remains not obtained.", "x": 1844.0, "y": 1309.5, "degree": 23 }, { "id": "w28-access-nominet-members", "label": "Nominet corporate members", "group": "reference-access-resources", "role": "Group of participants", "summary": "Members elect up to four non-executive directors and reappoint Board-appointed non-executives under article 30. Current-linked dated poll rules weight commercial domain involvement; current weights and exercised grant influence were not acquired.", "x": 1872.7, "y": 1212.1, "degree": 1 }, { "id": "w28-access-nominet-board", "label": "Nominet Board", "group": "reference-access-resources", "role": "Governance body", "summary": "Controls company affairs/property and may delegate under conditions it can change or revoke. Former Public Benefit Committee work moved to the main Board. General authority does not identify a specific UKSIC approval vote.", "x": 1861.1, "y": 1253.7, "degree": 3 }, { "id": "w28-access-nominet-public-benefit", "label": "Nominet public-benefit commitments", "group": "reference-access-resources", "role": "Funding programme", "summary": "Up to GBP 65m committed in 2021 could use reserves or annual surplus. FY 2025 public-benefit expense GBP 10.979m exceeded trading operating profit GBP 5.463m. Programme expense, beneficiary grants, CAF transfers and cash are different accounting states.", "x": 1857.9, "y": 1267.3, "degree": 2 }, { "id": "w28-access-caf-account", "label": "Nominet's CAF distribution account", "group": "reference-access-resources", "role": "Account", "summary": "Account described as Nominet’s own CAF account in FY2025 reporting: GBP 1.6m transferred for onward distribution in FY 2026/27. The UKSIC grant table excludes this transfer; it is not an additional UKSIC beneficiary payment.", "x": 1828.5, "y": 1279.1, "degree": 1 }, { "id": "w28-access-uksic", "label": "UK Safer Internet Centre (UKSIC)", "group": "reference-access-resources", "role": "Consortium", "summary": "Consortium of Childnet, Internet Watch Foundation and SWGfL in the 2021 donor announcement. Keep consortium grant totals separate from each charity, its other Nominet programmes and recipient restricted-fund recognition.", "x": 1855.7, "y": 1329.7, "degree": 6 }, { "id": "w28-access-childnet", "label": "Childnet", "group": "reference-access-resources", "role": "Organisation", "summary": "Company 03958416, charity 1080173. Recipient accounts identify a UKSIC partnership role part-funded by Nominet; the complete composition of its Nominet UKSIC 9 fund income remains unestablished.", "x": 1877.2, "y": 1323.7, "degree": 4 }, { "id": "w28-access-uksic-pledge2021", "label": "Nominet 2021 UKSIC three-year pledge", "group": "reference-access-resources", "role": "Funding programme", "summary": "GBP 5.1m pledge announced 29 October 2021, intended to replace lost EU support. Retrospective says the three-year grant ended in March 2025 in its existing form. Exact commencement, instalment, allocation and renewal instruments were not acquired.", "x": 1857.6, "y": 1299.9, "degree": 3 }, { "id": "w28-access-uksic-consortium-agreement", "label": "UKSIC formal partnership agreement — text unacquired", "group": "reference-access-resources", "role": "Agreement", "summary": "Childnet trustees report a formal agreement with SWGfL and IWF for UKSIC projects. This is an agreement-existence disclosure, not signed text or a grant allocation/approval schedule.", "x": 1850.6, "y": 1376.3, "degree": 4 }, { "id": "w28-access-iwf-project-officer", "label": "IWF Project Officer grant reporting role — 2022 specification", "group": "reference-access-resources", "role": "Decision-making role", "summary": "Recruitment pack specifies financial-return/report liaison through UKSIC, communications with Nominet-identified evaluators, and quarterly progress reports to IWF leadership. No appointed person or accepted return acquired.", "x": 1793.8, "y": 1330.7, "degree": 2 }, { "id": "w28-access-sid-research2026", "label": "Safer Internet Day 2026 commissioned study", "group": "reference-access-resources", "role": "Publication", "summary": "Public report credits Nominet commissioning/funding and design input, Childnet study-design lead and Opinium surveys. Distinct from the earlier three-year grant; no amount, signed brief, publication veto or contractual acceptance record acquired.", "x": 1879.3, "y": 1282.1, "degree": 3 }, { "id": "w28-access-opinium", "label": "Opinium", "group": "reference-access-resources", "role": "Organisation", "summary": "Named survey provider in the Safer Internet Day 2026 commissioning/method statement. No separate corporate-identity reconciliation or ownership claim is made.", "x": 1897.3, "y": 1242.9, "degree": 1 }, { "id": "w28-access-iwf-nominet-technical", "label": "IWF Nominet technical-project fund", "group": "reference-access-resources", "role": "Restricted fund", "summary": "Separate from the UKSIC Hotline fund. FY 2025 Note3 records GBP 226352 technical grant income; Note18a records GBP 226351 for the corresponding fund. The original tables differ by GBP 1; no cause or correction established.", "x": 1862.7, "y": 1275.9, "degree": 1 }, { "id": "w28-access-sid2025", "label": "Safer Internet Day — 11 February 2025", "group": "reference-access-resources", "role": "Event", "summary": "Childnet reports delivered activities and Nominet support alongside other funders. Donor-exclusive attribution and independently audited programme impact are not established by the narrative.", "x": 1910.0, "y": 1298.9, "degree": 1 }, { "id": "w29-enf-cpp2026", "label": "Nominet Criminal Practices Policy — July2026", "group": "domain-enforcement", "role": "Administrative policy", "summary": "Posted policy dated6July2026. Listed agencies certify criminal use; Nominet checks and implements registry measures. Suspension affects domain access, not hosted-content deletion.", "x": 1845.1, "y": 1266.1, "degree": 2 }, { "id": "w29-enf-registrant-terms", "label": "Nominet registrant terms — March2024", "group": "domain-enforcement", "role": "Contract terms", "summary": "Unlawful use, discretionary serious-risk suspension and valid binding court orders are distinct grounds. Posted general terms do not establish any particular suspension or enforceability result.", "x": 1853.2, "y": 1274.3, "degree": 1 }, { "id": "w29-enf-listed-agencies", "label": "CPP listed enforcement agencies", "group": "domain-enforcement", "role": "Eligibility role", "summary": "The complete July2026 section4 list has15 entries and includes the Gambling Commission. IWF and UKSIC are absent from this specific list; separate cooperation routes exist.", "x": 1839.4, "y": 1269.2, "degree": 2 }, { "id": "w29-enf-registrant", "label": "Affected domain registrant", "group": "domain-enforcement", "role": "Regulated role", "summary": "Abstract role for notice, corrective action and challenge rights. No named domain, private contact or content is added.", "x": 1831.1, "y": 1256.3, "degree": 6 }, { "id": "w29-enf-registrar", "label": "Domain registrar in Ashiko process", "group": "domain-enforcement", "role": "Operational role", "summary": "The 2020 police policy asks the registrar to investigate and act under its terms before police escalation to Nominet. This is distinct from registry, hosting provider and police authority.", "x": 1775.4, "y": 1110.1, "degree": 1 }, { "id": "w29-enf-reported-actions", "label": "Nominet reported suspensions and reversals", "group": "domain-enforcement", "role": "Reported activity", "summary": "Selected non-continuous periods only: 2019/20,2022/23,2023/24. IWF requests are not proved completed suspensions; reversals are not an audited false-positive rate. Units and reporting periods are preserved.", "x": 1832.0, "y": 1270.5, "degree": 2 }, { "id": "w29-enf-statutory-account", "label": "Nominet July2026 account of future court-backed route", "group": "domain-enforcement", "role": "Reported policy state", "summary": "Nominet reports April2026 Royal Assent and commencement still needed, with indicative end-November timing. Underlying Act and current commencement were not independently read in this run; voluntary CPP remains distinct.", "x": 1839.6, "y": 1277.0, "degree": 1 }, { "id": "w29-enf-city-police", "label": "City of London Police", "group": "domain-enforcement", "role": "Police force", "summary": "Public police force within which PIPCU operates. Unit funding, broad remit, case acceptance and individual suspension authorisation are separately evidenced.", "x": 1801.9, "y": 1162.2, "degree": 5 }, { "id": "w29-enf-ipo", "label": "UK Intellectual Property Office", "group": "domain-enforcement", "role": "Government body", "summary": "The Patent Office, operating as the Intellectual Property Office, is an executive agency and trading fund. Funding announcements, intelligence intake and policy coordination have different functions.", "x": 1768.7, "y": 1177.2, "degree": 13 }, { "id": "w29-enf-pipcu", "label": "Police Intellectual Property Crime Unit (PIPCU)", "group": "domain-enforcement", "role": "Police unit", "summary": "City of London Police specialist unit. Historical account describes operational independence; industry evidence, historical remit participation and secondment do not establish private control of individual cases.", "x": 1808.5, "y": 1215.2, "degree": 23 }, { "id": "w29-enf-ashiko", "label": "Operation Ashiko — dated suspension workflow", "group": "domain-enforcement", "role": "Operational programme", "summary": "2020 policy links rightsholder evidence, police investigation/DCI approval, registrar action and possible Nominet escalation. Agreement is described but not itself acquired. Delivered-operation numbers are police-reported.", "x": 1789.8, "y": 1154.0, "degree": 6 }, { "id": "w29-enf-pipcu-dci", "label": "PIPCU Detective Chief Inspector — suspension authoriser", "group": "domain-enforcement", "role": "Decision role", "summary": "Section9 of the June2020 Ashiko policy requires authorisation for every police domain suspension. This does not transfer registrar/registry implementation or court powers to the officer.", "x": 1785.9, "y": 1112.7, "degree": 1 }, { "id": "w29-enf-gateway", "label": "IPO Intelligence Gateway", "group": "domain-enforcement", "role": "Administrative intake", "summary": "Announced12August2025 as effective11August, replacing the previous PIPCU submission form for rightsholders/enforcement partners. Initial assessment and prioritisation may develop intelligence or lead to law-enforcement referral; acceptance is not guaranteed.", "x": 1791.9, "y": 1181.2, "degree": 3 }, { "id": "w29-enf-referrers", "label": "Ashiko rightsholder and enforcement referrers", "group": "domain-enforcement", "role": "Input role", "summary": "Brands, representatives/trade bodies and enforcement partners may supply evidence/intelligence. Submission, infringement confirmation and private research are inputs, not police certification or a right to compel registry action.", "x": 1799.6, "y": 1172.0, "degree": 4 }, { "id": "w29-enf-bpi", "label": "BPI — recorded-music trade body", "group": "business-identity-policy", "role": "Organisation", "summary": "Named in2014 as supplying an internet-investigator secondee to PIPCU and in2020 as a KYBC proponent. Neither record establishes a current secondment, its terms or a paid enforcement vote.", "x": 1840.8, "y": 1151.2, "degree": 2 }, { "id": "w29-enf-cofunding-pilot", "label": "Proposed PIPCU industry co-funding pilot — April2027", "group": "domain-enforcement", "role": "Proposed funding programme", "summary": "March2026 announcement plans additional industry-supported enforcement from April2027; IPO and PIPCU are to develop priorities and contribution levels. No contributor, payment, selection rule or executed agreement identified.", "x": 1784.7, "y": 1178.6, "degree": 2 }, { "id": "w29-enf-steering-group", "label": "PIPCU historical Steering Group", "group": "domain-enforcement", "role": "Governance body", "summary": "Posted undated referral guide says IPO, police, rightsholders and industry bodies agreed broad priorities. No voting/quorum instrument read. This is not the later IP Crime Group structure or current Gateway.", "x": 1794.4, "y": 1176.1, "degree": 4 }, { "id": "w29-enf-ipo-board", "label": "IPO Board — strategy delivery role", "group": "domain-enforcement", "role": "Governance body", "summary": "The 2022 strategy assigns setting delivery milestones to the IPO Board in consultation with ministers, partners and stakeholders. A published design is not an inspected board decision or delegated-case authority.", "x": 1741.0, "y": 1136.3, "degree": 1 }, { "id": "w29-enf-ipcg-delivery", "label": "IP Crime Group delivery group — FY2022/23 account", "group": "domain-enforcement", "role": "Governance body", "summary": "IPO reports a mixed enforcement-and-industry delivery group as implemented coordination machinery. Exact governing charter, votes and current membership were not acquired.", "x": 1748.9, "y": 1137.6, "degree": 1 }, { "id": "w29-enf-ipcg-industry-advisory", "label": "IP Crime Group industry advisory committee — FY2022/23", "group": "domain-enforcement", "role": "Advisory body", "summary": "Reported industry advisory channel, kept distinct from delivery group, senior executive board and PIPCU Steering Group. No case-approval right inferred.", "x": 1759.1, "y": 1134.1, "degree": 1 }, { "id": "w29-enf-ipcg-executive", "label": "IP Crime Group executive board — FY2022/23 account", "group": "domain-enforcement", "role": "Governance body", "summary": "IPO reports senior industry/enforcement executive coordination for tactical and strategic direction. A report of establishment is not a constitution or proof that industry can approve individual police actions.", "x": 1751.0, "y": 1131.8, "degree": 1 }, { "id": "w29-enf-alliance", "label": "Alliance for Intellectual Property", "group": "business-identity-policy", "role": "Coalition", "summary": "Coalition identified by its own substantive About page and published submissions. Trade association membership is distinct from corporate legal entity or government appointment. No funder control assumed.", "x": 1838.9, "y": 1104.7, "degree": 5 }, { "id": "w29-enf-sky-role", "label": "Sky anti-piracy participation — attributed event role", "group": "domain-enforcement", "role": "Organizational role", "summary": "Public signing post attributes attendance to Sky Anti-Piracy Director Matthew Hibbert. The exact corporate contracting entity, payment or approval role is not established.", "x": 1789.7, "y": 1081.8, "degree": 1 }, { "id": "w29-enf-signing2026", "label": "Reported 2026/27 PIPCU Service Agreement signing", "group": "domain-enforcement", "role": "Reported event", "summary": "Main post metadata29May2026 describes signing that week and industry attendance. Exact event date and agreement text not obtained; displayed account alias not independently reconciled to the government-linked profile.", "x": 1796.2, "y": 1126.0, "degree": 4 }, { "id": "w29-enf-cma-taskforce", "label": "CMA Digital Markets Taskforce — 2020 consultation", "group": "business-identity-policy", "role": "Government process", "summary": "Official case page published Alliance/BPI responses on8December2020. Earlier31July deadline is not a submission date. Publication is not adoption of KYBC proposals.", "x": 1872.7, "y": 1055.7, "degree": 2 }, { "id": "w29-enf-alliance-kybc2020", "label": "Alliance KYBC proposal — internal August2020", "group": "business-identity-policy", "role": "Policy proposal", "summary": "Proposes robust verification of business customers across critical online services and loss of hosting/payment access for illicit businesses withholding identity. Document predates RUSI2021; not a universal consumer age-ID scheme.", "x": 1855.4, "y": 1062.6, "degree": 3 }, { "id": "w29-enf-bpi-kybc2020", "label": "BPI business-verification proposal — redacted November2020", "group": "business-identity-policy", "role": "Policy proposal", "summary": "Section4.2.2 seeks name/address/contact and entity-responsible-person checks, bank details where appropriate, and stopping services for withheld/incorrect identity. That clause has no separate proved-infringement prerequisite.", "x": 1862.7, "y": 1095.7, "degree": 2 }, { "id": "w29-enf-alliance-osb2021", "label": "Alliance OSB0016 evidence — September2021", "group": "business-identity-policy", "role": "Parliamentary submission", "summary": "Advocates business verification across infrastructure and asks parliament/government to find a legislative vehicle. Citation of CMA G22(13b) about unlawful activity/economic detriment does not establish CMA adoption of this KYBC wording.", "x": 1840.7, "y": 1060.3, "degree": 1 }, { "id": "w29-enf-cic-rightsholders", "label": "Rightsholders on CIC IP subgroup — briefing authors", "group": "business-identity-policy", "role": "Group of participants", "summary": "Authorship attributed in MPA annex to rightsholder organisations on the Creative Industries Council IP subgroup. It is not demonstrated endorsement by the full CIC, all its members or government.", "x": 1894.4, "y": 1096.5, "degree": 1 }, { "id": "w29-enf-cic-briefing2021", "label": "Rightsholder KYBC briefing — annex-labelled September2021", "group": "business-identity-policy", "role": "Policy proposal", "summary": "Commercial-customer verification across platforms, hosting/CDNs, DNS/registrars, proxies, email, finance and advertising; failure or inability to verify would stop service. Detailed challenge mechanics not specified in the inspected Q1–5.", "x": 1877.9, "y": 1138.4, "degree": 2 }, { "id": "w29-enf-mpa-evidence2022", "label": "MPA FDF0068 evidence — internal10May2022", "group": "business-identity-policy", "role": "Parliamentary submission", "summary": "Carries rightsholder briefing, asks departments to join IPO forum, and quotes RUSI2021. Internal document date is not independently established publication day or proof of acceptance.", "x": 1849.1, "y": 1194.2, "degree": 4 }, { "id": "w29-enf-rusi", "label": "Royal United Services Institute (RUSI)", "group": "business-identity-policy", "role": "Research organization", "summary": "Publisher of the March2021 corrected Whitehall report. Funding and research methodology are disclosed by the report; commissioning contracts, funder clearance or item-level direction were not acquired.", "x": 1859.3, "y": 1130.1, "degree": 1 }, { "id": "w29-enf-industry-trust", "label": "Industry Trust — report-disclosed funder", "group": "business-identity-policy", "role": "Named funder", "summary": "Name printed in the RUSI funding disclosure. Exact legal entity/award agreement not obtained; no automatic reconciliation to another similarly named body.", "x": 1845.6, "y": 1128.8, "degree": 1 }, { "id": "w29-enf-premier-league", "label": "Premier League — report-disclosed funder", "group": "business-identity-policy", "role": "Named funder", "summary": "Named RUSI co-funder; amount, payment date and editorial/approval rights are not disclosed in the inspected report.", "x": 1864.9, "y": 1132.9, "degree": 1 }, { "id": "w29-enf-rusi-report2021", "label": "Taking the Profit Out of IP Crime — corrected March2021 report", "group": "business-identity-policy", "role": "Research output", "summary": "Authors Janjeva, Reid and Moiseienko disclose five co-funders and claim independent findings. Reported77 interviews Feb–Sep2020; two January2021 validation workshops excluded funders. These are reported methods, not an external audit.", "x": 1842.5, "y": 1172.9, "degree": 8 }, { "id": "w29-enf-rusi-recommendations", "label": "RUSI business-verification and financial-disruption recommendations", "group": "business-identity-policy", "role": "Policy recommendations", "summary": "Recommendation4 seeks identity recording/verification to extent possible and compliance resources;14 recommends financial institutions seek named lists and take appropriate action including termination/SAR;15 suggests reasonable identity best practice even without formal KYBC. Recommendations do not prove list access, enforcement or adoption.", "x": 1852.4, "y": 1130.5, "degree": 1 }, { "id": "w29-enf-kybc-forum", "label": "IPO cross-departmental KYBC consideration", "group": "business-identity-policy", "role": "Government process", "summary": "Government response/strategy and FY2022/23 report describe consideration and several Whitehall forums. Membership invitation and policy exploration are distinct from approved wording, statutory power or private authorship of a decision.", "x": 1810.0, "y": 1146.8, "degree": 3 }, { "id": "w29-enf-music-response2021", "label": "Government music-streaming response — recommendation23", "group": "business-identity-policy", "role": "Government response", "summary": "Received15September/published22September2021: records KYBC concerns, government attention and IPO cross-department consideration. It does not create a verifier, sanction or appeal route.", "x": 1802.7, "y": 1100.1, "degree": 1 }, { "id": "w29-enf-sector-vision2023", "label": "Sector Vision June2023 — explore KYBC potential", "group": "business-identity-policy", "role": "Government policy", "summary": "Following roundtables, government says it will engage technology/creative sectors to explore potential KYBC solutions. Adjacent departmental work does not allocate final legal powers or establish adopted text.", "x": 2046.8, "y": 1305.0, "degree": 1 }, { "id": "w29-enf-dcms-current", "label": "Department for Culture, Media and Sport — 2023/26 policy role", "group": "business-identity-policy", "role": "Government department", "summary": "Named in current-period advertising/creative-policy records. Kept separate from the inherited historically named Digital,Culture,Media and Sport buyer node because this run did not acquire institutional succession/contract novation records.", "x": 2019.2, "y": 1340.8, "degree": 3 }, { "id": "w29-enf-oap2023", "label": "Online Advertising Programme response — July2023 proposals", "group": "business-identity-policy", "role": "Government proposal", "summary": "Proposed statutory systems duties on platforms, intermediaries and publishers for paid/reciprocal illegal advertising and a separate under18 product strand. Advertisers excluded from direct new regime; identity checks would be necessary/appropriate and designed by regulators. Not enacted by this response.", "x": 2061.9, "y": 1289.6, "degree": 2 }, { "id": "w29-enf-oat", "label": "Online Advertising Taskforce", "group": "business-identity-policy", "role": "Coordination body", "summary": "Public minutes record members agreeing draft working-group terms/year plan, not statutory enforcement authority. Advertising dispositions are kept separate from Alliance/IPO KYBC pathways without an explicit transmission record.", "x": 1958.1, "y": 1409.6, "degree": 5 }, { "id": "w29-enf-iab-uk", "label": "IAB UK", "group": "business-identity-policy", "role": "Industry body", "summary": "Named partner in the anti-fraud advertising initiative and presenter of working-group design. Industry participation and an agreed workplan do not confer regulator or court powers.", "x": 1935.0, "y": 1331.8, "degree": 3 }, { "id": "w29-enf-ad-fraud-working-group", "label": "Ad Fraud and Standards Working Group", "group": "business-identity-policy", "role": "Working group", "summary": "26February2026 minutes record taskforce members agreeing draft terms/year plan presented by IAB UK. Actual terms not acquired; future work to map standards and disrupt malicious ads is not a demonstrated outcome.", "x": 1959.4, "y": 1362.5, "degree": 2 }, { "id": "w29-enf-fraud-strategy2026", "label": "Fraud Strategy — March2026 advertising account", "group": "business-identity-policy", "role": "Government strategy", "summary": "Paragraph55 describes Home Office/DCMS/IAB partnership reporting to ministers early2027 and conditional legislation if inadequate. Paragraph54 separately describes enacted OSA duties and then-future timetable; this source is not a fresh September2026 commencement audit.", "x": 1986.3, "y": 1394.3, "degree": 3 }, { "id": "w29-enf-advertising-check-design", "label": "OAP proposed identity-check design", "group": "business-identity-policy", "role": "Proposed regulatory design", "summary": "July2023 response leaves the extent of necessary/appropriate checks to future regulatory design; no universal business verification or current regulator requirement established.", "x": 2094.0, "y": 1253.5, "degree": 1 }, { "id": "w30-class-creative", "label": "Operation Creative", "group": "advertising-list-authority", "role": "Programme", "summary": "Dated police procedure and later portal design; current deployment and case-level correction remain unverified.", "x": 1819.7, "y": 1102.6, "degree": 2 }, { "id": "w30-class-iwl", "label": "PIPCU Infringing Website List", "group": "advertising-list-authority", "role": "Research resource", "summary": "Police-assessed list for advertising decisions; distinct from domain suspension and CODA's same-acronym list.", "x": 1884.2, "y": 1225.7, "degree": 4 }, { "id": "w30-class-creative-referrers", "label": "Creative rightsholders and nominated referrers", "group": "advertising-list-authority", "role": "Group of participants", "summary": "Role class supplying evidence, not decision-makers by referral alone.", "x": 1820.5, "y": 1179.9, "degree": 1 }, { "id": "w30-class-creative-head", "label": "PIPCU Head of Unit — Creative tactic role", "group": "advertising-list-authority", "role": "Decision-making role", "summary": "Decision role in the released guidance; not automatically Ashiko's DCI role or a current named officeholder.", "x": 1817.6, "y": 1058.2, "degree": 1 }, { "id": "w30-class-ad-recipients", "label": "IWL advertising recipients", "group": "advertising-list-authority", "role": "Group of participants", "summary": "Advertisers, agencies and intermediaries receiving/promoted to use IWL; choices and contractual duties require their own instruments.", "x": 1888.0, "y": 1256.2, "degree": 3 }, { "id": "w30-class-site-operators", "label": "Affected website operators — list review role", "group": "advertising-list-authority", "role": "Group of participants", "summary": "Applicants for challenge/removal under each specified national process; no individual site is designated here.", "x": 1787.5, "y": 1220.3, "degree": 3 }, { "id": "w30-class-prs", "label": "PRS for Music", "group": "advertising-list-authority", "role": "Organisation", "summary": "Reported supplier of138 Creative website referrals; counting period not specified in the report paragraph.", "x": 1818.6, "y": 1175.1, "degree": 1 }, { "id": "w30-class-portal-spec", "label": "2024 Op Creative Portal specification", "group": "advertising-list-authority", "role": "Policy", "summary": "Tender design places submission acceptance with police administrators; implementation and award unverified.", "x": 1808.6, "y": 1176.5, "degree": 1 }, { "id": "w30-class-whitebullet", "label": "White Bullet", "group": "advertising-list-authority", "role": "Organisation", "summary": "Vendor of official-list and proprietary risk resources; measurement and use claims retain vendor attribution.", "x": 1819.5, "y": 1166.2, "degree": 6 }, { "id": "w30-class-gpi", "label": "White Bullet Global Piracy Index", "group": "advertising-list-authority", "role": "Research resource", "summary": "Vendor-described combination of court-order and official infringement-list sources.", "x": 1820.0, "y": 1121.6, "degree": 1 }, { "id": "w30-class-ipi", "label": "White Bullet proprietary IPI Index", "group": "advertising-list-authority", "role": "Research resource", "summary": "Vendor-described risk assessment extending beyond court/enforcement-adjudicated sites.", "x": 1812.2, "y": 1121.9, "degree": 1 }, { "id": "w30-class-peer39", "label": "Peer39", "group": "advertising-list-authority", "role": "Organisation", "summary": "Advertises White Bullet piracy-domain exclusion offering in its marketplace; activation and transactions unverified.", "x": 1827.6, "y": 1122.2, "degree": 2 }, { "id": "w30-class-gambling-commission", "label": "Gambling Commission", "group": "advertising-list-authority", "role": "Organisation", "summary": "Issues the separately all-licence advertising condition and promotes IWL as an aid.", "x": 1988.3, "y": 1207.9, "degree": 2 }, { "id": "w30-class-gambling-licensees", "label": "Gambling licence holders — condition scope", "group": "advertising-list-authority", "role": "Group of participants", "summary": "Licensees under LCCP16.1.1; not a universal duty on all advertising users.", "x": 2150.8, "y": 1209.0, "degree": 1 }, { "id": "w30-class-gambling-condition", "label": "LCCP16.1.1 advertising condition", "group": "advertising-list-authority", "role": "Policy", "summary": "Reasonable-step duties and termination-capable contracts, qualified by reasonableness and dispute provisions; no IWL mandate in the condition.", "x": 2081.0, "y": 1203.9, "degree": 2 }, { "id": "w30-class-thinkbox-evidence", "label": "Thinkbox ADV0038 anonymous agency accounts", "group": "advertising-list-authority", "role": "Publication", "summary": "January2018 mediated practice descriptions; source notes caution post-purchase blocking can coexist with payment.", "x": 1910.9, "y": 1216.3, "degree": 1 }, { "id": "w30-class-payment-providers", "label": "Payment service providers / merchant banks", "group": "payment-notice-authority", "role": "Group of participants", "summary": "Unspecified recipient role in Creative requests and ALERT-PAY design; not an identified pilot participant.", "x": 1778.4, "y": 1254.1, "degree": 2 }, { "id": "w30-class-wipo", "label": "World Intellectual Property Organization", "group": "payment-notice-authority", "role": "Organisation", "summary": "Institutional distributor and programme administrator; no automatic authority over private provider termination.", "x": 1774.8, "y": 1390.8, "degree": 7 }, { "id": "w30-class-wipo-alert", "label": "WIPO ALERT — advertising distribution", "group": "payment-notice-authority", "role": "Programme", "summary": "National contributor assessments distributed for advertising-only use under described arrangements; distinct from ALERT-PAY.", "x": 1811.8, "y": 1273.0, "degree": 7 }, { "id": "w30-class-contributors", "label": "WIPO ALERT national contributors", "group": "payment-notice-authority", "role": "Group of participants", "summary": "Contributors retain lists, accuracy responsibility and national review; not every contributor is a ministry.", "x": 1787.7, "y": 1261.2, "degree": 4 }, { "id": "w30-class-alert-users", "label": "WIPO ALERT authorized advertising users", "group": "payment-notice-authority", "role": "Group of participants", "summary": "Screened advertising businesses; described user agreements, list selection and contributor permissions apply.", "x": 1787.3, "y": 1314.7, "degree": 4 }, { "id": "w30-class-pay", "label": "WIPO ALERT-PAY", "group": "payment-notice-authority", "role": "Programme", "summary": "Payment-notice initiative: manual pilot November2024–August2025, development at June2026 presentation; no public go-live or specific provider contract proved.", "x": 1764.3, "y": 1310.8, "degree": 4 }, { "id": "w30-class-pay-pilot", "label": "ALERT-PAY manual pilot", "group": "payment-notice-authority", "role": "Event", "summary": "WIPO reports six rightsholders, two unnamed PSPs,17actions/35sites and71% composite site-or-goods removal; no merchant closure count.", "x": 1757.3, "y": 1359.8, "degree": 1 }, { "id": "w30-class-pay-rightsholders", "label": "ALERT-PAY authorized rightsholders — design role", "group": "payment-notice-authority", "role": "Group of participants", "summary": "Rights claimants in proposed notice workflow; illustrative Adidas screens do not identify an actual participant.", "x": 1747.5, "y": 1273.2, "degree": 1 }, { "id": "w30-class-wipo-member-states", "label": "WIPO Member States — programme approval role", "group": "payment-notice-authority", "role": "Group of participants", "summary": "Approve WIPO work programme; no programme approval is treated as direct instruction to a merchant or PSP.", "x": 1725.3, "y": 1250.7, "degree": 1 }, { "id": "w30-class-wipo-programme", "label": "WIPO approved programme2026/27", "group": "payment-notice-authority", "role": "Policy", "summary": "Includes planned ALERT-PAY deployment within broader priorities. No exact Pay budget allocation acquired.", "x": 1742.1, "y": 1288.7, "degree": 3 }, { "id": "w30-class-wipo-er23", "label": "WIPO Expected Result2.3 — broad budget scope", "group": "payment-notice-authority", "role": "Research resource", "summary": "BRIP international dialogue/cooperation budget/expenditure. Figures are not an ALERT-PAY earmark or payment.", "x": 1755.3, "y": 1334.6, "degree": 2 }, { "id": "w30-class-match", "label": "Mastercard MATCH Pro", "group": "payment-record-authority", "role": "Research resource", "summary": "Shared terminated-merchant/owner information system; mandatory reporting/inquiry coexists with discretion over new admission.", "x": 1824.7, "y": 1518.6, "degree": 9 }, { "id": "w30-class-merchants", "label": "Merchants and owners — MATCH rule scope", "group": "payment-record-authority", "role": "Group of participants", "summary": "Role class affected by qualifying reports and information rights, not a finding about every merchant or all bank access.", "x": 1752.1, "y": 1514.6, "degree": 3 }, { "id": "w30-class-match-users", "label": "MATCH authorized acquirers and service providers", "group": "payment-record-authority", "role": "Group of participants", "summary": "Approved acquirers and registered/provisioned providers; service access requires acquirer approval. No actual authorization inspected.", "x": 1799.8, "y": 1510.2, "degree": 3 }, { "id": "w30-class-match-rules", "label": "MATCH Chapter11 — public August2026 edition", "group": "payment-record-authority", "role": "Policy", "summary": "Dated public manual reader, with currency caveat; not all underlying network contracts or historic rules.", "x": 1454.3, "y": 251.6, "degree": 0 }, { "id": "w30-class-stripe-acceptance", "label": "Stripe published MATCH acceptance practice", "group": "payment-record-authority", "role": "Policy", "summary": "General disqualification with exceptional-circumstance consideration, a provider choice rather than mandatory network-wide refusal.", "x": 1798.8, "y": 1529.2, "degree": 2 }, { "id": "w30-class-ip-claimants", "label": "IP owners and authorized associations — provider-policy role", "group": "payment-record-authority", "role": "Group of participants", "summary": "General provider-policy eligible claimant class; not proof of an accepted Creative or ALERT-PAY referral.", "x": 1735.2, "y": 1519.7, "degree": 2 }, { "id": "w30-class-egumball", "label": "eGumball, Inc.", "group": "commercial-classification-remedy", "role": "Organisation", "summary": "Merchant in2023/2025 appeals; allegations and procedural rulings do not establish completed delisting or restored processing.", "x": 2034.1, "y": 1529.4, "degree": 2 }, { "id": "w30-class-merrick", "label": "Merrick Bank Corporation", "group": "commercial-classification-remedy", "role": "Organisation", "summary": "Acquirer described in historical eGumball record; actual listing/removal role distinct from Visa and Mastercard.", "x": 1905.1, "y": 1528.8, "degree": 4 }, { "id": "w30-class-egumball-superior", "label": "Orange County Superior Court — eGumball case", "group": "commercial-classification-remedy", "role": "Organisation", "summary": "Trial forum30-2021-01216871; provisional2021 order described by appellate opinions, not final merits.", "x": 1978.3, "y": 1532.1, "degree": 2 }, { "id": "w30-class-appeal-fourth", "label": "California Court of Appeal, Fourth District, Division Three", "group": "commercial-classification-remedy", "role": "Organisation", "summary": "Issued the distinct2023 emergency-relief and2025 anti-SLAPP decisions.", "x": 2064.2, "y": 1518.4, "degree": 4 }, { "id": "w30-class-egumball2023", "label": "eGumball G060853 emergency-relief appeal", "group": "commercial-classification-remedy", "role": "Publication", "summary": "Unpublished17May2023 opinion reversed mandatory TRO relief and reserved merits.", "x": 2122.8, "y": 1530.7, "degree": 1 }, { "id": "w30-class-egumball2025", "label": "eGumball G062863M anti-SLAPP appeal", "group": "commercial-classification-remedy", "role": "Publication", "summary": "23April2025 decision as modified6May affirmed denial of Merrick's procedural motion and ordered a narrower public brief. No final liability/removal.", "x": 1992.7, "y": 1496.0, "degree": 4 }, { "id": "w30-class-ca-supreme", "label": "Supreme Court of California", "group": "commercial-classification-remedy", "role": "Organisation", "summary": "FilmOn2019 procedural decision; no classification correction or damages order.", "x": 1956.3, "y": 1370.6, "degree": 2 }, { "id": "w30-class-filmon", "label": "FilmOn.com Inc.", "group": "commercial-classification-remedy", "role": "Organisation", "summary": "Historical claimant disputing breadth/meaning and commercial effects of DV labels.", "x": 1546.0, "y": 1182.8, "degree": 2 }, { "id": "w30-class-dv-inc", "label": "DoubleVerify Inc. — named litigant", "group": "commercial-classification-remedy", "role": "Organisation", "summary": "Named in FilmOn litigation and2025 complaint; distinct from Holdings and the general product brand where contracting entity is not established.", "x": 1579.5, "y": 1219.4, "degree": 3 }, { "id": "w30-class-dv-holdings", "label": "DoubleVerify Holdings Inc.", "group": "commercial-classification-remedy", "role": "Organisation", "summary": "FY2025 consolidated issuer and2025 legal claimant; current service disclosures do not establish exact old contracting entity.", "x": 1578.9, "y": 1253.8, "degree": 7 }, { "id": "w30-class-dv", "label": "DoubleVerify — product and membership brand", "group": "commercial-classification-remedy", "role": "Organisation", "summary": "Public product, data-input and membership accounts; exact legal contract party must be separately established.", "x": 1634.0, "y": 1312.2, "degree": 14 }, { "id": "w30-class-historical-dv-reports", "label": "Historical DoubleVerify confidential website reports", "group": "commercial-classification-remedy", "role": "Research resource", "summary": "Court-described paid domain/subdomain, impressions and thematic-category reports; Adult Content was broader than pornography.", "x": 1553.3, "y": 1179.9, "degree": 2 }, { "id": "w30-class-filmon-opinion", "label": "FilmOn S244157 public-interest anti-SLAPP decision", "group": "commercial-classification-remedy", "role": "Publication", "summary": "6May2019 reversal uses context/audience/purpose; rejects categorical commercial-speech exclusion without ordering accuracy/remedy.", "x": 1985.1, "y": 1431.2, "degree": 2 }, { "id": "w30-class-filmon-amicus", "label": "2018 media amici brief supporting DoubleVerify", "group": "commercial-classification-remedy", "role": "Publication", "summary": "MPAA, Hearst, Tegna, California News Publishers Association and First Amendment Coalition requested affirmance; no funding or purchase proved.", "x": 1911.8, "y": 1319.5, "degree": 2 }, { "id": "w30-class-advertiser-customers", "label": "DV advertiser customers — disclosed class", "group": "commercial-classification-remedy", "role": "Group of participants", "summary": "Customers choose categories and have specified reporting eligibility; not all affected publishers or verified actual configurations.", "x": 1595.8, "y": 1272.0, "degree": 3 }, { "id": "w30-class-supply-customers", "label": "DV supply-side customers — disclosed class", "group": "commercial-classification-remedy", "role": "Group of participants", "summary": "Consolidated filing describes minimum guarantees/tiered pricing; no named customer contract or fee per exclusion acquired.", "x": 1546.1, "y": 1216.8, "degree": 1 }, { "id": "w30-class-dv-classifier", "label": "DV brand suitability classification service", "group": "commercial-classification-remedy", "role": "Research resource", "summary": "Vendor-described AI/human classification, customer categories and pre-bid/post-purchase products; not independently validated operation.", "x": 1590.2, "y": 1274.6, "degree": 4 }, { "id": "w30-class-hi-category", "label": "Highly Illicit: Do Not Monetize — DV category", "group": "commercial-classification-remedy", "role": "Policy", "summary": "Vendor category/default uses selected inputs and its own judgment; not equivalent to all NCMEC reports or IWF raw URL list.", "x": 1629.0, "y": 1323.5, "degree": 6 }, { "id": "w30-class-p2p-category", "label": "P2P Sharing and Streaming — DV category", "group": "commercial-classification-remedy", "role": "Policy", "summary": "Vendor-described avoidance category includes infrastructure with legitimate uses; no judicial illegality of every service inferred.", "x": 1587.5, "y": 1285.1, "degree": 2 }, { "id": "w30-class-dv-platforms", "label": "DV platform partners — reported implementation class", "group": "commercial-classification-remedy", "role": "Group of participants", "summary": "More than100 distribution partners announced February2025; nearly30 implementations reported April2025 without names/logs.", "x": 1601.5, "y": 1302.1, "degree": 2 }, { "id": "w30-class-ncmec-outbound", "label": "NCMEC public outbound-notification tables — examined2021–2023 family", "group": "advertising-input-and-pressure", "role": "Research resource", "summary": "Notice/response aggregates with repeated contacts, mixed provider responses and three2023content categories. Exact DV input files/filters remain unknown.", "x": 1580.5, "y": 1349.4, "degree": 2 }, { "id": "w30-class-ncmec-incoming", "label": "NCMEC2023 incoming CyberTipline report table", "group": "advertising-input-and-pressure", "role": "Research resource", "summary": "Reports submitted by providers, not outbound notices, unique incidents or confirmed criminal domains.", "x": 1555.9, "y": 1397.2, "degree": 1 }, { "id": "w30-class-dv-ncmec-input", "label": "DV-described public NCMEC input — versions unknown", "group": "advertising-input-and-pressure", "role": "Research resource", "summary": "DV says three years and over300domains/providers as a starting point, followed by its own assessment; exact dataset/columns and identity mapping not disclosed.", "x": 1608.0, "y": 1347.9, "degree": 3 }, { "id": "w30-class-iwf-url", "label": "IWF assessed URL List — DV input scope", "group": "advertising-input-and-pressure", "role": "Research resource", "summary": "Specialist URL-level resource with published upstream maintenance; not equivalent to an entire DV domain category.", "x": 1689.0, "y": 1342.7, "degree": 3 }, { "id": "w30-class-iwf-members", "label": "IWF funding members — governance opportunity class", "group": "advertising-input-and-pressure", "role": "Group of participants", "summary": "Members may apply/nominate to Funding Council; membership does not prove a particular member seat or assessment veto.", "x": 1712.7, "y": 1363.7, "degree": 3 }, { "id": "w30-class-iwf-council", "label": "IWF Funding Council — FY2025 account", "group": "advertising-input-and-pressure", "role": "Organisation", "summary": "Report describes industry advice and election of three industry trustees; not proof DV participated.", "x": 1684.8, "y": 1370.3, "degree": 2 }, { "id": "w30-class-iwf-board", "label": "IWF Board — FY2025 account", "group": "advertising-input-and-pressure", "role": "Organisation", "summary": "Dated report describes11-person composition and appointment approval processes; no unchanged2026 composition inferred.", "x": 1712.5, "y": 1369.9, "degree": 2 }, { "id": "w30-class-blackburn", "label": "Marsha Blackburn", "group": "advertising-input-and-pressure", "role": "Person", "summary": "Senator signing the February2025 requests; no enforcement order.", "x": 1535.0, "y": 1273.2, "degree": 1 }, { "id": "w30-class-blumenthal", "label": "Richard Blumenthal", "group": "advertising-input-and-pressure", "role": "Person", "summary": "Senator signing the February2025 requests; no enforcement order.", "x": 1532.5, "y": 1280.1, "degree": 1 }, { "id": "w30-class-senate-inquiry", "label": "February2025 Blackburn–Blumenthal advertising inquiry", "group": "advertising-input-and-pressure", "role": "Event", "summary": "Documented requests and14February response deadline, not subpoena, violation finding or actual sanction.", "x": 1583.6, "y": 1303.0, "degree": 9 }, { "id": "w30-class-cma", "label": "Check My Ads Institute", "group": "advertising-input-and-pressure", "role": "Organisation", "summary": "Claims advocacy contribution; recipient of preservation letter.", "x": 1685.9, "y": 1366.5, "degree": 14 }, { "id": "w30-class-adalytics", "label": "Adalytics Research LLC", "group": "advertising-input-and-pressure", "role": "Organisation", "summary": "Research source identified by Senate release and defendant in May2025 DV complaint; technical merits not adjudicated here.", "x": 1597.0, "y": 1286.6, "degree": 4 }, { "id": "w30-class-mrc", "label": "Media Rating Council", "group": "advertising-input-and-pressure", "role": "Organisation", "summary": "Recipient of Senate review/revocation questions; no compliance finding or revocation.", "x": 1542.3, "y": 1275.6, "degree": 1 }, { "id": "w30-class-tag", "label": "Trustworthy Accountability Group", "group": "advertising-input-and-pressure", "role": "Organisation", "summary": "Recipient of Senate pressure; later reports no BSC violations and additional training. Its determination is not independent exoneration.", "x": 1525.2, "y": 1260.0, "degree": 2 }, { "id": "w30-class-tag-threat-members", "label": "TAG threat-exchange members — training recipient class", "group": "advertising-input-and-pressure", "role": "Group of participants", "summary": "TAG describes members receiving NCMEC training; no operational shared database or participant list.", "x": 1554.0, "y": 1403.2, "degree": 1 }, { "id": "w30-class-tag-response", "label": "TAG2025 BSC inquiry determination — reported", "group": "advertising-input-and-pressure", "role": "Publication", "summary": "Visible2025report PDF20–21 says none of investigated companies violated BSC.23-page response and underlying findings not obtained.", "x": 1483.5, "y": 1230.1, "degree": 1 }, { "id": "w30-class-garcia-citation", "label": "Arielle Garcia article cited in TAG letter", "group": "advertising-input-and-pressure", "role": "Publication", "summary": "Citation-only input established by letter footnote7. Underlying article/testimony not read; no inference of letter authorship.", "x": 1545.4, "y": 1268.8, "degree": 1 }, { "id": "w30-class-bondi", "label": "Bradley J. Bondi", "group": "advertising-input-and-pressure", "role": "Person", "summary": "Named/signed counsel for DV in recipient-published preservation letter and listed counsel in later pleading. No legal-firm finance inferred.", "x": 1534.0, "y": 1224.1, "degree": 1 }, { "id": "w30-class-paul-hastings", "label": "Paul Hastings LLP", "group": "advertising-input-and-pressure", "role": "Organisation", "summary": "Firm on preservation letter and complaint; no fee, ownership or extra-client relation acquired.", "x": 1539.0, "y": 1219.5, "degree": 1 }, { "id": "w30-class-preservation-letter", "label": "14April2025 DV preservation demand", "group": "advertising-input-and-pressure", "role": "Publication", "summary": "Recipient-published redacted counsel letter seeks preservation and confirmation by16April; not production/retraction or court order.", "x": 1630.2, "y": 1306.2, "degree": 3 }, { "id": "w30-class-dv-complaint", "label": "DV v Adalytics complaint —12May2025", "group": "advertising-input-and-pressure", "role": "Publication", "summary": "Stamped Document1 in8:25-cv-01535 names Adalytics only as defendant; allegations and requested remedies unadjudicated.", "x": 1622.5, "y": 1285.5, "degree": 4 }, { "id": "w31-fa-ford", "label": "Ford Foundation", "group": "funding-authority", "role": "Organisation", "summary": "Named donor in its historical approved-grant record. Exact contracting entity and payment execution beyond that listing are not obtained.", "x": 1674.1, "y": 1330.2, "degree": 1 }, { "id": "w31-fa-democracy", "label": "Democracy Fund, Inc.", "group": "funding-authority", "role": "Organisation", "summary": "EIN38-3926408; donor-published2023/2024 Form990-PF schedules name CMA. Reader column alignment limits precise amounts, and recipient receipts were not read.", "x": 1649.4, "y": 1340.7, "degree": 2 }, { "id": "w31-fa-kr", "label": "KR Foundation", "group": "funding-authority", "role": "Organisation", "summary": "Danish grantmaker with published grant-specific climate/election purpose, terms and a board-governed selection structure. Funding does not establish approval of CMA publications.", "x": 1689.2, "y": 1283.1, "degree": 4 }, { "id": "w31-fa-krboard", "label": "KR Foundation board", "group": "funding-authority", "role": "Governance body", "summary": "Statutes assign grant distribution decisions and permit limited delegation. Historical first-board names are not a current roster or a reconstructed2023 vote.", "x": 1687.4, "y": 1221.3, "degree": 4 }, { "id": "w31-fa-villum", "label": "VILLUM FONDEN", "group": "funding-authority", "role": "Organisation", "summary": "Statutes name an appointment right for up to two of KR Foundation's five-to-seven directors. No individual grant veto or donor-to-CMA editorial instruction inferred.", "x": 1677.9, "y": 1177.5, "degree": 1 }, { "id": "w31-fa-krsign", "label": "KR Foundation authorized joint signatories / delegates", "group": "funding-authority", "role": "Decision-making role", "summary": "Statutory signing authority can bind the foundation. The document does not identify signatories to any particular CMA award.", "x": 1675.3, "y": 1242.0, "degree": 1 }, { "id": "w31-fa-mix2024", "label": "CMA reported funding mix — 2024", "group": "funding-authority", "role": "Research resource", "summary": "Annual chart: major donors59, foundations30, individuals11, without explicit percent sign, dollar denominator or accounting basis. Separate individual-donor chart is not share of total funding.", "x": 1670.3, "y": 1335.1, "degree": 1 }, { "id": "w31-fa-mix2025", "label": "CMA reported funding mix — 2025", "group": "funding-authority", "role": "Research resource", "summary": "Annual report explicitly gives foundations82.8%, major donors12.3%, individuals4.9%; dollar totals, definitions and accounting basis not supplied. Changes do not identify cause or each donor.", "x": 1660.5, "y": 1334.0, "degree": 1 }, { "id": "w31-fa-cmaboard2024", "label": "CMA board — 2024 reported roster", "group": "funding-authority", "role": "Governance body", "summary": "Historical annual-report board; founders Claire Atkin and Nandini Jammi are co-chairs. Roster is not proof of donor appointment, internal voting rules or current office.", "x": 1653.5, "y": 1299.6, "degree": 3 }, { "id": "w31-fa-atkin", "label": "Claire Atkin", "group": "funding-authority", "role": "Person", "summary": "CMA founder and co-chair in its2024 annual report; no later office or donor appointment inferred.", "x": 1634.3, "y": 1258.8, "degree": 1 }, { "id": "w31-fa-jammi", "label": "Nandini Jammi", "group": "funding-authority", "role": "Person", "summary": "CMA founder and co-chair in its2024 annual report; no later office or donor appointment inferred.", "x": 1624.6, "y": 1261.6, "degree": 1 }, { "id": "w31-fa-cmasub", "label": "CMA additional-safety response — 2025", "group": "funding-authority", "role": "Publication", "summary": "A31-page Ofcom-hosted response, listed11December2025; exact submission date unestablished. It asks for monetization-related checks and sanctions, not universal viewer identification.", "x": 1850.7, "y": 1357.7, "degree": 8 }, { "id": "w31-fa-creatorchecks", "label": "Creator / advertiser verification — CMA proposal", "group": "funding-authority", "role": "Policy", "summary": "Proposed KYC/KYBC for creators and advertisers concerns monetized participation. No specified identity document, shared database or universal viewer check; other dispositions unverified.", "x": 1882.5, "y": 1320.2, "degree": 1 }, { "id": "w31-fa-premonetization", "label": "Pre-monetization risk clearance — CMA proposal", "group": "funding-authority", "role": "Policy", "summary": "CMA proposes withholding monetization for indications of illegal-content risk until platform assessment clears it, including manual review of possible illegality. This is not a prior court-finding requirement.", "x": 1866.6, "y": 1322.0, "degree": 1 }, { "id": "w31-fa-sanctions", "label": "Monetizer return bans and financial sanctions — CMA proposal", "group": "funding-authority", "role": "Policy", "summary": "Proposed specified-offender/repeat-violator bans, advertiser clawbacks and related sanctions. These remain requested measures; actual closure, refund and adoption are unverified.", "x": 1885.1, "y": 1326.2, "degree": 1 }, { "id": "w31-fa-crisis-money", "label": "Crisis revenue accounting and refunds — CMA proposal", "group": "funding-authority", "role": "Policy", "summary": "Requested crisis monetization changes, revenue information and advertiser refunds where possible. Ofcom expressly declined recommending specific monetization practices in this crisis measure.", "x": 1916.7, "y": 1370.1, "degree": 2 }, { "id": "w31-fa-adincident", "label": "Harmful-ad incident reporting — CMA proposal", "group": "funding-authority", "role": "Policy", "summary": "CMA asks for advertising incident reporting to Ofcom and advertisers without a minimum threshold, using harmful/unsafe language broader than established illegality. No implemented reporting channel established.", "x": 1879.4, "y": 1329.9, "degree": 1 }, { "id": "w31-fa-fullfact", "label": "Full Fact", "group": "funding-authority", "role": "Organisation", "summary": "Consultation respondent seeking broader crisis scope, stronger supervision and meaningful data access for accredited fact-checkers; proposals and Ofcom-attributed input are distinct from granted authority.", "x": 2095.5, "y": 1304.1, "degree": 1 }, { "id": "w31-fa-ffsub", "label": "Full Fact additional-safety response — 17 October 2025", "group": "funding-authority", "role": "Publication", "summary": "Dated20-page response; official-reader text has four marked line omissions. Ofcom later cited itsp17 alongside an earlier DSIT meeting. No Full Fact participation in that meeting inferred.", "x": 2050.4, "y": 1333.1, "degree": 6 }, { "id": "w31-fa-ffscope", "label": "Broader crisis scope / systemic misinformation — Full Fact requests", "group": "funding-authority", "role": "Policy", "summary": "Requests align crisis scope with s175 health/national-security language, propose transparent criteria and systemic-harm legislative review. It does not expressly demand an automatic ministerial activation switch.", "x": 2085.0, "y": 1297.3, "degree": 1 }, { "id": "w31-fa-ffaccess", "label": "Accredited fact-checker data access — Full Fact request", "group": "funding-authority", "role": "Policy", "summary": "Full Fact acknowledges Ofcom cannot mandate provider partnerships, while requesting meaningful data access and consequences for failure to share. This could benefit its institutional category; no entitlement or appointment.", "x": 2088.0, "y": 1302.8, "degree": 1 }, { "id": "w31-fa-ffoversight", "label": "Central coordination, records and public summaries — Full Fact requests", "group": "funding-authority", "role": "Policy", "summary": "Requests Ofcom coordination, government/police/civil channels, start/end notices, evidence logs, stronger oversight and routine analyses/public summaries, with safeguards. Regulator adopted only selected elements.", "x": 2077.8, "y": 1295.4, "degree": 1 }, { "id": "w31-fa-noticechange", "label": "Ministerial-notice indicator — Ofcom-attributed change", "group": "funding-authority", "role": "Policy", "summary": "June2026 statement says it amended recommendations in response to Full Factp17 and questions at an Ofcom/DSIT meeting4August2025. Specific contribution shares, drafting authorship and underlying minute remain not obtained.", "x": 2057.6, "y": 1342.2, "degree": 3 }, { "id": "w31-fa-dsitmeeting", "label": "Ofcom / DSIT meeting — 4 August 2025", "group": "funding-authority", "role": "Event", "summary": "Existence/date and questions are attributed by Ofcom's statement footnote29. Attendees, agenda, content, initiation and approval rights are not established by a minute.", "x": 2021.2, "y": 1366.6, "degree": 2 }, { "id": "w31-fa-protocol", "label": "Ofcom crisis protocol recommendations — June2026", "group": "funding-authority", "role": "Policy", "summary": "Published statement and laid children-code draft are distinct stages. Provider decision criteria, channels and records are specified; final crisis-code commencement is unresolved in records.", "x": 2041.5, "y": 1359.7, "degree": 8 }, { "id": "w31-fa-providers", "label": "Providers within crisis-measure scope", "group": "funding-authority", "role": "Audience", "summary": "A regulated-provider role, not identified deployed systems. Scope and size matter: smaller high-risk services may be in protocol scope while dedicated police-channel scope is for large services.", "x": 2047.9, "y": 1363.7, "degree": 8 }, { "id": "w31-fa-indicators", "label": "External and internal crisis indicators", "group": "funding-authority", "role": "Research resource", "summary": "Examples include police/intelligence, NGOs, contingency bodies, news, other services, trusted flaggers and internal reports. These are possible inputs, not a named actor's automatic trigger or a shared compulsory database.", "x": 2093.2, "y": 1334.5, "degree": 1 }, { "id": "w31-fa-team", "label": "Provider senior cross-functional crisis team", "group": "funding-authority", "role": "Decision-making role", "summary": "The recommendation places a sufficiently senior cross-functional team inside the provider's process. Actual people, delegated limits and a deployment are not obtained.", "x": 2097.1, "y": 1340.4, "degree": 1 }, { "id": "w31-fa-secretary", "label": "Secretary of State — OSA section175 role", "group": "funding-authority", "role": "Decision-making role", "summary": "Enacted statutory role for directions on reasonable grounds of public-health/safety or national-security threat. Not a named minister, direct lawful-content takedown right or independently verified current commencement.", "x": 2005.3, "y": 1376.8, "degree": 4 }, { "id": "w31-fa-notice", "label": "Section175 public statement / information notice", "group": "funding-authority", "role": "Policy", "summary": "Ofcom notice requires a provider public statement by the specified date; subsection5 permits information needed to respond to the threat. Notice and direction can be varied/revoked; actual worked notice not obtained.", "x": 2016.7, "y": 1375.5, "degree": 3 }, { "id": "w31-fa-lechannel", "label": "Dedicated crisis law-enforcement channel", "group": "funding-authority", "role": "Research resource", "summary": "Recommended for large in-scope services when the provider determines a crisis. Fast incoming information is distinct from routine legal data requests or a general duty to disclose to police.", "x": 2076.7, "y": 1344.1, "degree": 2 }, { "id": "w31-fa-analysis", "label": "Provider post-crisis analysis and retained records", "group": "funding-authority", "role": "Research resource", "summary": "Statement and draft specify records/analysis; no universal public ledger or automatic submission of every analysis adopted. Ofcom may request records. The90-day analysis/protocol language is not an individual restoration guarantee.", "x": 2020.6, "y": 1381.3, "degree": 3 }, { "id": "w31-fa-safeguards", "label": "Existing moderation appeals, expression and privacy safeguards", "group": "funding-authority", "role": "Policy", "summary": "The crisis measure retains existing safeguards rather than creating a freestanding universal new remedy. Ofcom acknowledges potential false positives; no actual over-removal case or repair result acquired.", "x": 2082.8, "y": 1329.0, "degree": 1 }, { "id": "w31-fa-childdraft", "label": "Laid children-code crisis amendments — June2026 draft", "group": "funding-authority", "role": "Policy", "summary": "Submitted12June and laid18June2026. Blank signatures and bracketed commencement remain. C11.7 sets analysis timing, distinct from the statement's protocol-end explanation.", "x": 2012.8, "y": 1390.9, "degree": 2 }, { "id": "w31-fa-planning", "label": "Initial Ofcom crisis planning document — existence confirmed", "group": "funding-authority", "role": "Research resource", "summary": "FOI response says initial planning began23October2024 after17October parliamentary and22October Secretary correspondence. Underlying plan and correspondence were not acquired.", "x": 1996.5, "y": 1370.9, "degree": 1 }, { "id": "w31-fa-approval", "label": "Crisis recommendations approval — 21–26 May2026 window", "group": "funding-authority", "role": "Event", "summary": "Ofcom FOI response reports final approval in this range. The approval record, named approver and clearance conditions are not retained.", "x": 2006.6, "y": 1385.9, "degree": 1 }, { "id": "w31-fa-nciimeasure", "label": "NCII hash-matching measure — separate September2026 status", "group": "funding-authority", "role": "Policy", "summary": "Official September9 sources give30September2026 for a distinct NCII measure. This is not crisis-code commencement, an exclusive StopNCII procurement requirement or evidence of implementation by a named service.", "x": 1982.7, "y": 1373.5, "degree": 1 }, { "id": "w31-fa-statewatch", "label": "Statewatch", "group": "funding-authority", "role": "Organisation", "summary": "Named2025 Oak grantee; endorsed a2023 EUCSA objection and a2025 letter seeking OSA/protest-expression safeguards. Earlier advocacy is not caused by a later grant.", "x": 1779.1, "y": 1512.9, "degree": 3 }, { "id": "w31-fa-dff", "label": "Digital Freedom Fund", "group": "funding-authority", "role": "Organisation", "summary": "Oak2025 report lists three-year core support. DFF later reports a June2026 grant to Liberty for a separate Investigatory Powers Act TCN challenge; Oak-to-litigation allocation is untraced.", "x": 1720.6, "y": 1673.5, "degree": 4 }, { "id": "w31-fa-ihr", "label": "Oak International Human Rights programme", "group": "funding-authority", "role": "Programme", "summary": "Programme strategy supports plural media, reliable information, accountable digital space and rights defenders/intermediaries. Shared strategy does not establish identical recipient positions or donor approval of particular campaigns.", "x": 1635.1, "y": 1555.8, "degree": 1 }, { "id": "w31-fa-swcsa", "label": "EU CSA objection endorsed by Statewatch — 2023", "group": "funding-authority", "role": "Publication", "summary": "Statewatch published the EDRi-coordinated statement13September2023 opposing the proposal without stronger rights/security safeguards. This is a dated position on a then-current proposal, not the same final text as every other statement.", "x": 1752.7, "y": 1550.5, "degree": 2 }, { "id": "w31-fa-swletter", "label": "OSA protest-expression safeguards letter — August2025", "group": "funding-authority", "role": "Publication", "summary": "ORG-letterhead letter to Ofcom/technology companies, published by Statewatch4August2025; Statewatch signatory visible onPDF5. Requests, legal allegations, receipt and actual corrective outcomes are distinct.", "x": 1874.0, "y": 1456.5, "degree": 2 }, { "id": "w31-fa-liberty", "label": "Liberty — DFF grant recipient", "group": "funding-authority", "role": "Organisation", "summary": "Principal recipient and repeated narrative in DFF case/announcement identify Liberty. A trailing Women's Link Worldwide field is inconsistent; no second recipient inferred.", "x": 1664.6, "y": 1626.7, "degree": 6 }, { "id": "w31-fa-dffround", "label": "DFF June2026 grant round", "group": "funding-authority", "role": "Research resource", "summary": "DFF's18August2026 announcement reports16 grants approved in June, EUR500000 aggregate. Not a Liberty allocation, payout proof or a statement that every2025 decision was peer-selected.", "x": 1759.7, "y": 1759.1, "degree": 3 }, { "id": "w31-fa-peergroup", "label": "DFF external Community Peer Group — June2026 selection role", "group": "funding-authority", "role": "Governance body", "summary": "DFF attributes review and selection of the June2026 round to an external peer group. Its charter, historic adoption timing and reserved rights belong to a separate excluded governance lane.", "x": 1774.2, "y": 1806.4, "degree": 1 }, { "id": "w31-fa-tcn", "label": "Liberty TCN-regime challenge — DFF account", "group": "funding-authority", "role": "Event", "summary": "Grantor describes an ongoing challenge concerning Investigatory Powers Act2016 technical capability notice powers and encryption rights. No actual notice, pleading, court disposition or payment was in the accepted two-record DFF source wave; later Liberty proceedings are outside this candidate. Separate from OSA/age assurance.", "x": 1618.1, "y": 1642.6, "degree": 1 }, { "id": "w32-gdi-peer", "label": "DFF Community Peer Group", "group": "gdi-rights", "role": "Body", "summary": "DFF describes peers selecting grants within an institutional process and available budget. Actual appointment, recusal terms, legal binding force and payment authority are not disclosed. Member affiliations are not reserved corporate seats.", "x": 1786.3, "y": 1818.7, "degree": 4 }, { "id": "w32-gdi-dffboard", "label": "DFF board: adoption role", "group": "gdi-rights", "role": "Body", "summary": "DFF reports board adoption of the participatory grant-decision process in March2026. Resolution and award-by-award rights were not acquired.", "x": 1762.7, "y": 1805.9, "degree": 1 }, { "id": "w32-gdi-dffstaff", "label": "DFF grant-team facilitation", "group": "gdi-rights", "role": "Decision-making role", "summary": "Staff supply earlier-grantee information, assessments and budgets and facilitate discussion. No exclusive final selection or payment-signature power inferred.", "x": 1803.8, "y": 1866.7, "degree": 1 }, { "id": "w32-gdi-peerterms", "label": "Peer Group Terms of Reference", "group": "gdi-rights", "role": "Policy", "summary": "Current page says members sign confidentiality/conflict terms and are compensated. Actual terms, recusal rules, compensation amounts and appointment/removal provisions remain not obtained.", "x": 1795.4, "y": 1867.8, "degree": 1 }, { "id": "w32-gdi-peerprocess", "label": "DFF participatory grant-selection process", "group": "gdi-rights", "role": "Programme", "summary": "2024 proposal began with pre-litigation recommendations; DFF reports wider 2025 selection and March2026 board adoption. Strategy/criteria changes remained future work.", "x": 1754.7, "y": 1759.5, "degree": 3 }, { "id": "w32-gdi-pi", "label": "Privacy International", "group": "gdi-rights", "role": "Organisation", "summary": "Liberty names Privacy International as co-complainant in March2025; tribunal acknowledges its written arguments and separate complaint. Oak2023 lists core support; exact funding of this case remains unknown.", "x": 1648.3, "y": 1595.4, "degree": 3 }, { "id": "w32-gdi-org", "label": "Open Rights Group", "group": "gdi-rights", "role": "Organisation", "summary": "Oak2023 lists three-year support for work with refugee/migrant sectors concerning data and immigration technologies. No earmark to the later TCN case or OSA letter inferred.", "x": 1641.8, "y": 1560.3, "degree": 1 }, { "id": "w32-gdi-complaint", "label": "Liberty / Privacy International complaint, 13 March 2025", "group": "gdi-rights", "role": "Event", "summary": "Recipient account and tribunal paragraph33 establish a separate complaint filed before the identified June2026 DFF grant. Notice and original pleading not acquired; filing is not a merits victory or proof of initial financing.", "x": 1629.5, "y": 1623.3, "degree": 2 }, { "id": "w32-gdi-ipt", "label": "Investigatory Powers Tribunal", "group": "gdi-rights", "role": "Tribunal", "summary": "Issued the April2025 public procedural ruling. The judgment is a summary/extract of a fuller private judgment; no current docket or finality audit.", "x": 1557.3, "y": 1652.4, "degree": 3 }, { "id": "w32-gdi-judgment", "label": "Apple v Home Secretary: 7 April 2025 public judgment", "group": "gdi-rights", "role": "Publication", "summary": "[2025] UKIPTrib1, IPT/25/68/CH. Dismissed request to conceal basic case details, while keeping March hearing private and leaving other participation/disclosure issues unresolved. Does not confirm reported notice or decide TCN legality.", "x": 1647.5, "y": 1617.6, "degree": 4 }, { "id": "w32-gdi-respondent", "label": "Secretary of State for the Home Department: 2025 respondent", "group": "gdi-rights", "role": "Decision-making role", "summary": "Named respondent in Apple procedural judgment. This litigation role is distinct from the existing generic OSA section175 ministerial role.", "x": 1444.7, "y": 1698.2, "degree": 1 }, { "id": "w32-gdi-secrecyapplication", "label": "Application to conceal basic Apple case details", "group": "gdi-rights", "role": "Procedural application", "summary": "The tribunal rejected the government application. The ruling did not make every future hearing or pleading public.", "x": 1493.1, "y": 1679.6, "degree": 2 }, { "id": "w32-gdi-hearing", "label": "Private tribunal hearing, 14 March 2025", "group": "gdi-rights", "role": "Event", "summary": "March hearing remained private despite written Liberty/PI arguments; future attendance, intervention and disclosure were not settled by this ruling.", "x": 1510.1, "y": 1668.0, "degree": 1 }, { "id": "w32-gdi-leighday", "label": "Leigh Day", "group": "gdi-rights", "role": "Organisation", "summary": "Liberty account says the organisations instructed Leigh Day. No fee, retainer or grant-funded allocation was inspected.", "x": 1625.4, "y": 1648.7, "degree": 1 }, { "id": "w32-gdi-gdi", "label": "Global Disinformation Index: named operation", "group": "gdi-rights", "role": "Organisation", "summary": "Public-facing GDI operation named in grant and product records. The UK website operator and German Horizon participant are separately represented; a generic award/supplier name is not silently assigned to either legal entity.", "x": 1487.5, "y": 1633.8, "degree": 18 }, { "id": "w32-gdi-gdiuk", "label": "Disinformation Index Ltd (UK)", "group": "gdi-rights", "role": "Organisation", "summary": "Current privacy text identifies company11297397, limited by guarantee in England, as website operator/controller. Ownership of the separately named German participant and legal recipient of every historical award are unestablished.", "x": 1413.0, "y": 1626.2, "degree": 2 }, { "id": "w32-gdi-del", "label": "GDI Dynamic Exclusion List: historical licensing account", "group": "gdi-rights", "role": "Research resource", "summary": "Oak describes a licensed classification product used as an advertiser input. Exact historical licence, reviewer rules and current distribution are not disclosed by that account.", "x": 1638.0, "y": 1570.5, "degree": 1 }, { "id": "w32-gdi-oracle", "label": "Oracle", "group": "gdi-rights", "role": "Organisation", "summary": "Supplier published the2024 GDI segment guide and later Advertising end-of-life FAQ. Company policy is not a tested shutdown or GDI-specific continuing licence.", "x": 1388.1, "y": 1630.3, "degree": 3 }, { "id": "w32-gdi-segment", "label": "Oracle GDI segment gi_gdi_pfi", "group": "gdi-rights", "role": "Research resource", "summary": "2024 guide packages GDI high-risk news-site classifications. Advertiser selection/defaults and lost revenue unverified; Contextual Intelligence later included in announced30September2024 service end.", "x": 1426.1, "y": 1632.8, "degree": 4 }, { "id": "w32-gdi-buyers", "label": "Advertising licensees and buyers", "group": "gdi-rights", "role": "Decision-making role", "summary": "Generic recipient role: product users can select contextual exclusions and GDI says licensees retain final purchasing/data-use decisions. No named advertiser activation is established.", "x": 1427.0, "y": 1627.2, "degree": 4 }, { "id": "w32-gdi-publishers", "label": "Affected publishers: described review standing", "group": "gdi-rights", "role": "Decision-making role", "summary": "GDI2024 parliamentary evidence describes publisher ratings appeals. No sampled publisher case, successful appeal or recipient-restoration outcome acquired.", "x": 1360.0, "y": 1614.7, "degree": 2 }, { "id": "w32-gdi-ratingreview", "label": "GDI publisher rating review: 2024 description", "group": "gdi-rights", "role": "Policy", "summary": "FON0071 describes a linked review form, two-to-three-week review and reassessment. Form, operative criteria and current service guarantee were not acquired.", "x": 1414.5, "y": 1619.5, "degree": 2 }, { "id": "w32-gdi-researchcomplaints", "label": "GDI research-output representations: April2026 policy", "group": "gdi-rights", "role": "Policy", "summary": "Policy accepts concerns about research outputs and qualified review/clarification/update. Personal-data erasure rights are not automatic domain-rating deletion or recipient-cache replacement duties.", "x": 1358.4, "y": 1623.4, "degree": 2 }, { "id": "w32-gdi-ias", "label": "Integral Ad Science", "group": "gdi-rights", "role": "Organisation", "summary": "IAS describes its own Context Control Avoidance product and discontinuation of limited GDI input. Current wording has December2025 publication and July2026 modification dates; exact cutover unknown.", "x": 1370.0, "y": 1606.4, "degree": 1 }, { "id": "w32-gdi-iasproduct", "label": "IAS Context Control Avoidance", "group": "gdi-rights", "role": "Research resource", "summary": "Customer-selected contextual exclusions; IAS says it discontinued limited GDI input and past use will not affect this solution going forward. Not tested removal or a publisher appeal result.", "x": 1421.3, "y": 1616.2, "degree": 3 }, { "id": "w32-gdi-ohmd", "label": "Online Harms Monetisation Dashboard: stated plan", "group": "gdi-rights", "role": "Research resource", "summary": "Current GDI About page describes what this product will estimate. No deployed dashboard, live output or equivalence with AI4TRUST decision support is established.", "x": 1440.1, "y": 1643.8, "degree": 1 }, { "id": "w32-gdi-fcdo", "label": "Foreign, Commonwealth and Development Office", "group": "gdi-rights", "role": "Government", "summary": "Donor answer describes GDI funding outside the US and expected reports/reviews. Selected ledger row and parent programme summary do not identify executed GDI grant/IP terms.", "x": 1433.5, "y": 1639.7, "degree": 5 }, { "id": "w32-gdi-cabinet", "label": "UK Cabinet Office: CDMD programme partner", "group": "gdi-rights", "role": "Government", "summary": "2022-2023 retrospective programme summary names Cabinet Office alongside FCDO. No GDI-specific selection, payment or rating instruction inferred.", "x": 1359.4, "y": 1641.2, "degree": 1 }, { "id": "w32-gdi-cdmd", "label": "Counter Disinformation and Media Development Programme", "group": "gdi-rights", "role": "Programme", "summary": "CSSF-05-000006, April2016-March2023 programme. Published2025 summary describes regional/component aims and budget allocations without naming GDI. Programme objectives are not delivered outcomes.", "x": 1414.3, "y": 1637.5, "degree": 6 }, { "id": "w32-gdi-gdide", "label": "GDI Global Disinformation Index gUG (German participant)", "group": "gdi-rights", "role": "Organisation", "summary": "CORDIS names GDI GLOBAL DISINFORMATION INDEX GUGHAFTUNGSBESCHRANKT as funded German participant. No ownership/IP-transfer relationship with UK Disinformation Index Ltd, or dated cash ledger, was acquired.", "x": 1595.3, "y": 1773.4, "degree": 2 }, { "id": "w32-gdi-ai4trust", "label": "AI4TRUST, grant 101070190", "group": "gdi-rights", "role": "Programme", "summary": "Horizon research action, EC signature6December2022,1January2023-28February2026, CORDIS closed status as updated17June2026. Identified GDI project overlapping the unnamed2024 Horizon disclosure, not an explicit naming of that disclosure or proof of the only GDI Horizon grant.", "x": 1534.8, "y": 1768.6, "degree": 9 }, { "id": "w32-gdi-fbk", "label": "Fondazione Bruno Kessler: AI4TRUST coordinator", "group": "gdi-rights", "role": "Organisation", "summary": "CORDIS names coordinator; D2.1 modification history records FBK coordinator reviews. Coordination is not proof of Commission acceptance, ownership of all results or unilateral recipient-use authority.", "x": 1478.9, "y": 1755.6, "degree": 2 }, { "id": "w32-gdi-ucamreview", "label": "UCAM: D2.1 quality-assurance role", "group": "gdi-rights", "role": "Decision-making role", "summary": "D2.1 modification table records designated quality-manager review by UCAM on27October2023. Named document role, not a newly institution-wide control or approval charter.", "x": 1410.8, "y": 1731.0, "degree": 1 }, { "id": "w32-gdi-d21", "label": "AI4TRUST D2.1 methodological toolbox", "group": "gdi-rights", "role": "Publication", "summary": "Submitted1November2023; selected design pages, internal review history and later selected seed/annex reading. Sampling, annotation, inference and action remain distinct; no full90page or complete-annex audit.", "x": 1464.4, "y": 1718.7, "degree": 6 }, { "id": "w32-gdi-seeds", "label": "AI4TRUST retrieval seeds and sampling frame", "group": "gdi-rights", "role": "Research resource", "summary": "D2.1 blends research/media/fact-checker input, existing GDI lists and some ChatGPT suggestions. Climate, health and migration terms shape collection; no per-term attribution or falsehood verdict is supplied.", "x": 1411.7, "y": 1737.1, "degree": 1 }, { "id": "w32-gdi-framing", "label": "D2.1 intersectional methodological framing", "group": "gdi-rights", "role": "Policy", "summary": "Selected pages58-59 state attention to harm, discrimination and social identity. This is a design choice; its effect on annotations/thresholds requires rule-level evidence, not an assumed suppression finding.", "x": 1416.2, "y": 1742.2, "degree": 1 }, { "id": "w32-gdi-gdidss", "label": "GDI-reported AI4TRUST decision-support contribution", "group": "gdi-rights", "role": "Research resource", "summary": "GDI April2026 retrospective describes a decision-support system within the Monitoring Dashboard, using consortium fact-checker annotations and combined detection signals for human review. Exact crosswalk to commercial products and other component names is not acquired.", "x": 1411.9, "y": 1645.3, "degree": 2 }, { "id": "w32-gdi-factcheckers", "label": "AI4TRUST consortium fact-checker annotations", "group": "gdi-rights", "role": "Research resource", "summary": "GDI attributes training annotations to consortium fact-checkers. No annotation codebook, adjudication/override policy or sample correctness audit was acquired.", "x": 1357.2, "y": 1653.9, "degree": 1 }, { "id": "w32-gdi-accessrequests", "label": "GDI AI4TRUST research-access recommendations", "group": "gdi-rights", "role": "Policy", "summary": "April2026 article asks faster Article40 vetting, mandatory API compatibility with approved third-party detection and audit access. These are requests, not a grant of rights or confirmed platform compliance.", "x": 1443.4, "y": 1650.8, "degree": 2 }, { "id": "w32-gdi-platforms", "label": "Very large online platforms: research-access counterpart role", "group": "gdi-rights", "role": "Decision-making role", "summary": "GDI reports access refusals and proposes compatibility/audit obligations. Applications, decisions and adopted obligations are outside these accepted records; no named platform refusal is independently established here.", "x": 1436.3, "y": 1651.3, "degree": 2 }, { "id": "w32-gdi-warning", "label": "AI4TRUST Disinformation Warning System: reported final method", "group": "gdi-rights", "role": "Research resource", "summary": "Final-period report describes six-tool outputs combined with LightGBM classifiers and isotonic calibration for triage/feature attribution. Exact inputs, thresholds, approvals, aggregate validation and downstream action rights remain not obtained.", "x": 1490.6, "y": 1795.5, "degree": 1 }, { "id": "w32-gdi-textsignals", "label": "AI4TRUST textual disinformation-signals component", "group": "gdi-rights", "role": "Research resource", "summary": "Project reports39signals/six tactics/eight languages and about56percent weighted F1 for this component. It is not final warning-system accuracy, falsehood probability or measured advertising impact.", "x": 1484.9, "y": 1792.8, "degree": 1 }, { "id": "w32-gdi-validation", "label": "AI4TRUST Human Validation Dashboard", "group": "gdi-rights", "role": "Research resource", "summary": "Final report names enhanced human validation and iterative evaluation. The component name does not establish human veto, score replacement, public appeal or correction propagation.", "x": 1494.3, "y": 1800.2, "degree": 1 }, { "id": "w32-gdi-evaluation", "label": "AI4TRUST reported user evaluation", "group": "gdi-rights", "role": "Event", "summary": "Final-period report describes three rounds with206users from professional and student populations. Reported participation is not independent model performance or proof of production adoption.", "x": 1483.8, "y": 1787.3, "degree": 1 }, { "id": "w33-con-dsa", "label": "DSA Article40: distinct research-access channels", "group": "access-assurance", "role": "Policy", "summary": "Adopted eligibility and access duties. Public-data40(12), vetted40(4) and trusted-flagger22 are separate channels; no GDI qualification inferred.", "x": 1632.4, "y": 1861.1, "degree": 1 }, { "id": "w33-con-delegated", "label": "2025/2050: project-specific access procedure", "group": "access-assurance", "role": "Policy", "summary": "Operative prerequisites, modality, overview, amendment and delivery-notification rules;80working days is a decision clock with justified extension.", "x": 1648.3, "y": 1819.6, "degree": 2 }, { "id": "w33-con-estdsc", "label": "Provider-establishment Digital Services Coordinator", "group": "access-assurance", "role": "Statutory role", "summary": "Final project-specific vetting/reasoned-request authority, including access conditions and termination after response opportunity. No named actual vetting decision acquired.", "x": 1563.4, "y": 1922.4, "degree": 10 }, { "id": "w33-con-homedsc", "label": "Research-organisation home DSC — initial assessment", "group": "access-assurance", "role": "Statutory role", "summary": "May receive application, initially assess and forward; provider-establishment DSC retains final decision.", "x": 1540.4, "y": 1960.0, "degree": 1 }, { "id": "w33-con-vetted", "label": "Vetted researchers — project-specific role", "group": "access-assurance", "role": "Statutory role", "summary": "Eligibility includes affiliation, commercial independence, disclosed funding, security, necessity/proportionality and free public results. No named appointment here.", "x": 1534.9, "y": 1947.7, "degree": 4 }, { "id": "w33-con-providers", "label": "DSA designated VLOP/VLOSE data-provider role", "group": "access-assurance", "role": "Statutory role", "summary": "Bounded statutory provider duties; this role does not assert that every platform in the wider case has a particular designation.", "x": 1540.0, "y": 1926.7, "degree": 5 }, { "id": "w33-con-request", "label": "Reasoned data-access request and conditions", "group": "access-assurance", "role": "Administrative instrument", "summary": "Specified research, modalities, access-start and termination dates. No actual applicant request acquired; overview and delivery are distinct records.", "x": 1532.6, "y": 1956.3, "degree": 3 }, { "id": "w33-con-portal", "label": "DSA data-access portal — public overview channel", "group": "access-assurance", "role": "Infrastructure", "summary": "Required public overviews do not necessarily name researchers/funders. Retained shell yielded zero overview rows; no absence inference.", "x": 1615.4, "y": 1846.9, "degree": 2 }, { "id": "w33-con-publicresearch", "label": "Public-data researchers under40(12)", "group": "access-assurance", "role": "Statutory role", "summary": "Qualifying researchers, including nonprofit affiliations, obtain public-data access without the paragraph4 vetting route; purpose and40(8)(b)–(e) conditions remain.", "x": 1559.2, "y": 1885.8, "degree": 2 }, { "id": "w33-con-experts", "label": "Independent experts consulted by a DSC", "group": "access-assurance", "role": "Statutory role", "summary": "Optional expert consultation with impartiality/conflict declarations and AGORA record; no named engagement or donor approval inferred.", "x": 1548.2, "y": 1961.8, "degree": 1 }, { "id": "w33-con-flaggerdsc", "label": "Applicant-establishment DSC — trusted-flagger role", "group": "access-assurance", "role": "Statutory role", "summary": "Separate Article22 awarding/suspension/revocation jurisdiction, not the researcher-provider establishment rule.", "x": 2270.9, "y": 392.2, "degree": 2 }, { "id": "w33-con-flagger", "label": "Article22 trusted-flagger status", "group": "access-assurance", "role": "Statutory status", "summary": "Priority processing of qualified illegal-content notices, with independence/reporting/review obligations; each notice is not a binding removal order.", "x": 2278.5, "y": 368.0, "degree": 3 }, { "id": "w33-con-noticeproviders", "label": "Online-platform providers — Article22 notice role", "group": "access-assurance", "role": "Statutory role", "summary": "Article22 applies to online platforms; distinct from the VLOP/VLOSE Article40 data-provider role.", "x": 2286.2, "y": 343.6, "degree": 1 }, { "id": "w33-con-ali", "label": "AliExpress — service under examined proceedings", "group": "access-assurance", "role": "Service", "summary": "Covered service; commitments address Netherlands and Singapore legal parties. No applicant identity or use of its API is inferred.", "x": 1589.0, "y": 1831.5, "degree": 7 }, { "id": "w33-con-alinl", "label": "AliExpress International (Netherlands) B.V.", "group": "access-assurance", "role": "Company", "summary": "Exact Article3 addressee of the18June2025 decision. No parent/successor identity substitution.", "x": 1647.4, "y": 1797.6, "degree": 1 }, { "id": "w33-con-alisg", "label": "Alibaba.com Singapore E-commerce Private Limited", "group": "access-assurance", "role": "Company", "summary": "Exact decision addressee and reported audit-engaging provider; different from a prospective successor mentioned in procedural history.", "x": 1705.0, "y": 1845.2, "degree": 2 }, { "id": "w33-con-commitments", "label": "AliExpress binding commitments — C(2025)3982", "group": "access-assurance", "role": "Policy", "summary": "Public edited Article71 decision; public-data accessB.V and monitoringC. Receipt-based effective term unknown; access suspicions not adjudicated infringement/non-admission retained.", "x": 1543.9, "y": 1880.6, "degree": 2 }, { "id": "w33-con-ankura", "label": "Ankura Consulting (Europe), Limited", "group": "access-assurance", "role": "Company", "summary": "Commission approval23September2025 establishes named monitoring role; mandate, paid fee and actual delivered reports not obtained.", "x": 1618.7, "y": 1823.4, "degree": 8 }, { "id": "w33-con-legalteam", "label": "AliExpress Legal team — application process role", "group": "access-assurance", "role": "Operational role", "summary": "FTI describes tracked SOP eligibility/clarification decisions managed by Legal during evaluation1July2024–30June2025. No current-policy backdating or own test.", "x": 1786.5, "y": 2021.8, "degree": 2 }, { "id": "w33-con-qa", "label": "AliExpress internal access QA/reporting", "group": "access-assurance", "role": "Operational role", "summary": "B.V.6–7 specifies metrics, quarterly technical tests and semiannual material to trustee. Actual reports and continuing-duration interpretation remain not obtained.", "x": 1592.4, "y": 1852.5, "degree": 2 }, { "id": "w33-con-trustreport", "label": "Ankura monitoring reports and approved mandate — expected records", "group": "access-assurance", "role": "Record family", "summary": "Workplan/initial report/follow-up response are required or expected but not acquired. Private delivery is plausible; absence from main-activity index proves no failure.", "x": 1509.5, "y": 1914.6, "degree": 1 }, { "id": "w33-con-auditrole", "label": "AliExpress Article37 auditor — role", "group": "access-assurance", "role": "Statutory role", "summary": "Role with permitted two-way trustee evidence sharing; not assumed to be FTI for every future audit or the Ankura monitoring role.", "x": 1595.1, "y": 1864.5, "degree": 2 }, { "id": "w33-con-fti", "label": "FTI Consulting LLP —2024/25 audit firm", "group": "access-assurance", "role": "Company", "summary": "Implementation report identifies provider engagement; exact LLP, not generic FTI-branded group. Fees/independence are report declarations, not engagement or payment records.", "x": 1746.6, "y": 1918.2, "degree": 2 }, { "id": "w33-con-ftiaudit", "label": "FTI2024/25 Article40(12) audit opinion", "group": "access-assurance", "role": "Evaluation", "summary": "Positive intention/capability conclusion with no recommendation; no approved applications and all inspected applications judged incomplete/invalid coexist. Evaluation1July2024–30June2025, not laterAnkura performance.", "x": 1780.4, "y": 1981.7, "degree": 3 }, { "id": "w33-con-appcohort", "label": "AliExpress2024/25 public-data applications — redacted audit cohort", "group": "access-assurance", "role": "Reported population", "summary": "Count confidential. FTI says it inspected all and found incomplete/invalid submissions; no originals, identities or cure exchanges available. No GDI or current tracker join.", "x": 1794.9, "y": 2020.4, "degree": 2 }, { "id": "w33-con-collab", "label": "DSA40 Data Access Collaboratory", "group": "access-assurance", "role": "Initiative", "summary": "Author-led research/support initiative gathering voluntary application reports and informing regulation. Not a regulator, random sample or registry of every platform request.", "x": 1513.0, "y": 1712.1, "degree": 6 }, { "id": "w33-con-tracker", "label": "DSA40 voluntary application tracker —8September2026", "group": "access-assurance", "role": "Dataset summary", "summary": "Headline70 complete registrations; actualtable68=24accepted27rejected17open. Homepage61public differs from table60public8nonpublic. Unreconciled; complete registration not legal validity, accepted not delivery.", "x": 1501.1, "y": 1759.8, "degree": 4 }, { "id": "w33-con-aliopen", "label": "One tracked AliExpress public-data request — open", "group": "access-assurance", "role": "Reported application", "summary": "Current anonymous undated entry marked open. No exact paragraph, requested fields, correspondence, GDI identity or FTI-cohort match.", "x": 1531.5, "y": 1804.2, "degree": 2 }, { "id": "w33-con-paper14", "label": "Weizenbaum Policy Paper14 — September2025", "group": "access-assurance", "role": "Publication", "summary": "Author-led network accounts and proposals, including dated40(12) obstacles. Paper cites earlier research rather than furnishing original decision bundles.", "x": 1601.2, "y": 1744.4, "degree": 3 }, { "id": "w33-con-mercator", "label": "Mercator Stiftung — disclosed project funder", "group": "access-assurance", "role": "Organisation", "summary": "Name as printed; exact legal suffix, amount, award agreement, paid funds and reserved approval rights not obtained. No logo-based legal-entity merge.", "x": 1465.3, "y": 1726.9, "degree": 1 }, { "id": "w33-con-weizenbaum", "label": "Weizenbaum Institute — institutional host", "group": "access-assurance", "role": "Organisation", "summary": "Policy paper describes separately funded joint research institute; host funding is not a proved project grant or editorial mandate.", "x": 1671.6, "y": 1762.9, "degree": 3 }, { "id": "w33-con-bmftr", "label": "German BMFTR — disclosed institutional funder", "group": "access-assurance", "role": "Government department", "summary": "German Federal Ministry of Research, Technology and Space as credited inSeptember2025 paper. No specific study allocation or approval right disclosed.", "x": 1720.2, "y": 1766.4, "degree": 1 }, { "id": "w33-con-berlin", "label": "State of Berlin — disclosed institutional funder", "group": "access-assurance", "role": "Government body", "summary": "Separate host-institution funding credit, without amount, programme or study-specific allocation.", "x": 1720.1, "y": 1773.0, "degree": 1 }, { "id": "w33-con-applicants", "label": "Collaboratory participants — voluntary researcher accounts", "group": "access-assurance", "role": "Participant role", "summary": "Researchers/nonprofit organisations supplying accounts and receiving application support; identities of particular tracked records not acquired.", "x": 1477.4, "y": 1742.2, "degree": 4 }, { "id": "w33-con-treasury", "label": "HM Treasury — programme funding/approval role", "group": "access-assurance", "role": "Government department", "summary": "Treasury-facing reporting and intended phase gate are specified; announced funding is not paid supplier revenue or an approval minute.", "x": 1473.3, "y": 1364.3, "degree": 2 }, { "id": "w33-con-oliver", "label": "Oliver Wyman — inventory-identified CDDP evaluator", "group": "access-assurance", "role": "Organisation", "summary": "Department inventory names brand for102384; supplier-name/registration fields redacted in released order. Exact legal contracting entity and execution unresolved.", "x": 1421.1, "y": 1409.4, "degree": 1 }, { "id": "w33-con-evalcontract", "label": "CDDP evaluation contract102384 — released copy", "group": "access-assurance", "role": "Contractual instrument", "summary": "GBP656000excludingVAT tender/contract price, payment after satisfactory deliverables; visible signature fields empty. Order14January2022 vs printed start20December2022/schedules unresolved.", "x": 1399.8, "y": 1402.7, "degree": 3 }, { "id": "w33-con-fundgate", "label": "CDDP phase continuation — intended Treasury gate", "group": "access-assurance", "role": "Decision process", "summary": "Specification makes subsequent funding conditional on feasibility assessment and HMT approval; actual decision, criteria and submitted evaluation not acquired.", "x": 1412.1, "y": 1374.6, "degree": 2 }, { "id": "w33-con-evaloutput", "label": "CDDP final evaluation and acceptance — commissioned records", "group": "access-assurance", "role": "Record family", "summary": "Correct identification of mis/disinformation trends, value-for-money and delivery evaluation are specified. Detailed success/acceptance cells redacted; final report/acceptance not obtained.", "x": 1342.7, "y": 1404.6, "degree": 1 }, { "id": "w33-con-public", "label": "PUBLIC — programme/supplier brand", "group": "access-assurance", "role": "Organisation", "summary": "Reports OSDI and CDDP participation; current controller is PUBLICGROUPINTERNATIONALLIMITED10608507, without backdating historical contractparty.", "x": 1466.2, "y": 1388.8, "degree": 6 }, { "id": "w33-con-publiclegal", "label": "PUBLIC GROUP INTERNATIONAL LIMITED", "group": "access-assurance", "role": "Company", "summary": "Registry10608507; current websitecontroller association established, historical CDDP contracting entity not.", "x": 1409.6, "y": 1400.1, "degree": 1 }, { "id": "w33-con-subapproval", "label": "102548 key-subcontractor approval field — redacted", "group": "access-assurance", "role": "Contractual record", "summary": "Buyer written approval is specified by entry into orderform; names blacked out. Cannot identify PUBLIC, count entities from bars or call these emptyfields.", "x": 1422.7, "y": 1402.1, "degree": 1 }, { "id": "w33-con-osdi", "label": "Online Safety Data Initiative — OSDI", "group": "access-assurance", "role": "Programme", "summary": "Fifteen-month SharedOutcomesFund programme with government,OSTIA,Faculty andPUBLIC. Separate from CDDP andSTCF2; fund-wideGBP200m not projectallocation.", "x": 1496.6, "y": 1407.9, "degree": 7 }, { "id": "w33-con-cdei", "label": "Centre for Data Ethics and Innovation — OSDI role", "group": "access-assurance", "role": "Government body", "summary": "Convened/chairs10-person2021OnlineHarmsExpertGroup for scope, data-sharing and stewardship advice. No procurement/deployment veto acquired.", "x": 1366.8, "y": 1439.3, "degree": 1 }, { "id": "w33-con-expertgroup", "label": "OSDI Online Harms Expert Group — July2021", "group": "access-assurance", "role": "Advisory group", "summary": "Official dated advisory group; government reports input on userexperience alongside industryneeds. Advice dispositions/minutes not acquired.", "x": 1424.3, "y": 1432.4, "degree": 3 }, { "id": "w33-con-longlist", "label": "OSDI23 candidate solutions — June2021 longlist", "group": "access-assurance", "role": "Research output", "summary": "Nearly70discoveryinterviews,industryworkshops andWhitehallroundtable underpin candidate access/taxonomy/benchmarking/procurement responses. Candidates, not23deliveredtools.", "x": 1443.0, "y": 1419.3, "degree": 1 }, { "id": "w33-con-sandbox", "label": "OSDI TRL4 evaluation sandbox — PUBLIC account", "group": "access-assurance", "role": "Prototype", "summary": "Supplier says text suicide/self-harm classifier sandbox, dataset/framework,DPIA,3pilotfirms and scalingrecommendations delivered toDCMS forreview/testing. Productionoperation/acceptance unverified.", "x": 1446.0, "y": 1399.9, "degree": 2 }, { "id": "w33-con-stcf", "label": "Safety Tech Challenge Fund Round2 evaluation —2024", "group": "access-assurance", "role": "Evaluation", "summary": "SeparatePUBLIC-authoredDSITcommission, attributing part ofmethod toOSDI. Boundedtechnicalfindings rely on supplierperformance information and callforindependenttests.", "x": 1447.6, "y": 1405.7, "degree": 2 }, { "id": "w33-con-landscape", "label": "Data Landscape Review2024–2025 — market study", "group": "access-assurance", "role": "Publication", "summary": "PUBLICforDSIT;13interviews28Jan–5Mar2025,9surveyresponses6relevant and55plus sources. Smallpurposivestudy reportspersistentdata/cost/demand/taxonomyfriction; notcausalOSDIevaluation.", "x": 1407.4, "y": 1389.5, "degree": 1 } ], "edges": [ { "id": "e0", "source": "kidron", "target": "5rights", "label": "founded", "status": "documented", "type": "personnel" }, { "id": "e1", "source": "5rights", "target": "ab2273", "label": "Sponsored California adaptation", "status": "documented", "type": "policy" }, { "id": "e2", "source": "kidron", "target": "am84", "label": "was listed as a sponsor of the unmoved amendment", "status": "documented", "type": "policy" }, { "id": "e3", "source": "controlai", "target": "am84", "label": "was credited in Hansard with the shutdown proposal", "status": "reported", "type": "policy" }, { "id": "e4", "source": "collective-shout", "target": "payment-campaign", "label": "Organised open letter", "status": "documented", "type": "narrative" }, { "id": "e5", "source": "ncose", "target": "payment-campaign", "label": "Co-signed through Haley McNamara", "status": "documented", "type": "narrative" }, { "id": "e6", "source": "rouse", "target": "payment-campaign", "label": "Co-signatory", "status": "documented", "type": "narrative" }, { "id": "e7", "source": "tankard-reist", "target": "payment-campaign", "label": "Signed for Collective Shout", "status": "documented", "type": "narrative" }, { "id": "e8", "source": "tankard-reist", "target": "collective-shout", "label": "Movement Director", "status": "documented", "type": "personnel" }, { "id": "e9", "source": "payment-campaign", "target": "payment-processors", "label": "Requested processing withdrawal", "status": "documented", "type": "narrative" }, { "id": "e10", "source": "payment-campaign", "target": "steam", "label": "Targeted platform", "status": "documented", "type": "narrative" }, { "id": "e11", "source": "payment-campaign", "target": "itch-io", "label": "Platform reported campaign effects", "status": "reported", "type": "narrative" }, { "id": "e13", "source": "ncose", "target": "ai-moratorium-letter", "label": "Co-signatory", "status": "documented", "type": "narrative" }, { "id": "e14", "source": "fli", "target": "ai-moratorium-letter", "label": "Co-signatory", "status": "documented", "type": "narrative" }, { "id": "e15", "source": "au-infrastructure", "target": "accs", "label": "Awarded trial procurement", "status": "documented", "type": "money" }, { "id": "e16", "source": "accs", "target": "au-age-trial", "label": "Consortium lead", "status": "documented", "type": "policy" }, { "id": "e17", "source": "au-infrastructure", "target": "au-age-trial", "label": "Commissioned / published study", "status": "documented", "type": "policy" }, { "id": "e18", "source": "trial-board", "target": "au-age-trial", "label": "Stakeholder advice", "status": "documented", "type": "policy" }, { "id": "e19", "source": "rouse", "target": "trial-board", "label": "Chair / member appointment role", "status": "documented", "type": "personnel" }, { "id": "e20", "source": "tankard-reist", "target": "trial-board", "label": "Listed stakeholder adviser", "status": "documented", "type": "personnel" }, { "id": "e21", "source": "corby", "target": "au-age-trial", "label": "Stakeholder-engagement lead", "status": "documented", "type": "personnel" }, { "id": "e22", "source": "corby", "target": "avpa", "label": "Director / executive director roles", "status": "documented", "type": "personnel" }, { "id": "e23", "source": "corby", "target": "euconsent", "label": "Secretary general", "status": "documented", "type": "personnel" }, { "id": "e24", "source": "innovate-uk", "target": "defai", "label": "Project funding disclosed", "status": "documented", "type": "money" }, { "id": "e25", "source": "avpa", "target": "defai", "label": "Project delivery", "status": "documented", "type": "policy" }, { "id": "e26", "source": "accs", "target": "defai", "label": "Project delivery", "status": "documented", "type": "policy" }, { "id": "e27", "source": "illuminate-tech", "target": "defai", "label": "Ethics / data / outreach advice", "status": "documented", "type": "policy" }, { "id": "e28", "source": "digi", "target": "au-age-trial", "label": "Raised contemporaneous objections", "status": "documented", "type": "policy" }, { "id": "e29", "source": "au-infrastructure", "target": "social-research-centre", "label": "A$278,293.40 consumer research order", "status": "documented", "type": "money" }, { "id": "e30", "source": "inman-grant", "target": "esafety", "label": "Commissioner", "status": "documented", "type": "personnel" }, { "id": "e31", "source": "inman-grant", "target": "weprotect", "label": "Board role", "status": "documented", "type": "personnel" }, { "id": "e32", "source": "inman-grant", "target": "wef", "label": "Digital Safety coalition participation", "status": "documented", "type": "personnel" }, { "id": "e33", "source": "wef", "target": "wef-safety-principles", "label": "Convened principles work", "status": "documented", "type": "policy" }, { "id": "e34", "source": "weprotect", "target": "wef-safety-principles", "label": "Working-group participant", "status": "documented", "type": "policy" }, { "id": "e35", "source": "tech-principles-members", "target": "wef-safety-principles", "label": "Named working-group participants", "status": "documented", "type": "policy" }, { "id": "e36", "source": "ofcom", "target": "wef-safety-principles", "label": "Working-group participant", "status": "documented", "type": "policy" }, { "id": "e37", "source": "esafety", "target": "wef-safety-principles", "label": "Safety by Design framework connection", "status": "documented", "type": "policy" }, { "id": "e38", "source": "nij", "target": "ncose", "label": "US$239,946 research award", "status": "documented", "type": "money" }, { "id": "e39", "source": "jensen-project", "target": "ncose", "label": "Policy grant support; amount unknown", "status": "documented", "type": "money" }, { "id": "e40", "source": "pentecost", "target": "ncose", "label": "Named philanthropic support", "status": "documented", "type": "money" }, { "id": "e41", "source": "collective-shout", "target": "message-scanning-proposal", "label": "advocated message scanning in its submission", "status": "documented", "type": "policy" }, { "id": "e42", "source": "vitalik", "target": "fli", "label": "2021 support; total not established", "status": "documented", "type": "money" }, { "id": "e43", "source": "vitalik", "target": "fli", "label": "Reported disagreement over later strategy", "status": "reported", "type": "policy" }, { "id": "e44", "source": "fli", "target": "lightcone", "label": "$162.59m asset transfer", "status": "documented", "type": "money" }, { "id": "e45", "source": "lightcone", "target": "fli", "label": "$21.05m noncash contribution", "status": "documented", "type": "money" }, { "id": "e46", "source": "fli", "target": "flieurope", "label": "$180.30m asset transfer", "status": "documented", "type": "money" }, { "id": "e47", "source": "fli", "target": "flfoundation", "label": "$25.025m cash transfer", "status": "documented", "type": "money" }, { "id": "e48", "source": "tegmark", "target": "fli", "label": "Board / officer service", "status": "documented", "type": "personnel" }, { "id": "e49", "source": "tegmark", "target": "lightcone", "label": "Board / officer service", "status": "reported", "type": "personnel" }, { "id": "e50", "source": "aguirre", "target": "fli", "label": "Board / officer service", "status": "documented", "type": "personnel" }, { "id": "e51", "source": "aguirre", "target": "lightcone", "label": "Board / officer service", "status": "reported", "type": "personnel" }, { "id": "e52", "source": "chita", "target": "fli", "label": "Board / officer service", "status": "documented", "type": "personnel" }, { "id": "e53", "source": "chita", "target": "lightcone", "label": "Board / officer service", "status": "reported", "type": "personnel" }, { "id": "e54", "source": "tallinn", "target": "fli", "label": "Board member", "status": "documented", "type": "personnel" }, { "id": "e55", "source": "tallinn", "target": "cais", "label": "Board member", "status": "documented", "type": "personnel" }, { "id": "e56", "source": "tallinn", "target": "deepmind", "label": "Former investor-director", "status": "documented", "type": "personnel" }, { "id": "e57", "source": "fli", "target": "flfoundation", "label": "$286,738 expense reimbursement", "status": "documented", "type": "money" }, { "id": "e58", "source": "baif", "target": "fli", "label": "$73,577 expense reimbursement", "status": "documented", "type": "money" }, { "id": "e59", "source": "coefficient", "target": "goodventures", "label": "Primary philanthropic adviser", "status": "documented", "type": "personnel" }, { "id": "e60", "source": "tuna", "target": "goodventures", "label": "Cofounder / board chair", "status": "documented", "type": "personnel" }, { "id": "e61", "source": "moskovitz", "target": "goodventures", "label": "Cofounder / secretary / treasurer", "status": "documented", "type": "personnel" }, { "id": "e62", "source": "makan", "target": "goodventures", "label": "Board member", "status": "documented", "type": "personnel" }, { "id": "e64", "source": "svcf", "target": "cais", "label": "$1,455,200 cash grant", "status": "documented", "type": "money" }, { "id": "e65", "source": "fli", "target": "cais", "label": "$22,000 recommended grant", "status": "reported", "type": "money" }, { "id": "e66", "source": "fli", "target": "cht", "label": "$500,000 recommended grant", "status": "reported", "type": "money" }, { "id": "e67", "source": "hendrycks", "target": "cais", "label": "Executive / research director", "status": "documented", "type": "personnel" }, { "id": "e68", "source": "hendrycks", "target": "caisaction", "label": "President in FY2024 return", "status": "reported", "type": "personnel" }, { "id": "e69", "source": "hendrycks", "target": "xai", "label": "Professional adviser", "status": "documented", "type": "personnel" }, { "id": "e70", "source": "hendrycks", "target": "scale", "label": "Professional adviser", "status": "documented", "type": "personnel" }, { "id": "e71", "source": "fli", "target": "dma", "label": "Operates creator accelerator", "status": "documented", "type": "narrative" }, { "id": "e72", "source": "fli", "target": "pwh", "label": "announced an up-to-$8m campaign envelope", "status": "documented", "type": "money" }, { "id": "e73", "source": "peoplefirst", "target": "pwh", "label": "Recruits campaign creators", "status": "documented", "type": "narrative" }, { "id": "e74", "source": "peoplefirst", "target": "pwhcreators", "label": "Published creator offer / general terms", "status": "documented", "type": "narrative" }, { "id": "e75", "source": "pwhcreators", "target": "pwh", "label": "offered campaign-content work for", "status": "documented", "type": "narrative" }, { "id": "e76", "source": "cais", "target": "frame", "label": "was announced as a placement host by", "status": "reported", "type": "narrative" }, { "id": "e77", "source": "frame", "target": "sara", "label": "announced a prospective CAIS placement for", "status": "reported", "type": "narrative" }, { "id": "e78", "source": "cais", "target": "veronica", "label": "Creator contract reported by Frame", "status": "reported", "type": "narrative" }, { "id": "e79", "source": "cais", "target": "mikey", "label": "Creator contract reported by Frame", "status": "reported", "type": "narrative" }, { "id": "e80", "source": "fli", "target": "mateus", "label": "Creator funding reported by Frame", "status": "reported", "type": "money" }, { "id": "e81", "source": "frame", "target": "mateus", "label": "Fellow outcome reported", "status": "reported", "type": "narrative" }, { "id": "e82", "source": "winga", "target": "frame", "label": "Past mentor", "status": "documented", "type": "personnel" }, { "id": "e83", "source": "winga", "target": "controlai", "label": "Creator Outreach & Systems Lead", "status": "documented", "type": "personnel" }, { "id": "e84", "source": "sfr", "target": "controlai", "label": "UK operating / trading identity", "status": "documented", "type": "personnel" }, { "id": "e85", "source": "miotti", "target": "sfr", "label": ">25–50% votes; appointment rights", "status": "documented", "type": "personnel" }, { "id": "e86", "source": "behrens", "target": "sfr", "label": ">25–50% votes; appointment rights", "status": "documented", "type": "personnel" }, { "id": "e87", "source": "moes", "target": "sfr", "label": "Former director / voting controller", "status": "documented", "type": "personnel" }, { "id": "e88", "source": "miotti", "target": "controlai", "label": "Founder / CEO", "status": "documented", "type": "personnel" }, { "id": "e89", "source": "behrens", "target": "controlai", "label": "Special Projects Lead", "status": "documented", "type": "personnel" }, { "id": "e90", "source": "miotti", "target": "controlaius", "label": "Chief executive / representative", "status": "documented", "type": "personnel" }, { "id": "e91", "source": "leahy", "target": "controlaius", "label": "US Executive Director", "status": "documented", "type": "personnel" }, { "id": "e92", "source": "alfour", "target": "controlai", "label": "Adviser; founding assistance", "status": "documented", "type": "personnel" }, { "id": "e93", "source": "leahy", "target": "conjecture", "label": "Cofounder", "status": "documented", "type": "personnel" }, { "id": "e94", "source": "alfour", "target": "conjecture", "label": "Cofounder", "status": "documented", "type": "personnel" }, { "id": "e95", "source": "conbackers", "target": "conjecture", "label": "Disclosed 2022 venture backing", "status": "documented", "type": "money" }, { "id": "e96", "source": "miotti", "target": "conjecture", "label": "Former policy / strategy leadership", "status": "documented", "type": "personnel" }, { "id": "e97", "source": "behrens", "target": "conjecture", "label": "Operations work in 2025", "status": "documented", "type": "personnel" }, { "id": "e98", "source": "controlaius", "target": "balde", "label": "Commissioned German lobbying", "status": "documented", "type": "policy" }, { "id": "e99", "source": "controlai", "target": "controlcreators", "label": "Paid + unpaid creator collaborations", "status": "reported", "type": "narrative" }, { "id": "e100", "source": "controlai", "target": "controlcreators", "label": "Briefing and script drafting/review", "status": "documented", "type": "narrative" }, { "id": "e101", "source": "controlai", "target": "ukparliament", "label": "Briefings, support requests, drafting", "status": "documented", "type": "policy" }, { "id": "e102", "source": "controlcreators", "target": "ukparliament", "label": "described an intended constituent-pressure pathway to", "status": "documented", "type": "narrative" }, { "id": "e103", "source": "browne", "target": "controlai", "label": "Adviser", "status": "documented", "type": "personnel" }, { "id": "e104", "source": "controlai", "target": "browne", "label": "Emails encouraging parliamentary questions", "status": "documented", "type": "policy" }, { "id": "e105", "source": "controlai", "target": "hunt", "label": "Support acknowledged in Lords debate", "status": "documented", "type": "policy" }, { "id": "e106", "source": "browne", "target": "ukparliament", "label": "Former member of House of Lords", "status": "documented", "type": "personnel" }, { "id": "e107", "source": "miotti", "target": "ukparliament", "label": "Written evidence / professional biography", "status": "documented", "type": "policy" }, { "id": "e108", "source": "controlai", "target": "asiuk", "label": "published proposed ASI prohibition text", "status": "documented", "type": "policy" }, { "id": "e109", "source": "sobel", "target": "ukparliament", "label": "was scheduled to seek leave for an ASI bill in", "status": "documented", "type": "policy" }, { "id": "e110", "source": "controlai", "target": "nc12", "label": "backed the uncalled shutdown proposal", "status": "documented", "type": "policy" }, { "id": "e111", "source": "sobel", "target": "nc12", "label": "sponsored the uncalled Commons proposal", "status": "documented", "type": "policy" }, { "id": "e112", "source": "hunt", "target": "am84", "label": "sponsored the unmoved amendment", "status": "documented", "type": "policy" }, { "id": "e113", "source": "miotti", "target": "middlepowers", "label": "Coauthor", "status": "documented", "type": "policy" }, { "id": "e114", "source": "behrens", "target": "middlepowers", "label": "Coauthor", "status": "documented", "type": "policy" }, { "id": "e115", "source": "alfour", "target": "middlepowers", "label": "Coauthor", "status": "documented", "type": "policy" }, { "id": "e116", "source": "behrens", "target": "talos", "label": "Self-reported membership", "status": "reported", "type": "personnel" }, { "id": "e117", "source": "fli", "target": "talos", "label": "Support disclosed; amount/date unspecified", "status": "reported", "type": "money" }, { "id": "e118", "source": "coefficient", "target": "talos", "label": "Support disclosed; amount/date unspecified", "status": "reported", "type": "money" }, { "id": "e119", "source": "sff", "target": "talos", "label": "Support disclosed; amount/date unspecified", "status": "reported", "type": "money" }, { "id": "e120", "source": "behrens", "target": "cfg", "label": "Former policy researcher", "status": "documented", "type": "personnel" }, { "id": "e121", "source": "cfgfunders", "target": "cfg", "label": "Seed / general support disclosed", "status": "reported", "type": "money" }, { "id": "e122", "source": "behrens", "target": "cdi", "label": "Policy fellow", "status": "documented", "type": "personnel" }, { "id": "e123", "source": "behrens", "target": "oecd", "label": "Prior fiscal-foresight work", "status": "documented", "type": "personnel" }, { "id": "e124", "source": "behrens", "target": "bdi", "label": "Prior industrial-policy work", "status": "documented", "type": "personnel" }, { "id": "e125", "source": "behrens", "target": "eidproposal", "label": "Authored eID design position", "status": "documented", "type": "policy" }, { "id": "e126", "source": "dodds", "target": "fabians", "label": "Co-chair", "status": "documented", "type": "personnel" }, { "id": "e127", "source": "russell", "target": "fabians", "label": "Executive member", "status": "documented", "type": "personnel" }, { "id": "e128", "source": "hedges", "target": "fabians", "label": "Welsh Fabians executive representative", "status": "documented", "type": "personnel" }, { "id": "e129", "source": "asato", "target": "fabians", "label": "Chair, 2012–2014", "status": "documented", "type": "personnel" }, { "id": "e130", "source": "sobel", "target": "fabians", "label": "Executive member, 2011–2012", "status": "documented", "type": "personnel" }, { "id": "e131", "source": "knight", "target": "fabians", "label": "Research advisory panel, 2017", "status": "documented", "type": "personnel" }, { "id": "e132", "source": "lister", "target": "fabians", "label": "Named summit speaker", "status": "documented", "type": "personnel" }, { "id": "e133", "source": "dodds", "target": "controlai", "label": "Listed campaign supporter", "status": "documented", "type": "narrative" }, { "id": "e134", "source": "russell", "target": "controlai", "label": "Listed campaign supporter", "status": "documented", "type": "narrative" }, { "id": "e135", "source": "hedges", "target": "controlai", "label": "Listed campaign supporter", "status": "documented", "type": "narrative" }, { "id": "e136", "source": "asato", "target": "controlai", "label": "Listed campaign supporter", "status": "documented", "type": "narrative" }, { "id": "e137", "source": "sobel", "target": "controlai", "label": "Listed campaign supporter", "status": "documented", "type": "narrative" }, { "id": "e138", "source": "knight", "target": "controlai", "label": "Listed campaign supporter", "status": "documented", "type": "narrative" }, { "id": "e139", "source": "lister", "target": "controlai", "label": "Listed campaign supporter", "status": "documented", "type": "narrative" }, { "id": "e140", "source": "kidron", "target": "controlai", "label": "Listed campaign supporter", "status": "documented", "type": "narrative" }, { "id": "e141", "source": "browne", "target": "wef", "label": "Coauthored republished article", "status": "documented", "type": "personnel" }, { "id": "e142", "source": "browne", "target": "nti", "label": "Vice-chair in 2017 event report", "status": "documented", "type": "personnel" }, { "id": "e143", "source": "nti", "target": "wef", "label": "Co-hosted biological-risk roundtable", "status": "documented", "type": "narrative" }, { "id": "e144", "source": "a16z", "target": "ltf", "label": "$50m across two filed contributions", "status": "documented", "type": "money" }, { "id": "e145", "source": "gregbrockman", "target": "ltf", "label": "$12.5m individual contribution", "status": "documented", "type": "money" }, { "id": "e146", "source": "annabrockman", "target": "ltf", "label": "$12.5m individual contribution", "status": "documented", "type": "money" }, { "id": "e147", "source": "ltf", "target": "othercommittees", "label": "$40m committee-transfer aggregate", "status": "documented", "type": "money" }, { "id": "e148", "source": "anthropic", "target": "publicfirst", "label": "$40m total announced support", "status": "reported", "type": "money" }, { "id": "e149", "source": "ltf", "target": "buildai", "label": "Reported affiliation", "status": "reported", "type": "narrative" }, { "id": "e150", "source": "buildai", "target": "sm4", "label": "Reported intermediary pathway", "status": "reported", "type": "narrative" }, { "id": "e151", "source": "sm4", "target": "pitchcreator", "label": "Reported $5,000 pitch", "status": "reported", "type": "money" }, { "id": "e152", "source": "pierre-omidyar", "target": "omidyar", "label": "co-founded", "status": "documented", "type": "personnel" }, { "id": "e153", "source": "omidyar", "target": "5rights", "label": "$600k paid · 2021", "status": "documented", "type": "money" }, { "id": "e154", "source": "omidyar", "target": "5rights", "label": "$850k paid · 2023", "status": "documented", "type": "money" }, { "id": "e155", "source": "omidyar", "target": "ab2273", "label": "Acknowledged coalition support", "status": "documented", "type": "narrative" }, { "id": "e156", "source": "reset", "target": "ab2273", "label": "Acknowledged coalition support", "status": "documented", "type": "narrative" }, { "id": "e157", "source": "5rights", "target": "capitol-connection", "label": "$50k · paid lobbying contract", "status": "documented", "type": "money" }, { "id": "e158", "source": "capitol-connection", "target": "ab2273", "label": "Lobbying and amendment negotiation", "status": "documented", "type": "policy" }, { "id": "e159", "source": "alan-parker", "target": "oak", "label": "Source of wealth / trustee", "status": "documented", "type": "personnel" }, { "id": "e160", "source": "oak", "target": "5rights", "label": "$2m award / 3 years", "status": "documented", "type": "money" }, { "id": "e161", "source": "oak", "target": "weprotect", "label": "$2.368m award / 3 years", "status": "documented", "type": "money" }, { "id": "e162", "source": "prudence", "target": "5rights", "label": "£745,500 award · 2025–28", "status": "documented", "type": "money" }, { "id": "e163", "source": "rothschild-foundation", "target": "5rights", "label": "£150k award · AI / EdTech", "status": "documented", "type": "money" }, { "id": "e164", "source": "cross-street-films", "target": "kidron", "label": "Declared parliamentary assistance", "status": "documented", "type": "personnel" }, { "id": "e165", "source": "common-sense-media", "target": "kidron", "label": "Declared paid summit travel", "status": "documented", "type": "money" }, { "id": "e166", "source": "home-office", "target": "weprotect", "label": "Historical sole donor · 2021", "status": "documented", "type": "money" }, { "id": "e167", "source": "home-office", "target": "weprotect", "label": "Permanent board seat", "status": "documented", "type": "personnel" }, { "id": "e168", "source": "european-commission", "target": "weprotect", "label": "Board representation", "status": "documented", "type": "personnel" }, { "id": "e169", "source": "european-commission", "target": "weprotect", "label": "Briefed board on EU proposal", "status": "documented", "type": "policy" }, { "id": "e170", "source": "weprotect", "target": "eu-csa-proposal", "label": "Planned accessible advocacy material", "status": "documented", "type": "policy" }, { "id": "e171", "source": "google", "target": "weprotect", "label": "Institutional board representation", "status": "documented", "type": "personnel" }, { "id": "e172", "source": "snap", "target": "weprotect", "label": "Institutional board representation", "status": "documented", "type": "personnel" }, { "id": "e173", "source": "thorn", "target": "weprotect", "label": "Institutional board representation", "status": "documented", "type": "personnel" }, { "id": "e174", "source": "casir", "target": "controlai", "label": "claimed membership of", "status": "reported", "type": "narrative" }, { "id": "e175", "source": "casir", "target": "torchbearer", "label": "claimed membership of", "status": "reported", "type": "narrative" }, { "id": "e176", "source": "casir", "target": "pauseai-us", "label": "claimed membership of", "status": "reported", "type": "narrative" }, { "id": "e177", "source": "casir", "target": "alliance-secure-ai-action", "label": "claimed a partnership with", "status": "reported", "type": "narrative" }, { "id": "e178", "source": "asi-message-grants", "target": "controlai", "label": "used its plan as a reference in the draft grant design", "status": "documented", "type": "policy" }, { "id": "e179", "source": "casir", "target": "asi-message-grants", "label": "proposed eligibility and grant conditions for", "status": "documented", "type": "policy" }, { "id": "e180", "source": "ogara", "target": "longview", "label": "AI Program Officer", "status": "documented", "type": "personnel" }, { "id": "e181", "source": "ogara", "target": "cais", "label": "Affiliation on 2023 paper", "status": "documented", "type": "personnel" }, { "id": "e182", "source": "ogara", "target": "ai-deception-paper", "label": "Coauthor / origin-authentication proposals", "status": "documented", "type": "policy" }, { "id": "e183", "source": "esafety", "target": "weprotect", "label": "Institutional board representation", "status": "documented", "type": "personnel" }, { "id": "e184", "source": "google", "target": "wef-safety-principles", "label": "Named working-group participant", "status": "documented", "type": "policy" }, { "id": "e185", "source": "goodventures", "target": "longview", "label": "$595,426 indexed awards · Jan 2024", "status": "reported", "type": "money" }, { "id": "e186", "source": "goodventures", "target": "longview", "label": "$15.163m indexed award · Oct 2024", "status": "reported", "type": "money" }, { "id": "e187", "source": "hendrycks", "target": "ai-deception-paper", "label": "Coauthor", "status": "documented", "type": "policy" }, { "id": "e188", "source": "fli", "target": "dma", "label": "described planned aggregate support of up to $100k/month for", "status": "documented", "type": "money" }, { "id": "e63a", "source": "goodventures", "target": "cais", "label": "$5,160,000 · Historical grant award", "status": "reported", "type": "money" }, { "id": "e63b", "source": "goodventures", "target": "cais", "label": "$1,433,000 · Historical grant award", "status": "reported", "type": "money" }, { "id": "e63c", "source": "goodventures", "target": "cais", "label": "$4,025,729 · Historical grant award", "status": "reported", "type": "money" }, { "id": "e63d", "source": "goodventures", "target": "cais", "label": "$1,866,559 · Exit grant award", "status": "reported", "type": "money" }, { "id": "n01", "source": "singh", "target": "frame", "label": "Executive director", "status": "documented", "type": "personnel" }, { "id": "n02", "source": "singh", "target": "cais", "label": "Earlier creator campaigns reported", "status": "reported", "type": "narrative" }, { "id": "n03", "source": "austin-chen", "target": "frame", "label": "Adviser; cofounding described", "status": "reported", "type": "personnel" }, { "id": "n04", "source": "austin-chen", "target": "manifund", "label": "Cofounder listed", "status": "documented", "type": "personnel" }, { "id": "n05", "source": "ryan-kidd", "target": "frame", "label": "Adviser listed", "status": "documented", "type": "personnel" }, { "id": "n06", "source": "drew-spartz", "target": "frame", "label": "Adviser listed", "status": "documented", "type": "personnel" }, { "id": "n07", "source": "mozie", "target": "frame", "label": "Recruitment engagement reported", "status": "reported", "type": "infrastructure" }, { "id": "n08", "source": "frame-selection", "target": "frame", "label": "Selection authority reported", "status": "reported", "type": "governance" }, { "id": "n09", "source": "mox", "target": "frame", "label": "In-house incubation reported", "status": "reported", "type": "infrastructure" }, { "id": "n10", "source": "manifund", "target": "mox", "label": "$500,000 grant funding reported", "status": "reported", "type": "money" }, { "id": "n11", "source": "craig-eaif", "target": "mox", "label": "$300,000 combined funding reported", "status": "reported", "type": "money" }, { "id": "n12", "source": "sff", "target": "mox", "label": "$20,000 for Mox via Manifund", "status": "reported", "type": "money" }, { "id": "n13", "source": "manifund", "target": "grantmaking-ai", "label": "Fiscal sponsorship and payouts", "status": "reported", "type": "infrastructure" }, { "id": "n14", "source": "ryan-kidd", "target": "grantmaking-ai", "label": "Initial reviewer named", "status": "reported", "type": "governance" }, { "id": "n15", "source": "frame", "target": "frame-amplifier", "label": "announced an embedded-creator track", "status": "documented", "type": "narrative" }, { "id": "n16", "source": "eclag", "target": "child-safety-on", "label": "Participating campaign coalition", "status": "documented", "type": "narrative" }, { "id": "n17", "source": "ecpat", "target": "eclag", "label": "Steering-group member listed", "status": "documented", "type": "governance" }, { "id": "n18", "source": "eurochild", "target": "eclag", "label": "Steering-group member listed", "status": "documented", "type": "governance" }, { "id": "n19", "source": "missing-children-europe", "target": "eclag", "label": "Steering-group member listed", "status": "documented", "type": "governance" }, { "id": "n20", "source": "iwf", "target": "eclag", "label": "Steering-group member listed", "status": "documented", "type": "governance" }, { "id": "n21", "source": "terre-des-hommes", "target": "eclag", "label": "Steering-group member listed", "status": "documented", "type": "governance" }, { "id": "n22", "source": "thorn", "target": "eclag", "label": "Steering-group member listed", "status": "documented", "type": "governance" }, { "id": "n23", "source": "5rights", "target": "child-safety-on", "label": "Listed campaign supporter", "status": "documented", "type": "narrative" }, { "id": "n24", "source": "weprotect", "target": "child-safety-on", "label": "Campaign launch recorded", "status": "documented", "type": "narrative" }, { "id": "n25", "source": "wp-secretariat", "target": "eu-csa-proposal", "label": "Published supporting consultation", "status": "documented", "type": "policy" }, { "id": "n26", "source": "weprotect", "target": "wp-secretariat", "label": "Alliance Secretariat; authorship scope stated", "status": "documented", "type": "governance" }, { "id": "n27", "source": "wp-coo", "target": "weprotect", "label": "Manages expenditure and assets", "status": "documented", "type": "governance" }, { "id": "n28", "source": "wp-ed", "target": "wp-coo", "label": "Authority over financial management", "status": "documented", "type": "governance" }, { "id": "n29", "source": "wp-coo", "target": "wp-frac", "label": "Accountable to FRAC", "status": "documented", "type": "governance" }, { "id": "n30", "source": "wp-ed", "target": "wp-frac", "label": "Accountable to FRAC", "status": "documented", "type": "governance" }, { "id": "n31", "source": "wp-frac", "target": "wp-boards", "label": "Reports to both boards", "status": "documented", "type": "governance" }, { "id": "n32", "source": "wp-development", "target": "weprotect", "label": "Fundraising advice and support", "status": "documented", "type": "infrastructure" }, { "id": "n33", "source": "home-office", "target": "weprotect", "label": "Briefed board on five-country work", "status": "documented", "type": "policy" }, { "id": "n34", "source": "five-country-working-group", "target": "weprotect", "label": "Subject of Home Office briefing", "status": "documented", "type": "policy" }, { "id": "n35", "source": "pa-consulting", "target": "global-threat-assessment", "label": "Research briefing recorded", "status": "documented", "type": "infrastructure" }, { "id": "n36", "source": "weprotect", "target": "global-threat-assessment", "label": "Board feedback on scope recorded", "status": "documented", "type": "governance" }, { "id": "n37", "source": "stripe-bank-reference", "target": "stripe", "label": "Bank restriction relayed by itch", "status": "reported", "type": "policy" }, { "id": "n38", "source": "stripe", "target": "itch-io", "label": "Adult-content payment restriction reported", "status": "reported", "type": "policy" }, { "id": "n39", "source": "paypal", "target": "itch-io", "label": "Platform payout dependency reported", "status": "reported", "type": "infrastructure" }, { "id": "n40", "source": "itch-io", "target": "itch-reindex", "label": "No Payments condition published", "status": "documented", "type": "policy" }, { "id": "n41", "source": "mastercard", "target": "mc-acquirers", "label": "Network compliance obligations", "status": "documented", "type": "policy" }, { "id": "n42", "source": "mastercard", "target": "mastercard-statement", "label": "Denied directing game restrictions", "status": "reported", "type": "policy" }, { "id": "n43", "source": "cais", "target": "cais-reserve", "label": "$5,259,802 potential-refund reserve", "status": "documented", "type": "money" }, { "id": "n44", "source": "cais", "target": "ftx-estate", "label": "$5.2m resolution reported by debtors", "status": "reported", "type": "money" }, { "id": "n45", "source": "cais", "target": "cais-2024-close", "label": "Refund reserve no longer listed", "status": "documented", "type": "money" }, { "id": "i01", "source": "icmec", "target": "icmec-financial-coalition", "label": "operates", "status": "documented", "type": "governance" }, { "id": "i02", "source": "mastercard", "target": "icmec-financial-coalition", "label": "is listed as a programme donor to", "status": "reported", "type": "money" }, { "id": "i03", "source": "paypal", "target": "icmec-financial-coalition", "label": "is listed as a programme donor to", "status": "reported", "type": "money" }, { "id": "i04", "source": "visa", "target": "icmec-financial-coalition", "label": "is listed as a programme donor to", "status": "reported", "type": "money" }, { "id": "i05", "source": "icmec", "target": "ab1043", "label": "was named as a sponsor of", "status": "documented", "type": "policy" }, { "id": "i06", "source": "children-now", "target": "ab1043", "label": "was named as a sponsor of", "status": "documented", "type": "policy" }, { "id": "i07", "source": "bob-cunningham", "target": "icmec", "label": "was identified as policy-engagement director at", "status": "documented", "type": "personnel" }, { "id": "i08", "source": "icmec", "target": "wicks", "label": "reported collaborating on legislative drafting with", "status": "reported", "type": "coordination" }, { "id": "i09", "source": "wicks", "target": "5rights", "label": "cited a 2021 report by", "status": "documented", "type": "policy" }, { "id": "i10", "source": "icmec", "target": "icmec-age-whitepaper", "label": "hosts the technical proposal", "status": "documented", "type": "policy" }, { "id": "i11", "source": "icmec-age-whitepaper", "target": "api-admission-role", "label": "proposes OS review and approval of website API integrations", "status": "documented", "type": "policy" }, { "id": "i12", "source": "microsoft", "target": "windows-age-signals", "label": "documents planned age-group and verification-status APIs", "status": "documented", "type": "infrastructure" }, { "id": "i13", "source": "windows-age-signals", "target": "microsoft-account", "label": "requires a signed-in user session with", "status": "documented", "type": "infrastructure" }, { "id": "i14", "source": "google", "target": "ab1043", "label": "publicly endorsed", "status": "documented", "type": "policy" }, { "id": "i15", "source": "meta", "target": "ab1043", "label": "publicly endorsed OS and app-store age-assurance allocation in", "status": "documented", "type": "policy" }, { "id": "i16", "source": "discord", "target": "discord-age-assurance", "label": "documents regional methods and limited tests for", "status": "documented", "type": "infrastructure" }, { "id": "i17", "source": "discord-age-assurance", "target": "k-id", "label": "uses verification services from", "status": "reported", "type": "infrastructure" }, { "id": "i18", "source": "k-id", "target": "veratad", "label": "names as an ID-authentication provider", "status": "reported", "type": "infrastructure" }, { "id": "i19", "source": "k-id", "target": "stripe", "label": "routes credit-card verification through", "status": "reported", "type": "infrastructure" }, { "id": "i20", "source": "discord-age-assurance", "target": "google-wallet", "label": "tests receiving age-group and country signals from", "status": "reported", "type": "infrastructure" }, { "id": "i21", "source": "google", "target": "google-wallet", "label": "documents prior passport and face proofing for its ID-pass route", "status": "reported", "type": "infrastructure" }, { "id": "i22", "source": "discord", "target": "discord-age-appeals", "label": "reported possible exposure of government-ID photos in", "status": "reported", "type": "infrastructure" }, { "id": "i23", "source": "5ca", "target": "discord-age-appeals", "label": "disputed the incident attribution concerning", "status": "reported", "type": "infrastructure" }, { "id": "i24", "source": "idscan", "target": "idscan-incident", "label": "acknowledged possible exposure of customer-account data in", "status": "reported", "type": "infrastructure" }, { "id": "i25", "source": "idscan", "target": "veriscan", "label": "describes operating", "status": "reported", "type": "infrastructure" }, { "id": "i26", "source": "planet13", "target": "veriscan", "label": "was described as using", "status": "reported", "type": "infrastructure" }, { "id": "i27", "source": "veriscan", "target": "dutchie", "label": "was described as passing parsed ID data and images into", "status": "reported", "type": "infrastructure" }, { "id": "w1-icmec-e-singapore", "source": "icmec", "target": "w1-icmec-icmec-singapore", "label": "sole member and funds", "status": "documented", "type": "governance" }, { "id": "w1-icmec-e-swissloan", "source": "w1-icmec-icmec-switzerland", "target": "icmec", "label": "provides loans", "status": "documented", "type": "money" }, { "id": "w1-icmec-e-interest-humer", "source": "w1-icmec-humer", "target": "w1-icmec-loans24", "label": "named interested person", "status": "documented", "type": "personnel" }, { "id": "w1-icmec-e-interest-paul", "source": "w1-icmec-paul", "target": "w1-icmec-loans24", "label": "named interested person", "status": "documented", "type": "personnel" }, { "id": "w1-icmec-e-interest-li", "source": "w1-icmec-li", "target": "w1-icmec-loans24", "label": "named interested person", "status": "documented", "type": "personnel" }, { "id": "w1-icmec-e-paypal", "source": "paypal", "target": "w1-icmec-apfc-roundtables", "label": "grant enables", "status": "reported", "type": "money" }, { "id": "w1-icmec-e-auswg", "source": "w1-icmec-apfc-roundtables", "target": "w1-icmec-apfc-australia", "label": "establishes working group", "status": "reported", "type": "policy" }, { "id": "w1-icmec-e-aushome", "source": "w1-icmec-auspaynet", "target": "w1-icmec-apfc-australia", "label": "institutional home", "status": "reported", "type": "policy" }, { "id": "w1-icmec-e-hostmodel", "source": "icmec", "target": "w1-icmec-model", "label": "publishes hosted proposal", "status": "documented", "type": "policy" }, { "id": "w1-icmec-e-wicks-scope", "source": "wicks", "target": "ab1043", "label": "agrees scope narrowing", "status": "reported", "type": "policy" }, { "id": "w1-icmec-e-mpa", "source": "w1-icmec-mpa", "target": "ab1856", "label": "requests design changes", "status": "reported", "type": "policy" }, { "id": "w1-icmec-e-wicks-alternative", "source": "wicks", "target": "ab1856", "label": "chooses alternative clarification", "status": "reported", "type": "policy" }, { "id": "w1-controlai-e-us-payable", "source": "controlaius", "target": "w1-controlai-unnamed-related-org", "label": "reports amount due to", "status": "documented", "type": "money" }, { "id": "w1-controlai-e-miotti-us-officer", "source": "miotti", "target": "controlaius", "label": "reported president treasurer and secretary", "status": "documented", "type": "personnel" }, { "id": "w1-controlai-e-uk-formal-authority", "source": "w1-controlai-uk-board", "target": "sfr", "label": "formal management and budget authority", "status": "documented", "type": "governance" }, { "id": "w1-payments-edge-steam-processor-objection", "source": "w1-payments-steam-processors", "target": "steam", "label": "reportedly cited Rule 5.12.7 after rejecting Valve's response", "status": "reported", "type": "infrastructure" }, { "id": "w1-payments-edge-mastercard-relay-attribution", "source": "mastercard", "target": "w1-payments-steam-processors", "label": "reportedly communicated through processors and their acquiring banks", "status": "reported", "type": "infrastructure" }, { "id": "w1-payments-edge-steam-retirement-choice", "source": "steam", "target": "w1-payments-steam-retirements-202507", "label": "said it retired games to preserve payment methods", "status": "reported", "type": "policy" }, { "id": "w1-payments-edge-itch-payout-review", "source": "itch-io", "target": "w1-payments-itch-payout-eligibility", "label": "described selected account exclusions from collected payouts", "status": "reported", "type": "infrastructure" }, { "id": "w1-payments-edge-itch-payout-terms", "source": "itch-io", "target": "w1-payments-itch-payout-eligibility", "label": "publishes disqualification and final-payout transition terms", "status": "documented", "type": "policy" }, { "id": "w1-payments-edge-itch-geoblocking", "source": "itch-io", "target": "w1-payments-itch-uk-restrictions-202507", "label": "announced tag-based geographic blocks with manual review and appeals", "status": "reported", "type": "policy" }, { "id": "w1-payments-edge-itch-id-proposal", "source": "itch-io", "target": "w1-payments-itch-digital-id-exploration", "label": "announced exploration of unnamed third-party digital-ID providers", "status": "documented", "type": "infrastructure" }, { "id": "w1-icmec-fin-deficit", "source": "icmec", "target": "icmec-2024-accounts", "label": "reports a consolidated net-asset deficit of", "status": "documented", "type": "money" }, { "id": "w1-icmec-fin-restricted", "source": "icmec", "target": "icmec-2024-accounts", "label": "reports programme-restricted net assets of", "status": "documented", "type": "money" }, { "id": "w1-icmec-fin-conditional", "source": "icmec", "target": "icmec-2024-accounts", "label": "reports unrecognized future conditional contributions of", "status": "documented", "type": "money" }, { "id": "w1-icmec-e-going-concern", "source": "icmec", "target": "icmec-financing-dependence", "label": "discloses substantial doubt and dependence on unrestricted fundraising", "status": "documented", "type": "money" }, { "id": "w2-fli-e-lightcone25", "source": "lightcone", "target": "fli", "label": "related party grant received", "status": "documented", "type": "money" }, { "id": "w2-fli-e-flf-receivable", "source": "flfoundation", "target": "fli", "label": "year end amount due under cost sharing", "status": "documented", "type": "money" }, { "id": "w2-fli-e-flf-payable", "source": "fli", "target": "flfoundation", "label": "year end amount due for consulting costs", "status": "documented", "type": "money" }, { "id": "w2-fli-e-board-tallinn", "source": "tallinn", "target": "flieurope", "label": "foundation board member with joint signature by two", "status": "reported", "type": "governance" }, { "id": "w2-fli-e-board-aguirre", "source": "aguirre", "target": "flieurope", "label": "foundation board member with joint signature by two", "status": "reported", "type": "governance" }, { "id": "w2-fli-e-board-chita", "source": "chita", "target": "flieurope", "label": "foundation board member with joint signature by two", "status": "reported", "type": "governance" }, { "id": "w2-fli-e-board-fihn", "source": "w2-fli-fihn", "target": "flieurope", "label": "foundation board member with joint signature by two", "status": "reported", "type": "governance" }, { "id": "w2-fli-e-board-tegmark", "source": "tegmark", "target": "flieurope", "label": "foundation board member with joint signature by two", "status": "reported", "type": "governance" }, { "id": "w2-fli-e-echo", "source": "w2-fli-echo", "target": "flieurope", "label": "appointed auditor", "status": "reported", "type": "governance" }, { "id": "w2-fli-e-supervision", "source": "w2-fli-supervisor", "target": "flieurope", "label": "supervises foundation", "status": "documented", "type": "governance" }, { "id": "w2-fli-e-reservation", "source": "w2-fli-founder-role", "target": "flieurope", "label": "reserved purpose modification right mentioned", "status": "reported", "type": "governance" }, { "id": "w2-fli-e-supporting", "source": "fli", "target": "flfoundation", "label": "amended purpose supports", "status": "documented", "type": "governance" }, { "id": "w1-creator-e-output-fli", "source": "fli", "target": "w1-creator-futuretense-output", "label": "creator acknowledges grant support for output", "status": "documented", "type": "narrative" }, { "id": "w1-creator-e-output-frame", "source": "frame", "target": "w1-creator-futuretense-output", "label": "creator acknowledges grant support for output", "status": "documented", "type": "narrative" }, { "id": "w1-creator-e-output-bluedot", "source": "w1-creator-bluedot", "target": "w1-creator-futuretense-output", "label": "creator acknowledges grant support for output", "status": "documented", "type": "narrative" }, { "id": "w1-creator-e-frame-ltff", "source": "w1-creator-ltff", "target": "frame", "label": "programme names as lead funder", "status": "documented", "type": "narrative" }, { "id": "w1-creator-e-frame-aitof", "source": "w1-creator-aitof", "target": "frame", "label": "programme names as lead funder", "status": "documented", "type": "narrative" }, { "id": "w1-creator-e-rational-partnership", "source": "controlai", "target": "rational", "label": "jointly acknowledged published video collaboration", "status": "documented", "type": "narrative" }, { "id": "w1-creator-e-pf-terms", "source": "peoplefirst", "target": "w1-creator-peoplefirst-terms", "label": "publishes general creator contract terms", "status": "documented", "type": "narrative" }, { "id": "w1-creator-e-pf-dots", "source": "peoplefirst", "target": "w1-creator-dots", "label": "terms name payout provider", "status": "documented", "type": "narrative" }, { "id": "w1-australia-edge-march-objections", "source": "trial-board", "target": "au-age-trial", "label": "recorded objections on sampling, safeguarding and circumvention", "status": "documented", "type": "governance" }, { "id": "w1-australia-edge-scout-recruitment", "source": "w1-australia-scout-insights", "target": "au-age-trial", "label": "said mystery shoppers would come from its registered panel", "status": "reported", "type": "infrastructure" }, { "id": "w1-australia-edge-mou-signed", "source": "esafety", "target": "w1-australia-esafety-mou", "label": "reportedly signed trial impartiality boundary agreement", "status": "reported", "type": "governance" }, { "id": "w1-australia-edge-regulator-draft-input", "source": "esafety", "target": "w1-australia-trial-draft-review", "label": "reportedly provided preliminary-report feedback under the MOU", "status": "reported", "type": "policy" }, { "id": "w1-australia-edge-corby-review-commitment", "source": "corby", "target": "w1-australia-trial-draft-review", "label": "committed to reviewing Parts B–J for consistency", "status": "documented", "type": "governance" }, { "id": "w1-australia-edge-unequal-draft-access", "source": "au-age-trial", "target": "trial-board", "label": "reported withholding detailed vendor sections during legal review", "status": "reported", "type": "governance" }, { "id": "w1-australia-edge-responses-reported", "source": "au-age-trial", "target": "w1-australia-protocol-revisions", "label": "reported amending methodology in response to stakeholder concerns", "status": "reported", "type": "policy" }, { "id": "w2-fli-due-baif", "source": "baif", "target": "fli", "label": "owed year-end payroll and other cost reimbursements", "status": "documented", "type": "money" }, { "id": "w2-fli-due-fli-action-research", "source": "fli-action-research", "target": "fli", "label": "owed year-end payroll and other cost reimbursements", "status": "documented", "type": "money" }, { "id": "w2-fli-ed-board-flieurope", "source": "fli-ed-audit-role", "target": "flieurope", "label": "audit reports overlapping board service", "status": "documented", "type": "personnel" }, { "id": "w2-fli-ed-board-flfoundation", "source": "fli-ed-audit-role", "target": "flfoundation", "label": "audit reports overlapping board service", "status": "documented", "type": "personnel" }, { "id": "w2-fli-ed-board-fli-asia", "source": "fli-ed-audit-role", "target": "fli-asia", "label": "audit reports overlapping board service", "status": "documented", "type": "personnel" }, { "id": "w2-fli-ed-board-lightcone", "source": "fli-ed-audit-role", "target": "lightcone", "label": "audit reports overlapping board service", "status": "documented", "type": "personnel" }, { "id": "w2-fli-ed-board-fli-action-research", "source": "fli-ed-audit-role", "target": "fli-action-research", "label": "audit reports overlapping board service", "status": "documented", "type": "personnel" }, { "id": "w2-fli-ed-board-wise-ancestors", "source": "fli-ed-audit-role", "target": "wise-ancestors", "label": "audit reports overlapping board service", "status": "documented", "type": "personnel" }, { "id": "w2-fli-ed-board-windfall-trust", "source": "fli-ed-audit-role", "target": "windfall-trust", "label": "audit reports overlapping board service", "status": "documented", "type": "personnel" }, { "id": "w2-creator-published-output", "source": "mateus", "target": "w1-creator-futuretense-output", "label": "published the grant-acknowledging video", "status": "documented", "type": "narrative" }, { "id": "w3-5r-e-kidron-portantino", "source": "kidron", "target": "w3-5r-portantino", "label": "asked to retain under-18 coverage; reported amendment work", "status": "documented", "type": "policy" }, { "id": "w3-5r-e-kidron-stern", "source": "kidron", "target": "w3-5r-stern", "label": "sought support against lower age cutoffs", "status": "documented", "type": "policy" }, { "id": "w3-5r-e-background-drafts", "source": "capitol-connection", "target": "w3-5r-mckinley", "label": "reported sending committee-background drafts", "status": "documented", "type": "policy" }, { "id": "w3-5r-e-proposed-amendments", "source": "capitol-connection", "target": "w3-5r-skaggs", "label": "reported sending proposed author amendments", "status": "documented", "type": "policy" }, { "id": "w3-5r-e-comparison-chart", "source": "capitol-connection", "target": "w3-5r-dang", "label": "reported sending UK/California comparison chart", "status": "documented", "type": "policy" }, { "id": "w3-5r-e-governor-amendments", "source": "capitol-connection", "target": "w3-5r-aurre", "label": "reported sharing amendment summaries and opposition-response updates", "status": "documented", "type": "policy" }, { "id": "w3-5r-e-coalition-submission", "source": "capitol-connection", "target": "w3-5r-judiciary", "label": "reported submitting coalition support letter", "status": "documented", "type": "policy" }, { "id": "w3-5r-e-prose-uptake", "source": "w3-5r-coalition-letter", "target": "w3-5r-concurrence-analysis", "label": "distinctive passages reproduced as arguments in support", "status": "documented", "type": "policy" }, { "id": "w3-oak-e-senior-decision", "source": "w3-oak-senior-decision", "target": "oak", "label": "receive recommendations and approve or decline grants", "status": "documented", "type": "governance" }, { "id": "w3-oak-e-nl-uk-control", "source": "w3-oak-wp-nl", "target": "w3-oak-wp-uk", "label": "reports full control of its UK company", "status": "documented", "type": "governance" }, { "id": "w3-oak-e-budget-permission", "source": "w3-oak-wp-nl", "target": "w3-oak-wp-uk", "label": "requires prior written approval for spending outside annual budget", "status": "documented", "type": "governance" }, { "id": "w3-oak-e-underspend", "source": "w3-oak-wp-nl", "target": "w3-oak-wp-uk", "label": "may deduct underspend from a later annual grant", "status": "documented", "type": "governance" }, { "id": "w3-oak-e-subgrant2023", "source": "w3-oak-wp-nl", "target": "w3-oak-wp-uk", "label": "reports 2023 subgrant contribution", "status": "documented", "type": "money" }, { "id": "w3-oak-e-oak-income2023", "source": "oak", "target": "w3-oak-wp-nl", "label": "names Oak in consolidated recognised grant income", "status": "documented", "type": "money" }, { "id": "w3-oak-e-tfg-amendment", "source": "w3-oak-together", "target": "w3-oak-wp-nl", "label": "reports amended Safe Futures commitment via Together for Girls", "status": "documented", "type": "money" }, { "id": "w3-oak-e-swiss-grant", "source": "w3-oak-swiss-philanthropy", "target": "w3-oak-wp-nl", "label": "reports new Safe Futures grant agreement", "status": "documented", "type": "money" }, { "id": "w3-oak-e-budget-approved", "source": "w3-oak-policy-board", "target": "weprotect", "label": "approved 2024 budget", "status": "documented", "type": "governance" }, { "id": "w3-oak-e-oak-hub", "source": "oak", "target": "w3-oak-safe-futures", "label": "acknowledges support for the joint initiative", "status": "documented", "type": "money" }, { "id": "w3-oak-e-hub-wp", "source": "weprotect", "target": "w3-oak-safe-futures", "label": "co-leads joint initiative", "status": "documented", "type": "governance" }, { "id": "w3-oak-e-hub-svri", "source": "w3-oak-svri", "target": "w3-oak-safe-futures", "label": "co-leads joint initiative", "status": "documented", "type": "governance" }, { "id": "w3-oak-e-hub-tfg", "source": "w3-oak-together", "target": "w3-oak-safe-futures", "label": "co-leads joint initiative", "status": "documented", "type": "governance" }, { "id": "w3-oak-e-duch-role", "source": "w3-oak-helena-duch", "target": "oak", "label": "heads Solutions sub-programme within Prevent Child Sexual Abuse", "status": "documented", "type": "personnel" }, { "id": "w3-oak-e-duch-adviser", "source": "w3-oak-helena-duch", "target": "w3-oak-safe-futures", "label": "listed on advisory group", "status": "documented", "type": "governance" }, { "id": "w3-oak-e-review-commission", "source": "w3-oak-safe-futures", "target": "w3-oak-evidence-review", "label": "commissioned evidence review with Oak support", "status": "documented", "type": "resources" }, { "id": "w3-oak-e-duch-input", "source": "w3-oak-helena-duch", "target": "w3-oak-evidence-review", "label": "credited for input shaping the review", "status": "documented", "type": "resources" }, { "id": "w3-oak-e-guidance", "source": "w3-oak-safe-futures", "target": "w3-oak-guidance", "label": "developed Oak-supported guidance through consultation and draft review", "status": "documented", "type": "resources" }, { "id": "w3-oak-e-survey-commission", "source": "weprotect", "target": "w3-oak-eu-survey", "label": "commissioned research for EU-legislation advocacy", "status": "documented", "type": "resources" }, { "id": "w3-oak-e-survey-provider", "source": "w3-oak-economist", "target": "w3-oak-eu-survey", "label": "conducted retrospective survey", "status": "documented", "type": "resources" }, { "id": "w3-oak-e-eu-deployment", "source": "w3-oak-eu-survey", "target": "eu-csa-proposal", "label": "used by WeProtect to support legislative advocacy", "status": "documented", "type": "policy" }, { "id": "w3-opp-e-ltf-thinkbig", "source": "ltf", "target": "w3-opp-thinkbig", "label": "reported committee contribution", "status": "documented", "type": "money" }, { "id": "w3-opp-e-thinkbig-lantern-bores", "source": "w3-opp-thinkbig", "target": "w3-opp-lantern", "label": "reported independent expenditures opposing Bores", "status": "documented", "type": "money" }, { "id": "w3-opp-e-thinkbig-lantern-operations", "source": "w3-opp-thinkbig", "target": "w3-opp-lantern", "label": "reported operating disbursements", "status": "documented", "type": "money" }, { "id": "w3-opp-e-lantern-operation", "source": "w3-opp-lantern", "target": "w3-opp-march-operation", "label": "named vendor for reported production and distribution", "status": "documented", "type": "narrative" }, { "id": "w3-opp-e-operation-bores", "source": "w3-opp-march-operation", "target": "w3-opp-bores", "label": "opposed federal primary candidacy", "status": "documented", "type": "narrative" }, { "id": "w3-icmec-e-affiliate", "source": "w3-icmec-australia", "target": "icmec", "label": "was established as an affiliate of", "status": "reported", "type": "coordination" }, { "id": "w3-icmec-e-grant", "source": "w3-icmec-westpac", "target": "w3-icmec-australia", "label": "committed funding for data sharing", "status": "documented", "type": "money" }, { "id": "w3-icmec-e-lighthouse", "source": "w3-icmec-australia", "target": "w3-icmec-lighthouse", "label": "operates investigative data application", "status": "reported", "type": "resources" }, { "id": "w3-icmec-e-coalition", "source": "w3-icmec-australia", "target": "w3-icmec-saferai", "label": "sets participation and guides projects", "status": "documented", "type": "governance" }, { "id": "w3-icmec-e-partners", "source": "w3-icmec-saferai", "target": "w3-icmec-partner-category", "label": "permits contributions without public attribution", "status": "documented", "type": "governance" }, { "id": "w3-icmec-e-arthur-bank", "source": "w3-icmec-jodie-arthur", "target": "w3-icmec-westpac", "label": "held financial-crime executive role at", "status": "documented", "type": "personnel" }, { "id": "w3-icmec-e-arthur-board", "source": "w3-icmec-jodie-arthur", "target": "w3-icmec-australia", "label": "served as non-executive director of", "status": "documented", "type": "governance" }, { "id": "w3-icmec-e-gannon", "source": "w3-icmec-gannon", "target": "w3-icmec-australia", "label": "was identified as CEO of", "status": "documented", "type": "personnel" }, { "id": "w3-icmec-e-advocacy", "source": "w3-icmec-gannon", "target": "w3-icmec-device-advocacy", "label": "published device-centred age proposal", "status": "documented", "type": "narrative" }, { "id": "w3-icmec-e-presentation", "source": "bob-cunningham", "target": "w3-icmec-jcots", "label": "presented device-based age proposal to", "status": "documented", "type": "policy" }, { "id": "w3-icmec-e-draft", "source": "w3-icmec-craig", "target": "w3-icmec-draft-request", "label": "submitted legislative drafting request", "status": "documented", "type": "policy" }, { "id": "w3-icmec-e-uptake", "source": "icmec", "target": "w3-icmec-draft-request", "label": "presentation prompted drafting request", "status": "documented", "type": "policy" }, { "id": "w4-fin-e-control-ended", "source": "icmec", "target": "w3-icmec-australia", "label": "ceased sole membership and control", "status": "documented", "type": "governance" }, { "id": "w4-fin-e-accounting-removal", "source": "w3-icmec-australia", "target": "w4-fin-deconsolidation", "label": "balances removed from US consolidated statements", "status": "documented", "type": "finance" }, { "id": "w4-pol-e-publication", "source": "w3-icmec-australia", "target": "w4-pol-paper", "label": "published and promoted", "status": "documented", "type": "distribution" }, { "id": "w4-pol-e-collective-writing", "source": "w3-icmec-saferai", "target": "w4-pol-paper", "label": "reports collaborative writing with member insights", "status": "reported", "type": "coordination" }, { "id": "w4-pol-e-device-ask", "source": "w4-pol-paper", "target": "w4-pol-device-proposal", "label": "recommends considering", "status": "documented", "type": "policy" }, { "id": "w4-pol-e-tas-uptake", "source": "w4-pol-paper", "target": "w4-pol-tas-review", "label": "is cited and substantively used in", "status": "documented", "type": "policy" }, { "id": "w4-pol-e-reviewer", "source": "w4-pol-woolcott", "target": "w4-pol-tas-review", "label": "is named independent reviewer of", "status": "documented", "type": "authorship" }, { "id": "w4-pol-e-addressee", "source": "w4-pol-tas-review", "target": "w4-pol-premier-office", "label": "is formally addressed to", "status": "documented", "type": "policy" }, { "id": "w4-pol-e-toolkit-publisher", "source": "esafety", "target": "w4-pol-esafety-toolkit", "label": "publishes guidance in", "status": "documented", "type": "policy" }, { "id": "w4-pol-e-esafety-citation", "source": "w4-pol-paper", "target": "w4-pol-esafety-toolkit", "label": "is cited to support AI investigative triage in", "status": "documented", "type": "policy" }, { "id": "w4-pol-e-jago-role", "source": "w4-pol-jago", "target": "w3-icmec-australia", "label": "is listed as Senior Capacity and Prevention Coordinator by", "status": "reported", "type": "personnel" }, { "id": "w4-lh-e-build", "source": "w4-lh-kablamo", "target": "w3-icmec-lighthouse", "label": "reports building the production application", "status": "reported", "type": "infrastructure" }, { "id": "w4-lh-e-support", "source": "w3-icmec-australia", "target": "w4-lh-kablamo", "label": "reportedly engaged ongoing Product Care support", "status": "reported", "type": "resources" }, { "id": "w4-lh-e-owner", "source": "w4-lh-product-owner", "target": "w3-icmec-lighthouse", "label": "reportedly made technical delivery decisions", "status": "reported", "type": "governance" }, { "id": "w4-lh-e-feeds", "source": "w4-lh-data-suppliers", "target": "w3-icmec-lighthouse", "label": "reportedly supplied suspected-activity data", "status": "reported", "type": "resources" }, { "id": "w4-lh-e-local-use", "source": "w3-icmec-australia", "target": "w3-icmec-lighthouse", "label": "stated Australian-soil use restriction", "status": "documented", "type": "governance" }, { "id": "w4-lh-e-huntsman-austrac", "source": "w4-lh-austrac", "target": "w4-lh-huntsman", "label": "named as joint initiative participant", "status": "documented", "type": "coordination" }, { "id": "w4-lh-e-huntsman-accce", "source": "w4-lh-accce", "target": "w4-lh-huntsman", "label": "named as joint initiative participant", "status": "documented", "type": "coordination" }, { "id": "w4-lh-e-huntsman-closures", "source": "w4-lh-huntsman", "target": "w4-lh-huntsman-closures", "label": "reportedly led to bank-account closures", "status": "reported", "type": "operations" }, { "id": "w4-pol-e-esafety-encryption", "source": "w4-pol-esafety-toolkit", "target": "w4-pol-esafety-encryption-guidance", "label": "advises industry to consider", "status": "documented", "type": "policy" }, { "id": "w5-gov-e-us-renewal-exception", "source": "w5-gov-constitution", "target": "icmec", "label": "exempts from membership renewal", "status": "documented", "type": "governance" }, { "id": "w5-gov-e-admission", "source": "w5-gov-au-board", "target": "w5-gov-au-members", "label": "decides admission applications to", "status": "documented", "type": "governance" }, { "id": "w5-gov-e-election", "source": "w5-gov-au-members", "target": "w5-gov-au-board", "label": "elects and may remove directors of", "status": "documented", "type": "governance" }, { "id": "w5-gov-e-board-approvals", "source": "w5-gov-au-board", "target": "w3-icmec-australia", "label": "approves strategy budgets major partnerships and material programs for", "status": "documented", "type": "governance" }, { "id": "w5-gov-e-member-register", "source": "w5-gov-au-secretary", "target": "w5-gov-au-members", "label": "maintains register of", "status": "documented", "type": "governance" }, { "id": "w5-gov-e-conflict-records", "source": "w5-gov-au-secretary", "target": "w5-gov-au-board", "label": "maintains interests register and recorded conflict decisions for", "status": "documented", "type": "governance" }, { "id": "w5-gov-e-arthur-board", "source": "w3-icmec-jodie-arthur", "target": "w5-gov-au-board", "label": "is publicly listed as non-executive director on", "status": "documented", "type": "governance" }, { "id": "w5-gov-e-component-received", "source": "w5-gov-grf", "target": "w3-icmec-australia", "label": "reports receipt of component auditor financial statements and reports for", "status": "documented", "type": "governance" }, { "id": "w5-enc-e-ukri", "source": "w5-enc-ukri", "target": "w5-enc-rephrain", "label": "is acknowledged as funding", "status": "reported", "type": "money" }, { "id": "w5-enc-e-project-funding", "source": "w5-enc-rephrain", "target": "w5-enc-toolbox", "label": "funds through its fourth Strategic Funding Call", "status": "reported", "type": "money" }, { "id": "w5-enc-e-workpackage", "source": "w5-enc-toolbox", "target": "w5-enc-paper", "label": "includes as work package3 deliverable", "status": "documented", "type": "coordination" }, { "id": "w5-enc-e-centre-funding", "source": "5rights", "target": "w5-enc-dfc", "label": "is acknowledged as funding the centre", "status": "reported", "type": "money" }, { "id": "w5-enc-e-publisher", "source": "w5-enc-dfc", "target": "w5-enc-paper", "label": "publishes", "status": "documented", "type": "narrative" }, { "id": "w5-enc-e-lead", "source": "w5-enc-orben", "target": "w5-enc-toolbox", "label": "is identified as project lead of", "status": "documented", "type": "personnel" }, { "id": "w5-enc-e-author-spuy", "source": "w5-enc-spuy", "target": "w5-enc-paper", "label": "is named author of", "status": "documented", "type": "authorship" }, { "id": "w5-enc-e-author-witting", "source": "w5-enc-witting", "target": "w5-enc-paper", "label": "is named author of", "status": "documented", "type": "authorship" }, { "id": "w5-enc-e-author-burton", "source": "w5-enc-burton", "target": "w5-enc-paper", "label": "is named author of", "status": "documented", "type": "authorship" }, { "id": "w5-enc-e-author-day", "source": "w5-enc-day", "target": "w5-enc-paper", "label": "is named author of", "status": "documented", "type": "authorship" }, { "id": "w5-enc-e-author-livingstone", "source": "w5-enc-livingstone", "target": "w5-enc-paper", "label": "is named author of", "status": "documented", "type": "authorship" }, { "id": "w5-enc-e-author-sylwander", "source": "w5-enc-sylwander", "target": "w5-enc-paper", "label": "is named author of", "status": "documented", "type": "authorship" }, { "id": "w5-enc-e-review", "source": "w5-enc-nicholas", "target": "w5-enc-paper", "label": "is named independent peer reviewer of", "status": "documented", "type": "coordination" }, { "id": "w5-enc-e-affiliation", "source": "w5-enc-nicholas", "target": "esafety", "label": "is affiliated with in the report acknowledgement", "status": "reported", "type": "personnel" }, { "id": "w5-enc-e-fn29", "source": "w5-enc-paper", "target": "w4-pol-esafety-toolkit", "label": "is linked by footnote29 beside homomorphic-encryption guidance in", "status": "documented", "type": "policy" }, { "id": "w5-enc-e-fn26", "source": "w5-enc-paper", "target": "w4-pol-esafety-toolkit", "label": "is cited at footnote26 for platform-hopping context in", "status": "documented", "type": "policy" }, { "id": "w5-enc-e-workshop", "source": "w5-enc-workshop", "target": "w5-enc-consultation", "label": "is identified as an input to toolkit drafting before", "status": "documented", "type": "coordination" }, { "id": "w5-enc-e-consultation", "source": "esafety", "target": "w5-enc-consultation", "label": "invites written edits and verbal feedback through", "status": "documented", "type": "coordination" }, { "id": "w5-enc-e-x-access", "source": "w5-enc-consultation", "target": "w5-enc-x", "label": "provides draft-toolkit review opportunity to", "status": "documented", "type": "coordination" }, { "id": "w5-enc-e-x-response", "source": "w5-enc-x", "target": "w5-enc-consultation", "label": "replies that it has no additional feedback to", "status": "documented", "type": "coordination" }, { "id": "w5-fund-e-grant2021", "source": "w3-icmec-westpac", "target": "w5-fund-project2021", "label": "announced research award", "status": "documented", "type": "money" }, { "id": "w5-fund-e-recipient", "source": "w5-fund-unsw", "target": "w5-fund-project2021", "label": "named institution for project", "status": "documented", "type": "governance" }, { "id": "w5-fund-e-study-funding", "source": "w3-icmec-westpac", "target": "w5-fund-study2025", "label": "research funding acknowledged", "status": "reported", "type": "money" }, { "id": "w5-fund-e-publication", "source": "w5-fund-childlight", "target": "w5-fund-study2025", "label": "published in Searchlight 2025", "status": "documented", "type": "narrative" }, { "id": "w5-fund-e-core-funding", "source": "w5-fund-hdf", "target": "w5-fund-childlight", "label": "core grant acknowledged", "status": "reported", "type": "money" }, { "id": "w5-fund-e-advice", "source": "w5-fund-advisory", "target": "w5-fund-study2025", "label": "advisory committee listed", "status": "reported", "type": "coordination" }, { "id": "w5-fund-e-bank-adviser", "source": "w3-icmec-westpac", "target": "w5-fund-advisory", "label": "executive named as adviser", "status": "reported", "type": "personnel" }, { "id": "w5-fund-e-icmec-adviser", "source": "w3-icmec-australia", "target": "w5-fund-advisory", "label": "capacity lead named as adviser", "status": "reported", "type": "personnel" }, { "id": "w5-fund-e-jss-partner", "source": "w5-fund-jss", "target": "w5-fund-study2025", "label": "study partner listed", "status": "reported", "type": "coordination" }, { "id": "w5-fund-e-recommendation", "source": "w5-fund-study2025", "target": "w5-fund-monitoring", "label": "recommends expanded monitoring and sharing", "status": "documented", "type": "policy" }, { "id": "w5-fund-e-salter-author", "source": "w5-fund-salter", "target": "w5-fund-study2025", "label": "is named author of", "status": "documented", "type": "authorship" }, { "id": "w5-fund-e-salter-review", "source": "w5-fund-salter", "target": "w5-enc-paper", "label": "is named independent peer reviewer of", "status": "documented", "type": "coordination" }, { "id": "w5-enc-e-x-dispute", "source": "w5-enc-x", "target": "esafety", "label": "contests standard applicability in a recorded court dispute with", "status": "documented", "type": "policy" }, { "id": "w6-ref-e-levy-role", "source": "w6-ref-levy", "target": "w6-ref-ncsc", "label": "is identified as technical director of", "status": "documented", "type": "personnel" }, { "id": "w6-ref-e-robinson-role", "source": "w6-ref-robinson", "target": "w6-ref-gchq", "label": "is identified as technical director for cryptanalysis at", "status": "documented", "type": "personnel" }, { "id": "w6-ref-e-levy-author", "source": "w6-ref-levy", "target": "w6-ref-paper", "label": "coauthors", "status": "documented", "type": "authorship" }, { "id": "w6-ref-e-robinson-author", "source": "w6-ref-robinson", "target": "w6-ref-paper", "label": "coauthors", "status": "documented", "type": "authorship" }, { "id": "w6-ref-e-eu-citation", "source": "w6-ref-ec-defence", "target": "w6-ref-paper", "label": "cites to support possible encrypted-service detection", "status": "documented", "type": "policy" }, { "id": "w6-ref-e-eu-publisher", "source": "european-commission", "target": "w6-ref-ec-defence", "label": "publishes", "status": "documented", "type": "policy" }, { "id": "w6-ref-e-au-citation", "source": "w6-ref-position", "target": "w6-ref-paper", "label": "cites in discussion of client-side scanning", "status": "documented", "type": "policy" }, { "id": "w6-ref-e-au-publisher", "source": "esafety", "target": "w6-ref-position", "label": "publishes", "status": "documented", "type": "policy" }, { "id": "w6-ref-e-rephrain", "source": "w6-ref-paper", "target": "w5-enc-rephrain", "label": "points to separate evaluation work at", "status": "documented", "type": "coordination" }, { "id": "w6-ref-e-wp-input", "source": "wp-secretariat", "target": "w6-ref-consultation", "label": "advocates detection options in submission to", "status": "documented", "type": "policy" }, { "id": "w6-ref-e-drw-input", "source": "w6-ref-drw", "target": "w6-ref-consultation", "label": "opposes encrypted-content detection proposals in submission to", "status": "documented", "type": "policy" }, { "id": "w6-draft-e-may-a", "source": "w6-draft-may-a", "target": "w5-enc-consultation", "label": "is named core-content attachment to", "status": "documented", "type": "coordination" }, { "id": "w6-draft-e-may-b", "source": "w6-draft-may-b", "target": "w5-enc-consultation", "label": "is named tools attachment to", "status": "documented", "type": "coordination" }, { "id": "w6-draft-e-csm-workshop", "source": "common-sense-media", "target": "w5-enc-workshop", "label": "reports participation in", "status": "reported", "type": "coordination" }, { "id": "w6-draft-e-csm-publication", "source": "common-sense-media", "target": "w6-draft-csm-comments", "label": "authors and publicly lists", "status": "documented", "type": "narrative" }, { "id": "w6-draft-e-csm-recipient", "source": "w6-draft-csm-comments", "target": "esafety", "label": "addresses requested draft feedback to", "status": "documented", "type": "coordination" }, { "id": "w6-draft-e-csm-draft", "source": "w6-draft-csm-comments", "target": "w6-draft-october-draft", "label": "comments on", "status": "documented", "type": "coordination" }, { "id": "w6-draft-e-csm-age-detail", "source": "common-sense-media", "target": "w6-draft-october-draft", "label": "requests clearer age-assurance performance and privacy requirements in", "status": "documented", "type": "policy" }, { "id": "w6-member-e-tax-related-return", "source": "icmec", "target": "w3-icmec-australia", "label": "reports a related-corporation characterization of", "status": "documented", "type": "governance" }, { "id": "w6-member-e-services-category", "source": "icmec", "target": "w3-icmec-australia", "label": "reported transaction in services, membership or fundraising category with", "status": "documented", "type": "financial" }, { "id": "w6-award-e-registration", "source": "w6-award-survey", "target": "w6-award-protocol", "label": "has a publicly recorded post-analysis registration", "status": "documented", "type": "coordination" }, { "id": "w6-award-e-recruitment", "source": "w6-award-cloudresearch", "target": "w6-award-survey", "label": "is reported to have recruited and administered participants", "status": "reported", "type": "coordination" }, { "id": "w6-award-e-reuse", "source": "w6-award-survey", "target": "w5-fund-study2025", "label": "supplies the survey data for a later financial-behaviour analysis", "status": "reported", "type": "coordination" }, { "id": "w6-award-e-deliverable", "source": "w5-enc-toolbox", "target": "w6-award-wp2", "label": "produced a publicly available work-package2 analytical report", "status": "documented", "type": "coordination" }, { "id": "w6-award-e-wp2-author", "source": "w5-enc-orben", "target": "w6-award-wp2", "label": "is a credited co-author of", "status": "documented", "type": "authorship" }, { "id": "w7-impl-e-reddit-gate", "source": "w7-impl-reddit", "target": "w7-impl-reddit-age", "label": "introduced an age check for UK NSFW access", "status": "reported", "type": "policy" }, { "id": "w7-impl-e-processor", "source": "w7-impl-reddit-age", "target": "w7-impl-persona", "label": "uses a named verification processor", "status": "reported", "type": "infrastructure" }, { "id": "w7-impl-e-output", "source": "w7-impl-persona", "target": "w7-impl-reddit", "label": "says it returns birthdate or estimated age to", "status": "reported", "type": "infrastructure" }, { "id": "w7-impl-e-penalty", "source": "w7-impl-ico", "target": "w7-impl-reddit", "label": "imposed an appealed UK GDPR penalty on", "status": "documented", "type": "policy" }, { "id": "w7-impl-e-meetings", "source": "w7-impl-ico", "target": "ofcom", "label": "reports regular inter-regulator meetings with", "status": "reported", "type": "coordination" }, { "id": "w7-impl-e-government-proposal", "source": "w7-impl-government", "target": "w7-impl-national-id", "label": "consulted on and later cancelled", "status": "documented", "type": "policy" }, { "id": "w7-impl-e-historical-age-use", "source": "w7-impl-national-id", "target": "w7-impl-age-use", "label": "included a non-exclusive online-age use proposal", "status": "documented", "type": "policy" }, { "id": "w7-impl-e-discord-settings", "source": "discord", "target": "w7-impl-discord-settings", "label": "describes age checks for specified UK features", "status": "reported", "type": "policy" }, { "id": "w7-policy-e-authority", "source": "w7-policy-osa", "target": "ofcom", "label": "requires age-assurance guidance and relevant code measures from", "status": "documented", "type": "policy" }, { "id": "w7-policy-e-guidance", "source": "ofcom", "target": "w7-policy-heaa", "label": "publishes and applies", "status": "documented", "type": "policy" }, { "id": "w7-policy-e-issued-code", "source": "ofcom", "target": "w7-policy-child-code", "label": "issued", "status": "documented", "type": "policy" }, { "id": "w7-policy-e-code-age", "source": "w7-policy-child-code", "target": "w7-policy-heaa", "label": "recommends for specified access and content controls", "status": "documented", "type": "policy" }, { "id": "w7-policy-e-email-accepted", "source": "ofcom", "target": "w7-policy-email-estimation", "label": "added to methods capable of highly effective age assurance", "status": "documented", "type": "policy" }, { "id": "w7-policy-e-avpa-evidence", "source": "avpa", "target": "w7-policy-email-estimation", "label": "supplied evidence expressly relied upon in method acceptance", "status": "documented", "type": "policy" }, { "id": "w7-policy-e-thresholds-declined", "source": "ofcom", "target": "w7-policy-heaa", "label": "declined proposed fixed numerical thresholds in initial guidance for", "status": "documented", "type": "policy" }, { "id": "w7-policy-e-certification-conditional", "source": "ofcom", "target": "w7-policy-heaa", "label": "allows certification as evidence without making it mandatory or sufficient for", "status": "documented", "type": "policy" }, { "id": "w7-policy-e-age-fine", "source": "ofcom", "target": "w7-policy-xgroovy", "label": "reports imposing an age-assurance duty penalty on", "status": "reported", "type": "financial" }, { "id": "w7-policy-e-information-fine", "source": "ofcom", "target": "w7-policy-xgroovy", "label": "reports imposing a separate information-request penalty on", "status": "reported", "type": "financial" }, { "id": "w7-market-e-allen-avid", "source": "w7-market-allen", "target": "w7-market-avid", "label": "holds disclosed majority share/voting band and director appointment rights in", "status": "documented", "type": "governance" }, { "id": "w7-market-e-bate-avid", "source": "w7-market-bate", "target": "w7-market-avid", "label": "holds disclosed minority share and voting band in", "status": "documented", "type": "governance" }, { "id": "w7-market-e-avid-scheme-control", "source": "w7-market-avid", "target": "w7-market-scheme", "label": "holds director appointment/removal rights in", "status": "documented", "type": "governance" }, { "id": "w7-market-e-avid-scheme-sla", "source": "w7-market-avid", "target": "w7-market-scheme", "label": "supplies audit, certification and test purchasing services under a stated SLA to", "status": "reported", "type": "operations" }, { "id": "w7-market-e-scheme-rules", "source": "w7-market-scheme", "target": "w7-market-avid", "label": "sets published certification decision, fee, appeal and impartiality rules for", "status": "documented", "type": "governance" }, { "id": "w7-market-e-allen-iso", "source": "w7-market-allen", "target": "w7-market-iso27566", "label": "is listed by BSI as lead editor of", "status": "documented", "type": "authorship" }, { "id": "w7-market-e-allen-panel", "source": "w7-market-allen", "target": "w7-market-bsi-panel", "label": "is listed by BSI as chair of", "status": "documented", "type": "governance" }, { "id": "w7-market-e-panel-iso", "source": "w7-market-bsi-panel", "target": "w7-market-iso27566", "label": "coordinated UK input into development of", "status": "documented", "type": "policy" }, { "id": "w7-market-e-avid-iso-service", "source": "w7-market-avid", "target": "w7-market-iso27566", "label": "offers paid certification services incorporating", "status": "reported", "type": "operations" }, { "id": "w7-market-e-verifymy-cert", "source": "w7-market-verifymy", "target": "w7-market-scheme", "label": "reports its platform has certification incorporating ISO27566-1 from", "status": "reported", "type": "operations" }, { "id": "w7-market-e-safe-phase2", "source": "w7-market-safeonline", "target": "euconsent", "label": "is acknowledged as funding Phase II delivered by the later ASBL", "status": "reported", "type": "finance" }, { "id": "w7-market-e-euconsent-ageaware", "source": "euconsent", "target": "w7-market-ageaware", "label": "published a proposed admission, billing and governance design for", "status": "documented", "type": "governance" }, { "id": "w7-market-e-corby-ageaware", "source": "corby", "target": "w7-market-ageaware", "label": "is a credited author of the proposed specification for", "status": "documented", "type": "authorship" }, { "id": "w7-market-e-allen-ageaware", "source": "w7-market-allen", "target": "w7-market-ageaware", "label": "is named as the required ACCS approver for the draft specification of", "status": "documented", "type": "governance" }, { "id": "w7-market-e-verifymy-ageaware", "source": "w7-market-verifymy", "target": "w7-market-ageaware", "label": "is credited with detailed design work for", "status": "reported", "type": "resources" }, { "id": "w7-market-e-yoti-ageaware", "source": "w7-market-yoti", "target": "w7-market-ageaware", "label": "is credited with detailed design work for", "status": "reported", "type": "resources" }, { "id": "w7-market-e-agechecked-ageaware", "source": "w7-market-agechecked", "target": "w7-market-ageaware", "label": "is credited with detailed design work for", "status": "reported", "type": "resources" }, { "id": "w7-market-e-avpa-ageaware", "source": "avpa", "target": "w7-market-ageaware", "label": "describes the project's demonstration mission as fulfilled alongside private alternatives", "status": "reported", "type": "narrative" }, { "id": "w7-adv-e-chair", "source": "kidron", "target": "5rights", "label": "declared chair role in debate", "status": "documented", "type": "personnel" }, { "id": "w7-adv-e-credit", "source": "w7-adv-parkinson", "target": "kidron", "label": "credited work on age-assurance principles", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-cosponsor", "source": "kidron", "target": "w7-adv-amend124", "label": "co-sponsored", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-agreed", "source": "ukparliament", "target": "w7-adv-amend124", "label": "agreed amendment", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-privacy-amend", "source": "kidron", "target": "w7-adv-amend125", "label": "tabled privacy addition", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-declined-amend", "source": "w7-adv-parkinson", "target": "w7-adv-amend125", "label": "declined proposed privacy addition", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-proposed-privacy", "source": "5rights", "target": "w7-adv-privacy-criterion-proposal", "label": "requested express privacy criterion", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-declined-privacy", "source": "ofcom", "target": "w7-adv-privacy-criterion-proposal", "label": "declined separate criterion", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-proposed-token", "source": "5rights", "target": "w7-adv-token-proposal", "label": "recommended age tokens in broader assurance ecosystem", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-tokens-disposition", "source": "ofcom", "target": "w7-adv-token-proposal", "label": "added conditional references after token submissions", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-certification", "source": "5rights", "target": "w7-adv-certification-proposal", "label": "urged standards and certification consideration", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-accs-citation", "source": "w7-adv-certification-proposal", "target": "accs", "label": "cited technical requirements as certification example", "status": "documented", "type": "policy" }, { "id": "w7-adv-e-commission", "source": "euconsent", "target": "w7-adv-ageaware-rights-review", "label": "reported commissioning", "status": "reported", "type": "coordination" }, { "id": "w7-adv-e-review-day", "source": "w5-enc-day", "target": "w7-adv-ageaware-rights-review", "label": "was named as intended review co-lead", "status": "reported", "type": "coordination" }, { "id": "w7-adv-e-review-witting", "source": "w5-enc-witting", "target": "w7-adv-ageaware-rights-review", "label": "was named as intended review co-lead", "status": "reported", "type": "coordination" }, { "id": "w7-adv-e-review-livingstone", "source": "w5-enc-livingstone", "target": "w7-adv-ageaware-rights-review", "label": "was named among team academics through whom review was commissioned", "status": "reported", "type": "coordination" }, { "id": "w7-adv-e-review-scope", "source": "w7-adv-ageaware-rights-review", "target": "w7-market-ageaware", "label": "was commissioned to examine rights implications of", "status": "reported", "type": "coordination" }, { "id": "w7-policy-e-verifymy-evidence", "source": "w7-market-verifymy", "target": "w7-policy-email-estimation", "label": "supplied evidence expressly relied upon in method acceptance", "status": "documented", "type": "policy" }, { "id": "w8-red-e-organizer-cesia", "source": "w8-red-cesia", "target": "w8-red-call", "label": "named campaign co-organizer", "status": "documented", "type": "coordination" }, { "id": "w8-red-e-organizer-tfs", "source": "w8-red-tfs", "target": "w8-red-call", "label": "named campaign co-organizer", "status": "documented", "type": "coordination" }, { "id": "w8-red-e-organizer-chai", "source": "w8-red-chai", "target": "w8-red-call", "label": "named campaign co-organizer", "status": "documented", "type": "coordination" }, { "id": "w8-red-e-work-segerie", "source": "w8-red-segerie", "target": "w8-red-call", "label": "credited with initial drafting", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-work-niki", "source": "w8-red-iliadis", "target": "w8-red-call", "label": "credited with TFS project management", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-work-tereza", "source": "w8-red-zoumpalova", "target": "w8-red-call", "label": "credited with TFS management and signatory work", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-work-anoush", "source": "w8-red-tatevossian", "target": "w8-red-call", "label": "credited with UN-engagement advice", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-work-grimonpont", "source": "w8-red-grimonpont", "target": "w8-red-call", "label": "credited with press-conference coordination", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-work-shimi", "source": "w8-red-shimi", "target": "w8-red-call", "label": "credited with drafting feedback", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-work-miotti", "source": "miotti", "target": "w8-red-call", "label": "credited with drafting feedback", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-work-controlai", "source": "controlai", "target": "w8-red-call", "label": "credited with partner feedback", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-work-fli", "source": "fli", "target": "w8-red-call", "label": "credited with partner feedback", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-work-apollo", "source": "w8-red-apollo-pr", "target": "w8-red-call", "label": "credited with press-conference preparation and moderation", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-ressa-approach", "source": "w8-red-grimonpont", "target": "w8-red-ressa", "label": "credited with pitching the campaign", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-signature-hubinger", "source": "w8-red-hubinger", "target": "w8-red-call", "label": "listed individual signatory", "status": "documented", "type": "narrative" }, { "id": "w8-red-e-signature-kidron", "source": "kidron", "target": "w8-red-call", "label": "listed individual signatory", "status": "documented", "type": "narrative" }, { "id": "w8-red-e-signature-miotti", "source": "miotti", "target": "w8-red-call", "label": "listed individual signatory", "status": "documented", "type": "narrative" }, { "id": "w8-red-e-ctrl-interpretation", "source": "controlai", "target": "w8-red-narrow-path", "label": "promotes ASI prohibition as red-lines implementation", "status": "documented", "type": "policy" }, { "id": "w8-red-e-ctrl-partner", "source": "controlai", "target": "w8-red-call", "label": "self-identifies as official partner", "status": "documented", "type": "coordination" }, { "id": "w8-red-e-survey-tfs", "source": "w8-red-tfs", "target": "w8-red-survey", "label": "consolidated fifteen priorities into ten survey options", "status": "documented", "type": "policy" }, { "id": "w8-red-e-survey-renaissance", "source": "w8-red-renaissance", "target": "w8-red-survey", "label": "consolidated fifteen priorities into ten survey options", "status": "documented", "type": "policy" }, { "id": "w8-red-e-survey-reference", "source": "w8-red-survey", "target": "w8-red-call", "label": "survey cited in campaign development credits", "status": "documented", "type": "narrative" }, { "id": "w8-red-e-longview-cesia", "source": "longview", "target": "w8-red-cesia", "label": "recommended organizational support via Effektiv Spenden", "status": "documented", "type": "money" }, { "id": "w8-red-e-eff-cesia", "source": "w8-red-effektiv", "target": "w8-red-cesia", "label": "reported forwarding organizational support", "status": "reported", "type": "money" }, { "id": "w8-red-e-longview-tfs", "source": "longview", "target": "w8-red-tfs", "label": "recommended organizational support via Effektiv Spenden", "status": "documented", "type": "money" }, { "id": "w8-red-e-eff-tfs", "source": "w8-red-effektiv", "target": "w8-red-tfs", "label": "reported forwarding organizational support", "status": "reported", "type": "money" }, { "id": "w8-red-e-sff-cesia", "source": "sff", "target": "w8-red-cesia", "label": "recommended general support; EffiSciences receiving charity", "status": "documented", "type": "money" }, { "id": "w8-red-e-effisciences", "source": "w8-red-effisciences", "target": "w8-red-cesia", "label": "listed receiving charity for SFF recommendation", "status": "documented", "type": "money" }, { "id": "w8-red-e-sff-tfs-2024", "source": "sff", "target": "w8-red-tfs", "label": "recommended general support from Jaan Tallinn", "status": "documented", "type": "money" }, { "id": "w8-red-e-sff-tfs-2025", "source": "sff", "target": "w8-red-tfs", "label": "recommended general support from Jaan Tallinn", "status": "documented", "type": "money" }, { "id": "w8-red-e-sff-tfs-2023-fli", "source": "fli", "target": "w8-red-tfs", "label": "named funding source in SFF general-support recommendation", "status": "documented", "type": "money" }, { "id": "w8-red-e-sff-tfs-2023-tallinn", "source": "tallinn", "target": "w8-red-tfs", "label": "named funding source in SFF general-support recommendation", "status": "documented", "type": "money" }, { "id": "w8-red-e-solicitation", "source": "w8-red-cesia", "target": "w8-red-sprint", "label": "requested funding for diplomatic follow-through", "status": "documented", "type": "money" }, { "id": "w8-red-e-sprint-call", "source": "w8-red-sprint", "target": "w8-red-call", "label": "proposes converting campaign support into agreements", "status": "documented", "type": "policy" }, { "id": "w8-red-e-longview-2m", "source": "longview", "target": "w8-red-cesia", "label": "recipient reports two-year recommendation", "status": "reported", "type": "money" }, { "id": "w8-red-e-tfs-subgroup", "source": "w8-red-tfs", "target": "w8-red-unesco-subgroup", "label": "listed subgroup co-lead", "status": "documented", "type": "governance" }, { "id": "w8-red-e-ada-subgroup", "source": "w8-red-ada", "target": "w8-red-unesco-subgroup", "label": "listed subgroup co-lead", "status": "documented", "type": "governance" }, { "id": "w8-red-e-cesia-subgroup", "source": "w8-red-cesia", "target": "w8-red-unesco-subgroup", "label": "listed subgroup member", "status": "documented", "type": "coordination" }, { "id": "w8-red-e-subgroup-network", "source": "w8-red-unesco-subgroup", "target": "w8-red-unesco-network", "label": "works within network framework", "status": "documented", "type": "coordination" }, { "id": "w8-red-e-unesco-network", "source": "w8-red-unesco", "target": "w8-red-unesco-network", "label": "hosts civil-society and academic network", "status": "documented", "type": "coordination" }, { "id": "w8-red-e-tereza-representative", "source": "w8-red-zoumpalova", "target": "w8-red-unesco-subgroup", "label": "named subgroup representative in coordination group", "status": "documented", "type": "governance" }, { "id": "w8-red-e-mcgovern-network", "source": "w8-red-mcgovern", "target": "w8-red-unesco-network", "label": "initial network funding attributed by UNESCO", "status": "reported", "type": "money" }, { "id": "w8-red-e-ukri-pilot", "source": "w5-enc-ukri", "target": "w8-red-unesco-network", "label": "funded preceding pilot per UNESCO history", "status": "reported", "type": "money" }, { "id": "w8-red-e-ressa-call", "source": "w8-red-ressa", "target": "w8-red-call", "label": "advocated named campaign in UN-hosted prepared speech", "status": "documented", "type": "policy" }, { "id": "w8-red-e-ressa-venue", "source": "w8-red-ressa", "target": "w8-red-unga80", "label": "submitted campaign-bearing speech", "status": "documented", "type": "policy" }, { "id": "w8-red-e-bengio-venue", "source": "w8-red-bengio", "target": "w8-red-unsc10005", "label": "advocated international red lines in recorded intervention", "status": "reported", "type": "policy" }, { "id": "w8-red-e-clement-call", "source": "w8-red-clement-jones", "target": "w8-red-call", "label": "self-reported signatory and parliamentary advocate", "status": "documented", "type": "policy" }, { "id": "w8-red-e-india-tfs", "source": "w8-red-tfs", "target": "w8-red-india-workshop", "label": "reported workshop co-organization", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-india-cesia", "source": "w8-red-cesia", "target": "w8-red-india-workshop", "label": "reported workshop co-organization", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-athens-tfs", "source": "w8-red-tfs", "target": "w8-red-athens2025", "label": "organized policy roundtable", "status": "documented", "type": "coordination" }, { "id": "w8-red-e-athens-call", "source": "w8-red-athens2025", "target": "w8-red-call", "label": "developed agenda building on campaign", "status": "documented", "type": "policy" }, { "id": "w8-red-e-athens-unesco", "source": "w8-red-unesco", "target": "w8-red-athens2025", "label": "credited as event partner", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-athens-oecd", "source": "oecd", "target": "w8-red-athens2025", "label": "credited as event partner", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-athens-support-caf", "source": "w8-red-cooperative-ai", "target": "w8-red-athens2025", "label": "credited event support; form and amount unspecified", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-athens-support-arnold", "source": "w8-red-arnold-porter", "target": "w8-red-athens2025", "label": "credited event support; form and amount unspecified", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-athens-support-fathom", "source": "w8-red-fathom", "target": "w8-red-athens2025", "label": "credited event support; form and amount unspecified", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-athens-support-deliberaide", "source": "w8-red-deliberaide", "target": "w8-red-athens2025", "label": "credited event support; form and amount unspecified", "status": "reported", "type": "coordination" }, { "id": "w8-red-e-leanda-role", "source": "w8-red-barrington-leach", "target": "5rights", "label": "identified as executive director at event", "status": "documented", "type": "personnel" }, { "id": "w8-red-e-leanda-athens", "source": "w8-red-barrington-leach", "target": "w8-red-athens2025", "label": "presented child-rights AI safeguards", "status": "reported", "type": "policy" }, { "id": "w8-red-e-leanda-code", "source": "w8-red-barrington-leach", "target": "w8-red-childrens-code", "label": "advocated as model for AI safety by design", "status": "reported", "type": "policy" }, { "id": "w8-follow-e-fathom-tfs", "source": "w8-red-fathom", "target": "w8-red-tfs", "label": "publicly describes verified-redlines partnership", "status": "documented", "type": "coordination" }, { "id": "w8-follow-e-fathom-series", "source": "w8-red-fathom", "target": "w8-follow-athens-series", "label": "financial support acknowledged by event series", "status": "reported", "type": "money" }, { "id": "w8-follow-e-annual-event", "source": "w8-red-athens2025", "target": "w8-follow-athens-series", "label": "seventh edition of event series", "status": "documented", "type": "coordination" }, { "id": "w8-follow-e-reception", "source": "w8-red-cooperative-ai", "target": "w8-red-athens2025", "label": "reports hosting preceding speakers reception", "status": "reported", "type": "resources" }, { "id": "w8-follow-e-dialogue-part", "source": "w8-red-athens2025", "target": "w8-follow-athens-dialogue", "label": "included private red-lines discussion", "status": "reported", "type": "coordination" }, { "id": "w8-follow-e-caif-monitoring", "source": "w8-red-cooperative-ai", "target": "w8-follow-athens-dialogue", "label": "reports advocating multi-agent risk monitoring", "status": "reported", "type": "policy" }, { "id": "w8-follow-e-macroscopic-commitment", "source": "w8-follow-macroscopic", "target": "w8-red-cooperative-ai", "label": "philanthropic commitment reported by recipient", "status": "reported", "type": "money" }, { "id": "w8-follow-e-clifton-appointment", "source": "w8-follow-clifton", "target": "w8-red-cooperative-ai", "label": "appointed director", "status": "documented", "type": "governance" }, { "id": "w8-follow-e-clifton-donor-role", "source": "w8-follow-clifton", "target": "w8-follow-macroscopic", "label": "listed as grantmaking officer", "status": "reported", "type": "personnel" }, { "id": "w8-follow-e-deliberaide-workflow", "source": "w8-red-deliberaide", "target": "w8-red-athens2025", "label": "platform use reported in attributed client testimonial", "status": "reported", "type": "resources" }, { "id": "w8-follow-e-cesia-france", "source": "w8-red-cesia", "target": "w8-follow-france-g7", "label": "targeted as champion-state opportunity", "status": "documented", "type": "policy" }, { "id": "w8-follow-e-cesia-event", "source": "w8-red-cesia", "target": "w8-follow-sciencespo-march", "label": "named event co-organizer", "status": "documented", "type": "coordination" }, { "id": "w8-follow-e-mirlesse-event", "source": "w8-follow-mirlesse", "target": "w8-follow-sciencespo-march", "label": "scheduled moderator in institutional programme", "status": "documented", "type": "policy" }, { "id": "w8-follow-e-g7-label", "source": "w8-follow-france-g7", "target": "w8-follow-sciencespo-march", "label": "G7 label awarded according to organizer", "status": "reported", "type": "coordination" }, { "id": "w8-follow-e-g7-comparability", "source": "w8-follow-g7-ministers", "target": "w8-follow-risk-comparability", "label": "committed to risk-framework comparability discussions", "status": "documented", "type": "policy" }, { "id": "w8-follow-e-g7-openness", "source": "w8-follow-g7-ministers", "target": "w8-follow-g7-open-ai", "label": "promoted AI openness and broader adoption", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-fund-caif", "source": "w8-red-cooperative-ai", "target": "w9-assure-multi-agent-fund", "label": "named joint supporting party considering applications", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-fund-schmidt", "source": "w9-assure-schmidt-sciences", "target": "w9-assure-multi-agent-fund", "label": "named joint supporting party considering applications", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-fund-aria", "source": "w9-assure-aria", "target": "w9-assure-multi-agent-fund", "label": "named joint supporting party considering applications", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-fund-deepmind", "source": "deepmind", "target": "w9-assure-multi-agent-fund", "label": "named joint supporting party considering applications", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-fund-admin", "source": "w9-assure-schmidt-sciences", "target": "w9-assure-multi-agent-fund", "label": "hosts application and administrative front door", "status": "documented", "type": "resources" }, { "id": "w9-assure-e-fund-google-route", "source": "w9-assure-multi-agent-fund", "target": "w9-assure-google-option", "label": "offers optional differentiated funding route", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-fund-google-conditions", "source": "w9-assure-google-option", "target": "w9-assure-multi-agent-fund", "label": "conditions prospective awards on openness and AI Principles", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-fund-schmidt-renewal", "source": "w9-assure-schmidt-sciences", "target": "w9-assure-multi-agent-fund", "label": "describes progress-dependent annual funding", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-fund-agenda", "source": "w9-assure-multi-agent-fund", "target": "w9-assure-agent-infrastructure", "label": "solicits research on coordination and admission mechanisms", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-sb813-sponsor", "source": "w8-red-fathom", "target": "w9-assure-sb813", "label": "sponsors legislative proposal", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-fathom-prospective-mro", "source": "w8-red-fathom", "target": "w9-assure-old-mro", "label": "reported seeking prospective recognized-operator role", "status": "reported", "type": "governance" }, { "id": "w9-assure-e-sb813-version-change", "source": "w9-assure-old-mro", "target": "w9-assure-sb813", "label": "earlier certification framework superseded in enrolled text", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-sb813-govops", "source": "w9-assure-sb813", "target": "w9-assure-govops", "label": "would delegate verifier qualification and suspension procedures", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-govops-ivo", "source": "w9-assure-govops", "target": "w9-assure-ivo-role", "label": "would apply qualification and independence requirements", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-sb813-permitted-fees", "source": "w9-assure-assessed-role", "target": "w9-assure-ivo-role", "label": "would permit non-outcome-contingent market-rate fees", "status": "documented", "type": "money" }, { "id": "w9-assure-e-sb813-voluntary", "source": "w9-assure-sb813", "target": "w9-assure-assessed-role", "label": "would preserve operation without this verification route", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-bsa-pressure", "source": "w9-assure-bsa", "target": "w9-assure-sb813", "label": "objects that voluntary audits could become litigation expectations", "status": "reported", "type": "policy" }, { "id": "w9-assure-e-g7-ministers", "source": "w8-follow-g7-ministers", "target": "w9-assure-g7-principles", "label": "adopt shared minors principles", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-g7-call", "source": "w9-assure-g7-leaders", "target": "w9-assure-g7-call", "label": "adopt call including conversational-AI age assurance", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-g7-welcome", "source": "w9-assure-g7-call", "target": "w9-assure-g7-principles", "label": "welcomes preceding ministerial principles", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-g7-review", "source": "w9-assure-g7-leaders", "target": "w8-follow-g7-ministers", "label": "requests regular meetings and year-end progress review", "status": "documented", "type": "governance" }, { "id": "w9-assure-e-g7-privacy", "source": "w9-assure-g7-dpas", "target": "w9-assure-g7-privacy", "label": "adopts contextual and privacy-preserving assurance guidance", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-g7-privacy-reference", "source": "w9-assure-g7-privacy", "target": "w9-assure-g7-principles", "label": "expressly references ministerial assurance commitments", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-g7-unicef", "source": "w9-assure-unicef", "target": "w8-follow-g7-ministers", "label": "reports providing advance evidence", "status": "reported", "type": "policy" }, { "id": "w9-assure-e-g7-premarket", "source": "5rights", "target": "w9-assure-5rights-premarket", "label": "demands premarket certification beyond G7 outcome", "status": "documented", "type": "policy" }, { "id": "w9-assure-e-fathom-paris", "source": "w8-red-fathom", "target": "w9-assure-paris-simulation", "label": "publishes reported verification simulation findings", "status": "reported", "type": "narrative" }, { "id": "w9-assure-e-fathom-ashby", "source": "w8-red-fathom", "target": "w9-assure-ashby-proposals", "label": "publishes verification and commercial-incentive proposals", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-carnegie-review", "source": "w10-duty-carnegie", "target": "w10-duty-review2025", "label": "earlier model work cited in review", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-woods-review", "source": "w10-duty-woods", "target": "w10-duty-review2025", "label": "credited with duty-of-care conception", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-perrin-review", "source": "w10-duty-perrin", "target": "w10-duty-review2025", "label": "credited with duty-of-care conception", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-woods-network", "source": "w10-duty-woods", "target": "w10-duty-network", "label": "identified as principal adviser", "status": "documented", "type": "person" }, { "id": "w10-duty-e-network-review", "source": "w10-duty-network", "target": "w10-duty-review2025", "label": "submits systems-focused policy recommendations", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-farthing-reset", "source": "w10-duty-farthing", "target": "w10-duty-reset-australia", "label": "identified with Australian organization in byline", "status": "documented", "type": "person" }, { "id": "w10-duty-e-woods-article", "source": "w10-duty-woods", "target": "w10-duty-joint-output", "label": "coauthors singular-duty argument", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-farthing-article", "source": "w10-duty-farthing", "target": "w10-duty-joint-output", "label": "coauthors singular-duty argument", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-reset-review", "source": "w10-duty-reset-australia", "target": "w10-duty-review2025", "label": "submission70 argument quoted in recommendation", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-review-draft", "source": "w10-duty-review2025", "target": "w10-duty-draft2026", "label": "earlier overarching-duty recommendation reflected in proposal", "status": "inferred", "type": "policy" }, { "id": "w10-duty-e-draft-provider", "source": "w10-duty-draft2026", "target": "w10-duty-provider-role", "label": "would impose reasonably practicable safe-environment duty", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-draft-minister", "source": "w10-duty-draft2026", "target": "w10-duty-minister-role", "label": "would delegate specified scope and harm-category rules", "status": "documented", "type": "governance" }, { "id": "w10-duty-e-minister-provider", "source": "w10-duty-minister-role", "target": "w10-duty-provider-role", "label": "would specify exemptions and some user-empowerment requirements", "status": "documented", "type": "governance" }, { "id": "w10-duty-e-draft-esafety", "source": "w10-duty-draft2026", "target": "esafety", "label": "would extend assessment and compliance powers", "status": "documented", "type": "governance" }, { "id": "w10-duty-e-esafety-provider", "source": "esafety", "target": "w10-duty-provider-role", "label": "would issue compliance directions on reasonable belief", "status": "documented", "type": "governance" }, { "id": "w10-duty-e-draft-adult", "source": "w10-duty-draft2026", "target": "w10-duty-private-adult-limit", "label": "limits what duty requires for private adult communications", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-draft-privacy", "source": "w10-duty-draft2026", "target": "w10-duty-privacy-factor", "label": "includes privacy in proportionality assessment", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-draft-review", "source": "w10-duty-draft2026", "target": "w10-duty-merits-review", "label": "makes specified decisions reviewable", "status": "documented", "type": "governance" }, { "id": "w10-duty-e-review-esafety", "source": "w10-duty-merits-review", "target": "esafety", "label": "would permit review of specified Commissioner decisions", "status": "documented", "type": "governance" }, { "id": "w10-duty-e-reset-research-request", "source": "w10-duty-reset-australia", "target": "w10-duty-third-sector-request", "label": "requests vetted access including third-sector organizations", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-draft-research", "source": "w10-duty-draft2026", "target": "w10-duty-research-role", "label": "would restrict eligibility to approved university researchers", "status": "documented", "type": "governance" }, { "id": "w10-duty-e-research-narrower", "source": "w10-duty-research-role", "target": "w10-duty-third-sector-request", "label": "university requirement narrows requested independent eligibility", "status": "documented", "type": "policy" }, { "id": "w10-duty-e-minister-research", "source": "w10-duty-minister-role", "target": "w10-duty-research-role", "label": "would establish access and approval scheme through rules", "status": "documented", "type": "governance" }, { "id": "w10-duty-e-research-provider", "source": "w10-duty-research-role", "target": "w10-duty-provider-role", "label": "could obtain scheme-governed research access and test accounts", "status": "documented", "type": "resources" }, { "id": "w10-duty-e-public-draft", "source": "w10-duty-minister-role", "target": "w10-duty-draft2026", "label": "announces targeted consultation alongside public release", "status": "documented", "type": "policy" }, { "id": "w11-fund-e-luminate2021", "source": "w11-fund-luminate", "target": "w11-fund-carnegie-programme", "label": "accounts report restricted partnership income", "status": "documented", "type": "money" }, { "id": "w11-fund-e-minderoo2021", "source": "w11-fund-minderoo", "target": "w11-fund-carnegie-programme", "label": "accounts report restricted grant income", "status": "documented", "type": "money" }, { "id": "w11-fund-e-luminate2023", "source": "w11-fund-luminate", "target": "w11-fund-carnegie-programme", "label": "accounts report restricted partnership income", "status": "documented", "type": "money" }, { "id": "w11-fund-e-carnegie-programme", "source": "w10-duty-carnegie", "target": "w11-fund-carnegie-programme", "label": "operates financed duty-model programme", "status": "documented", "type": "resources" }, { "id": "w11-fund-e-programme-advice", "source": "w11-fund-carnegie-programme", "target": "w11-fund-un-advice", "label": "includes identifiable international advice output", "status": "documented", "type": "policy" }, { "id": "w11-fund-e-minderoo-advice", "source": "w11-fund-minderoo", "target": "w11-fund-un-advice", "label": "financial support credited through Frontier Technology initiative", "status": "reported", "type": "money" }, { "id": "w11-fund-e-carnegie-board", "source": "w11-fund-carnegie-board", "target": "w10-duty-carnegie", "label": "sets strategy and agrees annual budgets", "status": "documented", "type": "governance" }, { "id": "w11-fund-e-perrin-board", "source": "w10-duty-perrin", "target": "w11-fund-carnegie-board", "label": "listed as trustee and Audit and Risk Committee member", "status": "documented", "type": "person" }, { "id": "w11-fund-e-carnegie-network", "source": "w10-duty-carnegie", "target": "w10-duty-network", "label": "reports support for programme succession", "status": "reported", "type": "resources" }, { "id": "w11-fund-e-reset-network", "source": "reset", "target": "w10-duty-network", "label": "identified as supporter in dated transition account", "status": "reported", "type": "resources" }, { "id": "w11-fund-e-mannifera-acf", "source": "w11-fund-mannifera", "target": "w11-fund-acf", "label": "uses pooled ACF sub-fund route", "status": "documented", "type": "resources" }, { "id": "w11-fund-e-mannifera-reset", "source": "w11-fund-mannifera", "target": "w10-duty-reset-australia", "label": "records2024 grant for transparency advocacy", "status": "documented", "type": "money" }, { "id": "w11-fund-e-reset-purpose", "source": "w10-duty-reset-australia", "target": "w11-fund-reset-advocacy", "label": "named grantee for transparency-advocacy purpose", "status": "documented", "type": "policy" }, { "id": "w11-fund-e-mannifera-hrlc", "source": "w11-fund-mannifera", "target": "w11-fund-hrlc", "label": "records2024 grant for expert alignment and advocacy", "status": "documented", "type": "money" }, { "id": "w11-fund-e-hrlc-purpose", "source": "w11-fund-hrlc", "target": "w11-fund-alignment-work", "label": "named grantee for expert alignment and advocate training", "status": "documented", "type": "policy" }, { "id": "w11-fund-e-mckinnon-functioning", "source": "w11-fund-mckinnon", "target": "w11-fund-functioning", "label": "reports commissioning misinformation-code research", "status": "reported", "type": "money" }, { "id": "w11-fund-e-reset-functioning", "source": "w10-duty-reset-australia", "target": "w11-fund-functioning", "label": "publishes research with duty and assessment recommendations", "status": "documented", "type": "policy" }, { "id": "w11-fund-e-mckinnon-transparency", "source": "w11-fund-mckinnon", "target": "w11-fund-transparency", "label": "reports commissioning comparative transparency research", "status": "reported", "type": "money" }, { "id": "w11-fund-e-reset-transparency", "source": "w10-duty-reset-australia", "target": "w11-fund-transparency", "label": "publishes comparison proposing EU and UK mechanisms", "status": "documented", "type": "policy" }, { "id": "w11-fund-e-isf-award", "source": "w11-fund-isf", "target": "w11-fund-privacy-project", "label": "lists phase-two children privacy research award", "status": "documented", "type": "money" }, { "id": "w11-fund-e-reset-privacy", "source": "w10-duty-reset-australia", "target": "w11-fund-privacy-project", "label": "named operator of intended phase-two research", "status": "documented", "type": "resources" }, { "id": "w11-fund-e-isf-annual-support", "source": "w11-fund-isf", "target": "w10-duty-reset-australia", "label": "annual return reports monetary support to named legal recipient", "status": "documented", "type": "money" }, { "id": "w12-gov-e-reset-operates-network", "source": "reset", "target": "w10-duty-network", "label": "notice identifies operator of website, newsletter and shared inbox", "status": "reported", "type": "resources" }, { "id": "w12-gov-e-uk-notice-entity", "source": "w12-gov-reset-uk", "target": "reset", "label": "listed among notice-covered federated legal entities", "status": "reported", "type": "coordination" }, { "id": "w12-gov-e-uk-board-powers", "source": "w12-gov-uk-director-members", "target": "w12-gov-reset-uk", "label": "holds constitutional management and revocable delegation powers", "status": "documented", "type": "governance" }, { "id": "w12-gov-e-scott-board", "source": "w12-gov-scott", "target": "w12-gov-uk-director-members", "label": "listed as company director", "status": "documented", "type": "person" }, { "id": "w12-gov-e-wood-board", "source": "w12-gov-wood", "target": "w12-gov-uk-director-members", "label": "listed as company director", "status": "documented", "type": "person" }, { "id": "w12-gov-e-scott-adoption", "source": "w12-gov-scott", "target": "w12-gov-reset-uk", "label": "agrees to May2025 replacement constitution as voting member", "status": "documented", "type": "governance" }, { "id": "w12-gov-e-uk-member-limits", "source": "w12-gov-uk-director-members", "target": "w12-gov-uk-constitutional-limits", "label": "must obtain member unanimity for specified entrenched amendments", "status": "documented", "type": "governance" }, { "id": "w12-gov-e-uk-australia-group", "source": "w12-gov-reset-uk", "target": "w10-duty-reset-australia", "label": "expressly lists Australian entity within defined Group Company category", "status": "documented", "type": "governance" }, { "id": "w12-gov-e-action-receivable", "source": "w12-gov-reset-action-ggmbh", "target": "w12-gov-reset-uk", "label": "UK accounts report year-end amount receivable from named entity", "status": "documented", "type": "money" }, { "id": "w12-gov-e-member-selection", "source": "w12-gov-mann-members", "target": "w11-fund-mannifera", "label": "published historical process selects projects by staged consensus", "status": "documented", "type": "governance" }, { "id": "w12-gov-e-acf-trustee", "source": "w11-fund-acf", "target": "w11-fund-mannifera", "label": "retains trustee approval over hosted grant recommendations", "status": "documented", "type": "governance" }, { "id": "w12-gov-e-myer-membership", "source": "w12-gov-myer", "target": "w11-fund-mannifera", "label": "reports historical three-year membership award", "status": "documented", "type": "money" }, { "id": "w12-gov-e-myer-representation", "source": "w12-gov-myer", "target": "w12-gov-mann-members", "label": "names foundation representative as Mannifera board member", "status": "documented", "type": "governance" }, { "id": "w12-gov-e-reset-hrlc-partner", "source": "w10-duty-reset-australia", "target": "w11-fund-hrlc", "label": "grantor names HRLC partner in funded transparency advocacy", "status": "reported", "type": "coordination" }, { "id": "w12-gov-e-reset-drw-partner", "source": "w10-duty-reset-australia", "target": "w6-ref-drw", "label": "grantor names Digital Rights Watch partner in transparency advocacy", "status": "reported", "type": "coordination" }, { "id": "w12-gov-e-mann-rights-first", "source": "w11-fund-mannifera", "target": "w12-gov-rights-first", "label": "publication acknowledges grant support", "status": "reported", "type": "money" }, { "id": "w12-gov-e-mejia-authorship", "source": "w12-gov-mejia", "target": "w12-gov-rights-first", "label": "credited as report author", "status": "documented", "type": "policy" }, { "id": "w12-gov-e-hrlc-launch", "source": "w11-fund-hrlc", "target": "w12-gov-rights-first", "label": "launches related institutional regulatory recommendations", "status": "documented", "type": "policy" }, { "id": "w12-gov-e-hrlc-parliament", "source": "w11-fund-hrlc", "target": "w12-gov-hrlc-engagement", "label": "reports taking framework to Parliament and government", "status": "reported", "type": "policy" }, { "id": "w12-gov-e-author-parliament", "source": "w12-gov-mejia", "target": "w12-gov-hrlc-engagement", "label": "named representative in HRLC engagement account", "status": "reported", "type": "person" }, { "id": "w12-gov-e-mckinnon-entity-trust", "source": "w11-fund-mckinnon", "target": "w12-gov-mckinnon-trust", "label": "funding page identifies distinct entity within organization", "status": "reported", "type": "coordination" }, { "id": "w12-gov-e-mckinnon-entity-research", "source": "w11-fund-mckinnon", "target": "w12-gov-research-centre", "label": "funding page identifies distinct entity within organization", "status": "reported", "type": "coordination" }, { "id": "w12-gov-e-mckinnon-entity-charitable", "source": "w11-fund-mckinnon", "target": "w12-gov-charitable-foundation", "label": "funding page identifies distinct entity within organization", "status": "reported", "type": "coordination" }, { "id": "w12-gov-e-trust-research-funding", "source": "w12-gov-mckinnon-trust", "target": "w12-gov-research-centre", "label": "both entity reports record one FY2024 intra-group donation", "status": "documented", "type": "money" }, { "id": "w12-gov-e-committee-oversight", "source": "w12-gov-research-committee", "target": "w12-gov-research-centre", "label": "oversees application of research funds and publication/dissemination", "status": "documented", "type": "governance" }, { "id": "w12-gov-e-mckinnon-briefing", "source": "w11-fund-mckinnon", "target": "w12-gov-april-briefing", "label": "briefing acknowledges support", "status": "reported", "type": "resources" }, { "id": "w12-gov-e-reset-convening", "source": "w10-duty-reset-australia", "target": "w12-gov-march-roundtable", "label": "briefing reports convening academic and civil-society experts", "status": "reported", "type": "coordination" }, { "id": "w12-gov-e-convening-output", "source": "w12-gov-march-roundtable", "target": "w12-gov-april-briefing", "label": "discussion informs published regulatory recommendations", "status": "reported", "type": "policy" }, { "id": "w13-disp-e-network-feb", "source": "w10-duty-network", "target": "w13-disp-osan-feb", "label": "submitted identified illegal-harms consultation response", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-feb-inference", "source": "w13-disp-osan-feb", "target": "w13-disp-civil-inference", "label": "proposed contextual civil interpretation", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-woods-inference", "source": "w10-duty-woods", "target": "w13-disp-civil-inference", "label": "publicly argued contextual civil approach", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-ofcom-declines-inference", "source": "ofcom", "target": "w13-disp-civil-inference", "label": "declined proposed interpretive departure", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-feb-search", "source": "w13-disp-osan-feb", "target": "w13-disp-search-risk", "label": "raised mismatch in risk discussion and search measures", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-ofcom-search-amendment", "source": "ofcom", "target": "w13-disp-search-risk", "label": "reports amending risk register while retaining statutory limit", "status": "reported", "type": "policy" }, { "id": "w13-disp-e-network-july", "source": "w10-duty-network", "target": "w13-disp-osan-july", "label": "published named children-consultation response", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-woods-july", "source": "w10-duty-woods", "target": "w13-disp-osan-july", "label": "named as co-author", "status": "documented", "type": "personnel" }, { "id": "w13-disp-e-walsh-july", "source": "w13-disp-maeve-walsh", "target": "w13-disp-osan-july", "label": "named as co-author", "status": "documented", "type": "personnel" }, { "id": "w13-disp-e-ofcom-age-clarification", "source": "ofcom", "target": "w13-disp-child-risk-ages", "label": "published clarified risk-assessment assumptions", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-osan-credited-feedback", "source": "w10-duty-network", "target": "w13-disp-child-risk-ages", "label": "Ofcom credits network among contributing civil-society feedback", "status": "reported", "type": "policy" }, { "id": "w13-disp-e-hrlc-submission", "source": "w11-fund-hrlc", "target": "w13-disp-hrlc155", "label": "issued review submission", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-hrlc-reset-endorsement", "source": "w13-disp-hrlc155", "target": "w13-disp-reset-five-pillars", "label": "expressly recommends government adopt Reset risk-based model", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-reset-model-attribution", "source": "w10-duty-reset-australia", "target": "w13-disp-reset-five-pillars", "label": "HRLC explicitly attributes five-pillar model to Reset", "status": "reported", "type": "policy" }, { "id": "w13-disp-e-hrlc-taxonomy", "source": "w13-disp-hrlc155", "target": "w13-disp-wef-typology", "label": "expressly adopts typology as harm framework", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-commitment-review", "source": "w13-disp-au-commitment", "target": "w10-duty-review2025", "label": "ministerial announcement expressly grounds direction in received review", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-brief-awareness-hrlc", "source": "w13-disp-au-briefing", "target": "w11-fund-hrlc", "label": "official briefing records public advocacy in media summary", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-brief-awareness-reset", "source": "w13-disp-au-briefing", "target": "w10-duty-reset-australia", "label": "official briefing records public advocacy in media summary", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-government-rec4", "source": "w13-disp-au-response", "target": "w10-duty-review2025", "label": "supports review recommendation4: overarching duty", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-government-rec7", "source": "w13-disp-au-response", "target": "w10-duty-review2025", "label": "does not support recommendation7: automatic highest tier", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-government-rec9", "source": "w13-disp-au-response", "target": "w10-duty-review2025", "label": "supports recommendation9 in principle with altered rule-maker", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-government-rec13", "source": "w13-disp-au-response", "target": "w10-duty-review2025", "label": "supports recommendation13 researcher access in principle", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-government-rec25", "source": "w13-disp-au-response", "target": "w10-duty-review2025", "label": "supports recommendation25: accessible internal complaints", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-government-rec26", "source": "w13-disp-au-response", "target": "w10-duty-review2025", "label": "notes recommendation26: digital Ombuds scheme", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-hrlc-civil-access", "source": "w13-disp-hrlc155", "target": "w13-disp-hrlc-civil-research", "label": "requests access for vetted researchers and civil-society organizations", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-draft-narrows-hrlc-access", "source": "w10-duty-research-role", "target": "w13-disp-hrlc-civil-research", "label": "proposed university eligibility is narrower than HRLC direct-access request", "status": "inferred", "type": "policy" }, { "id": "w13-disp-e-rights-safeguards", "source": "w12-gov-rights-first", "target": "w13-disp-rights-safeguards", "label": "sets out reasons, remedy and review safeguards", "status": "documented", "type": "policy" }, { "id": "w13-disp-e-draft-safeguard-comparison", "source": "w10-duty-draft2026", "target": "w13-disp-rights-safeguards", "label": "inspected clauses do not enact full proposed rights framework", "status": "inferred", "type": "policy" }, { "id": "w14-tax-e-published", "source": "wef", "target": "w13-disp-wef-typology", "label": "publishes coalition typology", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-hildreth-cochair", "source": "w14-tax-hildreth", "target": "w13-disp-wef-typology", "label": "participant reports co-chairing compilation workstream", "status": "reported", "type": "governance" }, { "id": "w14-tax-e-inman-cochair", "source": "inman-grant", "target": "w13-disp-wef-typology", "label": "participant reports co-chairing compilation workstream", "status": "reported", "type": "governance" }, { "id": "w14-tax-e-hanna-drafting", "source": "w14-tax-hanna", "target": "w13-disp-wef-typology", "label": "biography reports drafting-committee participation", "status": "reported", "type": "personnel" }, { "id": "w14-tax-e-hildreth-foreword", "source": "w14-tax-hildreth", "target": "w13-disp-wef-typology", "label": "credited as foreword signatory", "status": "documented", "type": "personnel" }, { "id": "w14-tax-e-inman-foreword", "source": "inman-grant", "target": "w13-disp-wef-typology", "label": "credited as foreword signatory", "status": "documented", "type": "personnel" }, { "id": "w14-tax-e-dobrygowski-foreword", "source": "w14-tax-dobrygowski", "target": "w13-disp-wef-typology", "label": "credited as foreword signatory", "status": "documented", "type": "personnel" }, { "id": "w14-tax-e-bantourakis-foreword", "source": "w14-tax-bantourakis", "target": "w13-disp-wef-typology", "label": "credited as foreword signatory", "status": "documented", "type": "personnel" }, { "id": "w14-tax-e-core-lineage", "source": "w13-disp-wef-typology", "target": "w14-tax-core4cs", "label": "participant account identifies prior research as antecedent", "status": "reported", "type": "policy" }, { "id": "w14-tax-e-official-use", "source": "w14-tax-issues2024", "target": "w13-disp-wef-typology", "label": "cites typology for descriptive harm framing", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-guidance-use", "source": "w14-tax-esafety-harms-guidance", "target": "w13-disp-wef-typology", "label": "explicitly uses typology for educational categories", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-esafety-guidance", "source": "esafety", "target": "w14-tax-esafety-harms-guidance", "label": "publishes industry Safety by Design guidance", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-model-scope", "source": "w10-duty-reset-australia", "target": "w13-disp-reset-five-pillars", "label": "publishes broader risk model with named international comparators", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-review-rec5", "source": "w10-duty-review2025", "target": "w14-tax-review-categories", "label": "proposes harm families in recommendation5", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-review-defers", "source": "w10-duty-review2025", "target": "w14-tax-deferred-consolidation", "label": "defers broader consolidation to other regimes or future review", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-government-rec5", "source": "w13-disp-au-response", "target": "w14-tax-review-categories", "label": "supports recommendation5 in principle", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-draft-initial", "source": "w10-duty-draft2026", "target": "w14-tax-draft-categories", "label": "specifies initial harm and child/feature categories", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-minister-additions", "source": "w10-duty-minister-role", "target": "w14-tax-draft-categories", "label": "would add harm or feature categories under differing tests", "status": "documented", "type": "governance" }, { "id": "w14-tax-e-draft-assessment", "source": "w10-duty-draft2026", "target": "w14-tax-draft-risk-scope", "label": "would assess foreseeable risks including combined and non-user harms", "status": "documented", "type": "policy" }, { "id": "w14-tax-e-earlier-law", "source": "w14-tax-act2021", "target": "w14-tax-draft-categories", "label": "provides earlier substantive harm-test comparator", "status": "inferred", "type": "policy" }, { "id": "w14-tax-e-version-callouts", "source": "w14-tax-review-version-brief", "target": "w10-duty-review2025", "label": "records post-delivery changes with conflicting summaries", "status": "documented", "type": "policy" }, { "id": "w15-impl-e-group-uk", "source": "w15-impl-crisp-group", "target": "w15-impl-crisp-uk", "label": "registered corporate PSC holds75%+ shares", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-ki-group", "source": "w15-impl-kroll-international", "target": "w15-impl-crisp-group", "label": "recorded historical75%+ shares and voting rights", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-ka-group", "source": "w15-impl-kroll-advisory", "target": "w15-impl-crisp-group", "label": "recorded current75%+ shares and voting rights", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-group-brand", "source": "w15-impl-crisp-group", "target": "w15-impl-crisp-brand", "label": "privacy policy identifies company trading as Crisp", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-hildreth-brand", "source": "w14-tax-hildreth", "target": "w15-impl-crisp-brand", "label": "named founder in dated participant account", "status": "reported", "type": "personnel" }, { "id": "w15-impl-e-brand-alignment", "source": "w15-impl-crisp-brand", "target": "w15-impl-resolver-brand", "label": "announces operation under Resolver banner", "status": "reported", "type": "coordination" }, { "id": "w15-impl-e-terms-provider", "source": "w15-impl-resolver-brand", "target": "w15-impl-resolver-inc", "label": "served standard terms name provider and applicable affiliates", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-service-instructions", "source": "w15-impl-crisp-brand", "target": "w15-impl-customer-instructions", "label": "describes moderation subject to customer instructions", "status": "reported", "type": "infrastructure" }, { "id": "w15-impl-e-dsit-payment", "source": "w15-impl-dsit", "target": "w15-impl-crisp-uk", "label": "publishes supplier payment for professional services", "status": "documented", "type": "money" }, { "id": "w15-impl-e-illuminate-partnership", "source": "w15-impl-resolver-brand", "target": "illuminate-tech", "label": "announces compliance-service collaboration", "status": "reported", "type": "coordination" }, { "id": "w15-impl-e-illuminate-oscar", "source": "illuminate-tech", "target": "w15-impl-oscar", "label": "announcement identifies tool supplier", "status": "reported", "type": "infrastructure" }, { "id": "w15-impl-e-review-licence", "source": "w10-duty-review2025", "target": "w15-impl-licensing45", "label": "recommends considering licensing as operation condition", "status": "documented", "type": "policy" }, { "id": "w15-impl-e-licence-rejected", "source": "w13-disp-au-response", "target": "w15-impl-licensing45", "label": "does not support recommendation45 licensing proposal", "status": "documented", "type": "policy" }, { "id": "w15-impl-e-government-rec42", "source": "w13-disp-au-response", "target": "w10-duty-review2025", "label": "notes recommendation42 business-disruption powers", "status": "documented", "type": "policy" }, { "id": "w15-impl-e-government-rec43", "source": "w13-disp-au-response", "target": "w10-duty-review2025", "label": "does not support recommendation43 mandatory local presence", "status": "documented", "type": "policy" }, { "id": "w15-impl-e-government-rec44", "source": "w13-disp-au-response", "target": "w10-duty-review2025", "label": "supports recommendation44 Australian service contact", "status": "documented", "type": "policy" }, { "id": "w15-impl-e-review-phone", "source": "w10-duty-review2025", "target": "w15-impl-identification48", "label": "proposes phone-at-opening, confidentiality and preservation", "status": "documented", "type": "policy" }, { "id": "w15-impl-e-government-phone", "source": "w13-disp-au-response", "target": "w15-impl-identification48", "label": "supports recommendation48 in principle", "status": "documented", "type": "policy" }, { "id": "w15-impl-e-draft-phone-comparison", "source": "w10-duty-draft2026", "target": "w15-impl-identification48", "label": "inspected amendments do not expressly implement phone proposal", "status": "inferred", "type": "policy" }, { "id": "w15-impl-e-dsit-unit", "source": "w15-impl-dsit", "target": "w15-impl-nsoit", "label": "identifies department operational team", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-minister-remit", "source": "w15-impl-dsit-ministers", "target": "w15-impl-nsoit", "label": "government says ministers agree and review remit", "status": "reported", "type": "governance" }, { "id": "w15-impl-e-draft-policy", "source": "w15-impl-nsoit", "target": "w15-impl-nsoit-protocol", "label": "disclosed draft describes operating and escalation rules", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-partner-directions", "source": "w15-impl-nsoit-protocol", "target": "w15-impl-nsoit-partners", "label": "would require buyer-agreed instructions and methods", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-draft-escalation", "source": "w15-impl-nsoit-protocol", "target": "w15-impl-platform-review", "label": "describes authorized referrals with platform discretion", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-crisp-output", "source": "w15-impl-crisp-brand", "target": "w15-impl-crisp-report2024", "label": "ICO confirms supplier-produced report was held", "status": "documented", "type": "infrastructure" }, { "id": "w15-impl-e-nsoit-report", "source": "w15-impl-nsoit", "target": "w15-impl-crisp-report2024", "label": "department held and partly disclosed report", "status": "documented", "type": "infrastructure" }, { "id": "w15-impl-e-lawful-referrals", "source": "w15-impl-nsoit", "target": "w15-impl-platform-review", "label": "testimony describes lawful-content warnings and review requests", "status": "reported", "type": "policy" }, { "id": "w15-impl-e-ico-decision", "source": "w7-impl-ico", "target": "w15-impl-crisp-report2024", "label": "upholds remaining national-security redactions", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-gsn-appointed", "source": "w15-impl-nsoit", "target": "w15-impl-gsn", "label": "current privacy notice names analysis contractor", "status": "reported", "type": "money" }, { "id": "w15-impl-e-oversight-request", "source": "w15-impl-science-committee", "target": "w15-impl-nsoit", "label": "recommends statutory footing and intelligence-committee oversight", "status": "documented", "type": "governance" }, { "id": "w15-impl-e-oversight-answer", "source": "w15-impl-dsit", "target": "w15-impl-science-committee", "label": "defends existing review channels without committing to requested change", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-barrett-gsn", "source": "w16-supp-barrett", "target": "w15-impl-gsn", "label": "register reports more than25% through50% shares and votes", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-king-gsn", "source": "w16-supp-king", "target": "w15-impl-gsn", "label": "register reports more than25% through50% shares and votes", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-gsn-payment", "source": "w15-impl-dsit", "target": "w15-impl-gsn", "label": "publishes payment to named analysis supplier", "status": "documented", "type": "money" }, { "id": "w16-supp-e-gsn-analysis", "source": "w15-impl-gsn", "target": "w15-impl-nsoit", "label": "ministerial answer identifies narrative-analysis function", "status": "reported", "type": "infrastructure" }, { "id": "w16-supp-e-faculty-platform", "source": "w16-supp-faculty", "target": "w16-supp-cddp", "label": "ministerial answer identifies platform development function", "status": "reported", "type": "infrastructure" }, { "id": "w16-supp-e-tender-platform", "source": "w16-supp-procurement26087", "target": "w16-supp-cddp", "label": "specifies development,hosting andmaintenance procurement", "status": "documented", "type": "infrastructure" }, { "id": "w16-supp-e-government-priorities", "source": "w15-impl-dsit", "target": "w16-supp-cddp", "label": "tender assigns government team technical priorities and testing", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-faculty-incumbency", "source": "w16-supp-faculty", "target": "w16-supp-procurement26087", "label": "procurement names incumbent in extendedAlpha", "status": "reported", "type": "infrastructure" }, { "id": "w16-supp-e-business-case", "source": "w16-supp-dcms", "target": "w16-supp-cddp", "label": "business case proposes cross-government data-sharing capacity", "status": "documented", "type": "policy" }, { "id": "w16-supp-e-dsit-variation", "source": "w15-impl-dsit", "target": "w16-supp-faculty102548", "label": "disclosed variation names department asbuyer", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-variation-cap", "source": "w16-supp-faculty102548", "target": "w16-supp-faculty", "label": "specifies capped fee for additional services", "status": "documented", "type": "money" }, { "id": "w16-supp-e-ip-variation", "source": "w16-supp-faculty102548", "target": "w16-supp-licensed-ip", "label": "varies ownership and licensing across two call-offs", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-supplier-ip", "source": "w16-supp-faculty", "target": "w16-supp-licensed-ip", "label": "variation allocates new IP and software ownership to supplier", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-buyer-licence", "source": "w15-impl-dsit", "target": "w16-supp-licensed-ip", "label": "variation provides perpetual buyer licence with misuse obligations", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-authorise-maintainers", "source": "w15-impl-dsit", "target": "w16-supp-authorised-maintainers", "label": "variation permits buyer to authorise third-party services", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-maintainer-limits", "source": "w16-supp-authorised-maintainers", "target": "w16-supp-licensed-ip", "label": "permitted use serves buyer with commercial restrictions", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-source-access", "source": "w16-supp-faculty", "target": "w15-impl-dsit", "label": "variation requires accessible source code and documentation", "status": "documented", "type": "infrastructure" }, { "id": "w16-supp-e-dcms-trail", "source": "w16-supp-dcms", "target": "w16-supp-faculty102887", "label": "named buyer in separate102887 variation trail", "status": "documented", "type": "governance" }, { "id": "w16-supp-e-fixed-fee-acceptance", "source": "w16-supp-faculty102887", "target": "w16-supp-faculty", "label": "specifies fixed fee payable against delivery and buyer acceptance", "status": "documented", "type": "money" }, { "id": "w16-supp-e-newswhip-contract", "source": "w16-supp-dcms", "target": "w16-supp-newswhip", "label": "disclosed contract identifies monitoring-tool supplier", "status": "documented", "type": "infrastructure" }, { "id": "w22-sig-e-funding", "source": "w22-sig-tc", "target": "w22-sig-lantern", "label": "reports member-supported programme funding", "status": "reported", "type": "money" }, { "id": "w22-sig-e-hosting", "source": "meta", "target": "w22-sig-lantern", "label": "provides hosting in kind, according to reports", "status": "reported", "type": "infrastructure" }, { "id": "w22-sig-e-admission", "source": "w22-sig-tc", "target": "w22-sig-lantern", "label": "administers entry, compliance and database access", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-taxonomy-admin", "source": "w22-sig-tc", "target": "w22-sig-taxonomy", "label": "maintains taxonomy with participants", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-taxonomy-input", "source": "w22-sig-participants", "target": "w22-sig-taxonomy", "label": "may propose taxonomy updates", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-taxonomy-rules", "source": "w22-sig-taxonomy", "target": "w22-sig-lantern", "label": "supplies required investigation labels", "status": "documented", "type": "policy" }, { "id": "w22-sig-e-own-review", "source": "w22-sig-lantern", "target": "w22-sig-participants", "label": "requires review and own-policy decisions", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-appeal-removal", "source": "w22-sig-lantern", "target": "w22-sig-participants", "label": "requires recourse, relevant signals and TC notification", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-remove-own", "source": "w22-sig-lantern", "target": "w22-sig-participants", "label": "limits participant deletion to own uploads", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-compliance-consequence", "source": "w22-sig-tc", "target": "w22-sig-noncontinuation", "label": "reported a participant would not continue", "status": "reported", "type": "governance" }, { "id": "w22-sig-e-pilot", "source": "w22-sig-lantern", "target": "w22-sig-pilot", "label": "reports receive-only financial pilot", "status": "reported", "type": "coordination" }, { "id": "w22-sig-e-pilot-cashapp", "source": "w22-sig-cashapp", "target": "w22-sig-pilot", "label": "named as pilot participant", "status": "reported", "type": "coordination" }, { "id": "w22-sig-e-pilot-paypal", "source": "paypal", "target": "w22-sig-pilot", "label": "named as pilot participant", "status": "reported", "type": "coordination" }, { "id": "w22-sig-e-finance-expansion", "source": "w22-sig-lantern", "target": "w22-sig-financial-eligibility", "label": "announced worldwide eligibility expansion", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-meta-case", "source": "meta", "target": "w22-sig-case2025", "label": "conducted investigation, operator reports", "status": "reported", "type": "coordination" }, { "id": "w22-sig-e-case-signals", "source": "w22-sig-case2025", "target": "w22-sig-lantern", "label": "generated shared Meta signals, operator reports", "status": "reported", "type": "infrastructure" }, { "id": "w22-sig-e-block-case", "source": "w22-sig-lantern", "target": "w22-sig-block", "label": "informed case investigation, operator reports", "status": "reported", "type": "infrastructure" }, { "id": "w22-sig-e-cashapp-outcome", "source": "w22-sig-block", "target": "w22-sig-cashapp", "label": "closed 18 accounts in the reported case", "status": "reported", "type": "governance" }, { "id": "w22-sig-e-xbox-outcome", "source": "microsoft", "target": "w22-sig-case2025", "label": "acted on Xbox accounts, operator reports", "status": "reported", "type": "governance" }, { "id": "w22-sig-e-block-use", "source": "w22-sig-block", "target": "w22-sig-lantern", "label": "reports investigative use through Cash App", "status": "reported", "type": "infrastructure" }, { "id": "w22-sig-e-block-member", "source": "w22-sig-block", "target": "w22-sig-tc", "label": "reports Tech Coalition membership", "status": "reported", "type": "coordination" }, { "id": "w22-sig-e-icmec-involvement", "source": "w22-sig-cashapp", "target": "icmec-financial-coalition", "label": "reports regular involvement in US coalition", "status": "reported", "type": "coordination" }, { "id": "w22-sig-e-hria-commission", "source": "w22-sig-tc", "target": "w22-sig-bsr", "label": "commissioned Lantern human-rights assessment", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-hria-output", "source": "w22-sig-bsr", "target": "w22-sig-hria2023", "label": "published human-rights risk assessment", "status": "documented", "type": "narrative" }, { "id": "w22-sig-e-hria-response", "source": "w22-sig-tc", "target": "w22-sig-hria2023", "label": "reported nine recommendations implemented", "status": "reported", "type": "governance" }, { "id": "w22-sig-e-classifier-workflow", "source": "w22-sig-sightengine", "target": "w22-sig-checkstep", "label": "describes supplying classification outputs", "status": "documented", "type": "infrastructure" }, { "id": "w22-sig-e-workflow", "source": "w22-sig-checkstep", "target": "w22-sig-customer-workflow", "label": "documents configurable moderation workflow", "status": "documented", "type": "infrastructure" }, { "id": "w22-sig-e-policy-choice", "source": "w22-sig-customer-role", "target": "w22-sig-customer-workflow", "label": "selects policy configuration in described workflow", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-webhook-decisions", "source": "w22-sig-checkstep", "target": "w22-sig-customer-role", "label": "can return content and account action instructions", "status": "documented", "type": "infrastructure" }, { "id": "w22-sig-e-appeal-events", "source": "w22-sig-checkstep", "target": "w22-sig-customer-role", "label": "can return appeal and restoration events", "status": "documented", "type": "infrastructure" }, { "id": "w22-sig-e-checkstep-training", "source": "w22-sig-checkstep", "target": "w22-sig-checkstep-training", "label": "publishes opt-out customer-data ML permission", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-checkstep-optout", "source": "w22-sig-customer-role", "target": "w22-sig-checkstep-training", "label": "may opt out under published terms", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-checkstep-controller", "source": "w22-sig-customer-role", "target": "w22-sig-checkstep", "label": "retains controller role under published terms", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-feedback-training", "source": "w22-sig-sightengine", "target": "w22-sig-feedback", "label": "may use voluntarily submitted media for model work", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-feedback-controller", "source": "w22-sig-sightengine", "target": "w22-sig-feedback", "label": "acts as independent controller under feedback exception", "status": "documented", "type": "governance" }, { "id": "w22-sig-e-sightengine-processor", "source": "w22-sig-customer-role", "target": "w22-sig-sightengine", "label": "instructs ordinary processing under published DPA", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-terms", "source": "w22-sig-block", "target": "w23-cash-terms", "label": "publishes contracting terms", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-incorporates-aup", "source": "w23-cash-terms", "target": "w23-cash-aup", "label": "incorporates acceptable-use policy", "status": "documented", "type": "policy" }, { "id": "w23-cash-e-external-rules", "source": "w23-cash-aup", "target": "w23-cash-external-policies", "label": "references bank and card-network policies", "status": "documented", "type": "policy" }, { "id": "w23-cash-e-termination", "source": "w22-sig-block", "target": "w23-cash-customers", "label": "reserves suspension and termination discretion", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-conditional-balance", "source": "w22-sig-block", "target": "w23-cash-customers", "label": "reserves conditional holds and withdrawal access", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-legal-process", "source": "w22-sig-block", "target": "w23-cash-customers", "label": "reserves responses to legal process and government requests", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-custody-access", "source": "w23-cash-customers", "target": "w22-sig-cashapp", "label": "must use Cash App to access custodial balances under terms", "status": "documented", "type": "infrastructure" }, { "id": "w23-cash-e-help-route", "source": "w22-sig-cashapp", "target": "w23-cash-help", "label": "publishes Support and eligible-case review route", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-cfpb-issued", "source": "w23-cash-cfpb", "target": "w23-cash-cfpb-order", "label": "entered consent order against Block", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-historical-recourse", "source": "w23-cash-cfpb", "target": "w22-sig-block", "label": "found inadequate account-restriction recourse", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-issuer-disputes", "source": "w22-sig-block", "target": "w23-cash-cfpb-order", "label": "disputed characterization and claimed improvements", "status": "reported", "type": "narrative" }, { "id": "w23-cash-e-restriction-duty", "source": "w23-cash-cfpb-order", "target": "w23-cash-restriction-procedure", "label": "requires restriction notice and appeal procedure", "status": "documented", "type": "policy" }, { "id": "w23-cash-e-restriction-block", "source": "w22-sig-block", "target": "w23-cash-restriction-procedure", "label": "must implement account-restriction safeguards", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-error-duty", "source": "w23-cash-cfpb-order", "target": "w23-cash-error-procedure", "label": "requires distinct transfer-error investigation and evidence access", "status": "documented", "type": "policy" }, { "id": "w23-cash-e-fraud-controls", "source": "w23-cash-cfpb-order", "target": "w22-sig-block", "label": "requires fraud controls including account exclusion", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-working-groups", "source": "w22-sig-block", "target": "w23-cash-working-groups", "label": "must maintain specified internal oversight roles", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-board-cfpb", "source": "w23-cash-block-board", "target": "w23-cash-compliance-plan", "label": "has ultimate compliance responsibility", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-compliance-create", "source": "w22-sig-block", "target": "w23-cash-compliance-plan", "label": "must create and implement within 120 days", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-compliance-inspect", "source": "w23-cash-cfpb", "target": "w23-cash-compliance-plan", "label": "may obtain plan and compliance records on request", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-redress-review", "source": "w23-cash-cfpb-enforcement", "target": "w23-cash-redress-plan", "label": "reviews for non-objection and may require revision", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-redress-reserve", "source": "w22-sig-block", "target": "w23-cash-redress-plan", "label": "must reserve funds for ordered redress", "status": "documented", "type": "money" }, { "id": "w23-cash-e-penalty-paid", "source": "w22-sig-block", "target": "w23-cash-cfpb", "label": "was ordered to pay a penalty and reports paying it", "status": "reported", "type": "money" }, { "id": "w23-cash-e-state-issued", "source": "w23-cash-state-regulators", "target": "w23-cash-state-order", "label": "entered multistate consent instrument with Block", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-state-committee", "source": "w23-cash-state-regulators", "target": "w23-cash-state-exec", "label": "act through specified Executive Committee", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-company-committee", "source": "w22-sig-block", "target": "w23-cash-company-committee", "label": "must appoint company compliance committee", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-state-board", "source": "w23-cash-block-board", "target": "w23-cash-state-cap", "label": "must formally approve corrective action plan", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-consultant-assess", "source": "w23-cash-consultant", "target": "w23-cash-state-cap", "label": "must assess, prioritize and validate corrections", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-consultant-access", "source": "w22-sig-block", "target": "w23-cash-consultant", "label": "must provide personnel, vendor and record access", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-consultant-replacement", "source": "w22-sig-block", "target": "w23-cash-consultant", "label": "selects replacement subject to independence and notice", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-state-report-access", "source": "w23-cash-state-exec", "target": "w23-cash-consultant", "label": "receives reports and may inspect workpapers", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-state-extensions", "source": "w23-cash-state-exec", "target": "w23-cash-state-cap", "label": "may extend reporting deadlines on good cause", "status": "documented", "type": "governance" }, { "id": "w23-cash-e-state-obligation", "source": "w22-sig-block", "target": "w23-cash-state-regulators", "label": "owes separate penalty and administrative costs", "status": "documented", "type": "money" }, { "id": "w24-ag-e-july-announced", "source": "w24-ag-july-ags", "target": "w24-ag-july-resolution", "label": "announced agreed multistate resolution", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-july-assent", "source": "w22-sig-block", "target": "w24-ag-july-resolution", "label": "signed Oregon party assent to agreed terms", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-july-service", "source": "w24-ag-july-resolution", "target": "w22-sig-block", "label": "sets agreed support and escalation duties", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-july-restriction", "source": "w24-ag-july-resolution", "target": "w22-sig-block", "label": "sets agreed restriction-review procedures", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-july-errors-fraud", "source": "w24-ag-july-resolution", "target": "w22-sig-block", "label": "sets distinct transfer-error and anti-fraud duties", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-july-corporate", "source": "w22-sig-block", "target": "w24-ag-july-compliance", "label": "agreed to maintain internal compliance oversight", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-july-board-report", "source": "w24-ag-july-compliance", "target": "w23-cash-block-board", "label": "must provide quarterly compliance reporting", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-july-redress", "source": "w24-ag-july-resolution", "target": "w24-ag-july-redress", "label": "sets credit and conditional fallback mechanism", "status": "documented", "type": "policy" }, { "id": "w24-ag-e-credits-cfpb", "source": "w24-ag-july-redress", "target": "w23-cash-redress-plan", "label": "credits payments under federal redress order", "status": "documented", "type": "policy" }, { "id": "w24-ag-e-fallback-approval", "source": "w24-ag-july-exec", "target": "w24-ag-july-redress", "label": "must approve conditional fallback plan before payment", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-july-enforcement", "source": "w24-ag-july-ags", "target": "w22-sig-block", "label": "agreed a conditional notice and response process", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-naag-reimbursement", "source": "w22-sig-block", "target": "w24-ag-naag", "label": "agreed investigation-expense reimbursement allocation", "status": "documented", "type": "money" }, { "id": "w24-ag-e-naag-governance", "source": "w24-ag-naag-exec", "target": "w24-ag-naag", "label": "governs funds, property and management per current page", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-naag-director", "source": "w24-ag-naag-exec", "target": "w24-ag-naag-director", "label": "employs executive director for daily operations", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-consolidated-remit", "source": "w24-ag-naag", "target": "w24-ag-consolidated", "label": "describes separate fund-oversight committee", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-member-capacity", "source": "w24-ag-naag", "target": "w24-ag-member-offices", "label": "describes grants, expertise and collaboration support", "status": "reported", "type": "infrastructure" }, { "id": "w24-ag-e-resource-policy", "source": "w24-ag-member-offices", "target": "w24-ag-naag", "label": "supply service fees under stated operating model", "status": "reported", "type": "money" }, { "id": "w24-ag-e-vw-provision", "source": "w24-ag-vw-allocation", "target": "w24-ag-naag", "label": "rules describe specified settlement allocation", "status": "reported", "type": "money" }, { "id": "w24-ag-e-vw-appoint", "source": "w24-ag-naag-president", "target": "w24-ag-vw-committee", "label": "appoints three-AG committee subject to approval", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-vw-appointment-approval", "source": "w24-ag-naag-exec", "target": "w24-ag-vw-committee", "label": "approves presidential committee appointments", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-vw-grants", "source": "w24-ag-vw-committee", "target": "w24-ag-vw-fund", "label": "approves principal grants by two affirmative votes", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-vw-purpose", "source": "w24-ag-vw-fund", "target": "w24-ag-member-offices", "label": "may finance purpose-limited work with conditional repayment", "status": "documented", "type": "money" }, { "id": "w24-ag-e-vw-fee", "source": "w24-ag-vw-fund", "target": "w24-ag-naag", "label": "authorizes inception and annual administration fee", "status": "documented", "type": "money" }, { "id": "w24-ag-e-vw-dissolution", "source": "w24-ag-naag-exec", "target": "w24-ag-vw-fund", "label": "may dissolve upon specified purpose determination", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-fsf-appoint", "source": "w24-ag-naag-president", "target": "w24-ag-fsf-committee", "label": "appoints five-AG committee subject to approval", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-fsf-appointment-approval", "source": "w24-ag-naag-exec", "target": "w24-ag-fsf-committee", "label": "approves presidential committee appointments", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-fsf-grants", "source": "w24-ag-fsf-committee", "target": "w24-ag-fsf", "label": "approves grants using size and party thresholds", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-fsf-purpose", "source": "w24-ag-fsf", "target": "w24-ag-member-offices", "label": "may finance qualifying investigations and training", "status": "documented", "type": "money" }, { "id": "w24-ag-e-fsf-fee", "source": "w24-ag-fsf", "target": "w24-ag-naag", "label": "authorizes annual administration fee at fiscal close", "status": "documented", "type": "money" }, { "id": "w24-ag-e-foundation-support", "source": "w24-ag-foundation", "target": "w24-ag-naag", "label": "reports other support to its sole supported organization", "status": "reported", "type": "money" }, { "id": "w24-ag-e-policy-procedure", "source": "w24-ag-naag", "target": "w24-ag-procedure", "label": "publishes sponsorship and circulation requirements", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-jurisdiction-signoff", "source": "w24-ag-member-offices", "target": "w24-ag-procedure", "label": "confirm or reject each circulated draft under process", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-letterhead-threshold", "source": "w24-ag-procedure", "target": "w24-ag-naag", "label": "sets 36-state threshold for association distribution", "status": "documented", "type": "governance" }, { "id": "w24-ag-e-payment-endorsement", "source": "w24-ag-letter-coalition", "target": "w24-ag-payment-letter", "label": "endorsed requests to payment-platform counsel", "status": "documented", "type": "coordination" }, { "id": "w24-ag-e-search-endorsement", "source": "w24-ag-letter-coalition", "target": "w24-ag-search-letter", "label": "endorsed separate search-provider requests", "status": "documented", "type": "coordination" }, { "id": "w24-ag-e-letterhead", "source": "w24-ag-naag", "target": "w24-ag-payment-letter", "label": "appears as letterhead and coordinating identity", "status": "documented", "type": "coordination" }, { "id": "w24-ag-e-clark-authorship", "source": "w24-ag-clark", "target": "w24-ag-payment-letter", "label": "official announcement attributes authorship to her", "status": "reported", "type": "personnel" }, { "id": "w24-ag-e-demand-visa", "source": "w24-ag-payment-letter", "target": "visa", "label": "requests proactive NCII seller payment-authorization removal", "status": "documented", "type": "policy" }, { "id": "w24-ag-e-demand-mastercard", "source": "w24-ag-payment-letter", "target": "mastercard", "label": "requests proactive NCII seller payment-authorization removal", "status": "documented", "type": "policy" }, { "id": "w24-ag-e-demand-amex", "source": "w24-ag-payment-letter", "target": "w24-ag-amex", "label": "requests proactive NCII seller payment-authorization removal", "status": "documented", "type": "policy" }, { "id": "w24-ag-e-demand-paypal", "source": "w24-ag-payment-letter", "target": "paypal", "label": "requests proactive NCII seller payment-authorization removal", "status": "documented", "type": "policy" }, { "id": "w24-ag-e-demand-google", "source": "w24-ag-payment-letter", "target": "google", "label": "requests proactive NCII seller payment-authorization removal", "status": "documented", "type": "policy" }, { "id": "w24-ag-e-demand-apple", "source": "w24-ag-payment-letter", "target": "w24-ag-apple", "label": "requests proactive NCII seller payment-authorization removal", "status": "documented", "type": "policy" }, { "id": "w24-ag-e-idaho-warning", "source": "w24-ag-labrador", "target": "w24-ag-payment-letter", "label": "reiterated demand and threatened legal consequences", "status": "documented", "type": "narrative" }, { "id": "w25-sf-e-chiu-office", "source": "w25-sf-chiu", "target": "w25-sf-city-attorney", "label": "acted as City Attorney in this proceeding", "status": "documented", "type": "governance" }, { "id": "w25-sf-e-sponsor", "source": "w25-sf-chiu", "target": "w25-sf-ab621", "label": "was identified as sponsor of", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-author", "source": "w25-sf-bauer-kahan", "target": "w25-sf-ab621", "label": "authored and offered primary-purpose clarification for", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-opposition", "source": "w25-sf-civil-liberties-advocacy", "target": "w25-sf-ab621", "label": "opposed unless amended", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-litigation-rationale", "source": "w25-sf-city-attorney", "target": "w25-sf-ab621", "label": "supplied litigation-based rationale for facilitator liability", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-facilitation-rule", "source": "w25-sf-ab621", "target": "w25-sf-facilitator-class", "label": "creates conditional civil exposure for specified facilitation", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-presumption-design", "source": "w25-sf-ab621", "target": "w25-sf-notice-presumption", "label": "establishes a receipt-based conditional presumption", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-public-remedies", "source": "w25-sf-ab621", "target": "w25-sf-prosecutor-role", "label": "authorizes civil enforcement and court-awarded remedies", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-notice-issuance", "source": "w25-sf-city-attorney", "target": "w25-sf-july-notices", "label": "reportedly sent notices to Apple and Google", "status": "reported", "type": "policy" }, { "id": "w25-sf-e-prior-letter-citation", "source": "w24-ag-payment-letter", "target": "w25-sf-july-notices", "label": "was reportedly cited as earlier notice", "status": "reported", "type": "narrative" }, { "id": "w25-sf-e-demand-apple", "source": "w25-sf-july-notices", "target": "w24-ag-apple", "label": "reportedly requested action against identified nudify apps", "status": "reported", "type": "policy" }, { "id": "w25-sf-e-demand-google", "source": "w25-sf-july-notices", "target": "google", "label": "reportedly requested action against identified nudify apps", "status": "reported", "type": "policy" }, { "id": "w25-sf-e-apple-response", "source": "w24-ag-apple", "target": "w25-sf-july-notices", "label": "reported responsive app enforcement through a spokesperson", "status": "reported", "type": "infrastructure" }, { "id": "w25-sf-e-google-response", "source": "google", "target": "w25-sf-july-notices", "label": "reported five identified app suspensions through a spokesperson", "status": "reported", "type": "infrastructure" }, { "id": "w25-sf-e-prosecution", "source": "w25-sf-city-attorney", "target": "w25-sf-case", "label": "prosecuted the action for the People", "status": "documented", "type": "governance" }, { "id": "w25-sf-e-sol-case", "source": "w25-sf-sol-ecom", "target": "w25-sf-case", "label": "was named in the alternative-service proceeding", "status": "documented", "type": "governance" }, { "id": "w25-sf-e-court-judgment", "source": "w25-sf-superior-court", "target": "w25-sf-briver-judgment", "label": "signed the Briver final judgment", "status": "documented", "type": "governance" }, { "id": "w25-sf-e-briver-settlement", "source": "w25-sf-briver", "target": "w25-sf-briver-judgment", "label": "stipulated to company-specific obligations without admissions", "status": "documented", "type": "governance" }, { "id": "w25-sf-e-operating-bar", "source": "w25-sf-briver-judgment", "target": "w25-sf-briver", "label": "permanently prohibits the specified image-transformation operation", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-domain-control", "source": "w25-sf-briver-judgment", "target": "w25-sf-briver", "label": "prohibits registration transfer and requires five years of domain inactivity", "status": "documented", "type": "infrastructure" }, { "id": "w25-sf-e-reporting-order", "source": "w25-sf-briver-judgment", "target": "w25-sf-briver-reporting", "label": "requires information and sworn documentary attestations", "status": "documented", "type": "governance" }, { "id": "w25-sf-e-reports-to-city", "source": "w25-sf-briver", "target": "w25-sf-city-attorney", "label": "must deliver network information and recurring compliance proof", "status": "documented", "type": "governance" }, { "id": "w25-sf-e-penalty", "source": "w25-sf-briver", "target": "w25-sf-city-attorney", "label": "owes scheduled civil penalties to the office", "status": "documented", "type": "money" }, { "id": "w25-sf-e-inspection", "source": "w25-sf-city-attorney", "target": "w25-sf-briver", "label": "may inspect domain-related records for judgment compliance", "status": "documented", "type": "governance" }, { "id": "w25-sf-e-further-enforcement", "source": "w25-sf-superior-court", "target": "w25-sf-briver-judgment", "label": "retains conditional enforcement and modification authority", "status": "documented", "type": "governance" }, { "id": "w25-sf-e-announced-shutdown", "source": "w25-sf-city-attorney", "target": "w25-sf-announced-outcome", "label": "reported the wider outcome of its investigation and lawsuit", "status": "reported", "type": "narrative" }, { "id": "w25-sf-e-production-cloudflare", "source": "w25-sf-cloudflare", "target": "w25-sf-city-attorney", "label": "reportedly produced responsive records to a subpoena", "status": "reported", "type": "governance" }, { "id": "w25-sf-e-production-namecheap", "source": "w25-sf-namecheap", "target": "w25-sf-city-attorney", "label": "reportedly produced responsive records to a subpoena", "status": "reported", "type": "governance" }, { "id": "w25-sf-e-production-namecom", "source": "w25-sf-namecom", "target": "w25-sf-city-attorney", "label": "reportedly produced responsive records to a subpoena", "status": "reported", "type": "governance" }, { "id": "w25-sf-e-production-google", "source": "google", "target": "w25-sf-city-attorney", "label": "reportedly produced responsive records to a subpoena", "status": "reported", "type": "governance" }, { "id": "w25-sf-e-correspondence-service", "source": "w25-sf-drawnudes", "target": "w25-sf-operator-claim", "label": "was the service associated with the correspondent by city counsel", "status": "reported", "type": "narrative" }, { "id": "w25-sf-e-operator-esafety", "source": "w25-sf-operator-claim", "target": "esafety", "label": "claims prior requirements and confirmation of compliance", "status": "reported", "type": "narrative" }, { "id": "w25-sf-e-operator-city", "source": "w25-sf-operator-claim", "target": "w25-sf-city-attorney", "label": "offered cooperation and requested written requirements", "status": "documented", "type": "narrative" }, { "id": "w25-sf-e-clothoff-outreach", "source": "esafety", "target": "w25-sf-clothoff", "label": "sent preparatory guidance and sought details of claimed changes", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-dis-made", "source": "inman-grant", "target": "w25-sf-dis-standard", "label": "made the industry standard under statutory authority", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-dis-clock", "source": "w25-sf-dis-standard", "target": "w25-sf-high-impact-genai", "label": "imposes category-specific duties from 22 December 2024", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-dis-contractual-rights", "source": "w25-sf-dis-standard", "target": "w25-sf-high-impact-genai", "label": "requires particular terms and appropriate enforcement", "status": "documented", "type": "policy" }, { "id": "w25-sf-e-dis-report-records", "source": "w25-sf-dis-standard", "target": "esafety", "label": "authorizes bounded reports and requires compliance records", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-ohf-implements", "source": "w26-ver-ohf", "target": "w26-ver-tat", "label": "implements the initiative", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-tat-operates", "source": "w26-ver-tat", "target": "w26-ver-tcap", "label": "operates the classification and alerting service", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-qs-controls-ohf", "source": "w26-ver-quantspark", "target": "w26-ver-ohf", "label": "holds registered voting and director-appointment control", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-hadley-controls", "source": "w26-ver-hadley", "target": "w26-ver-quantspark", "label": "holds registered share, voting and appointment control", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-hadley-leadership", "source": "w26-ver-hadley", "target": "w26-ver-tat", "label": "was described by QuantSpark as also running TAT", "status": "reported", "type": "personnel" }, { "id": "w26-ver-e-qs-invoicing", "source": "w26-ver-quantspark", "target": "w26-ver-ohf", "label": "reported consultancy and other services invoiced to OHF", "status": "reported", "type": "money" }, { "id": "w26-ver-e-ohf-invoicing-account", "source": "w26-ver-quantspark", "target": "w26-ver-ohf", "label": "was reported by OHF to have invoiced consultancy", "status": "reported", "type": "money" }, { "id": "w26-ver-e-ohf-balance", "source": "w26-ver-ohf", "target": "w26-ver-quantspark", "label": "reported an outstanding related-party balance", "status": "reported", "type": "money" }, { "id": "w26-ver-e-qs-loan-support", "source": "w26-ver-quantspark", "target": "w26-ver-ohf", "label": "provided loans on which OHF reports relying", "status": "reported", "type": "money" }, { "id": "w26-ver-e-ca21415-original-value", "source": "w26-ver-public-safety-canada", "target": "w26-ver-ca21415", "label": "disclosed the original contribution total", "status": "documented", "type": "money" }, { "id": "w26-ver-e-ca21415-amended-value", "source": "w26-ver-public-safety-canada", "target": "w26-ver-ca21415", "label": "disclosed an increased total for the same contribution", "status": "documented", "type": "money" }, { "id": "w26-ver-e-ca21415-recipient", "source": "w26-ver-ca21415", "target": "w26-ver-ohf", "label": "names the foundation under its former legal name", "status": "documented", "type": "money" }, { "id": "w26-ver-e-ca21415-execution-report", "source": "w26-ver-public-safety-canada", "target": "w26-ver-ca21415", "label": "reported signing an initial TCAP funding agreement in 2019–20", "status": "reported", "type": "money" }, { "id": "w26-ver-e-ca22841-value", "source": "w26-ver-public-safety-canada", "target": "w26-ver-ca22841", "label": "disclosed and announced the Phase 2 contribution", "status": "documented", "type": "money" }, { "id": "w26-ver-e-ca22841-recipient", "source": "w26-ver-ca22841", "target": "w26-ver-ohf", "label": "names OHF as the legal recipient", "status": "documented", "type": "money" }, { "id": "w26-ver-e-ca22841-scope", "source": "w26-ver-ca22841", "target": "w26-ver-tcap", "label": "specifies archive/classification and moderation-workflow development", "status": "documented", "type": "policy" }, { "id": "w26-ver-e-crf-approval-rules", "source": "w26-ver-public-safety-canada", "target": "w26-ver-crf-terms", "label": "publishes current budget, report and payment conditions", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-crf-ip-and-selection", "source": "w26-ver-crf-terms", "target": "w26-ver-public-safety-canada", "label": "limits the department’s stated IP and onward-funding rights", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-centre-nsc-coordination", "source": "w26-ver-canada-centre", "target": "w26-ver-us-nsc", "label": "coordinated engagement with TAT according to Canada’s status account", "status": "reported", "type": "governance" }, { "id": "w26-ver-e-centre-tat-engagement", "source": "w26-ver-canada-centre", "target": "w26-ver-tat", "label": "engaged TAT on support for smaller platforms", "status": "reported", "type": "policy" }, { "id": "w26-ver-e-canada-us-funding-report", "source": "w26-ver-public-safety-canada", "target": "w26-ver-us-funding-process", "label": "reported a US funding process linked to Christchurch commitments", "status": "reported", "type": "narrative" }, { "id": "w26-ver-e-us-process-beneficiary", "source": "w26-ver-us-funding-process", "target": "w26-ver-tat", "label": "identified TAT as intended funding recipient", "status": "reported", "type": "money" }, { "id": "w26-ver-e-centre-launch", "source": "w26-ver-canada-centre", "target": "w26-ver-tcap-launch2023", "label": "hosted the event with the joint TCAP 2.0 launch", "status": "reported", "type": "policy" }, { "id": "w26-ver-e-tat-launch", "source": "w26-ver-tat", "target": "w26-ver-tcap-launch2023", "label": "partnered in the Canadian launch", "status": "reported", "type": "campaign" }, { "id": "w26-ver-e-home-office-pilot", "source": "home-office", "target": "w26-ver-uk-pilot", "label": "published a pending direct-award pilot notice", "status": "documented", "type": "money" }, { "id": "w26-ver-e-pilot-tat-supplier", "source": "w26-ver-uk-pilot", "target": "w26-ver-tat", "label": "names TAT as proposed supplier", "status": "documented", "type": "money" }, { "id": "w26-ver-e-home-office-guide", "source": "home-office", "target": "w26-ver-uk-voluntary-guide", "label": "issued voluntary provider guidance", "status": "documented", "type": "policy" }, { "id": "w26-ver-e-guide-tcap", "source": "w26-ver-uk-voluntary-guide", "target": "w26-ver-tcap", "label": "recommends TCAP where applicable", "status": "documented", "type": "policy" }, { "id": "w26-ver-e-dis-example-tat", "source": "w25-sf-dis-standard", "target": "w26-ver-tat", "label": "names it as an example of independent expert verification tools", "status": "documented", "type": "policy" }, { "id": "w26-ver-e-dis-example-gifct", "source": "w25-sf-dis-standard", "target": "w26-ver-gifct", "label": "names it as an example of independent expert verification tools", "status": "documented", "type": "policy" }, { "id": "w26-ver-e-tat-verifies", "source": "w26-ver-tat", "target": "w26-ver-tcap-policy", "label": "applies source and content assessment through in-house experts", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-tcap-policy-scope", "source": "w26-ver-tcap-policy", "target": "w26-ver-tcap", "label": "sets inclusion, imitation and delisting parameters", "status": "documented", "type": "policy" }, { "id": "w26-ver-e-tcap-status-corrections", "source": "w26-ver-tat", "target": "w26-ver-tcap", "label": "reported correcting URL-availability statuses after platform disputes", "status": "reported", "type": "governance" }, { "id": "w26-ver-e-gifct-funds-tat", "source": "w26-ver-gifct", "target": "w26-ver-tat", "label": "contributes funding according to TAT’s account", "status": "reported", "type": "money" }, { "id": "w26-ver-e-board-governs", "source": "w26-ver-gifct-board", "target": "w26-ver-gifct", "label": "provides strategic oversight", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-meta-chairs", "source": "meta", "target": "w26-ver-gifct-board", "label": "chairs the Operating Board in 2026", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-gifct-dues", "source": "w26-ver-gifct-members", "target": "w26-ver-gifct", "label": "must pay tiered annual membership dues", "status": "documented", "type": "money" }, { "id": "w26-ver-e-gifct-access", "source": "w26-ver-gifct", "target": "w26-ver-gifct-members", "label": "conditions shared-tool access on membership and agreement", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-gifct-access-withdrawal", "source": "w26-ver-gifct", "target": "w26-ver-gifct-hsdb", "label": "reported taking management and withdrawing five legacy accesses", "status": "reported", "type": "governance" }, { "id": "w26-ver-e-members-contribute", "source": "w26-ver-gifct-members", "target": "w26-ver-gifct-hsdb", "label": "contribute qualified hashes and labels", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-gifct-contributes", "source": "w26-ver-gifct", "target": "w26-ver-gifct-hsdb", "label": "can create qualifying hashes itself", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-taxonomy-controls", "source": "w26-ver-gifct-taxonomy", "target": "w26-ver-gifct-hsdb", "label": "defines qualified inclusion routes", "status": "documented", "type": "policy" }, { "id": "w26-ver-e-tcap-gifct-filter", "source": "w26-ver-tcap", "target": "w26-ver-gifct-hsdb", "label": "supplies URL-hash inputs subject to GIFCT’s own parameters", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-hsdb-provider-discretion", "source": "w26-ver-gifct-hsdb", "target": "w26-ver-gifct-members", "label": "supplies matching information while platforms retain decisions", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-gifct-reported-corrections", "source": "w26-ver-gifct-members", "target": "w26-ver-gifct-hsdb", "label": "reported correcting labels and removing hashes", "status": "reported", "type": "governance" }, { "id": "w26-ver-e-iac-advice", "source": "w26-ver-gifct-iac", "target": "w26-ver-gifct-board", "label": "advises on priorities and performance", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-iac-draft-appointments", "source": "w26-ver-gifct-iac-draft", "target": "w26-ver-gifct-board", "label": "describes appointment endorsement and response duties in a draft", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-iac-draft-support", "source": "w26-ver-gifct-iac-draft", "target": "w26-ver-gifct-iac", "label": "describes funded participation and conflict safeguards in a draft", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-code-sanctions", "source": "w26-ver-gifct-code", "target": "w26-ver-gifct", "label": "sets leadership and Board consultation procedures for sanctions", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-member-appeal", "source": "w26-ver-gifct-members", "target": "w26-ver-gifct-board", "label": "may appeal organizational sanctions through the Code process", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-gifct-hria-commission", "source": "w26-ver-gifct", "target": "w26-ver-gifct-hria", "label": "commissioned and paid for the human-rights assessment", "status": "reported", "type": "money" }, { "id": "w26-ver-e-bsr-assessment", "source": "w22-sig-bsr", "target": "w26-ver-gifct-hria", "label": "conducted the assessment with editorial control", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-gifct-response", "source": "w26-ver-gifct", "target": "w26-ver-gifct-hria", "label": "responded with completed changes and further commitments", "status": "documented", "type": "policy" }, { "id": "w26-ver-e-gifct-policy-update", "source": "w26-ver-gifct", "target": "w26-ver-gifct-hr-policy", "label": "updated its public human-rights policy in May 2026", "status": "documented", "type": "policy" }, { "id": "w26-ver-e-map-transition", "source": "w26-ver-gifct", "target": "w26-ver-gifct-map", "label": "announced transfer of BSR-delivered rights services into MAP", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-tat-reported-archive", "source": "w26-ver-tat", "target": "w26-ver-tcap-archive", "label": "reported building an archive and reusable hash interfaces", "status": "reported", "type": "infrastructure" }, { "id": "w26-ver-e-tat-reported-relaunch", "source": "w26-ver-tat", "target": "w26-ver-tcap", "label": "reported a March 2025 interface relaunch and later sharing activity", "status": "reported", "type": "infrastructure" }, { "id": "w26-ver-e-microsoft-pilot", "source": "microsoft", "target": "w26-ver-tcap", "label": "was named in TAT’s account of an AI-detection pilot", "status": "reported", "type": "infrastructure" }, { "id": "w26-ver-e-archive-research-access", "source": "w26-ver-tcap-archive", "target": "w26-ver-archive-audience", "label": "advertises access for platforms and researchers", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-archive-hash-export", "source": "w26-ver-tcap-archive", "target": "w26-ver-tcap-hash-api", "label": "supplies the content fingerprints described by the API", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-api-onboarding", "source": "w26-ver-tcap", "target": "w26-ver-tcap-hash-api", "label": "requires prior onboarding for authenticated API use", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-api-feedback-route", "source": "w26-ver-tcap-hash-api", "target": "w26-ver-tcap-hash-feedback", "label": "provides authenticated per-hash error and inclusion feedback", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-tat-analyst-feedback", "source": "w26-ver-tat", "target": "w26-ver-tcap-hash-feedback", "label": "describes analyst review and an email-response aim", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-hash-api-deletions", "source": "w26-ver-tcap-hash-api", "target": "w26-ver-hash-recipient", "label": "documents withdrawal metadata and incremental update handling", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-tfp-referrals", "source": "w26-ver-trusted-flagger", "target": "w26-ver-tcap", "label": "advertises vetted partner referrals into the alerting workflow", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-tfp-archive", "source": "w26-ver-trusted-flagger", "target": "w26-ver-tcap-archive", "label": "advertises expansion of the archive and hash lists", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-google-altitude", "source": "google", "target": "w26-ver-altitude", "label": "was named as Jigsaw collaborator in Altitude development", "status": "reported", "type": "infrastructure" }, { "id": "w26-ver-e-tat-altitude-stewardship", "source": "w26-ver-tat", "target": "w26-ver-altitude", "label": "announced it would lead maintenance and development", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-altitude-software-rights", "source": "w26-ver-altitude", "target": "w26-ver-altitude-operator", "label": "permits software adaptation subject to Apache 2.0 conditions", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-altitude-source-input", "source": "w26-ver-tcap", "target": "w26-ver-altitude", "label": "supplies URL-alert metadata in the inspected adapter", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-altitude-gifct-input", "source": "w26-ver-gifct-hsdb", "target": "w26-ver-altitude", "label": "is a separately credentialed source described by Altitude", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-altitude-trust-priority", "source": "w26-ver-altitude", "target": "w26-ver-tcap", "label": "treats TCAP as trusted in its confidence heuristic", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-altitude-operator-choices", "source": "w26-ver-altitude-operator", "target": "w26-ver-altitude", "label": "configures access, source connections and moderation choices", "status": "documented", "type": "governance" }, { "id": "w26-ver-e-altitude-review-delivery", "source": "w26-ver-altitude", "target": "w26-ver-altitude-verdict-delivery", "label": "schedules reviewed verdict publication and conditional delivery", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-altitude-receiver-responsibility", "source": "w26-ver-altitude-verdict-delivery", "target": "w26-ver-altitude-operator", "label": "requires the platform’s implementation to give verdicts effect", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-altitude-optional-feedback", "source": "w26-ver-altitude", "target": "w26-ver-tcap", "label": "can return optional URL and decision diagnostics", "status": "documented", "type": "infrastructure" }, { "id": "w26-ver-e-altitude-redaction-boundary", "source": "w26-ver-altitude", "target": "w26-ver-altitude-verdict-delivery", "label": "contains a generic source-redaction hook but no demonstrated restoration chain", "status": "documented", "type": "infrastructure" }, { "id": "w27-roost-e-proposal-capacity", "source": "w27-roost-proposal2023", "target": "w27-roost-ccf", "label": "proposed a separate legal capacity and oversight split", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-leaders-direction", "source": "w27-roost-call-leaders", "target": "w27-roost-call", "label": "were described as retaining programme direction", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-call-entry", "source": "w27-roost-call-leaders", "target": "w27-roost-call", "label": "make the final entry decision in the published joining process", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-secretariat-handoff", "source": "w27-roost-support-secretariat", "target": "w27-roost-ccf", "label": "was announced and retrospectively reported as transitioned", "status": "reported", "type": "governance" }, { "id": "w27-roost-e-nz-support-end", "source": "w27-roost-mfat", "target": "w27-roost-support-secretariat", "label": "reported the support transition and government-funding endpoint", "status": "documented", "type": "funding" }, { "id": "w27-roost-e-mfat-costs", "source": "w27-roost-mfat", "target": "w27-roost-call", "label": "disclosed department-only historic expenditure", "status": "documented", "type": "funding" }, { "id": "w27-roost-e-ccan-core", "source": "w27-roost-ccan-core", "target": "w27-roost-ccan", "label": "governs its own admissions, statements and internal decisions", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-ccan-advice", "source": "w27-roost-ccan", "target": "w27-roost-ccf", "label": "requests consultation and reasons for rejected advice", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-ccan-funding", "source": "w27-roost-ccf", "target": "w27-roost-ccan", "label": "is described as providing advisory-network funding", "status": "reported", "type": "funding" }, { "id": "w27-roost-e-ccan-custody", "source": "w27-roost-ccan-fundholder", "target": "w27-roost-ccan", "label": "is assigned operational fund custody in published terms", "status": "documented", "type": "funding" }, { "id": "w27-roost-e-ccan-use", "source": "w27-roost-ccan-core", "target": "w27-roost-ccan-fundholder", "label": "approves funding applications and fund-use rules", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-ccan-review", "source": "w27-roost-ccf", "target": "w27-roost-ccan", "label": "reviews and gives feedback on amended terms", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-packard-support", "source": "w27-roost-packard", "target": "w27-roost-ccf", "label": "publishes a 2024 general-support grant entry", "status": "documented", "type": "funding" }, { "id": "w27-roost-e-carnegie2024", "source": "w27-roost-carnegie-ny", "target": "w27-roost-ccf", "label": "publishes a launch grant dated 12 December 2024", "status": "documented", "type": "funding" }, { "id": "w27-roost-e-carnegie2026", "source": "w27-roost-carnegie-ny", "target": "w27-roost-ccf", "label": "announced trustee approval of a further grant", "status": "documented", "type": "funding" }, { "id": "w27-roost-e-pivotal-support", "source": "w27-roost-pivotal-leaders-fund", "target": "w27-roost-ccf", "label": "is named by CCF as a support channel", "status": "reported", "type": "funding" }, { "id": "w27-roost-e-canada-award", "source": "w26-ver-public-safety-canada", "target": "w27-roost-canada26007", "label": "disclosed a contribution agreement value and expected results", "status": "documented", "type": "funding" }, { "id": "w27-roost-e-canada-recipient", "source": "w27-roost-canada26007", "target": "w27-roost-ccf", "label": "names CCF as recipient", "status": "documented", "type": "funding" }, { "id": "w27-roost-e-canada-workstreams", "source": "w27-roost-canada26007", "target": "w27-roost-ccf-elevate", "label": "specifies expected new-sector, ROOST and community work", "status": "documented", "type": "policy" }, { "id": "w27-roost-e-ccf-programme", "source": "w27-roost-ccf", "target": "w27-roost-ccf-elevate", "label": "describes its Elevate programme", "status": "reported", "type": "governance" }, { "id": "w27-roost-e-ccf-roost", "source": "w27-roost-ccf", "target": "w27-roost-roost", "label": "announced a technical and information collaboration", "status": "reported", "type": "campaign" }, { "id": "w27-roost-e-columbia-incubation", "source": "w27-roost-columbia-igp", "target": "w27-roost-roost", "label": "reported incubation and intended nonprofit independence", "status": "reported", "type": "governance" }, { "id": "w27-roost-e-employment", "source": "w27-roost-global-impact", "target": "w27-roost-roost", "label": "is the advertised US employer for a ROOST-linked vacancy", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-board-roost", "source": "w27-roost-roost-board", "target": "w27-roost-roost", "label": "is the published governing board", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-board-tdc", "source": "w27-roost-roost-board", "target": "w27-roost-tdc", "label": "appoints the Technical Design Committee", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-tdc-direction", "source": "w27-roost-tdc", "target": "w27-roost-maintainers", "label": "sets overarching technical direction and resolves technical ties", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-host-maintainers", "source": "w27-roost-roost", "target": "w27-roost-maintainers", "label": "initially appoints and may replace maintainers", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-maintainer-projects", "source": "w27-roost-maintainers", "target": "w27-roost-rmc", "label": "exercise project and merge authority under repository rules", "status": "inferred", "type": "governance" }, { "id": "w27-roost-e-community-proposals", "source": "w27-roost-community", "target": "w27-roost-maintainers", "label": "may propose changes without membership-based permissions", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-fork-check", "source": "w27-roost-roost", "target": "w27-roost-tool-operator", "label": "describes forking as a check on upstream project control", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-roadmap", "source": "w27-roost-tdc", "target": "w27-roost-roost", "label": "has a published roadmap-direction role with feedback routes", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-fundraising-account", "source": "w27-roost-roost", "target": "w27-roost-resource-account", "label": "published differing non-additive founding resource totals", "status": "reported", "type": "funding" }, { "id": "w27-roost-e-discord-founding", "source": "discord", "target": "w27-roost-roost", "label": "is named as a founding partner in the issuer release", "status": "reported", "type": "campaign" }, { "id": "w27-roost-e-google-founding", "source": "google", "target": "w27-roost-roost", "label": "is named as a founding partner in the issuer release", "status": "reported", "type": "campaign" }, { "id": "w27-roost-e-tc-expertise", "source": "w22-sig-tc", "target": "w27-roost-roost", "label": "was credited with child-safety expertise and capacity-building context", "status": "reported", "type": "campaign" }, { "id": "w27-roost-e-roost-rmc", "source": "w27-roost-roost", "target": "w27-roost-rmc", "label": "hosts an open model and policy integration project", "status": "documented", "type": "infrastructure" }, { "id": "w27-roost-e-ccf-template", "source": "w27-roost-ccf", "target": "w27-roost-ccf-tvec-pack", "label": "contributed a TVEC policy pack to RMC", "status": "documented", "type": "policy" }, { "id": "w27-roost-e-template-hosting", "source": "w27-roost-ccf-tvec-pack", "target": "w27-roost-rmc", "label": "is present in the inspected repository", "status": "documented", "type": "infrastructure" }, { "id": "w27-roost-e-licence-applies", "source": "w27-roost-rmc", "target": "w27-roost-rmc-licence", "label": "publishes CC BY 4.0 as its root content licence", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-template-reuse", "source": "w27-roost-rmc-licence", "target": "w27-roost-tool-operator", "label": "permits sharing and adaptation with attribution conditions", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-template-action", "source": "w27-roost-ccf-tvec-pack", "target": "w27-roost-tool-operator", "label": "offers removal and escalation language for voluntary adaptation", "status": "documented", "type": "policy" }, { "id": "w27-roost-e-template-context", "source": "w27-roost-ccf-tvec-pack", "target": "w27-roost-tool-operator", "label": "supplies contextual actor guidance with explicit limitations", "status": "documented", "type": "policy" }, { "id": "w27-roost-e-template-evaluation", "source": "w27-roost-ccf-tvec-pack", "target": "w27-roost-tool-operator", "label": "publishes a finite evaluation with mixed category results", "status": "reported", "type": "policy" }, { "id": "w27-roost-e-template-ambiguity", "source": "w27-roost-ccf-tvec-pack", "target": "w27-roost-tool-operator", "label": "retains an unresolved category reference and contextual tension", "status": "documented", "type": "policy" }, { "id": "w27-roost-e-guide-discretion", "source": "w27-roost-rmc", "target": "w27-roost-tool-operator", "label": "describes configurable model and human-review choices", "status": "documented", "type": "infrastructure" }, { "id": "w27-roost-e-discord-donation", "source": "discord", "target": "w27-roost-osprey", "label": "reported donating its production-origin rules engine", "status": "reported", "type": "infrastructure" }, { "id": "w27-roost-e-roost-osprey", "source": "w27-roost-roost", "target": "w27-roost-osprey", "label": "reported maintaining and developing the donated engine", "status": "reported", "type": "infrastructure" }, { "id": "w27-roost-e-discord-production", "source": "w27-roost-osprey", "target": "discord", "label": "was reported in production with local rules and controls", "status": "reported", "type": "infrastructure" }, { "id": "w27-roost-e-bluesky-planned", "source": "w27-roost-bluesky", "target": "w27-roost-osprey", "label": "was named as planning adoption in July 2025", "status": "reported", "type": "infrastructure" }, { "id": "w27-roost-e-bluesky-production", "source": "w27-roost-bluesky", "target": "w27-roost-osprey", "label": "was later reported using Osprey with locally authored rules", "status": "reported", "type": "infrastructure" }, { "id": "w27-roost-e-bluesky-output", "source": "w27-roost-osprey", "target": "w27-roost-bluesky-effects", "label": "was reported to send results through Kafka to a local receiver", "status": "reported", "type": "infrastructure" }, { "id": "w27-roost-e-bluesky-authority", "source": "w27-roost-bluesky", "target": "w27-roost-bluesky-effects", "label": "owns the reported effects receiver and local enforcement choices", "status": "reported", "type": "governance" }, { "id": "w27-roost-e-guide-phoebe", "source": "w27-roost-rmc", "target": "w27-roost-phoebe", "label": "links Phoebe as an investigation example", "status": "documented", "type": "infrastructure" }, { "id": "w27-roost-e-phoebe-config", "source": "w27-roost-tool-operator", "target": "w27-roost-phoebe", "label": "selects model, endpoint and repository configuration", "status": "documented", "type": "infrastructure" }, { "id": "w27-roost-e-phoebe-tools", "source": "w27-roost-phoebe", "target": "w27-roost-phoebe-rules", "label": "exposes local rule-save and separate dry-run validation tools", "status": "documented", "type": "infrastructure" }, { "id": "w27-roost-e-phoebe-validation", "source": "w27-roost-phoebe", "target": "w27-roost-phoebe-rules", "label": "instructs validation without enforcing it in the save function", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-phoebe-ozone-boundary", "source": "w27-roost-phoebe", "target": "w27-roost-tool-operator", "label": "does not implement Ozone label application or removal in the inspected adapter", "status": "documented", "type": "infrastructure" }, { "id": "w27-roost-e-phoebe-licence", "source": "w27-roost-phoebe", "target": "w27-roost-tool-operator", "label": "offers MIT software rights without service or data entitlement", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-tc-pathways", "source": "w22-sig-tc", "target": "w27-roost-pathways", "label": "launched and operates a capacity-building programme", "status": "reported", "type": "campaign" }, { "id": "w27-roost-e-pathways-gate", "source": "w27-roost-pathways", "target": "w27-roost-tool-operator", "label": "offers no-cost support after eligibility and registration review", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-pathways-conditions", "source": "w27-roost-pathways", "target": "w27-roost-pathways-terms", "label": "requires agreement to published non-member access conditions", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-pathways-information", "source": "w27-roost-pathways-terms", "target": "w27-roost-tool-operator", "label": "requires accurate registration and information cooperation", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-pathways-site-ip", "source": "w27-roost-pathways-terms", "target": "w27-roost-tool-operator", "label": "grants a limited revocable site licence and restricts copying", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-pathways-confidentiality", "source": "w27-roost-pathways-terms", "target": "w27-roost-tool-operator", "label": "restricts disclosure of mentoring information", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-pathways-discretion", "source": "w27-roost-pathways-terms", "target": "w27-roost-tool-operator", "label": "leaves implementation voluntary and in the user’s discretion", "status": "documented", "type": "governance" }, { "id": "w27-roost-e-tc-elevate", "source": "w22-sig-tc", "target": "w27-roost-tc-elevate", "label": "offers tailored paid advisory and implementation support", "status": "documented", "type": "funding" }, { "id": "w27-roost-e-elevate-complements", "source": "w27-roost-tc-elevate", "target": "w27-roost-pathways", "label": "is described as a deeper service that can complement free support", "status": "documented", "type": "campaign" }, { "id": "w27-roost-e-stability-pathways", "source": "w27-roost-stability-ai", "target": "w27-roost-pathways", "label": "reports participation with conflicting enrollment dates", "status": "reported", "type": "campaign" }, { "id": "w27-roost-e-stability-membership", "source": "w27-roost-stability-ai", "target": "w22-sig-tc", "label": "reported subsequent Coalition membership", "status": "reported", "type": "campaign" }, { "id": "w27-roost-e-clint-board", "source": "w27-roost-clint-smith", "target": "w27-roost-roost-board", "label": "is listed as a current co-chair", "status": "documented", "type": "personnel" }, { "id": "w27-roost-e-clint-discord", "source": "w27-roost-clint-smith", "target": "discord", "label": "is listed with a professional legal-leadership affiliation", "status": "documented", "type": "personnel" }, { "id": "w28-access-e-ms-software", "source": "microsoft", "target": "w28-access-photodna", "label": "describes qualified free software provision", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-ms-2012-license", "source": "microsoft", "target": "w28-access-photodna", "label": "reported named source-code licences in March 2012", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-ms-cloud-authority", "source": "microsoft", "target": "w28-access-photodna-cloud", "label": "retains eligibility, reverification and suspension powers", "status": "documented", "type": "governance" }, { "id": "w28-access-e-cloud-use-audit", "source": "w28-access-photodna-cloud", "target": "w28-access-cloud-customer", "label": "limits use and requires audit/reporting cooperation", "status": "documented", "type": "governance" }, { "id": "w28-access-e-cloud-report-ncmec", "source": "w28-access-photodna-cloud", "target": "w28-access-ncmec", "label": "authorizes customer-identified aggregate reporting", "status": "documented", "type": "infrastructure" }, { "id": "w28-access-e-ms-tc-delegation", "source": "microsoft", "target": "w28-access-tc-sublicensing", "label": "is reported to have delegated PhotoDNA sublicensing", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-tc-administers", "source": "w22-sig-tc", "target": "w28-access-tc-sublicensing", "label": "publishes an expanded member and non-member licensing route", "status": "reported", "type": "governance" }, { "id": "w28-access-e-tc-expanded-scope", "source": "w28-access-tc-sublicensing", "target": "w28-access-photodna", "label": "reports expanded permitted subject matter", "status": "reported", "type": "policy" }, { "id": "w28-access-e-tc-integration", "source": "w28-access-tc-sublicensing", "target": "w28-access-licensed-implementer", "label": "offers integration support to licensed companies", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-pathways-opportunity", "source": "w27-roost-pathways", "target": "w28-access-tc-sublicensing", "label": "offers an application opportunity separate from participation", "status": "documented", "type": "governance" }, { "id": "w28-access-e-ncmec-access", "source": "w28-access-ncmec", "target": "w28-access-ncmec-reference", "label": "issues environment-specific credentials and permissions", "status": "documented", "type": "governance" }, { "id": "w28-access-e-submitter-visibility", "source": "w28-access-ncmec-submitter", "target": "w28-access-ncmec-reference", "label": "can restrict visibility of its own entries", "status": "documented", "type": "governance" }, { "id": "w28-access-e-submitter-retraction", "source": "w28-access-ncmec-submitter", "target": "w28-access-ncmec-reference", "label": "alone can retract its entries through the described API", "status": "documented", "type": "infrastructure" }, { "id": "w28-access-e-client-polling", "source": "w28-access-ncmec-reference", "target": "w28-access-ncmec-client", "label": "supplies latest-state updates for locally maintained datasets", "status": "documented", "type": "infrastructure" }, { "id": "w28-access-e-provider-decisions", "source": "w28-access-ncmec-reference", "target": "w28-access-ncmec-client", "label": "is described as informing company-policy action", "status": "reported", "type": "policy" }, { "id": "w28-access-e-concentrix-audit", "source": "w28-access-concentrix", "target": "w28-access-ncmec-audit2023", "label": "attested a finite two-review audit", "status": "reported", "type": "governance" }, { "id": "w28-access-e-ncmec-correction", "source": "w28-access-ncmec", "target": "w28-access-ncmec-audit2023", "label": "was reported to have removed the 59 notified exceptions", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-audit-in-kind", "source": "w28-access-concentrix", "target": "w28-access-ncmec", "label": "was credited with a donated audit valued at $300000", "status": "reported", "type": "money" }, { "id": "w28-access-e-swgfl-stop", "source": "w28-access-swgfl", "target": "w28-access-stopncii", "label": "operates StopNCII", "status": "documented", "type": "governance" }, { "id": "w28-access-e-swgfl-rph", "source": "w28-access-swgfl", "target": "w28-access-rph", "label": "describes the Helpline as its support service", "status": "reported", "type": "governance" }, { "id": "w28-access-e-trustee-members", "source": "w28-access-swgfl-trustees", "target": "w28-access-swgfl", "label": "hold sole corporate membership and management rights", "status": "documented", "type": "governance" }, { "id": "w28-access-e-trustee-delegation", "source": "w28-access-swgfl-trustees", "target": "w28-access-swgfl", "label": "may impose budget limits and revoke recorded delegations", "status": "documented", "type": "governance" }, { "id": "w28-access-e-nain-trustee", "source": "w28-access-karuna-nain", "target": "w28-access-swgfl-trustees", "label": "has a filed director appointment and current trustee role", "status": "documented", "type": "personnel" }, { "id": "w28-access-e-nain-meta", "source": "w28-access-karuna-nain", "target": "meta", "label": "is described as a former safety-policy employee", "status": "reported", "type": "personnel" }, { "id": "w28-access-e-pershing-income", "source": "w28-access-pershing-fund", "target": "w28-access-swgfl", "label": "is recorded as restricted incoming and expended resources", "status": "documented", "type": "money" }, { "id": "w28-access-e-pershing-purpose", "source": "w28-access-pershing-fund", "target": "w28-access-stopncii", "label": "has an accounts purpose expressly supporting StopNCII", "status": "documented", "type": "money" }, { "id": "w28-access-e-meta-swgfl", "source": "meta", "target": "w28-access-swgfl", "label": "is labelled as comparative restricted support in the accounts", "status": "documented", "type": "money" }, { "id": "w28-access-e-homeoffice-swgfl", "source": "home-office", "target": "w28-access-swgfl", "label": "is recorded as restricted support for violence-against-women work", "status": "documented", "type": "money" }, { "id": "w28-access-e-user-entry", "source": "w28-access-stop-user", "target": "w28-access-stopncii", "label": "may create a case under the public-user eligibility rules", "status": "documented", "type": "governance" }, { "id": "w28-access-e-user-hash", "source": "w28-access-stop-user", "target": "w28-access-stop-reference", "label": "is described as submitting an on-device hash", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-user-recipient-scope", "source": "w28-access-stop-reference", "target": "w28-access-stop-partner", "label": "is shared with current and future participating platforms", "status": "documented", "type": "infrastructure" }, { "id": "w28-access-e-user-withdrawal", "source": "w28-access-stop-user", "target": "w28-access-stop-reference", "label": "can request whole-case withdrawal using credentials", "status": "documented", "type": "governance" }, { "id": "w28-access-e-partner-own-policy", "source": "w28-access-stop-partner", "target": "w28-access-stopncii", "label": "retains its own policy assessment and continuing-removal basis", "status": "documented", "type": "policy" }, { "id": "w28-access-e-swgfl-issued-api", "source": "w28-access-swgfl", "target": "w28-access-stop-reference", "label": "documents provider credentials and several input categories", "status": "documented", "type": "governance" }, { "id": "w28-access-e-partner-api-contribution", "source": "w28-access-stop-partner", "target": "w28-access-stop-reference", "label": "can contribute and withdraw its own submissions in the documented interface", "status": "documented", "type": "infrastructure" }, { "id": "w28-access-e-list-status-polling", "source": "w28-access-stop-reference", "target": "w28-access-stop-partner", "label": "documents status changes and limited deletion-marker availability", "status": "documented", "type": "infrastructure" }, { "id": "w28-access-e-stop-two-routes", "source": "w28-access-swgfl", "target": "w28-access-stop-approved", "label": "publishes an approved-participant route", "status": "documented", "type": "governance" }, { "id": "w28-access-e-stop-programme-route", "source": "w28-access-swgfl", "target": "w28-access-stop-programme", "label": "publishes a fuller programme-partner route", "status": "documented", "type": "governance" }, { "id": "w28-access-e-approved-fees", "source": "w28-access-stop-approved", "target": "w28-access-stop-partner", "label": "publishes revenue-tier annual GBP access fees", "status": "documented", "type": "money" }, { "id": "w28-access-e-programme-fees", "source": "w28-access-stop-programme", "target": "w28-access-stop-partner", "label": "publishes a different annual fee scale", "status": "documented", "type": "money" }, { "id": "w28-access-e-participation-conflict", "source": "w28-access-stop-approved", "target": "w28-access-stop-programme", "label": "has conflicting public benefit comparisons", "status": "documented", "type": "governance" }, { "id": "w28-access-e-terms-published", "source": "w28-access-swgfl", "target": "w28-access-stop-partner-terms", "label": "posts an access instrument with unresolved drafting and identity conflicts", "status": "documented", "type": "governance" }, { "id": "w28-access-e-terms-eligibility", "source": "w28-access-stop-partner-terms", "target": "w28-access-stop-partner", "label": "reserves vetting, reverification and access-suspension powers", "status": "documented", "type": "governance" }, { "id": "w28-access-e-terms-use", "source": "w28-access-stop-partner-terms", "target": "w28-access-stop-partner", "label": "restricts credentials, redistribution and model training", "status": "documented", "type": "governance" }, { "id": "w28-access-e-terms-security", "source": "w28-access-stop-partner-terms", "target": "w28-access-stop-partner", "label": "specifies scoped incident and implementation reporting", "status": "documented", "type": "governance" }, { "id": "w28-access-e-terms-termination", "source": "w28-access-stop-partner-terms", "target": "w28-access-stop-partner", "label": "requires cessation and deletion on partner termination", "status": "documented", "type": "governance" }, { "id": "w28-access-e-screen-admin", "source": "w28-access-swgfl", "target": "w28-access-stop-applicant-process", "label": "publishes human screening review and narrow challenge commitments", "status": "documented", "type": "governance" }, { "id": "w28-access-e-screen-deloitte", "source": "w28-access-stop-applicant-process", "target": "w28-access-deloitte", "label": "names a business-verification and process provider", "status": "documented", "type": "infrastructure" }, { "id": "w28-access-e-screen-neterium", "source": "w28-access-stop-applicant-process", "target": "w28-access-neterium", "label": "names sanctions and adverse-media screening", "status": "documented", "type": "infrastructure" }, { "id": "w28-access-e-stop-photodna2024", "source": "w28-access-stopncii", "target": "w28-access-photodna", "label": "announced on-device integration in March 2024", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-bing-donation", "source": "microsoft", "target": "w28-access-stopncii", "label": "reported donating the March2024 capability", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-bing-pilot", "source": "microsoft", "target": "w28-access-stop-reference", "label": "reported a Bing pilot and 268899 image-result actions", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-google-rollout", "source": "google", "target": "w28-access-stopncii", "label": "announced an intended following-months rollout", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-snap-filtering", "source": "snap", "target": "w28-access-stop-reference", "label": "reported daily updates and withdrawal-state exclusion on named surfaces", "status": "reported", "type": "infrastructure" }, { "id": "w28-access-e-fsc-promotes", "source": "w28-access-fsc", "target": "w28-access-stopncii", "label": "advocates adoption partly for payment-partner credibility", "status": "reported", "type": "narrative" }, { "id": "w28-access-e-nominet-member-governance", "source": "w28-access-nominet-members", "target": "w28-access-nominet-board", "label": "elect some non-executives and reappoint Board-appointed non-executives", "status": "documented", "type": "governance" }, { "id": "w28-access-e-nominet-board-management", "source": "w28-access-nominet-board", "target": "w28-access-nominet", "label": "controls affairs and property and may condition delegation", "status": "documented", "type": "governance" }, { "id": "w28-access-e-nominet-mixed-resources", "source": "w28-access-nominet", "target": "w28-access-nominet-public-benefit", "label": "describes commitments funded from reserves or annual surplus", "status": "documented", "type": "money" }, { "id": "w28-access-e-nominet-board-budget", "source": "w28-access-nominet-board", "target": "w28-access-nominet-public-benefit", "label": "oversees strategic alignment and agrees programme commitments", "status": "documented", "type": "governance" }, { "id": "w28-access-e-nominet-caf-transfer", "source": "w28-access-nominet", "target": "w28-access-caf-account", "label": "reported a transfer into its own account for later distribution", "status": "documented", "type": "money" }, { "id": "w28-access-e-nominet-uksic-annual", "source": "w28-access-nominet", "target": "w28-access-uksic", "label": "reported UKSIC amounts in annual grant tables", "status": "documented", "type": "money" }, { "id": "w28-access-e-nominet-uksic-pledge", "source": "w28-access-nominet", "target": "w28-access-uksic-pledge2021", "label": "announced a GBP 5.1m three-year pledge", "status": "reported", "type": "money" }, { "id": "w28-access-e-nominet-uksic-purpose", "source": "w28-access-uksic-pledge2021", "target": "w28-access-uksic", "label": "was described as funding the three-charity consortium", "status": "reported", "type": "money" }, { "id": "w28-access-e-nominet-programme-end", "source": "w28-access-nominet", "target": "w28-access-uksic-pledge2021", "label": "reported that the grant ended in March 2025 in its existing form", "status": "reported", "type": "money" }, { "id": "w28-access-e-nominet-continuing-support", "source": "w28-access-nominet", "target": "w28-access-uksic", "label": "stated that support and joint work would continue", "status": "reported", "type": "coordination" }, { "id": "w28-access-e-nominet-sid2022-commitment", "source": "w28-access-nominet", "target": "w28-access-uksic", "label": "reported a separate earlier commitment for Safer Internet Day 2022", "status": "reported", "type": "money" }, { "id": "w28-access-e-childnet-consortium-agreement", "source": "w28-access-childnet", "target": "w28-access-uksic-consortium-agreement", "label": "reported a formal partnership and a trustee pre-signature approval route", "status": "reported", "type": "governance" }, { "id": "w28-access-e-uksic-agreement-swgfl", "source": "w28-access-uksic-consortium-agreement", "target": "w28-access-swgfl", "label": "was disclosed as including SWGfL as a partner", "status": "reported", "type": "coordination" }, { "id": "w28-access-e-uksic-agreement-iwf", "source": "w28-access-uksic-consortium-agreement", "target": "iwf", "label": "was disclosed as including IWF as a partner", "status": "reported", "type": "coordination" }, { "id": "w28-access-e-uksic-agreement-purpose", "source": "w28-access-uksic-consortium-agreement", "target": "w28-access-uksic", "label": "was described as covering consortium projects", "status": "reported", "type": "coordination" }, { "id": "w28-access-e-iwf-recruitment-role", "source": "iwf", "target": "w28-access-iwf-project-officer", "label": "published a grant-management and reporting job specification", "status": "documented", "type": "governance" }, { "id": "w28-access-e-iwf-nominet-reporting", "source": "w28-access-iwf-project-officer", "target": "w28-access-nominet", "label": "was assigned financial-return liaison and funder-identified evaluator communication", "status": "documented", "type": "governance" }, { "id": "w28-access-e-nominet-swgfl-recognition", "source": "w28-access-nominet", "target": "w28-access-swgfl", "label": "was identified as the source of a recognized UKSIC grant", "status": "documented", "type": "money" }, { "id": "w28-access-e-nominet-iwf-recognition", "source": "w28-access-nominet", "target": "iwf", "label": "was explicitly named as funder of the UKSIC Hotline grant", "status": "documented", "type": "money" }, { "id": "w28-access-e-childnet-uksic-fund", "source": "w28-access-childnet", "target": "w28-access-uksic", "label": "reported income and expenditure in its Nominet UKSIC 9 restricted fund", "status": "documented", "type": "money" }, { "id": "w28-access-e-nominet-iwf-technical", "source": "w28-access-nominet", "target": "w28-access-iwf-nominet-technical", "label": "was named in separate technical-project grant accounting", "status": "documented", "type": "money" }, { "id": "w28-access-e-childnet-sid2025-delivery", "source": "w28-access-childnet", "target": "w28-access-sid2025", "label": "reported activity delivered with support from Nominet and other funders", "status": "reported", "type": "operations" }, { "id": "w28-access-e-nominet-research-commission", "source": "w28-access-nominet", "target": "w28-access-sid-research2026", "label": "commissioned and funded research with disclosed design input", "status": "documented", "type": "money" }, { "id": "w28-access-e-childnet-research-design", "source": "w28-access-childnet", "target": "w28-access-sid-research2026", "label": "was identified as the lead for study design", "status": "documented", "type": "operations" }, { "id": "w28-access-e-opinium-research-delivery", "source": "w28-access-opinium", "target": "w28-access-sid-research2026", "label": "was credited with conducting the November 2025 surveys", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-nominet-policy", "source": "w28-access-nominet", "target": "w29-enf-cpp2026", "label": "publishes and administers", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-agency-certification", "source": "w29-enf-listed-agencies", "target": "w28-access-nominet", "label": "may certify criminal use under listed-agency route", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-cpp-notice", "source": "w29-enf-cpp2026", "target": "w29-enf-registrant", "label": "specifies notice and corrective-action opportunity", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-cpp-measure", "source": "w28-access-nominet", "target": "w29-enf-registrant", "label": "may suspend or redirect under published CPP", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-cpp-merits-challenge", "source": "w29-enf-registrant", "target": "w29-enf-listed-agencies", "label": "is directed to agency merits-review routes", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-terms-discretion", "source": "w29-enf-registrant-terms", "target": "w28-access-nominet", "label": "reserves suspension and court-order grounds", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-iwf-historical-membership", "source": "w28-access-nominet", "target": "iwf", "label": "reported membership and prompt-removal cooperation in2013", "status": "reported", "type": "affiliation" }, { "id": "w29-enf-e-iwf-reported-requests", "source": "iwf", "target": "w28-access-nominet", "label": "was attributed one request in each selected later period", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-nominet-reported-actions", "source": "w28-access-nominet", "target": "w29-enf-reported-actions", "label": "reported suspensions, non-suspensions and reversals", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-nominet-reversal-basis", "source": "w29-enf-reported-actions", "target": "w29-enf-registrant", "label": "reports examples of restraint and lifting suspension", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-historical-review-promise", "source": "w28-access-nominet", "target": "w29-enf-registrant", "label": "announced separate review promises in2014", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-reported-future-statutory-route", "source": "w28-access-nominet", "target": "w29-enf-statutory-account", "label": "described future court-backed powers and voluntary-first practice", "status": "reported", "type": "policy" }, { "id": "w29-enf-e-police-unit", "source": "w29-enf-city-police", "target": "w29-enf-pipcu", "label": "hosts the specialist unit", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-police-ashiko", "source": "w29-enf-pipcu", "target": "w29-enf-ashiko", "label": "operates the published suspension process", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-private-evidence", "source": "w29-enf-referrers", "target": "w29-enf-ashiko", "label": "may provide intelligence and infringement evidence", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-dci-authorisation", "source": "w29-enf-pipcu-dci", "target": "w29-enf-ashiko", "label": "must authorise every police domain suspension", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-registrar-first", "source": "w29-enf-ashiko", "target": "w29-enf-registrar", "label": "requests registrar investigation and terms-based action first", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-police-registry-request", "source": "w29-enf-pipcu", "target": "w28-access-nominet", "label": "may escalate after registrar non-action under2020policy", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-ashiko-challenge", "source": "w29-enf-registrant", "target": "w29-enf-pipcu", "label": "may request reassessment within the initial notice window", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-ashiko-delivery2019", "source": "w29-enf-pipcu", "target": "w29-enf-ashiko", "label": "reported32995 website suspensions in2018/19", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-ashiko-resource-account2022", "source": "w29-enf-ipo", "target": "w29-enf-ashiko", "label": "reported one detective plus management and no referral charge", "status": "reported", "type": "resources" }, { "id": "w29-enf-e-gateway-operated", "source": "w29-enf-ipo", "target": "w29-enf-gateway", "label": "announced assessment and prioritisation of incoming intelligence", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-gateway-referrers", "source": "w29-enf-referrers", "target": "w29-enf-gateway", "label": "are directed to the replacement professional intake", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-gateway-referral", "source": "w29-enf-gateway", "target": "w29-enf-pipcu", "label": "may refer intelligence following assessment", "status": "documented", "type": "operations" }, { "id": "w29-enf-e-ipo-dsit-dividend", "source": "w29-enf-ipo", "target": "w15-impl-dsit", "label": "recorded an additional dividend paid to cover PIPCU costs", "status": "reported", "type": "money" }, { "id": "w29-enf-e-baseline-funding", "source": "w15-impl-dsit", "target": "w29-enf-pipcu", "label": "announced baseline support through IPO forApril2026–March2029", "status": "documented", "type": "money" }, { "id": "w29-enf-e-historic-funding2014", "source": "w29-enf-ipo", "target": "w29-enf-pipcu", "label": "announced further£3m funding to2017", "status": "documented", "type": "money" }, { "id": "w29-enf-e-bpi-secondment", "source": "w29-enf-bpi", "target": "w29-enf-pipcu", "label": "was reported to supply an internet-investigator secondee", "status": "reported", "type": "resources" }, { "id": "w29-enf-e-historical-unconfirmed-forecast", "source": "w29-enf-city-police", "target": "w29-enf-pipcu", "label": "published unconfirmed IPO income forecasts in2018", "status": "documented", "type": "money" }, { "id": "w29-enf-e-combined-budget-proposal", "source": "w29-enf-city-police", "target": "w29-enf-pipcu", "label": "included the unit in a combined2025 funding increase proposal", "status": "documented", "type": "money" }, { "id": "w29-enf-e-pilot-design-ipo", "source": "w29-enf-ipo", "target": "w29-enf-cofunding-pilot", "label": "is to develop priorities and contribution levels with PIPCU", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-pilot-design-police", "source": "w29-enf-pipcu", "target": "w29-enf-cofunding-pilot", "label": "is named joint developer of the future pilot", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-steering-ipo", "source": "w29-enf-ipo", "target": "w29-enf-steering-group", "label": "is named in the historical remit-setting group", "status": "reported", "type": "governance" }, { "id": "w29-enf-e-steering-police", "source": "w29-enf-city-police", "target": "w29-enf-steering-group", "label": "is named in the historical remit-setting group", "status": "reported", "type": "governance" }, { "id": "w29-enf-e-steering-industry", "source": "w29-enf-referrers", "target": "w29-enf-steering-group", "label": "are named among participants agreeing broad priorities", "status": "reported", "type": "governance" }, { "id": "w29-enf-e-steering-remit", "source": "w29-enf-steering-group", "target": "w29-enf-pipcu", "label": "is reported to have agreed broad unit priorities", "status": "reported", "type": "governance" }, { "id": "w29-enf-e-historical-case-acceptance", "source": "w29-enf-pipcu", "target": "w29-enf-referrers", "label": "published evidence requirements and police refusal review", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-ipo-board-delivery", "source": "w29-enf-ipo-board", "target": "w29-enf-ipo", "label": "was assigned setting strategy delivery milestones", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-ipcg-delivery", "source": "w29-enf-ipo", "target": "w29-enf-ipcg-delivery", "label": "reported establishing a mixed industry/enforcement delivery group", "status": "reported", "type": "governance" }, { "id": "w29-enf-e-ipcg-advisory", "source": "w29-enf-ipo", "target": "w29-enf-ipcg-industry-advisory", "label": "reported establishing an industry advisory committee", "status": "reported", "type": "governance" }, { "id": "w29-enf-e-ipcg-executive", "source": "w29-enf-ipo", "target": "w29-enf-ipcg-executive", "label": "reported an executive board for tactical/strategic coordination", "status": "reported", "type": "governance" }, { "id": "w29-enf-e-reported-agreement-signing", "source": "w29-enf-ipo", "target": "w29-enf-signing2026", "label": "was attributed participation in signing the2026/27Service Agreement", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-reported-police-signing", "source": "w29-enf-city-police", "target": "w29-enf-signing2026", "label": "was attributed participation in the annual agreement signing", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-alliance-attendance", "source": "w29-enf-alliance", "target": "w29-enf-signing2026", "label": "was attributed attendance through its Director General", "status": "reported", "type": "affiliation" }, { "id": "w29-enf-e-sky-attendance", "source": "w29-enf-sky-role", "target": "w29-enf-signing2026", "label": "was attributed attendance", "status": "reported", "type": "affiliation" }, { "id": "w29-enf-e-alliance-business-proposal", "source": "w29-enf-alliance", "target": "w29-enf-alliance-kybc2020", "label": "submitted a business-verification proposal in the CMA record", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-bpi-business-proposal", "source": "w29-enf-bpi", "target": "w29-enf-bpi-kybc2020", "label": "proposed stopping services for withheld or incorrect business identity", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-cma-published-alliance", "source": "w29-enf-cma-taskforce", "target": "w29-enf-alliance-kybc2020", "label": "published the response on8December2020", "status": "documented", "type": "publication" }, { "id": "w29-enf-e-cma-published-bpi", "source": "w29-enf-cma-taskforce", "target": "w29-enf-bpi-kybc2020", "label": "published the response on8December2020", "status": "documented", "type": "publication" }, { "id": "w29-enf-e-alliance-parliamentary-route", "source": "w29-enf-alliance", "target": "w29-enf-alliance-osb2021", "label": "advocated a legislative business-verification duty to parliament", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-alliance-current-position", "source": "w29-enf-alliance", "target": "w29-enf-alliance-kybc2020", "label": "retains a similar business-verification position in current-captured material", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-rightsholder-authorship", "source": "w29-enf-cic-rightsholders", "target": "w29-enf-cic-briefing2021", "label": "is attributed authorship of the annexed KYBC briefing", "status": "reported", "type": "policy" }, { "id": "w29-enf-e-mpa-parliamentary-evidence", "source": "w1-icmec-mpa", "target": "w29-enf-mpa-evidence2022", "label": "submitted evidence advocating KYBC and cross-department coordination", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-mpa-carried-briefing", "source": "w29-enf-mpa-evidence2022", "target": "w29-enf-cic-briefing2021", "label": "reproduced the rightsholder briefing as an annex", "status": "documented", "type": "publication" }, { "id": "w29-enf-e-rusi-publication", "source": "w29-enf-rusi", "target": "w29-enf-rusi-report2021", "label": "published the co-funded research report", "status": "documented", "type": "publication" }, { "id": "w29-enf-e-rusi-funder-ipo", "source": "w29-enf-ipo", "target": "w29-enf-rusi-report2021", "label": "is disclosed as a co-funder", "status": "reported", "type": "money" }, { "id": "w29-enf-e-rusi-funder-alliance", "source": "w29-enf-alliance", "target": "w29-enf-rusi-report2021", "label": "is disclosed as a co-funder", "status": "reported", "type": "money" }, { "id": "w29-enf-e-rusi-funder-mpa", "source": "w1-icmec-mpa", "target": "w29-enf-rusi-report2021", "label": "is disclosed as a co-funder", "status": "reported", "type": "money" }, { "id": "w29-enf-e-rusi-funder-industry-trust", "source": "w29-enf-industry-trust", "target": "w29-enf-rusi-report2021", "label": "is disclosed as a co-funder", "status": "reported", "type": "money" }, { "id": "w29-enf-e-rusi-funder-premier-league", "source": "w29-enf-premier-league", "target": "w29-enf-rusi-report2021", "label": "is disclosed as a co-funder", "status": "reported", "type": "money" }, { "id": "w29-enf-e-rusi-recommendations", "source": "w29-enf-rusi-report2021", "target": "w29-enf-rusi-recommendations", "label": "recommends verification and proportionate financial interventions", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-mpa-rusi-citation", "source": "w29-enf-mpa-evidence2022", "target": "w29-enf-rusi-report2021", "label": "quotes the2021 report in parliamentary advocacy", "status": "documented", "type": "publication" }, { "id": "w29-enf-e-music-kybc-consideration", "source": "w29-enf-music-response2021", "target": "w29-enf-kybc-forum", "label": "records government attention and IPO-led consideration", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-ipo-kybc-forums", "source": "w29-enf-ipo", "target": "w29-enf-kybc-forum", "label": "reported cross-departmental KYBC work", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-mpa-forum-request", "source": "w29-enf-mpa-evidence2022", "target": "w29-enf-kybc-forum", "label": "asks departments to join and advance KYBC work", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-sector-vision-exploration", "source": "w29-enf-dcms-current", "target": "w29-enf-sector-vision2023", "label": "published further exploration of potential KYBC solutions", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-advertising-proposal", "source": "w29-enf-dcms-current", "target": "w29-enf-oap2023", "label": "published proposed systems duties and regulatory design", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-advertising-check-design", "source": "w29-enf-oap2023", "target": "w29-enf-advertising-check-design", "label": "sets limits on proposed identity-check design", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-iab-working-group-proposal", "source": "w29-enf-iab-uk", "target": "w29-enf-ad-fraud-working-group", "label": "presented draft terms and year plan to the taskforce", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-taskforce-workplan-agreement", "source": "w29-enf-oat", "target": "w29-enf-ad-fraud-working-group", "label": "recorded members agreeing draft terms and year plan", "status": "documented", "type": "governance" }, { "id": "w29-enf-e-advertising-partnership-home-office", "source": "home-office", "target": "w29-enf-oat", "label": "was named a launcher of the anti-fraud advertising partnership", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-advertising-partnership-dcms", "source": "w29-enf-dcms-current", "target": "w29-enf-oat", "label": "was named a launcher alongside Home Office", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-advertising-partnership-iab", "source": "w29-enf-iab-uk", "target": "w29-enf-oat", "label": "was named industry partner for anti-fraud work", "status": "reported", "type": "operations" }, { "id": "w29-enf-e-conditional-advertising-legislation", "source": "w29-enf-fraud-strategy2026", "target": "w29-enf-oat", "label": "describes report-back and conditional future legislation", "status": "documented", "type": "policy" }, { "id": "w29-enf-e-fraud-strategy-osa-account", "source": "w29-enf-fraud-strategy2026", "target": "w7-policy-osa", "label": "separately describes enacted fraudulent-advertising duties", "status": "reported", "type": "policy" }, { "id": "w29-enf-e-march2026-ofcom-timetable", "source": "w29-enf-fraud-strategy2026", "target": "ofcom", "label": "reported then-planned consultation and expected2027commencement", "status": "reported", "type": "policy" }, { "id": "w30-class-e-creative-evidence", "source": "w30-class-creative-referrers", "target": "w29-enf-pipcu", "label": "submit evidence for police assessment", "status": "documented", "type": "policy" }, { "id": "w30-class-e-creative-screen", "source": "w29-enf-pipcu", "target": "w30-class-iwl", "label": "assesses qualifying referrals before distribution", "status": "documented", "type": "policy" }, { "id": "w30-class-e-creative-tactics", "source": "w30-class-creative-head", "target": "w30-class-creative", "label": "determines tactics case by case", "status": "documented", "type": "policy" }, { "id": "w30-class-e-creative-notice", "source": "w29-enf-pipcu", "target": "w30-class-site-operators", "label": "contacts operators before further disruption", "status": "documented", "type": "policy" }, { "id": "w30-class-e-creative-payment-request", "source": "w29-enf-pipcu", "target": "w30-class-payment-providers", "label": "requests withdrawal of payment services", "status": "documented", "type": "policy" }, { "id": "w30-class-e-creative-challenge", "source": "w30-class-site-operators", "target": "w29-enf-pipcu", "label": "may challenge police action", "status": "documented", "type": "policy" }, { "id": "w30-class-e-creative-removal", "source": "w29-enf-pipcu", "target": "w30-class-iwl", "label": "removes immediately upon evidence offences ceased", "status": "documented", "type": "policy" }, { "id": "w30-class-e-creative-portal-design", "source": "w30-class-portal-spec", "target": "w29-enf-pipcu", "label": "assigns submission acceptance to police administrators", "status": "documented", "type": "policy" }, { "id": "w30-class-e-prs-referrals", "source": "w30-class-prs", "target": "w29-enf-pipcu", "label": "reports138 website referrals", "status": "reported", "type": "coordination" }, { "id": "w30-class-e-iwl-distribution", "source": "w29-enf-pipcu", "target": "w30-class-ad-recipients", "label": "distributes IWL for advertising decisions", "status": "documented", "type": "coordination" }, { "id": "w30-class-e-iab-contracts", "source": "w29-enf-iab-uk", "target": "w30-class-ad-recipients", "label": "recommends IWL in contracts and processes", "status": "documented", "type": "policy" }, { "id": "w30-class-e-gambling-regulation", "source": "w30-class-gambling-commission", "target": "w30-class-gambling-condition", "label": "publishes all-licence placement condition", "status": "documented", "type": "policy" }, { "id": "w30-class-e-gambling-contract-right", "source": "w30-class-gambling-condition", "target": "w30-class-gambling-licensees", "label": "requires termination-capable third-party contracts", "status": "documented", "type": "policy" }, { "id": "w30-class-e-gambling-iwl", "source": "w30-class-gambling-commission", "target": "w30-class-iwl", "label": "encourages licensee access", "status": "documented", "type": "policy" }, { "id": "w30-class-e-mixed-ad-inputs", "source": "w30-class-thinkbox-evidence", "target": "w30-class-ad-recipients", "label": "reports mixed public and private list use", "status": "reported", "type": "coordination" }, { "id": "w30-class-e-wb-official-index", "source": "w30-class-whitebullet", "target": "w30-class-gpi", "label": "advertises combined official-list resource", "status": "reported", "type": "policy" }, { "id": "w30-class-e-wb-proprietary-index", "source": "w30-class-whitebullet", "target": "w30-class-ipi", "label": "advertises independent risk assessments", "status": "reported", "type": "policy" }, { "id": "w30-class-e-wb-peer-plan", "source": "w30-class-whitebullet", "target": "w30-class-peer39", "label": "announces prospective marketplace delivery", "status": "reported", "type": "coordination" }, { "id": "w30-class-e-peer-offering", "source": "w30-class-peer39", "target": "w30-class-whitebullet", "label": "currently advertises piracy exclusion offering", "status": "reported", "type": "coordination" }, { "id": "w30-class-e-creative-estimate", "source": "w30-class-whitebullet", "target": "w30-class-creative", "label": "supplies police-published advertising estimate", "status": "reported", "type": "narrative" }, { "id": "w30-class-e-wipo-precedent", "source": "w30-class-iwl", "target": "w30-class-wipo-alert", "label": "is cited as national precedent", "status": "documented", "type": "narrative" }, { "id": "w30-class-e-wipo-host", "source": "w30-class-wipo", "target": "w30-class-wipo-alert", "label": "provides distribution infrastructure", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-wipo-classification", "source": "w30-class-contributors", "target": "w30-class-wipo-alert", "label": "supply national assessments and maintain entries", "status": "documented", "type": "policy" }, { "id": "w30-class-e-wipo-user-approval", "source": "w30-class-wipo", "target": "w30-class-alert-users", "label": "screens and approves advertising access", "status": "documented", "type": "governance" }, { "id": "w30-class-e-wipo-contributor-permission", "source": "w30-class-contributors", "target": "w30-class-alert-users", "label": "may refuse particular users list access", "status": "documented", "type": "governance" }, { "id": "w30-class-e-wipo-user-selection", "source": "w30-class-alert-users", "target": "w30-class-wipo-alert", "label": "select which contributor lists to use", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-wipo-purpose", "source": "w30-class-wipo-alert", "target": "w30-class-alert-users", "label": "limits described use to preventing misplaced ads", "status": "documented", "type": "policy" }, { "id": "w30-class-e-wipo-national-review", "source": "w30-class-site-operators", "target": "w30-class-contributors", "label": "seek correction through applicable national process", "status": "documented", "type": "policy" }, { "id": "w30-class-e-wipo-cross-list-notification", "source": "w30-class-wipo-alert", "target": "w30-class-contributors", "label": "describes cross-list deletion notification", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-wb-wipo-use", "source": "w30-class-whitebullet", "target": "w30-class-wipo-alert", "label": "reports membership and use alongside own assessments", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-pay-hosting", "source": "w30-class-wipo", "target": "w30-class-pay", "label": "administers formalities, notices and review requests", "status": "documented", "type": "governance" }, { "id": "w30-class-e-pay-first-notice", "source": "w30-class-pay-rightsholders", "target": "w30-class-pay", "label": "submit claims after first notice interval", "status": "documented", "type": "policy" }, { "id": "w30-class-e-pay-provider-discretion", "source": "w30-class-pay", "target": "w30-class-payment-providers", "label": "transmits claims for provider-specific decisions", "status": "documented", "type": "policy" }, { "id": "w30-class-e-pay-named-mc", "source": "w30-class-wipo", "target": "mastercard", "label": "names provider in programme description", "status": "reported", "type": "coordination" }, { "id": "w30-class-e-pay-named-paypal", "source": "w30-class-wipo", "target": "paypal", "label": "names provider in programme description", "status": "reported", "type": "coordination" }, { "id": "w30-class-e-pay-pilot-outcome", "source": "w30-class-wipo", "target": "w30-class-pay-pilot", "label": "reports manual pilot and composite removals", "status": "reported", "type": "narrative" }, { "id": "w30-class-e-wipo-approval", "source": "w30-class-wipo-member-states", "target": "w30-class-wipo-programme", "label": "approved biennial programme", "status": "reported", "type": "governance" }, { "id": "w30-class-e-wipo-priority", "source": "w30-class-wipo-programme", "target": "w30-class-pay", "label": "includes planned deployment priority", "status": "documented", "type": "policy" }, { "id": "w30-class-e-wipo-er-budget", "source": "w30-class-wipo-programme", "target": "w30-class-wipo-er23", "label": "allocates broader result budget", "status": "documented", "type": "money" }, { "id": "w30-class-e-wipo-er-spend", "source": "w30-class-wipo", "target": "w30-class-wipo-er23", "label": "reports broader pre-IPSAS expenditure", "status": "reported", "type": "money" }, { "id": "w30-class-e-mc-referral", "source": "w30-class-ip-claimants", "target": "mastercard", "label": "may submit qualifying IP referrals under general policy", "status": "documented", "type": "policy" }, { "id": "w30-class-e-mc-investigation", "source": "mastercard", "target": "mc-acquirers", "label": "requires investigation and reporting", "status": "documented", "type": "governance" }, { "id": "w30-class-e-mc-product-cessation", "source": "mc-acquirers", "target": "w30-class-merchants", "label": "must end acceptance for affirmatively infringing products", "status": "documented", "type": "policy" }, { "id": "w30-class-e-mc-termination-match", "source": "mc-acquirers", "target": "w30-class-match", "label": "must report applicable merchant termination", "status": "documented", "type": "policy" }, { "id": "w30-class-e-mc-enforcement", "source": "mastercard", "target": "mc-acquirers", "label": "retains policy interpretation and membership sanctions", "status": "documented", "type": "governance" }, { "id": "w30-class-e-paypal-referral", "source": "paypal", "target": "w30-class-ip-claimants", "label": "invites infringement reports and may take appropriate action", "status": "documented", "type": "policy" }, { "id": "w30-class-e-match-access", "source": "mc-acquirers", "target": "w30-class-match-users", "label": "approve service-provider access within registration rules", "status": "documented", "type": "governance" }, { "id": "w30-class-e-match-entry", "source": "mc-acquirers", "target": "w30-class-match", "label": "must report under qualifying termination triggers", "status": "documented", "type": "policy" }, { "id": "w30-class-e-match-inquiry", "source": "w30-class-match-users", "target": "w30-class-match", "label": "must inquire before agreement or acceptance", "status": "documented", "type": "policy" }, { "id": "w30-class-e-match-admission", "source": "mc-acquirers", "target": "w30-class-merchants", "label": "may admit a listed merchant", "status": "documented", "type": "policy" }, { "id": "w30-class-e-match-accuracy", "source": "mc-acquirers", "target": "w30-class-match", "label": "retain accuracy and reasonable-correction responsibility", "status": "documented", "type": "governance" }, { "id": "w30-class-e-match-retroactive", "source": "w30-class-match", "target": "w30-class-match-users", "label": "provides retroactive alerts after inquiries", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-match-purge", "source": "mastercard", "target": "w30-class-match", "label": "provides five-year listing purge", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-match-correction", "source": "mastercard", "target": "w30-class-match", "label": "permits early removal on specified grounds", "status": "documented", "type": "policy" }, { "id": "w30-class-e-match-personal-rights", "source": "w30-class-merchants", "target": "mastercard", "label": "may seek applicable personal-information remedies", "status": "documented", "type": "policy" }, { "id": "w30-class-e-stripe-policy", "source": "stripe", "target": "w30-class-stripe-acceptance", "label": "publishes general disqualification with exceptions", "status": "documented", "type": "policy" }, { "id": "w30-class-e-stripe-match-effect", "source": "w30-class-match", "target": "w30-class-stripe-acceptance", "label": "is an input to restrictive underwriting", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-visa-investigation", "source": "visa", "target": "w30-class-merrick", "label": "initiated historical investigation route", "status": "documented", "type": "coordination" }, { "id": "w30-class-e-merrick-match", "source": "w30-class-merrick", "target": "w30-class-match", "label": "reported eGumball after termination", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-egumball-challenge", "source": "w30-class-egumball", "target": "w30-class-egumball-superior", "label": "sought provisional delisting relief", "status": "documented", "type": "policy" }, { "id": "w30-class-e-egumball-tro", "source": "w30-class-egumball-superior", "target": "w30-class-merrick", "label": "ordered removal requests in provisional TRO", "status": "documented", "type": "policy" }, { "id": "w30-class-e-egumball-tro-reversal", "source": "w30-class-appeal-fourth", "target": "w30-class-egumball2023", "label": "reversed emergency mandatory relief", "status": "documented", "type": "policy" }, { "id": "w30-class-e-merrick-slapp", "source": "w30-class-merrick", "target": "w30-class-egumball2025", "label": "sought anti-SLAPP protection for MATCH report", "status": "documented", "type": "policy" }, { "id": "w30-class-e-egumball-slapp-disposition", "source": "w30-class-appeal-fourth", "target": "w30-class-egumball2025", "label": "affirmed denial at protected-activity step", "status": "documented", "type": "policy" }, { "id": "w30-class-e-egumball-version", "source": "w30-class-appeal-fourth", "target": "w30-class-egumball2025", "label": "modified inverted sentence without changing judgment", "status": "documented", "type": "policy" }, { "id": "w30-class-e-egumball-public-brief", "source": "w30-class-appeal-fourth", "target": "w30-class-egumball", "label": "ordered revised public redacted brief", "status": "documented", "type": "policy" }, { "id": "w30-class-e-filmon-to-egumball", "source": "w30-class-filmon-opinion", "target": "w30-class-egumball2025", "label": "supplied expressly applied legal framework", "status": "documented", "type": "policy" }, { "id": "w30-class-e-historical-dv-authorship", "source": "w30-class-dv-inc", "target": "w30-class-historical-dv-reports", "label": "defined and sold confidential classifications", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-filmon-demand", "source": "w30-class-filmon", "target": "w30-class-dv-inc", "label": "asserted a correction and clarification demand", "status": "reported", "type": "policy" }, { "id": "w30-class-e-filmon-harm-allegation", "source": "w30-class-filmon", "target": "w30-class-historical-dv-reports", "label": "alleged false labels and advertising harm", "status": "reported", "type": "narrative" }, { "id": "w30-class-e-mpaa-amicus", "source": "w1-icmec-mpa", "target": "w30-class-filmon-amicus", "label": "appeared under historical MPAA name as amicus", "status": "documented", "type": "coordination" }, { "id": "w30-class-e-amici-request", "source": "w30-class-filmon-amicus", "target": "w30-class-ca-supreme", "label": "requested affirmance of DV procedural protection", "status": "documented", "type": "policy" }, { "id": "w30-class-e-filmon-reversal", "source": "w30-class-ca-supreme", "target": "w30-class-filmon-opinion", "label": "reversed protection for these confidential reports", "status": "documented", "type": "policy" }, { "id": "w30-class-e-dv-classifier-offer", "source": "w30-class-dv", "target": "w30-class-dv-classifier", "label": "describes proprietary classification service", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-dv-customer-choice", "source": "w30-class-advertiser-customers", "target": "w30-class-dv-classifier", "label": "may choose categories to avoid", "status": "documented", "type": "policy" }, { "id": "w30-class-e-default-announced-hi-category", "source": "w30-class-dv", "target": "w30-class-hi-category", "label": "announced future default enablement", "status": "reported", "type": "policy" }, { "id": "w30-class-e-default-reported-hi-category", "source": "w30-class-dv", "target": "w30-class-hi-category", "label": "reported enabled default", "status": "reported", "type": "policy" }, { "id": "w30-class-e-default-announced-p2p-category", "source": "w30-class-dv", "target": "w30-class-p2p-category", "label": "announced future default enablement", "status": "reported", "type": "policy" }, { "id": "w30-class-e-default-reported-p2p-category", "source": "w30-class-dv", "target": "w30-class-p2p-category", "label": "reported enabled default", "status": "reported", "type": "policy" }, { "id": "w30-class-e-dv-category-distribution", "source": "w30-class-hi-category", "target": "w30-class-dv-platforms", "label": "was reported distributed to over100 partners", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-dv-adoption", "source": "w30-class-dv", "target": "w30-class-dv-platforms", "label": "reported nearly30 category implementations", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-dv-review-route", "source": "w30-class-dv", "target": "w30-class-dv-classifier", "label": "describes publisher review and sales separation", "status": "reported", "type": "governance" }, { "id": "w30-class-e-dv-url-access", "source": "w30-class-dv", "target": "w30-class-advertiser-customers", "label": "announced requestable URL-level reporting", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-dv-advertiser-revenue", "source": "w30-class-advertiser-customers", "target": "w30-class-dv-holdings", "label": "accounted for reported90percent of group revenue", "status": "reported", "type": "money" }, { "id": "w30-class-e-dv-supply-revenue", "source": "w30-class-supply-customers", "target": "w30-class-dv-holdings", "label": "accounted for reported10percent of group revenue", "status": "reported", "type": "money" }, { "id": "w30-class-e-dv-business-reporting", "source": "w30-class-dv-holdings", "target": "w30-class-dv-classifier", "label": "reports consolidated configuration and service model", "status": "reported", "type": "narrative" }, { "id": "w30-class-e-ncmec-publishes-outbound", "source": "w28-access-ncmec", "target": "w30-class-ncmec-outbound", "label": "publishes reviewed-notice and response aggregates", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-ncmec-publishes-incoming", "source": "w28-access-ncmec", "target": "w30-class-ncmec-incoming", "label": "publishes provider-submitted report counts", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-dv-ncmec-origin", "source": "w28-access-ncmec", "target": "w30-class-dv-ncmec-input", "label": "is named as public-data origin by DV", "status": "reported", "type": "narrative" }, { "id": "w30-class-e-dv-ncmec-transformation", "source": "w30-class-dv", "target": "w30-class-dv-ncmec-input", "label": "reports source selection and further assessment", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-dv-input-category", "source": "w30-class-dv-ncmec-input", "target": "w30-class-hi-category", "label": "is described as an input to vendor category", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-iwf-maintenance", "source": "iwf", "target": "w30-class-iwf-url", "label": "describes daily rechecks and same-day obsolete-entry removal", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-iwf-dv-planned", "source": "w30-class-iwf-url", "target": "w30-class-hi-category", "label": "was announced as planned integration", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-iwf-dv-testimonial", "source": "w30-class-iwf-url", "target": "w30-class-hi-category", "label": "is described as integrated in DV testimonial", "status": "reported", "type": "infrastructure" }, { "id": "w30-class-e-dv-iwf-membership", "source": "w30-class-dv", "target": "iwf", "label": "is listed as member since1April2025", "status": "reported", "type": "coordination" }, { "id": "w30-class-e-dv-iwf-band", "source": "w30-class-dv", "target": "iwf", "label": "is disclosed in contribution band overGBP25000", "status": "reported", "type": "money" }, { "id": "w30-class-e-iwf-fee-framework", "source": "w30-class-iwf-members", "target": "iwf", "label": "support services under stated annual fee framework", "status": "documented", "type": "money" }, { "id": "w30-class-e-iwf-member-resources", "source": "iwf", "target": "w30-class-iwf-members", "label": "describes licensed lists, expertise and participation benefits", "status": "documented", "type": "infrastructure" }, { "id": "w30-class-e-iwf-council-application", "source": "w30-class-iwf-members", "target": "w30-class-iwf-council", "label": "may apply or nominate representatives", "status": "documented", "type": "governance" }, { "id": "w30-class-e-iwf-council-elections", "source": "w30-class-iwf-council", "target": "w30-class-iwf-board", "label": "elects three industry trustees in dated report", "status": "reported", "type": "governance" }, { "id": "w30-class-e-iwf-board-appointments", "source": "w30-class-iwf-board", "target": "iwf", "label": "approves appointments under described processes", "status": "reported", "type": "governance" }, { "id": "w30-class-e-senator-blackburn", "source": "w30-class-blackburn", "target": "w30-class-senate-inquiry", "label": "signed February7 requests", "status": "documented", "type": "policy" }, { "id": "w30-class-e-senator-blumenthal", "source": "w30-class-blumenthal", "target": "w30-class-senate-inquiry", "label": "signed February7 requests", "status": "documented", "type": "policy" }, { "id": "w30-class-e-adalytics-early-access", "source": "w30-class-adalytics", "target": "w30-class-senate-inquiry", "label": "provided research described as available before release", "status": "reported", "type": "narrative" }, { "id": "w30-class-e-cma-contribution", "source": "w30-class-cma", "target": "w30-class-senate-inquiry", "label": "claims advocacy contribution", "status": "reported", "type": "coordination" }, { "id": "w30-class-e-garcia-citation", "source": "w30-class-garcia-citation", "target": "w30-class-senate-inquiry", "label": "is expressly cited in TAG letter", "status": "documented", "type": "narrative" }, { "id": "w30-class-e-senate-dv-requests", "source": "w30-class-senate-inquiry", "target": "w30-class-dv", "label": "requests remediation, data-use and revenue answers", "status": "documented", "type": "policy" }, { "id": "w30-class-e-senate-ncmec-citation", "source": "w30-class-ncmec-outbound", "target": "w30-class-senate-inquiry", "label": "supplies specifically cited2021 notice-table reference", "status": "documented", "type": "narrative" }, { "id": "w30-class-e-senate-mrc-pressure", "source": "w30-class-senate-inquiry", "target": "w30-class-mrc", "label": "requests review, audit and possible revocation explanations", "status": "documented", "type": "policy" }, { "id": "w30-class-e-senate-tag-pressure", "source": "w30-class-senate-inquiry", "target": "w30-class-tag", "label": "questions certification and whole-website monetization", "status": "documented", "type": "policy" }, { "id": "w30-class-e-dv-critical-response", "source": "w30-class-dv", "target": "w30-class-adalytics", "label": "disputes interpretation and reports added safeguards", "status": "reported", "type": "narrative" }, { "id": "w30-class-e-tag-determination", "source": "w30-class-tag", "target": "w30-class-tag-response", "label": "reports no investigated BSC guideline violations", "status": "reported", "type": "policy" }, { "id": "w30-class-e-tag-ncmec-training", "source": "w28-access-ncmec", "target": "w30-class-tag-threat-members", "label": "was reported to train threat-exchange members", "status": "reported", "type": "coordination" }, { "id": "w30-class-e-dv-rebuttal-april", "source": "w30-class-dv", "target": "w30-class-adalytics", "label": "publishes technical and commercial rebuttal", "status": "reported", "type": "narrative" }, { "id": "w30-class-e-bondi-representation", "source": "w30-class-bondi", "target": "w30-class-dv-holdings", "label": "signed preservation letter as counsel", "status": "documented", "type": "personnel" }, { "id": "w30-class-e-firm-representation", "source": "w30-class-paul-hastings", "target": "w30-class-dv-holdings", "label": "is named counsel in demand and pleading", "status": "documented", "type": "personnel" }, { "id": "w30-class-e-dv-preservation", "source": "w30-class-dv-holdings", "target": "w30-class-preservation-letter", "label": "issued counsel demand under anticipated litigation", "status": "documented", "type": "policy" }, { "id": "w30-class-e-preservation-recipient", "source": "w30-class-preservation-letter", "target": "w30-class-cma", "label": "requests preservation and confirmation", "status": "documented", "type": "policy" }, { "id": "w30-class-e-cma-objection", "source": "w30-class-cma", "target": "w30-class-preservation-letter", "label": "publicly objects while asserting legal compliance", "status": "reported", "type": "narrative" }, { "id": "w30-class-e-complaint-plaintiff-dv-holdings", "source": "w30-class-dv-holdings", "target": "w30-class-dv-complaint", "label": "is named plaintiff in filed complaint", "status": "documented", "type": "policy" }, { "id": "w30-class-e-complaint-plaintiff-dv-inc", "source": "w30-class-dv-inc", "target": "w30-class-dv-complaint", "label": "is named plaintiff in filed complaint", "status": "documented", "type": "policy" }, { "id": "w30-class-e-complaint-defendant", "source": "w30-class-dv-complaint", "target": "w30-class-adalytics", "label": "seeks damages, retraction and corrective advertising", "status": "documented", "type": "policy" }, { "id": "w30-class-e-complaint-cma-allegation", "source": "w30-class-dv-complaint", "target": "w30-class-cma", "label": "alleges coordinated amplification", "status": "reported", "type": "narrative" }, { "id": "w31-fa-e-ford2022", "source": "w31-fa-ford", "target": "w30-class-cma", "label": "approved general support", "status": "documented", "type": "money" }, { "id": "w31-fa-e-democracy2023", "source": "w31-fa-democracy", "target": "w30-class-cma", "label": "reported paid general support", "status": "reported", "type": "money" }, { "id": "w31-fa-e-democracy2024", "source": "w31-fa-democracy", "target": "w30-class-cma", "label": "reported paid general support", "status": "reported", "type": "money" }, { "id": "w31-fa-e-kr2023", "source": "w31-fa-kr", "target": "w30-class-cma", "label": "reported a climate/election disinformation grant", "status": "reported", "type": "money" }, { "id": "w31-fa-e-kr2025", "source": "w31-fa-kr", "target": "w30-class-cma", "label": "listed a later grant term", "status": "reported", "type": "money" }, { "id": "w31-fa-e-krdistribution", "source": "w31-fa-krboard", "target": "w31-fa-kr", "label": "decides distribution, eligibility, scale and timing", "status": "documented", "type": "governance" }, { "id": "w31-fa-e-krappointments", "source": "w31-fa-villum", "target": "w31-fa-krboard", "label": "may appoint up to two directors", "status": "documented", "type": "governance" }, { "id": "w31-fa-e-krelections", "source": "w31-fa-krboard", "target": "w31-fa-krboard", "label": "elects remaining directors under a dissent limit", "status": "documented", "type": "governance" }, { "id": "w31-fa-e-krsigning", "source": "w31-fa-krsign", "target": "w31-fa-kr", "label": "can bind the foundation under joint-signature rules", "status": "documented", "type": "governance" }, { "id": "w31-fa-e-mix2024", "source": "w30-class-cma", "target": "w31-fa-mix2024", "label": "reported a funding-source chart", "status": "reported", "type": "money" }, { "id": "w31-fa-e-mix2025", "source": "w30-class-cma", "target": "w31-fa-mix2025", "label": "reported foundation grants as 82.8 percent", "status": "reported", "type": "money" }, { "id": "w31-fa-e-boardroster", "source": "w30-class-cma", "target": "w31-fa-cmaboard2024", "label": "published a historical board roster", "status": "reported", "type": "governance" }, { "id": "w31-fa-e-atkincochair", "source": "w31-fa-atkin", "target": "w31-fa-cmaboard2024", "label": "was listed as co-chair", "status": "reported", "type": "personnel" }, { "id": "w31-fa-e-jammicochair", "source": "w31-fa-jammi", "target": "w31-fa-cmaboard2024", "label": "was listed as co-chair", "status": "reported", "type": "personnel" }, { "id": "w31-fa-e-oakcma", "source": "oak", "target": "w30-class-cma", "label": "reported a three-year core grant", "status": "reported", "type": "money" }, { "id": "w31-fa-e-cmapublish", "source": "w30-class-cma", "target": "w31-fa-cmasub", "label": "provided a consultation response", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-cmasubmission", "source": "w31-fa-cmasub", "target": "ofcom", "label": "submitted proposals for additional safety measures", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-checks", "source": "w31-fa-cmasub", "target": "w31-fa-creatorchecks", "label": "requested verification of creators and advertisers", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-clearance", "source": "w31-fa-cmasub", "target": "w31-fa-premonetization", "label": "requested withholding monetization pending platform assessment", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-sanctionrequest", "source": "w31-fa-cmasub", "target": "w31-fa-sanctions", "label": "requested return bans, repeat-violator sanctions and clawbacks", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-crisismoney", "source": "w31-fa-cmasub", "target": "w31-fa-crisis-money", "label": "requested crisis monetization controls and conditional refunds", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-adincident", "source": "w31-fa-cmasub", "target": "w31-fa-adincident", "label": "requested advertising incident reporting without a minimum threshold", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-declinemonetization", "source": "ofcom", "target": "w31-fa-crisis-money", "label": "declined to recommend specific monetization practices", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-ffpublish", "source": "w31-fa-fullfact", "target": "w31-fa-ffsub", "label": "published a dated consultation response", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-ffsubmission", "source": "w31-fa-ffsub", "target": "ofcom", "label": "made consultation requests", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-ffscope", "source": "w31-fa-ffsub", "target": "w31-fa-ffscope", "label": "requested broader crisis and legislative scope", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-ffaccess", "source": "w31-fa-ffsub", "target": "w31-fa-ffaccess", "label": "requested meaningful access for accredited fact-checkers", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-ffoversight", "source": "w31-fa-ffsub", "target": "w31-fa-ffoversight", "label": "requested coordination and stronger accountability", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-ffinput", "source": "w31-fa-ffsub", "target": "w31-fa-noticechange", "label": "was cited as an input to a recommendation amendment", "status": "reported", "type": "policy" }, { "id": "w31-fa-e-meetinginput", "source": "w31-fa-dsitmeeting", "target": "w31-fa-noticechange", "label": "was cited as another input to the amendment", "status": "reported", "type": "policy" }, { "id": "w31-fa-e-ofcommeeting", "source": "ofcom", "target": "w31-fa-dsitmeeting", "label": "reported a meeting with DSIT", "status": "reported", "type": "coordination" }, { "id": "w31-fa-e-protocoladoption", "source": "ofcom", "target": "w31-fa-protocol", "label": "published crisis protocol recommendations", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-recommendprovider", "source": "w31-fa-protocol", "target": "w31-fa-providers", "label": "recommends criteria-based crisis processes", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-inputs", "source": "w31-fa-indicators", "target": "w31-fa-providers", "label": "may inform a provider determination", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-providerdecision", "source": "w31-fa-providers", "target": "w31-fa-protocol", "label": "determines activation using relevant criteria", "status": "documented", "type": "governance" }, { "id": "w31-fa-e-teamselection", "source": "w31-fa-providers", "target": "w31-fa-team", "label": "assembles a sufficiently senior cross-functional team", "status": "documented", "type": "governance" }, { "id": "w31-fa-e-noticerecommendation", "source": "w31-fa-noticechange", "target": "w31-fa-protocol", "label": "added consideration of a ministerial-notice indicator", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-ministerliteracy", "source": "w31-fa-secretary", "target": "ofcom", "label": "may direct media-literacy priorities for a specified period", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-ministernotice", "source": "w31-fa-secretary", "target": "ofcom", "label": "may require public statement notices", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-ofcomnotice", "source": "ofcom", "target": "w31-fa-notice", "label": "must issue a notice under a qualifying direction", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-publicaccount", "source": "w31-fa-notice", "target": "w31-fa-providers", "label": "can require a public statement by a specified date", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-information", "source": "ofcom", "target": "w31-fa-providers", "label": "may require threat-response information", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-reasonspublish", "source": "w31-fa-secretary", "target": "w31-fa-notice", "label": "must publish reasons for health/safety-ground directions", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-directionrevoke", "source": "w31-fa-secretary", "target": "ofcom", "label": "may vary or revoke the direction", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-policechannel", "source": "w31-fa-protocol", "target": "w31-fa-lechannel", "label": "recommends a dedicated channel for large services", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-channeloperation", "source": "w31-fa-providers", "target": "w31-fa-lechannel", "label": "would operate a conditional incoming-information channel", "status": "documented", "type": "infrastructure" }, { "id": "w31-fa-e-declineformalngo", "source": "ofcom", "target": "w31-fa-protocol", "label": "declined extending formal crisis-contact obligations to NGOs/fact-checkers", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-declinecentral", "source": "ofcom", "target": "w31-fa-protocol", "label": "declined a central coordinating role and routine analysis submission", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-analysisrecords", "source": "w31-fa-providers", "target": "w31-fa-analysis", "label": "would retain post-crisis analysis and key records", "status": "documented", "type": "governance" }, { "id": "w31-fa-e-analysisrequest", "source": "ofcom", "target": "w31-fa-analysis", "label": "may seek records for supervision", "status": "documented", "type": "governance" }, { "id": "w31-fa-e-draftanalysistime", "source": "w31-fa-childdraft", "target": "w31-fa-analysis", "label": "sets analysis timing at crisis end or90days if earlier", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-retainappeals", "source": "w31-fa-protocol", "target": "w31-fa-safeguards", "label": "retains existing appeal, expression and privacy safeguards", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-laidchild", "source": "ofcom", "target": "w31-fa-childdraft", "label": "submitted crisis code amendments for parliamentary process", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-remainingforecast", "source": "ofcom", "target": "w31-fa-cmasub", "label": "forecast remaining additional-measures decisions for autumn2026", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-planningreported", "source": "ofcom", "target": "w31-fa-planning", "label": "confirmed an initial planning document and chronology", "status": "reported", "type": "governance" }, { "id": "w31-fa-e-approvalreported", "source": "ofcom", "target": "w31-fa-approval", "label": "reported a final approval window", "status": "reported", "type": "governance" }, { "id": "w31-fa-e-nciidate", "source": "ofcom", "target": "w31-fa-nciimeasure", "label": "announced a separate30September2026 NCII measure", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-nciirecommend", "source": "ofcom", "target": "w28-access-stopncii", "label": "recommended StopNCII as a suitable option", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-swgflexpertise", "source": "w28-access-swgfl", "target": "ofcom", "label": "was described as providing expertise through a partnership", "status": "reported", "type": "coordination" }, { "id": "w31-fa-e-oakstatewatch", "source": "oak", "target": "w31-fa-statewatch", "label": "reported a three-year information and analysis grant", "status": "reported", "type": "money" }, { "id": "w31-fa-e-oakdff", "source": "oak", "target": "w31-fa-dff", "label": "reported three-year core support", "status": "reported", "type": "money" }, { "id": "w31-fa-e-oakstrategy", "source": "oak", "target": "w31-fa-ihr", "label": "publishes a human-rights funding strategy", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-swcsaendorse", "source": "w31-fa-statewatch", "target": "w31-fa-swcsa", "label": "published and endorsed a proposal objection", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-swcsaposition", "source": "w31-fa-swcsa", "target": "eu-csa-proposal", "label": "opposed agreement without stronger rights/security safeguards", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-swlettersign", "source": "w31-fa-statewatch", "target": "w31-fa-swletter", "label": "was a signatory to a published safeguards letter", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-swrequest", "source": "w31-fa-swletter", "target": "ofcom", "label": "requested clarification, transparency and conditional remedies", "status": "documented", "type": "policy" }, { "id": "w31-fa-e-dffround", "source": "w31-fa-dff", "target": "w31-fa-dffround", "label": "reported16 grants approved in June2026", "status": "reported", "type": "money" }, { "id": "w31-fa-e-peerselection", "source": "w31-fa-peergroup", "target": "w31-fa-dffround", "label": "was credited with review and grant selection", "status": "reported", "type": "governance" }, { "id": "w31-fa-e-libertyaward", "source": "w31-fa-dff", "target": "w31-fa-liberty", "label": "reported EUR36000 litigation support", "status": "reported", "type": "money" }, { "id": "w31-fa-e-libertychallenge", "source": "w31-fa-liberty", "target": "w31-fa-tcn", "label": "was described as pursuing a TCN-regime challenge", "status": "reported", "type": "policy" }, { "id": "w32-gdi-e-pilot", "source": "w31-fa-dff", "target": "w32-gdi-peerprocess", "label": "announced a pre-litigation-only pilot", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-peerselection", "source": "w32-gdi-peer", "target": "w32-gdi-peerprocess", "label": "reportedly selected projects in two 2025 rounds", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-boardadoption", "source": "w32-gdi-dffboard", "target": "w32-gdi-peerprocess", "label": "reportedly adopted the process in March 2026", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-staffsupport", "source": "w32-gdi-dffstaff", "target": "w32-gdi-peer", "label": "provides review information and facilitation", "status": "reported", "type": "resources" }, { "id": "w32-gdi-e-terms", "source": "w32-gdi-peerterms", "target": "w32-gdi-peer", "label": "is described as requiring confidentiality and conflict commitments", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-juneselection", "source": "w32-gdi-peer", "target": "w31-fa-dffround", "label": "is credited with selecting the June 2026 grants", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-libertyfiled", "source": "w31-fa-liberty", "target": "w32-gdi-complaint", "label": "filed the separate complaint", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-pifiled", "source": "w32-gdi-pi", "target": "w32-gdi-complaint", "label": "co-filed the separate complaint", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-representation", "source": "w31-fa-liberty", "target": "w32-gdi-leighday", "label": "reported instructing counsel for the complaint", "status": "reported", "type": "resources" }, { "id": "w32-gdi-e-courtissued", "source": "w32-gdi-ipt", "target": "w32-gdi-judgment", "label": "issued the public procedural judgment", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-govapplication", "source": "w32-gdi-respondent", "target": "w32-gdi-secrecyapplication", "label": "asked to withhold basic case information", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-applicationdenied", "source": "w32-gdi-ipt", "target": "w32-gdi-secrecyapplication", "label": "dismissed the confidentiality application", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-writtenarguments", "source": "w31-fa-liberty", "target": "w32-gdi-judgment", "label": "had written arguments considered", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-piarguments", "source": "w32-gdi-pi", "target": "w32-gdi-judgment", "label": "had written arguments considered", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-privatehearing", "source": "w32-gdi-ipt", "target": "w32-gdi-hearing", "label": "kept the March hearing private", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-appleparty", "source": "w24-ag-apple", "target": "w32-gdi-judgment", "label": "is the applicant named in the public ruling", "status": "documented", "type": "affiliation" }, { "id": "w32-gdi-e-ukoperator", "source": "w32-gdi-gdiuk", "target": "w32-gdi-gdi", "label": "operates the public website and acts as controller", "status": "documented", "type": "governance" }, { "id": "w32-gdi-e-oakgdi", "source": "oak", "target": "w32-gdi-gdi", "label": "reported a three-year 2023 grant", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-oakliberty", "source": "oak", "target": "w31-fa-liberty", "label": "reported a three-year 2023 grant", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-oakorg", "source": "oak", "target": "w32-gdi-org", "label": "reported a three-year 2023 grant", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-oakpi", "source": "oak", "target": "w32-gdi-pi", "label": "reported a three-year 2023 grant", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-donorpurpose", "source": "oak", "target": "w32-gdi-del", "label": "described support for a commercial-incentive intervention", "status": "reported", "type": "resources" }, { "id": "w32-gdi-e-ratingsupply", "source": "w32-gdi-gdi", "target": "w32-gdi-segment", "label": "supplied ratings described in the Oracle product guide", "status": "reported", "type": "infrastructure" }, { "id": "w32-gdi-e-oraclepackages", "source": "w32-gdi-oracle", "target": "w32-gdi-segment", "label": "offered the GDI contextual segment", "status": "reported", "type": "infrastructure" }, { "id": "w32-gdi-e-buyerchoice", "source": "w32-gdi-segment", "target": "w32-gdi-buyers", "label": "offered a selectable contextual input", "status": "reported", "type": "infrastructure" }, { "id": "w32-gdi-e-oracleend", "source": "w32-gdi-oracle", "target": "w32-gdi-segment", "label": "announced an end to service and supporting access", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-priorrights", "source": "w32-gdi-oracle", "target": "w32-gdi-buyers", "label": "described conditional rights for previously supplied data", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-licensedata", "source": "w32-gdi-gdi", "target": "w32-gdi-buyers", "label": "reported charging for risk data licensed to buyers", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-appealstanding", "source": "w32-gdi-publishers", "target": "w32-gdi-ratingreview", "label": "can appeal ratings in the described process", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-ratingrevision", "source": "w32-gdi-gdi", "target": "w32-gdi-ratingreview", "label": "described review and changeable ratings", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-researchreview", "source": "w32-gdi-gdiuk", "target": "w32-gdi-researchcomplaints", "label": "offers qualified review of research-output representations", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-researchlimits", "source": "w32-gdi-researchcomplaints", "target": "w32-gdi-publishers", "label": "provides a representation route without automatic restoration", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-iashistorical", "source": "w32-gdi-gdi", "target": "w32-gdi-iasproduct", "label": "was identified as a limited historical input", "status": "reported", "type": "infrastructure" }, { "id": "w32-gdi-e-iasdiscontinues", "source": "w32-gdi-ias", "target": "w32-gdi-iasproduct", "label": "announced discontinuation of the limited GDI input", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-iaschoices", "source": "w32-gdi-iasproduct", "target": "w32-gdi-buyers", "label": "allows category and risk-level exclusions", "status": "reported", "type": "infrastructure" }, { "id": "w32-gdi-e-dashboardplan", "source": "w32-gdi-gdi", "target": "w32-gdi-ohmd", "label": "describes a forthcoming revenue-estimation dashboard", "status": "reported", "type": "infrastructure" }, { "id": "w32-gdi-e-fcdo2019", "source": "w32-gdi-fcdo", "target": "w32-gdi-gdi", "label": "reported providing funding between 2019 and 2022", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-fcdo2022", "source": "w32-gdi-fcdo", "target": "w32-gdi-gdi", "label": "reported further funding in the then-current financial year", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-ledger", "source": "w32-gdi-fcdo", "target": "w32-gdi-gdi", "label": "records a supplier invoice line with a 5 September 2022 GL date", "status": "documented", "type": "money" }, { "id": "w32-gdi-e-programmepartnership", "source": "w32-gdi-fcdo", "target": "w32-gdi-cdmd", "label": "is named as a programme partner", "status": "documented", "type": "governance" }, { "id": "w32-gdi-e-cabinetpartner", "source": "w32-gdi-cabinet", "target": "w32-gdi-cdmd", "label": "is named as a programme partner", "status": "documented", "type": "governance" }, { "id": "w32-gdi-e-programmebudget", "source": "w32-gdi-fcdo", "target": "w32-gdi-cdmd", "label": "published programme-level budget allocations", "status": "documented", "type": "money" }, { "id": "w32-gdi-e-fundedremit", "source": "w32-gdi-cdmd", "target": "w32-gdi-gdi", "label": "was identified as supporting outside-US activities", "status": "reported", "type": "policy" }, { "id": "w32-gdi-e-expectedreviews", "source": "w32-gdi-cdmd", "target": "w32-gdi-gdi", "label": "was described as having an expected reporting and review relationship", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-recipientperiod", "source": "w32-gdi-gdi", "target": "w32-gdi-cdmd", "label": "described historical funding and a later grant expiry", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-ecproject", "source": "european-commission", "target": "w32-gdi-ai4trust", "label": "published the whole-project EU contribution", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-ecgerman", "source": "european-commission", "target": "w32-gdi-gdide", "label": "published a participant contribution of EUR 385275", "status": "reported", "type": "money" }, { "id": "w32-gdi-e-participant", "source": "w32-gdi-gdide", "target": "w32-gdi-ai4trust", "label": "is listed as a funded participant", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-coordination", "source": "w32-gdi-fbk", "target": "w32-gdi-ai4trust", "label": "is named as project coordinator", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-reportedclosed", "source": "european-commission", "target": "w32-gdi-ai4trust", "label": "lists the research project as closed", "status": "reported", "type": "governance" }, { "id": "w32-gdi-e-d21editor", "source": "w32-gdi-fbk", "target": "w32-gdi-d21", "label": "edited and recorded coordinator review before submission", "status": "documented", "type": "governance" }, { "id": "w32-gdi-e-qualityreview", "source": "w32-gdi-ucamreview", "target": "w32-gdi-d21", "label": "is credited with the designated quality-manager review", "status": "documented", "type": "governance" }, { "id": "w32-gdi-e-gdicontribution", "source": "w32-gdi-gdi", "target": "w32-gdi-d21", "label": "contributed pre-existing keyword resources to the design", "status": "documented", "type": "resources" }, { "id": "w32-gdi-e-sampling", "source": "w32-gdi-d21", "target": "w32-gdi-seeds", "label": "defines a mixed-source collection perimeter", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-framing", "source": "w32-gdi-d21", "target": "w32-gdi-framing", "label": "expressly adopts an intersectional methodological frame", "status": "documented", "type": "policy" }, { "id": "w32-gdi-e-researchdesign", "source": "w32-gdi-d21", "target": "w32-gdi-ai4trust", "label": "describes purpose-limited collection and agreement-dependent access", "status": "documented", "type": "governance" }, { "id": "w32-gdi-e-dsscontribution", "source": "w32-gdi-gdi", "target": "w32-gdi-gdidss", "label": "reported developing the decision-support contribution", "status": "reported", "type": "resources" }, { "id": "w32-gdi-e-annotations", "source": "w32-gdi-factcheckers", "target": "w32-gdi-gdidss", "label": "are reported as supplying training annotations", "status": "reported", "type": "resources" }, { "id": "w32-gdi-e-requests", "source": "w32-gdi-gdi", "target": "w32-gdi-accessrequests", "label": "requests new vetting, compatibility and audit arrangements", "status": "reported", "type": "policy" }, { "id": "w32-gdi-e-requestedplatformrole", "source": "w32-gdi-accessrequests", "target": "w32-gdi-platforms", "label": "would require compatibility and audit access if adopted", "status": "reported", "type": "policy" }, { "id": "w32-gdi-e-refusalsreported", "source": "w32-gdi-gdi", "target": "w32-gdi-platforms", "label": "reported research-access refusals", "status": "reported", "type": "policy" }, { "id": "w32-gdi-e-warningmethod", "source": "w32-gdi-ai4trust", "target": "w32-gdi-warning", "label": "reports a six-tool calibrated ensemble for triage", "status": "reported", "type": "infrastructure" }, { "id": "w32-gdi-e-componentmetric", "source": "w32-gdi-ai4trust", "target": "w32-gdi-textsignals", "label": "reports a component-specific weighted F1 of about 56 percent", "status": "reported", "type": "infrastructure" }, { "id": "w32-gdi-e-humanvalidation", "source": "w32-gdi-ai4trust", "target": "w32-gdi-validation", "label": "reports an enhanced human-validation dashboard", "status": "reported", "type": "infrastructure" }, { "id": "w32-gdi-e-userevaluation", "source": "w32-gdi-ai4trust", "target": "w32-gdi-evaluation", "label": "reports three evaluation rounds involving 206 users", "status": "reported", "type": "resources" }, { "id": "w33-con-e-delegationbasis", "source": "w33-con-dsa", "target": "w33-con-delegated", "label": "provides the basis for technical data-sharing conditions", "status": "documented", "type": "policy" }, { "id": "w33-con-e-adoptprocedure", "source": "european-commission", "target": "w33-con-delegated", "label": "adopted the project-specific research-access procedures", "status": "documented", "type": "policy" }, { "id": "w33-con-e-vetting", "source": "w33-con-estdsc", "target": "w33-con-vetted", "label": "grants project-specific vetted status when conditions are met", "status": "documented", "type": "governance" }, { "id": "w33-con-e-homeassessment", "source": "w33-con-homedsc", "target": "w33-con-estdsc", "label": "forwards application and initial assessment", "status": "documented", "type": "governance" }, { "id": "w33-con-e-fundingdisclosure", "source": "w33-con-vetted", "target": "w33-con-request", "label": "must demonstrate eligibility and disclose project funding", "status": "documented", "type": "governance" }, { "id": "w33-con-e-modalities", "source": "w33-con-estdsc", "target": "w33-con-request", "label": "sets technical legal and organisational access conditions", "status": "documented", "type": "governance" }, { "id": "w33-con-e-clock", "source": "w33-con-estdsc", "target": "w33-con-request", "label": "must formulate the request or explain refusal within80working days", "status": "documented", "type": "policy" }, { "id": "w33-con-e-provide", "source": "w33-con-providers", "target": "w33-con-vetted", "label": "must provide access under a reasoned request", "status": "documented", "type": "infrastructure" }, { "id": "w33-con-e-amend", "source": "w33-con-providers", "target": "w33-con-estdsc", "label": "may request amendment on specified access or security grounds", "status": "documented", "type": "governance" }, { "id": "w33-con-e-mediation", "source": "w33-con-providers", "target": "w33-con-estdsc", "label": "may seek mediation after an amendment decision", "status": "documented", "type": "governance" }, { "id": "w33-con-e-experts", "source": "w33-con-estdsc", "target": "w33-con-experts", "label": "may consult independent impartial experts", "status": "documented", "type": "governance" }, { "id": "w33-con-e-publish", "source": "w33-con-estdsc", "target": "w33-con-portal", "label": "must publish an overview of an issued reasoned request", "status": "documented", "type": "policy" }, { "id": "w33-con-e-portalhost", "source": "european-commission", "target": "w33-con-portal", "label": "must establish and host the portal", "status": "documented", "type": "infrastructure" }, { "id": "w33-con-e-fulfilnotify", "source": "w33-con-providers", "target": "w33-con-estdsc", "label": "must notify actual access provision or termination within3working days", "status": "documented", "type": "policy" }, { "id": "w33-con-e-terminate", "source": "w33-con-estdsc", "target": "w33-con-vetted", "label": "terminates access if eligibility is lost after investigation", "status": "documented", "type": "governance" }, { "id": "w33-con-e-publicaccess", "source": "w33-con-providers", "target": "w33-con-publicresearch", "label": "must give qualifying researchers public-data access without undue delay", "status": "documented", "type": "policy" }, { "id": "w33-con-e-flagaward", "source": "w33-con-flaggerdsc", "target": "w33-con-flagger", "label": "awards status on demonstrated expertise independence and diligent work", "status": "documented", "type": "governance" }, { "id": "w33-con-e-priority", "source": "w33-con-flagger", "target": "w33-con-noticeproviders", "label": "receives priority processing for notices within designated expertise", "status": "documented", "type": "policy" }, { "id": "w33-con-e-flagreview", "source": "w33-con-flaggerdsc", "target": "w33-con-flagger", "label": "must suspend or revoke when specified evidentiary conditions are met", "status": "documented", "type": "governance" }, { "id": "w33-con-e-bindnl", "source": "european-commission", "target": "w33-con-alinl", "label": "made annexed commitments binding on the named addressee", "status": "documented", "type": "policy" }, { "id": "w33-con-e-bindsg", "source": "european-commission", "target": "w33-con-alisg", "label": "made annexed commitments binding on the named addressee", "status": "documented", "type": "policy" }, { "id": "w33-con-e-commitservice", "source": "w33-con-commitments", "target": "w33-con-ali", "label": "requires a qualifying-public-data access process", "status": "documented", "type": "policy" }, { "id": "w33-con-e-assessment", "source": "w33-con-ali", "target": "w33-con-publicresearch", "label": "must assess eligibility within15business days", "status": "documented", "type": "governance" }, { "id": "w33-con-e-qa", "source": "w33-con-ali", "target": "w33-con-qa", "label": "must monitor API accuracy stability and processing metrics", "status": "documented", "type": "governance" }, { "id": "w33-con-e-nominatepay", "source": "w33-con-ali", "target": "w33-con-ankura", "label": "must fund the approved trustee on noncontingent terms", "status": "documented", "type": "money" }, { "id": "w33-con-e-approveankura", "source": "european-commission", "target": "w33-con-ankura", "label": "approved the named monitoring trustee", "status": "documented", "type": "governance" }, { "id": "w33-con-e-mandate", "source": "european-commission", "target": "w33-con-ankura", "label": "approves mandate and may instruct or replace the trustee", "status": "documented", "type": "governance" }, { "id": "w33-con-e-monitor", "source": "w33-con-ankura", "target": "w33-con-ali", "label": "must monitor and report without business decision powers", "status": "documented", "type": "governance" }, { "id": "w33-con-e-qareport", "source": "w33-con-qa", "target": "w33-con-ankura", "label": "must send semiannual access counts tests and feedback", "status": "documented", "type": "policy" }, { "id": "w33-con-e-trustreport", "source": "w33-con-ankura", "target": "european-commission", "label": "must submit monitoring reports and promptly flag concluded failure", "status": "documented", "type": "policy" }, { "id": "w33-con-e-reportscope", "source": "w33-con-commitments", "target": "w33-con-trustreport", "label": "defines reporting cycles separate from five-year substantive term", "status": "documented", "type": "policy" }, { "id": "w33-con-e-toauditor", "source": "w33-con-ankura", "target": "w33-con-auditrole", "label": "must permit relevant trustee-document access on auditor request", "status": "documented", "type": "infrastructure" }, { "id": "w33-con-e-fromauditor", "source": "w33-con-auditrole", "target": "w33-con-ankura", "label": "may supply already collected evidence to reduce repeated requests", "status": "documented", "type": "infrastructure" }, { "id": "w33-con-e-fine", "source": "european-commission", "target": "w33-con-ali", "label": "announced a separate EUR550million illegal-product-risk fine", "status": "reported", "type": "money" }, { "id": "w33-con-e-engagefti", "source": "w33-con-alisg", "target": "w33-con-fti", "label": "reported engaging FTI for the2024/25 audit", "status": "reported", "type": "governance" }, { "id": "w33-con-e-ftiopinion", "source": "w33-con-fti", "target": "w33-con-ftiaudit", "label": "issued a positive40(12) readiness conclusion", "status": "reported", "type": "policy" }, { "id": "w33-con-e-noapprovals", "source": "w33-con-ftiaudit", "target": "w33-con-appcohort", "label": "reports zero approvals and all inspected applications incomplete or invalid", "status": "reported", "type": "policy" }, { "id": "w33-con-e-legalprocess", "source": "w33-con-ftiaudit", "target": "w33-con-legalteam", "label": "describes Legal-managed eligibility and clarification workflow", "status": "reported", "type": "governance" }, { "id": "w33-con-e-staffdecisions", "source": "w33-con-legalteam", "target": "w33-con-appcohort", "label": "is reported as managing the tracked review process", "status": "reported", "type": "governance" }, { "id": "w33-con-e-gdilink", "source": "w32-gdi-gdi", "target": "w33-con-collab", "label": "links readers to the DSA40 Collaboratory", "status": "documented", "type": "narrative" }, { "id": "w33-con-e-runs", "source": "w33-con-collab", "target": "w33-con-tracker", "label": "collects voluntary application reports and publishes summaries", "status": "reported", "type": "infrastructure" }, { "id": "w33-con-e-trackeroutcomes", "source": "w33-con-tracker", "target": "w33-con-applicants", "label": "reports application statuses with unreconciled denominators", "status": "reported", "type": "policy" }, { "id": "w33-con-e-alientry", "source": "w33-con-tracker", "target": "w33-con-aliopen", "label": "lists one AliExpress public-data entry as open", "status": "reported", "type": "policy" }, { "id": "w33-con-e-aliprovider", "source": "w33-con-aliopen", "target": "w33-con-ali", "label": "names the platform counterpart", "status": "reported", "type": "coordination" }, { "id": "w33-con-e-reasons", "source": "w33-con-tracker", "target": "w33-con-applicants", "label": "publishes partial reported refusal categories and organiser assessments", "status": "reported", "type": "policy" }, { "id": "w33-con-e-support", "source": "w33-con-collab", "target": "w33-con-applicants", "label": "offers application support and recruits submitted experiences", "status": "reported", "type": "coordination" }, { "id": "w33-con-e-paperbasis", "source": "w33-con-collab", "target": "w33-con-paper14", "label": "supplies network experience for an author-led study", "status": "documented", "type": "narrative" }, { "id": "w33-con-e-mercator", "source": "w33-con-mercator", "target": "w33-con-collab", "label": "is disclosed as funding the joint project since early2024", "status": "reported", "type": "money" }, { "id": "w33-con-e-bmftr", "source": "w33-con-bmftr", "target": "w33-con-weizenbaum", "label": "is credited with institutional funding", "status": "reported", "type": "money" }, { "id": "w33-con-e-berlin", "source": "w33-con-berlin", "target": "w33-con-weizenbaum", "label": "is credited with institutional funding", "status": "reported", "type": "money" }, { "id": "w33-con-e-paperpublication", "source": "w33-con-weizenbaum", "target": "w33-con-paper14", "label": "publishes the policy paper", "status": "documented", "type": "narrative" }, { "id": "w33-con-e-resourcesproposal", "source": "w33-con-paper14", "target": "european-commission", "label": "advocates guidance and support for researcher intermediaries", "status": "documented", "type": "policy" }, { "id": "w33-con-e-issues", "source": "w33-con-collab", "target": "w33-con-applicants", "label": "reports checked practical issues separate from application decisions", "status": "reported", "type": "narrative" }, { "id": "w33-con-e-owcommission", "source": "w16-supp-dcms", "target": "w33-con-oliver", "label": "is identified as commissioning evaluation through102384", "status": "documented", "type": "money" }, { "id": "w33-con-e-evalscope", "source": "w33-con-evalcontract", "target": "w33-con-evaloutput", "label": "commissions process impact correctness and value-for-money evaluation", "status": "documented", "type": "policy" }, { "id": "w33-con-e-satisfaction", "source": "w16-supp-dcms", "target": "w33-con-evalcontract", "label": "conditions payment in arrears on satisfactory deliverables", "status": "documented", "type": "governance" }, { "id": "w33-con-e-gatecondition", "source": "w33-con-evalcontract", "target": "w33-con-fundgate", "label": "makes later-phase funding conditional on feasibility assessment", "status": "documented", "type": "policy" }, { "id": "w33-con-e-treasurygate", "source": "w33-con-treasury", "target": "w33-con-fundgate", "label": "is assigned approval of subsequent-phase funding", "status": "documented", "type": "governance" }, { "id": "w33-con-e-pilotfunding", "source": "w33-con-treasury", "target": "w16-supp-cddp", "label": "announced GBP8.4million pilot funding", "status": "reported", "type": "money" }, { "id": "w33-con-e-publicidentity", "source": "w33-con-publiclegal", "target": "w33-con-public", "label": "is identified as the current website controller", "status": "documented", "type": "governance" }, { "id": "w33-con-e-publiccddp", "source": "w33-con-public", "target": "w16-supp-cddp", "label": "reports participation in delivery", "status": "reported", "type": "infrastructure" }, { "id": "w33-con-e-subapproval", "source": "w16-supp-dcms", "target": "w33-con-subapproval", "label": "provides written approval through the order-form mechanism", "status": "documented", "type": "governance" }, { "id": "w33-con-e-chairs", "source": "w33-con-cdei", "target": "w33-con-expertgroup", "label": "convened and chaired the expert group", "status": "documented", "type": "governance" }, { "id": "w33-con-e-woodgroup", "source": "w12-gov-wood", "target": "w33-con-expertgroup", "label": "was listed as Reset Tech Senior Advisor in the group", "status": "documented", "type": "personnel" }, { "id": "w33-con-e-advises", "source": "w33-con-expertgroup", "target": "w33-con-osdi", "label": "advises scope direction and legally robust data stewardship", "status": "documented", "type": "governance" }, { "id": "w33-con-e-discovery", "source": "w33-con-osdi", "target": "w33-con-longlist", "label": "published candidate solutions from discovery input", "status": "documented", "type": "policy" }, { "id": "w33-con-e-facultyosdi", "source": "w16-supp-faculty", "target": "w33-con-osdi", "label": "is named in the delivery consortium", "status": "documented", "type": "infrastructure" }, { "id": "w33-con-e-publicosdi", "source": "w33-con-public", "target": "w33-con-osdi", "label": "is named in the delivery consortium", "status": "documented", "type": "infrastructure" }, { "id": "w33-con-e-paevaluate", "source": "w16-supp-dcms", "target": "pa-consulting", "label": "appointed PA as external OSDI evaluator", "status": "documented", "type": "governance" }, { "id": "w33-con-e-jointscope", "source": "pa-consulting", "target": "w33-con-osdi", "label": "had responsibilities and priorities agreed with government and consortium", "status": "documented", "type": "coordination" }, { "id": "w33-con-e-sandboxdelivery", "source": "w33-con-public", "target": "w33-con-sandbox", "label": "reports delivering a TRL4 prototype for DCMS review and testing", "status": "reported", "type": "infrastructure" }, { "id": "w33-con-e-sandboxprogramme", "source": "w33-con-sandbox", "target": "w33-con-osdi", "label": "is identified as the programme case study", "status": "reported", "type": "coordination" }, { "id": "w33-con-e-stcfcommission", "source": "w33-con-public", "target": "w33-con-stcf", "label": "authored a separately DSIT-commissioned evaluation", "status": "reported", "type": "policy" }, { "id": "w33-con-e-methodreuse", "source": "w33-con-osdi", "target": "w33-con-stcf", "label": "is expressly credited as a source of evaluation practices", "status": "reported", "type": "narrative" }, { "id": "w33-con-e-landscapework", "source": "w33-con-public", "target": "w33-con-landscape", "label": "produced a DSIT-commissioned sector study", "status": "documented", "type": "policy" } ], "groups": { "risk": "AI foundations & creators", "control": "ControlAI & political access", "industry": "Industry & electoral advocacy", "rights": "Rights advocacy & funders", "safety": "Content & payment campaigns", "state": "Government & international policy", "access-assurance": "Research access & assurance", "funding-authority": "Funding & policy influence", "gdi-rights": "GDI & digital rights", "reference-access-resources": "Standards, access & resources", "credentials": "Age checks & identity", "custody": "Reports & institutional records", "private-verification": "Private safety tools", "christchurch-roost": "Christchurch Call & ROOST", "business-identity-policy": "Business identity proposals", "domain-enforcement": "Domain names & enforcement", "advertising-input-and-pressure": "Advertising pressure & inputs", "implementation": "Putting policies into practice", "cashapp-authority": "Cash App & government powers", "commercial-classification-remedy": "Commercial classifications & appeals", "signals": "Sharing safety information", "governance": "Boards & decision rights", "dutycare": "Australian duty of care", "advertising-list-authority": "Advertising lists & discretion", "funding": "Who funds the work", "disposition": "Government responses", "taxonomy": "Who defines the categories", "platform": "Platforms & intermediaries", "payment-notice-authority": "Notices to payment providers", "payment-record-authority": "Merchant records & access", "redlines": "AI Red Lines campaign" } } }