Part of the research through 9 September 2026. This dated note preserves its original findings; later developments are discussed in the synthesis and linked profiles.

Conjecture's faction history and ControlAI's policy pathways

Observed 5 September 2026. This is a factual follow-up to the Miotti/Behrens connection map. The user's phrase "control injection node" is a hypothesis about institutional function, not a finding of covert control.

Completed workstreams

Conjecture chronology and faction evidence: the documented rupture is with parts of core EA, contemporaneously visible in June 2023. The reviewed e/acc connection is public disagreement. No earlier e/acc membership, departure, or investor withdrawal was established.

Institutional uptake and current legislative status: official records distinguish parliamentary acknowledgments, lodged evidence, tabled amendments and future scheduled business. No enacted provision attributable to ControlAI was established.

Upstream funding follow-up: direct donor identities and transfer conditions remain unresolved. A surfaced grantmaking entry is an application, not an award. Original US 990-EZ retrieval failed; secondary financial extraction is not treated as original-filing verification.

Published influence method

ControlAI's Direct Institutional Plan links policy drafting with persuasion across legislatures, government, media and civil society. Its current funding pitch describes sales-style measurement: briefings, signed endorsements, then public statements after briefings. The same pitch connects advertising and creator work to constituent messages and reusable supporter lists. These are the organization's stated methods; sequence alone does not establish the counterfactual effect of each contact.

Sources: Direct Institutional Plan, funding and operating strategy. The funding page is dated 20 April 2026, but its current counters may have been updated. Budget scenarios are fundraising plans, not verified spending.

flowchart LR
  P["Policy drafting"] -->|"published strategy"| C["ControlAI advocacy"]
  C -->|"briefings and support requests"| L["Legislators"]
  C -->|"creator work and advertising"| A["Public audiences"]
  A -->|"campaign tools"| M["Constituent messages"]
  M --> L
  L -->|"official records"| T["Debates and tabled amendments"]
  T -.->|"adoption not established"| E["Enacted authority"]
  D["Upstream donors and conditions unknown"] -.->|"unresolved funding path"| C

The chart describes documented activities and an intended endpoint. It does not establish that all audience members changed their views because of the campaign or that legislators were obligated to act.

Current legislative positions

Instrument	Evidence at the cutoff
Commons NC12, cyber-resilience bill	Official decision: Not called. Sobel said on 16 June that he would not push it to a vote.
Lords Amendment 84, same bill	Official decision: Not moved. Sponsors include Clement-Jones, Kidron, Harding and Hunt.
Sobel ten-minute-rule motion on ASI	Official schedule lists 8 September 2026. Future business, not completed introduction or passage.
ControlAI's own ASI draft	Current download exists; the exact text ultimately introduced by an MP remains to be checked after introduction.

Primary status sources: NC12, Amendment 84, Parliament's 4 September schedule. July 14 Hansard explicitly attributes shutdown proposals to ControlAI. The committee stage continues; the amendment's current procedural status does not close the campaign.

The shutdown amendment and the ASI-prohibition draft are separate instruments. The former includes an AI-contributed systems compromise and catastrophic-risk trigger, parliamentary reporting and High Court relief; its subject is emergency intervention. It does not itself enact the latter's prohibition.

Direct reading of the current ControlAI draft

The 22-page download defines ASI through serious UK-security harm from capability to override specified state authorities (clause 3). It proposes compute-recipient identity records (14), model/data seizure and destruction following detected ASI progress (15), and an expert panel appointed by the Comptroller and Auditor General with the Home Affairs chair's agreement (16). Recent AI-developer employment/funding disqualifies panel candidates. Parliamentary review and civil-penalty appeals appear in clauses 15 and 23. This is proposed text, not enacted authority or an NGO appointment entitlement.

Source: current downloadable draft. More detailed clause references are included so the propositions can be checked without reproducing the draft.

The landing page currently supplies both a 22-page download and a different 19-page embedded PDF. Both were captured, with URL, time, size and SHA-256, in capture receipts. The 22-page download is the version described above. No claim is made that it exactly matches the text of a bill not yet introduced.

Evidence boundary and next records

The reviewed record establishes an organization producing policy proposals, assembling public support, and obtaining parliamentary sponsorship. It does not establish the source or conditions of its major funding, covert direction of politicians, or appointments to a future regulator. Comparative centrality among all AI-policy organizations was not measured, so the word "major" should not be treated as a quantified ranking.

The next records that would change the map are an actual introduced bill and its sponsor acknowledgments, decisions on subsequent amendments, specific donor/grant disclosures, inter-entity resource agreements, and any future appointment or procurement record. No recurring monitoring, outreach, or publication has been initiated. This standalone supplement is outside the earlier core report's structural citation audit.