Detailed research note

DV category authority: a later adoption claim and an upstream correction boundary

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.

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.

Dated records and their limits

DV'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.

The [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.

IWF'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.

The [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.

Authority remains split

The 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.

The 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.

Consequently, 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.

Finite search and exact reopening record

Ten 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.

Source custody and integration candidates

Four 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 SHA256
category-adalytics-response2025.html d243a7439049bcfa7f8ce3611bdb12d4276187bd0c72d77bd5b1acd8943c9ba9
category-iwf-progress2025.html c9ea0b9f1d8c6a2253382cb3cf037d95a94634e5326561b85690c02cc9765c83
category-iwf-partnership2025.html 64763eefe3e058b885e105dd3d9a9c538c4450a7f6c26bb9a735bdfdd3f178d2
category-iwf-dynamic2025.html 6a98771131af9c350f6f7a2cab2c0d868d9331a8193ac89758e125b6c6fad869

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.