AI Policy Wiki
Dashboard
⚲
AI Policy Wiki › Briefings

Gap Scan — 2026-07-06

Daily gap hunt — what the wiki is missing and how each gap was triaged.

Contents

Scanned

Recent window: 4 dev-log files (07-04 morning/evening, 07-05 morning/evening), ~40 pages edited in last 48h. Rotation slice: 4 — entities/ A–G (~110 pages).

Gaps actioned (7 of ~16 found)

New pages created (live)

  • China — Interim Measures for the Administration of AI Anthropomorphic Interaction Services — top live-thread gap (score ~6: +3 live, type-3 missing legislation page). The 07-05 evening dev-log's lead regulatory item — ByteDance/Alibaba/Tencent disabling humanlike agent features ahead of the July 15 effective date — referenced a law with zero wiki coverage ("anthropomorphic interaction" appeared nowhere). Built from Geopolitechs' full translation and draft-to-final analysis (April 10, 2026 joint issuance by CAC + NDRC/MIIT/MPS/SAMR; December 27, 2025 CAC draft), a Lexology law-firm alert, Digital Policy Alert's entry-into-force record, and SCMP's shutdown coverage. Distinguished from the 2023 China — Interim Measures for the Management of Generative AI Services (back-linked from its Relationships). confidence: medium.
  • Disney v. Midjourney (consolidated studio copyright suits) — live-thread gap (score ~5: +3 live, type-3 missing litigation page). The July 2 discovery-review motion (Midjourney asking Judge Kronstadt to overturn Magistrate Judge Richlin's June 15 ruling) ran in the 07-05 evening dev-log, but the consolidated studio suit — Disney/Universal (2:25-cv-05275, filed June 11, 2025) consolidated with Warner Bros. (2:25-cv-08376) in November 2025 — had no litigation page despite Midjourney carrying scattered mentions. Built from the CourtListener docket, the chatgptiseatingtheworld case tracker (consolidation, pretrial schedule, Aug 19 2026 mediation deadline, Segal as neutral, Richlin ruling), Variety, and Mashable. Back-linked from Midjourney. confidence: medium.

Pages expanded (live)

  • Center for Security and Emerging Technology (CSET) — slice thin anchor: 625 words, sources_count: 1, in-degree 32. Added the late-2025 leadership transition (Helen Toner executive director, Cara LaPointe director of analysis), CSET's self-description, the 2025 PLA-procurement paper series, the summer 2025 AI R&D automation workshop, Project ATLAS, and the stated 2026 research priorities — all from CSET's 2025 Annual Report (March 31, 2026, primary host). sources_count 1 → 4. Every prior fact and citation preserved. [1 new source, 4 new (Source: URL) cites]
  • Bureau of Industry and Security (BIS) — slice thin anchor in a core area (export controls): in-degree 30, sources_count missing, coverage ended May 2026. Added a section on the June 2026 Anthropic model controls — the June 12 restrictions on Claude Mythos 5 / Claude Fable 5 and the June 30 withdrawal with Lutnick's letter ending license requirements, contrasted with the rescinded Diffusion Rule's ECCN 4E091 country-tier approach. Frontmatter repaired (sources_count: 6). [3 new (Source: URL) cites: NYT, CNBC, WTTL]
  • None this run. The remaining high-count broken targets are either escaped-pipe table links (false positives, e.g. [[companies/anthropic\|Anthropic]]), legacy pre-v4.3 pointers to _meta/ artifacts ([[lint-report]], [[log]] — a lint task, not a content gap), or deliberately deferred (below).

Queued — foundational sources

  • "The Alignment Problem of 1776" (Peter Wildeford, July 4, 2026) — type-1 dangling foundational reference on a live thread: Peter Wildeford names and paraphrases the essay with only a supporting URL. Verified: author's own Substack (primary host). Saved: Raw Sources/The Alignment Problem of 1776.md (full text). Queued: INGEST-alignment-problem-1776-wildeford-2026-07-06.md.
  • "U.S. Policies Unintentionally Accelerated China's Open AI Ecosystems" (Jin, Kunievsky, Lou, Sun & Evans, arXiv 2606.15999, June 14, 2026) — type-1 dangling foundational reference in a core area: Export Controls (AI) (edited <48h) cites its 11× forking differential as a supporting URL. Verified: arXiv abstract page (DOI 10.48550/arXiv.2606.15999). Large open-access PDF → fetch fresh at ingest. Queued: INGEST-us-policies-china-open-ai-ecosystems-2026-07-06.md.

Authenticity-verification failures

  • Goldman Sachs Research, "An AI job apocalypse?" (Joseph Briggs) — sought as a type-1 queue candidate (AI Labor Disruption cites its 9%-temporary-displacement projection). The primary-host PDF (goldmansachs.com) rejects automated fetching, and no trustworthy secondary host carries the full text (proprietary research). Not queued — the finding remains covered as an attributed supporting cite. Revisit if Goldman publishes an open landing page.

Deferred backlog (over the daily cap — re-surfaces next run)

  • Duplicate pages entities/cdao.md / government/cdao.md — new slice finding: the CDAO appears in both folders, violating the no-duplication rule (basename in-degree 57). Merge/supersede is out of gap-identifier scope — flag for lint, same class as the companies/reflection* duplicate (open since 06-28). Score ~4.
  • Ramp × Revelio Labs AI jobs study (21,559 firms; June 30) — borderline foundational; already well-covered as supporting cites on AI Labor Disruption. Queue candidate if it keeps being named. Score ~5.
  • Slice thin anchors below the top two — entities/access-now (339w, in-deg 22), entities/aclu (335w, src 0, in-deg 16), entities/gillian-hadfield (288w, in-deg 18), entities/deirdre-mulligan (286w, in-deg 17). Score ~4 each.
  • entities/dario-amodei frontmatter — sources_count missing at in-degree 151; frontmatter repair for lint, not expansion. Score ~4. Same for entities/bureau-of-industry-and-security (fixed this run), entities/center-for-ai-safety, entities/future-of-life-institute, entities/american-innovators-network.
  • Low-confidence stubs in slice — entities/anna-neumann (70w, low, in-deg 7), entities/abundance-institute (151w, low, 7), entities/adele-lopez (264w, low, 6), entities/charles-sun (259w, low, 6). Confirm substance before expanding. Score ~3.
  • [[claude-code]] (in-deg 7, broken) — perennial; placement needs the user's products/-schema call. Score ~4.
  • Argentina "automated company" bill — recurring since 07-03; lint rule (page on formal advancement) holds; folds into AI Personhood meanwhile. Score ~4.
  • sources/anthropic-pwc-expanded-partnership (in-deg 6, broken) — supporting press release; lint/fold task, no sources/ page warranted. Score ~3.
  • Bender & Koller 2020 ("Climbing towards NLU") — carry from 07-05; queue candidate once the stochastic-parrots ingest lands. Score ~3.
  • arXiv 2606.26470 (China's scientific self-reliance in patents) — the companion paper from the 07-05 dev-log; less core than 2606.15999. Score ~3.
  • Person/org pages at in-degree 2 — pam-bondi, masahiro-mori, anil-seth, eric-horvitz, orin-kerr, alondra-nelson, palmer-luckey, tom-mitchell, soufan-center, federal-reserve. Confirm substance before creating. Score ~2–3.
  • concepts/foundation-models (in-deg 2, broken) — still needs disambiguation vs. General-Purpose AI (GPAI). Score ~2.
  • [[chinatalk]] (in-deg 1, broken, on Bureau of Industry and Security (BIS)) — no page under any name; an entities/ page for the ChinaTalk publication would be defensible if references recur. Score ~2.

One-line summary

Two live-thread holes closed with new pages (China's anthropomorphic-interaction measures, the consolidated Disney/Universal/WB v. Midjourney suit), the slice's two biggest thin anchors (CSET, BIS) expanded and repaired, and two named foundational sources (Wildeford's 1776 essay, the arXiv export-controls paper) verified and queued; the Goldman jobs report failed primary-host verification and stays a supporting cite.

Categories: Briefings | briefing | gap-scan