AI Policy Wiki
Dashboard
⚲
AI Policy Wiki › Briefings

Gap Scan — 2026-07-24

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

Contents

Scanned

Recent window: 4 dev-log files (07-22 20:07, 07-22 22:05, 07-23 08:05, 07-23 22:05 — the last unprocessed, its items left to tonight's ingest cycle per the 07-22 Ted Cruz precedent), ~90 pages edited in last 48h. Rotation slice: 8 — litigation/ (38 pages) + standards/ (7) + government/ (19). Firecrawl unavailable again this run; built-in web search/fetch used with the full authenticity protocol (9 pulls across 8 gaps, within cap).

Gaps actioned (8 of ~22 found)

New pages created (live)

None. Recent-window pageless candidates (Sierra, Black Forest Labs, and the 07-23 22:05 deal cluster) either sit below the reference threshold or belong to tonight's unprocessed dev-log fold; slice-8 pageless candidates were all covered under existing slugs after alias resolution. Quality gate — no thin page forced.

Pages expanded (live)

  • Department of Energy (DOE) (score ~7) — the slice's squarest thin anchor: 208 words, medium/1, in-degree 12, stale since 06-24, while the Genesis Mission thread ran live all window (Genesis Mission created 07-23; Golden Age report; DOD $150M; Microsoft $60M). Added a Genesis Mission section (November 2025 EO, 17 national labs, 26 Science and Technology Challenges, ~24 AI research partners, Lux AI cluster at ORNL with AMD, July 2026 $5B+ federal commitments) and a compute/national-labs section, all from the canonical energy.gov / whitehouse.gov / anl.gov hosts, cross-wired to Science: A New Golden Age (OSTP, 2026), Microsoft, DOD — Department of Defense (AI Deployer). LBNL 15%-projection section and every prior fact/citation preserved. sources_count 1 → 5; confidence stays medium. [5 new (Source: URL) cites]
  • Raine v. OpenAI, Inc. — the slice's most-relied litigation page (in-degree 77) with last_status_check 06-12 on an active case. Status-checked and expanded: September 15, 2025 congressional testimony (CBS); October 2025 amended complaint alleging GPT-4o launched after removal of auto-termination safety protocols (Time); the fall-2025 discovery episode (memorial-footage requests; Edelson "despicable" — The Atlantic); the November 26, 2025 answer (denial of responsibility, 100+ crisis referrals, other-sources and terms-of-use defenses — The Guardian); case number CGC-25-628528; July 2026 posture (active pretrial, no dispositive ruling since the April 22 MTD denial). Every prior fact and citation preserved. sources_count 5 → 9; last_status_check → 2026-07-24. [4 new (Source: URL) cites]
  • Mobley v. Workday, Inc. (score ~5) — slice thin anchor: in-degree 26, medium/2, stale since 06-06, and mischaracterized. Corrected "first class certification" to preliminary ADEA collective certification (§ 216(b), May 16, 2025; 2025 WL 1424347) and replaced an uncorroborated Ninth Circuit-appeal sentence with the documented record (Clearinghouse docket through Dec 2025). Added: case number 3:23-cv-00770 and Judge Lin; the July 12, 2024 order's actual sort (agent theory survives, employment-agency and intentional-discrimination claims dismissed; EEOC amicus); the July 2025 HiredScore expansion and customer-list order; the 1.1 billion-application notice scale; the March 6, 2026 ruling rejecting Workday's ADEA-doesn't-cover-applicants argument (Loper Bright/Skidmore reasoning) with FEHA and physical-disability claims dismissed with leave; the March 30, 2026 amended complaint; the January 2026 Eightfold AI FCRA companion suit. Every prior sourced fact and citation preserved (incl. the Slow AI cite). sources_count 2 → 7; last_status_check → 2026-07-24. [6 new (Source: URL) cites]
  • Kadrey v. Meta (score ~5) — slice thin anchor on the live copyright thread (AI Copyright Litigation — Analysis edited in-window): in-degree 9, medium/2, stale since 06-06, page ended at the June 2025 ruling. Added a Subsequent procedural history (2026) section: the March 25, 2026 order granting leave to add contributory-infringement and uploading-based claims (Ars-hosted order PDF); the July 2026 denial of interlocutory-appeal certification on the downloading ruling with the "tidy package" final-judgment reasoning (Prof. Lee's tracker); CourtListener docket link; infobox status refreshed. Every prior fact and citation preserved. sources_count 2 → 5; last_status_check → 2026-07-24. [3 new (Source: URL) cites]

None this run. The ≥2-inbound broken targets are table pipe-link artifacts (\| escapes), _meta cross-namespace links, literal illustrative text ([[sources/...]] inside provenance notes — confirmed intentional prose, not links to fix), already-queued sources, or standing carries — consistent with the 07-20 → 07-23 runs.

Queued — foundational sources

  • AI Kill Switch Act (Lieu–Moran) — press release + bill text (score ~8) — type-1 dangling primary text on the window's top thread: formally introduced July 23, 2026; AI Kill Switch Act (Lieu–Moran) (created 07-23, low/1) cites only an aggregator. Verified: lieu.house.gov (sponsor's official House site; published_time 07-23; quotes and supporter list confirmed verbatim; corroborated by CNBC/Politico/PYMNTS). Full press-release raw saved: Raw Sources/AI Kill Switch Act - Lieu Moran Press Release.md; the linked bill-text PDF queued URL-only. Queued: INGEST-ai-kill-switch-act-2026-07-24.md.
  • OSTP, "Science: A New Golden Age" (July 21, 2026) (score ~8) — the 07-23 deferred backlog's source-queue candidate, matured: Science: A New Golden Age (OSTP, 2026) now exists citing the release URL only. Verified: whitehouse.gov PDF (canonical; 123 pages; Kratsios report-to-the-President; corroborated by UW Federal Relations, ExecutiveGov, ResearchHub). Queued URL-only per the large-PDF pattern: INGEST-science-new-golden-age-report-2026-07-24.md.
  • Google, "AI & Economy ATLAS v1.0" (July 23, 2026) (score ~7) — type-1 dangling institutional report (the Anthropic Economic Index pattern): folded onto AI Labor Disruption, Google DeepMind, Center for Security and Emerging Technology (CSET), Index by the 07-23 cycle with supporting cites only. Verified: ai.google PDF + blog.google announcement (canonical; 14.65M interactions / 68% of occupations / 21% of tasks / <10% full automation corroborated across Axios, Fox Business, adgully). Queued URL-only: INGEST-google-atlas-v1-2026-07-24.md.
  • Federal Reserve, "The AI Buildout and the Economy" (FEDS Note, July 17, 2026) (score ~5) — type-1 dangling institutional note folded onto Transformative AI (TAI) with supporting cites only. Verified: federalreserve.gov (canonical FEDS Notes path; Fed in Print item fedgfn/103558 corroborates title/date). Queued URL-only: INGEST-fed-ai-buildout-feds-note-2026-07-24.md.

Authenticity-verification failures

  • Liang Wenfeng May 20, 2026 investor-call transcript (circulated July 23) — not pulled, deliberately: a leaked document with no canonical host; the circulating copies (Substack translations/summaries) cannot satisfy the primary-document requirement, and a transcript raw could not be authenticated against any trustworthy host. The reporting folds normally as supporting cites in tonight's cycle; the transcript itself is flagged for the curator should a verifiable text emerge.

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

One-line summary

Slice 8's audit refreshed the four stalest high-reliance litigation/government anchors (DOE + three active cases, incl. a certification-mischaracterization fix on Mobley) and queued the week's four primary texts (Kill Switch Act, Golden Age report, Google ATLAS, Fed FEDS note) — 4 INGEST tasks now await curator review; no new pages needed.

Categories: Briefings | briefing | gap-scan