Gap Scan — 2026-06-26

Scanned

Recent window: 4 dev-log files (2026-06-24 15:41, 2026-06-24 22:05, 2026-06-25 15:10, 2026-06-26 05:09) and ~40 wiki pages edited in the last 48h (companies, concepts, entities, government, legislation, litigation, analysis, sources). Rotation slice: 8 — litigation/ + standards/ + government/ (45 pages). Broken-link analysis ran wiki-wide (existing-slug set vs. every [[wikilink]] target, ranked by inbound count); dangling-reference, anchor-thinness, and recent-thread scans ran over the slice plus the recent window. Deduplicated against the open queue (48 INGEST- tasks, 35 proposed-sources records), the last 7 days of gap-scan reports, and Raw Sources/.

Gaps actioned (5 of ~12 surviving candidates)

New pages created (live)

• World Artificial Intelligence Cooperation Organization (WAICO) — recent-thread missing entity. The World Artificial Intelligence Cooperation Organization, China's proposed intergovernmental AI-governance body (announced by Premier Li Qiang at WAIC Shanghai, July 26, 2025, alongside the Global AI Governance Action Plan), named by name in the 2026-06-26 dev-log as the explicit target of the U.S. State Department's Pax Silica push ("dispel the chimera of AI global governance … in the UN and WAICO"). No page and no prior mentions anywhere in the wiki despite Pax Silica (US-Led AI Supply-Chain Initiative) and US-China AI Competition: Different Races, Different Metrics already existing. Built from the Carnegie Endowment analysis (Basu, May 2026), China's MFA primary statements, and the gov.cn announcement; US-opposition context from the dev-log source. entity_type: intergov. confidence: medium.
• Resolução CFM nº 2.454/2026 (Brazil — AI in medical practice) — recurring international sector reg (deferred 2026-06-25). Brazil's Federal Council of Medicine Resolution No. 2,454/2026, the first sector-specific standard of care for AI in Brazilian medicine: human-in-the-loop, physician final authority, a four-tier (low/medium/high/unacceptable) risk classification, a bar on AI autonomously communicating diagnoses, institutional AI/Telemedicine Commissions, and LGPD compliance. Built from the verified CFM primary announcement plus IBA/Veirano, Mayer Brown, Lefosse, and IAPP. Corrected a dev-log date error: the 2026-06-24 dev-log dated publication to June 24; the resolution was actually published Feb 27, 2026 (the June 24 item was IAPP coverage). confidence: medium. Depends on queued source (INGEST-cfm-resolution-2454-2026.md) — primary text cited inline.

Pages expanded (live)

• Chief Digital and AI Office (CDAO) — the single most-relied weak page in the slice (in-degree 50, confidence: medium, sources_count: 1, one source). Added a "2025 realignment and Department of War redesignation" section from the Congressional Research Service insight IN12615: the JAIC→CDAO lineage, the Aug 14, 2025 Feinberg memo moving CDAO under USD(R&E) (a CRS-characterized "effective demotion"), the EO 14347 "Department of War" secondary designation, and the Shanahan-vs-proponents debate. Stale-fact fix: the page asserted "led by Radha Plumb (as of 2025)"; revised to reflect that leadership was in transition as of early 2026, with Cameron Stanley reported as the frontrunner (DefenseScoop, Jan 2026). Added a reporting-line row to the infobox. sources_count 1 → 3; every prior fact and the existing [[cset-csis-us-china-natsec]] citation preserved. [3 new (Source: URL) cites — 1 CRS primary, repeated; 1 DefenseScoop]
• Five Eyes Joint Guidance on Secure Deployment of AI Agents (May 2026) — slice anchor, sources_count: 1 (single cyberscoop source), in-degree 6. Added the official document title ("Careful Adoption of Agentic AI Services"), the CISA primary release page, and the U.S. DoD-hosted PDF (document-dated Apr 30, 2026), plus CISA's own framing of the guide's audience. sources_count 1 → 3; existing facts and the cyberscoop citation preserved. [2 new primary (Source: URL) cites]

Link aliases fixed

• [[concepts/human-in-the-loop]] → [[concepts/meaningful-human-review]] in the new CFM page (pre-emptive: the intended target exists under another slug; corrected before commit so no new broken link was introduced).

Queued — foundational sources

• Resolução CFM nº 2.454/2026 (Conselho Federal de Medicina, Brazil; published in the Diário Oficial da União Feb 27, 2026) — primary regulation behind the new Resolução CFM nº 2.454/2026 (Brazil — AI in medical practice) page. Verified: portal.cfm.org.br (the issuing body's own domain; announcement scraped live, HTTP 200, confirms number/date/provisions), corroborated by six independent legal/industry references. Queued URL-only (full text is a Portuguese PDF best fetched at ingest): INGEST-cfm-resolution-2454-2026.md (+ verification record in proposed-sources/cfm-resolution-2454-2026.md).

Authenticity-verification failures

• None this run. The one queued source verified against its canonical primary host (the issuing council's own site). New pages and expansions used primary/supporting web sources folded as inline (Source: <URL>); the CRS (congress.gov) and CISA (cisa.gov / media.defense.gov) citations are primary-host government documents.

Deduplicated (already pending / already covered — not re-actioned)

• Pax Silica / second summit (June 25) — Pax Silica (US-Led AI Supply-Chain Initiative) and sources/pax-silica-state-dept-2025 already exist; the second-summit news is an update-class item for developments-log, not a gap. (Its WAICO counterpart had no page — that is the actioned gap above.)
• Hugging Face nudification tools (06-26) — Hugging Face already exists; news-cycle update.
• Anthropic–Alibaba distillation (06-24/25) — Adversarial Distillation (edited 06-25) and Alibaba / Qwen Team already cover it.
• House Science 10-bill markup / CREATE AI Act / NAIRR / NIST CAISI (06-24/25/26) — CREATE AI Act (Young) (updated 06-24) and NIST CAISI (Center for AI Standards and Innovation) exist; update-class.
• Five Eyes "Call to Action" (June 22) — already queued 2026-06-25 (INGEST-five-eyes-ai-call-to-action-2026-06-25.md); distinct from the May 1 agentic-AI guidance expanded above.
• OpenAI/Broadcom Jalapeño, Micron, Qualcomm/Modular, SK Hynix, Meta, FERC/DOE/DOD (06-24/25) — all surfaced from dev-logs; targets already have pages or are supporting-class. Broadcom/Qualcomm remain supporting mentions (see deferred).
• [[claude-code]] (in-degree 8, broken) — placement still genuinely ambiguous; held under quality gate again (deferred).

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

• GPT-5.6 government-restricted release (06-25/26) — the first preemptive U.S. request to restrict a model launch; currently update-class on OpenAI, no model page warranted while the model is unreleased and access-restricted. Revisit if it crystallizes into a named policy precedent. Score ~4.
• [[claude-code]] (in-degree 8) — highest-in-degree genuine broken target; needs a deliberate placement call (fold into Anthropic, add to AI Coding Agents, or a dedicated product page). Score ~4.
• LIFT AI Act (Literacy in Future Technologies Artificial Intelligence Act) (in-degree 11, confidence: high with sources_count: 1) — confidence overstated relative to single-source support; expand to 3 sources or downgrade next run (carried from 2026-06-25). Score ~4.
• government/cdao residual broken links — [[clawed]], [[dod-directive-3000-09]], [[autonomous-weapons]], [[ai-national-security]], [[ai-race-dynamics]], [[missing-coverage-2026-04-14]] are unprefixed/aliased targets; resolve as a small alias-fix batch (lint-adjacent) rather than singly. Score ~3.
• Broadcom / Qualcomm (custom-silicon partners recurring across 06-24/25 dev-logs) — still supporting mentions; revisit if either recurs as a tracked actor. Score ~3.
• Five Eyes co-issuer entity pages (entities/ncsc-uk, entities/acsc, entities/cccs, entities/nsa) — referenced by both Five Eyes products; build as a small cluster at the queued Call-to-Action ingest. Score ~3.
• litigation/taiwan-tsmc-trade-secrets (sources_count 1, in-degree 0) and litigation/kahn-v-anthropic (256 words, in-degree 0) — thin/weak but low-reliance; hold. Score ~2.
• Brazilian AI Bill (PL 2.338/2023) — referenced by the new CFM page as the pending comprehensive framework; candidate legislation/ page once it advances. Score ~2.

Needs user action

• Dashboard rebuild blocked (recurring since 2026-06-06). bin/build-dashboard.py again aborted at shutil.rmtree(Wiki/_meta/site/) with PermissionError: Operation not permitted: 'companies 2'; an explicit rm -rf of each 2 folder this run also returned Operation not permitted. iCloud conflict-copy folders inside the site tree cannot be deleted from this environment. Delete the 2 folders inside Wiki/_meta/site/ in Finder, then run python3 bin/build-dashboard.py. All of today's content changes (2 new pages, 2 expansions) are written to their real folders and will render on the next successful build. (Logged in needs-review/2026-06-26-dashboard-rebuild-failed.md.)

One-line summary

Created two live pages — World Artificial Intelligence Cooperation Organization (WAICO) (China's proposed WAICO, the explicit target of the State Department's Pax Silica push) and Resolução CFM nº 2.454/2026 (Brazil — AI in medical practice) (Brazil's first AI-in-medicine standard of care, with a dev-log publication-date error corrected) — expanded the heavily-relied single-source Chief Digital and AI Office (CDAO) (with the 2025 USD(R&E) realignment and a Radha-Plumb stale-fact fix) and the single-source Five Eyes Joint Guidance on Secure Deployment of AI Agents (May 2026) (adding the official title and CISA/DoD primary sources), and queued the verified CFM primary resolution for review; GPT-5.6, claude-code, and the over-confident LIFT AI Act (Literacy in Future Technologies Artificial Intelligence Act) lead the deferred backlog; the dashboard still needs the manual 2-folder cleanup before it will rebuild.