AI Policy Wiki
Dashboard
⚲
AI Policy Wiki › Briefings

Gap Scan — 2026-06-06

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

Contents

Scanned

Recent window: 5 dev-log files (2026-06-04-1505/2205, 2026-06-05-1505/2205, 2026-06-06-0505) and ~54 wiki pages edited in the last 48h. Rotation slice: 2 — concepts/ G–O (34 pages thinness-scanned against in-degree). The nightly Ingest-Reflect cycle (06-05 evening) had already folded the Anthropic "When AI builds itself" essay and most of the 06-04/06-05 dev-log items, so this run targeted the residue: the June 5 NSPM-11 primary text (not yet a page), live-thread entity/concept gaps, and slice-2 thin anchors.

Gaps actioned (6 of ~21 found)

New pages created (live)

  • NSPM-11 — Artificial Intelligence in the National Security Enterprise (Trump, June 5, 2026) — gap type 1+3, live thread (June 5–6 dev logs), core area (US AI national-security policy). Trump's NSPM-11 on AI in the national security enterprise had no page despite dominating the window. Built from the authentic whitehouse.gov primary text + fact sheet; documents the four pillars (Adoption/Adaptation/Assurance/Accountability), the 90-day DOD Directive 3000.09 rework, the AI National Security Strategic Reserve, the anti-kill-switch assurance clause, and the Sec. 3(b) FISMA-based vendor contract-termination mechanism (the lever that lands on the Anthropic/Pentagon dispute). confidence: high (single authoritative primary text, cross-corroborated). Depends on queued source: NSPM-11 (queued this run).
  • Victor Zhenyi Wang — gap type 3, dangling foundational reference (5 inbound; deferred since 2026-05-26). Co-author of the Knight Columbia Conceptual Model to Guide AI Risk Governance Strategies behind Sociotechnical AI Risk Governance. PhD student, UC Berkeley School of Information. 2 sources; confidence: medium.
  • Center for a New American Security (CNAS) — gap type 3 (2–4 inbound; named by name in Dominance by Understanding (Frazier-Rozenshtein policy frame) and the Lawfare source). Center for a New American Security — DC national-security think tank, AI Security and Stability project; source of the NAIRR public-vs-private investment-gap figure. Sits in a wiki core area (defense / compute / US-China). 2 sources; confidence: medium.
  • Jack Dorsey — gap type 3 (4 inbound; deferred repeatedly since 2026-05-26). The most-cited corporate exemplar of Intelligence Replaces Hierarchy (Block's AI flattening, ~175-report spans, all-report-to-Dorsey goal). Built from existing wiki anchors (Block (Square / Cash App), the Guardian manager-purge source) + the Block statement. 2 sources; confidence: medium.

Pages expanded (live)

  • LLM Fallacy — slice-2 thin anchor (442 words, confidence: low, sources_count: 1, in-degree 20). Added an Empirical corroboration section folding three supporting sources: the 2024 arXiv study "AI Makes You Smarter, But None The Wiser" (AI improves scores but degrades self-calibration), reporting on AI-and-metacognition experiments (up to 4-point self-overestimation), and the AI-literacy inversion of Dunning-Kruger. Raised low → medium; sources_count 1 → 3. [3 new (Source: URL) cites]

Queued — foundational sources

  • NSPM-11 — AI in the National Security Enterprise (June 5, 2026) — gap type 1, named across the June 5–6 dev logs and now anchoring the new legislation page. Verified: whitehouse.gov primary text + fact sheet, corroborated by Reuters and InsideAIPolicy (2026-06-05). Saved: Raw Sources/National Security Presidential Memorandum NSPM-11 — AI in the National Security Enterprise (5 June 2026).md (full verbatim text). Queued: INGEST-nsm-ai-national-security-2026.md. Verification record: NSPM-11 — Artificial Intelligence in the National Security Enterprise (Trump, signed June 5, 2026).

Authenticity-verification failures

  • None this run. The single pulled foundational source (NSPM-11) verified cleanly against its canonical .gov host.

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

  • Government equity stakes in AI labs / OpenAI "Public Wealth Fund" / Sanders Sovereign Wealth Fund (live thread, June 5) — conceptually already covered by AI Dividends (Universal Basic Capital, Digital Dividend, Global Dividend) (UBC / sovereign-wealth-fund variant) and AI Governance (umbrella); the policy-event belongs to developments-log / nightly ingest, not a new page. Flagged, not actioned. Re-evaluate if it hardens into a named program.
  • entities/nscai (National Security Commission on AI) (2–4 inbound; named alongside CNAS in dominance-by-understanding). Score 3 — strong next-run candidate now that CNAS exists.
  • entities/brian-armstrong (4 inbound) — sibling to the Dorsey founder-mode/manager-purge thread; create alongside a Coinbase pass. Score 3.
  • claude-code (6 inbound) — genuinely ambiguous placement (product, not company/entity per "Where does X go?"; candidate to fold into Anthropic or a concepts page). Quality-gate deferred pending a placement decision. Score 2.
  • Slice-2 thin anchors not reached: Normative Competence (1227w but src 1, deg 26), Levels of Autonomy for AI Agents (Feng-McDonald-Zhang framework) (897w, src 1, deg 25 — has a queued INGEST for the Feng paper), Machine Fluency (650w, src 1, deg 17), Monitorability Tax (539w, src 1, deg 13), Model Welfare (453w, src 0, deg 6). Score 2.
  • concepts/singularity (3 inbound; decide distinct page vs. fold into Superintelligence) and concepts/universal-basic-income (3; likely fold into AI Dividends (Universal Basic Capital, Digital Dividend, Global Dividend)). Score 3 — route to lint/crystallize for the fold decision.
  • Lint territory (route to lint, not gap-scan): sources/anthropic-pwc-expanded-partnership (6, likely supporting → fold into Anthropic); sources/trustworthy-agents-in-practice (5, logged created-then-deleted); the entities/pim-dewitte ↔ entities/pim-de-witte slug split (5 combined).

Dashboard rebuild

Failed (recurring) — bin/build-dashboard.py aborts on shutil.rmtree(SITE) because Wiki/site/ contains OS-locked iCloud duplicate-naming artifacts (concepts 2, entities 2, …) that can't be removed from the sandbox. Impact is cosmetic: the live wiki graph is fully updated; only the HTML mirror under Wiki/site/ is stale. Logged at 2026 06 06 Dashboard Rebuild Failed with a host-side fix.

One-line summary

Created NSPM-11 + three live-thread entity pages, expanded the llm-fallacy thin anchor with three empirical sources, fixed the aisi alias, and queued the verified NSPM-11 primary text for foundational ingest. User review needed: the queued NSPM-11 ingest (INGEST-nsm-ai-national-security-2026.md) and the recurring dashboard-rebuild failure (host-side fix in needs-review).

Categories: Briefings | briefing | gap-scan