AI Policy Wiki
Dashboard

Gap Scan — 2026-06-18

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

Scanned

Recent window: 4 dev-log files (2026-06-16 15:09; 2026-06-17 05:05, 15:09, 22:05) and 11 wiki pages edited in the last 48h. Rotation slice: 0 — companies/ (all 94 pages). Broken-link analysis run wiki-wide (~1,580 content pages); alias resolution applied to every type-2/3/8 candidate before flagging.

Gaps actioned (10 of ~16 found)

New pages created (live)

Pages expanded (live)

  • Zhipu AI — slice-0 company on a live thread. Folded in the Z.ai marketing name and the GLM-5.2 release (MIT-licensed, 753B, 1M-token context, coding benchmarks vs GPT-5.5 / Claude Opus 4.8, ~1/6 API price); reordered the model list to lead with GLM-5.2. last_updated → 2026-06-18, sources_count 2→3. [1 new (Source: URL) cite]
  • [[concepts/us-china-ai-competition]][[analysis/us-china-ai-competition]] (5 instances across 5 files)
  • [[entities/doge]][[government/doge]] (2 instances across 2 files)
  • [[concepts/ai-regulatory-sandboxes]][[concepts/ai-regulatory-sandbox]] (1 instance)
  • [[companies/zhipu-z-ai]][[companies/zhipu-ai]] (1 instance)

Queued — foundational sources

  • "From AGI to ASI" (Google DeepMind, June 12, 2026; Genewein, Franklin, … Dafoe, Hutter, Legg) — named in the 06-17 22:05 dev-log; no page, raw file, or open INGEST task. Verified: deepmind.google publications page + arXiv 2606.12683 (PDF resolvable). No raw file saved (large PDF, fetched fresh at ingest). Queued: INGEST-agi-to-asi-deepmind-2026-06-18.md. Verification trail: proposed-sources/agi-to-asi-deepmind-2026.md.

Authenticity-verification failures

  • None this run.

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

  • entities/federal-reserve, entities/ecb, entities/bank-of-england — central banks as AI / financial-stability actors (in-degree 2 each; G7 on 06-17 tasked finance officials with assessing frontier-model effects). Score ~3. Cluster better created together.
  • concepts/ai-interpretability — broken in 3 places but an ambiguous alias between Mechanistic Interpretability and Interpretability and Safety: From Microscopes to Arms Races; deferred for human judgment rather than risk a wrong merge.
  • entities/anthropic-institute — newly launched Anthropic research/policy arm referenced across 8 pages; borderline between its own entity page and folding into Anthropic. Its first report ("When AI builds itself") is already staged in proposed-sources/. Score ~5; revisit once ingest decides the source's landing page.
  • EU "Digital Omnibus on AI" (AI Act high-risk delay to Dec 2027) — likely folds into EU AI Act (Regulation 2024/1689) rather than a new page; developments-log territory. Score ~3.
  • concepts/foundation-models — distinct from Frontier Models but low in-degree (1). Score ~1.
  • GSA revised AI contract clause / collaborative combat aircraft (CCA) — government-deployer updates better handled by developments-log folding into GSA — General Services Administration (AI Deployer) / DOD — Department of Defense (AI Deployer).

One-line summary

Created 4 pages (Noam Shazeer, CREATOR Act, ADMT, GLM-5.2), expanded Z.ai with the GLM-5.2 release, fixed 4 alias groups (9 links), and queued the DeepMind "From AGI to ASI" report for review; the only item needing the user is the recurring iCloud dashboard-rebuild conflict (see needs-review).