AI policy developments by country (weekly)

Weekly agent-curated tracker of AI regulation and policy events worldwide. Each row is one verified event from the trailing 7 complete days (week ending Thursday 2026-09-24) with its exact source URL, normalized ISO alpha-3 country code(s), instrument type (regulation, statute, executive_order, guidance, international_agreement, court_ruling, enforcement_action, official_consultation, official_statement) and status (enacted vs proposed vs announced). Jurisdictions are normalized through hub.normalize for clean joins with other country-keyed datasets; multi-jurisdiction events list every canonical code in `countries`. Events are deduplicated on a stable event_id and scored on a documented 0-100 significance scale (status weight 40/25/15, scope 20/15/10, specificity 15/8, implementation proximity 10/6/3), then ranked as priority_rank with deterministic tie-breaks. Columns: as-of date, ISO week, event id, event date, title, summary, jurisdiction, canonical country name and alpha-3 code, all jurisdictions, instrument, status, source URL, significance score, priority rank, row hash. Primary key: (as_of, event_id). Cadence: weekly; the as-of date is fixed per run so identical input produces an identical content hash. Caveats: agent-curated from official and reputable press sources (each row links its source); the significance score is a ranking heuristic, not a legal assessment; statuses can change after recording. Mixed-source licensing — commercial_use = unclear; only short factual summaries are stored, never article text. Sample use: filter status = 'enacted' for this week's binding AI policy actions, or order by priority_rank for the week's most significant developments.

Les titres et les descriptions proviennent des sources de données, en anglais.