🇪🇺EU AI ActActive✓ PROVA covers
European Union · Regulation (EU) 2024/1689
high risk
Scope: All AI systems placed on EU market — risk-tiered obligations for providers, deployers, and importers
📅 Annex III high-risk: 2 Dec 2027 · Annex I: 2 Aug 2028 (Digital Omnibus, pending adoption)
Recent developments
2026-05-07Digital Omnibus on AI: Council + Parliament reach provisional political agreement. Annex III high-risk deferred to 2 Dec 2027, Annex I to 2 Aug 2028, Art. 50(2) transparency to 2 Dec 2026. Substantive obligations and penalties unchanged. Not yet law — pending formal adoption + Official Journal publication; original dates technically in force until then.
2025-08-01Tier 1 provisions in force: prohibited AI practices, GPAI obligations, market surveillance framework.
2025-02-13Commission published AI Office guidance on GPAI model classification and systemic risk thresholds.
🇺🇸Colorado AI ActStayed✓ PROVA covers
Colorado, US · SB 24-205
high risk
Scope: High-risk AI in consequential decisions — employment, credit, healthcare, education
📅 Est. Jan 1, 2027 (SB 26-189 pending)
Recent developments
2025-11-14Federal court stayed SB 24-205 enforcement pending constitutional challenge. SB 26-189 introduced as replacement with revised scope.
2026-03-01SB 26-189 passed committee. Estimated enforcement Jan 2027 if signed. Developer + deployer obligations retained.
🇺🇸Texas TRAIGAActive✓ PROVA covers
Texas, US · HB 4
medium risk
Scope: Government AI systems — law enforcement, justice, critical infrastructure; behavioral manipulation prohibited
📅 In force Jan 1, 2026
Recent developments
2026-01-01HB 4 took effect. Applies to state agencies and contractors using AI in law enforcement, courts, and critical infrastructure.
2026-02-15Texas DPS published first agency AI inventory under TRAIGA requirements. 43 systems catalogued.
🇺🇸Illinois Employment AI ActActive✓ PROVA covers
Illinois, US · HB 3773
medium risk
Scope: AI used in hiring and HR decisions — mandatory notice, opt-out rights, IDOL enforcement
📅 In force Jan 1, 2026
Recent developments
2026-01-01HB 3773 in effect. Illinois IDOL issued initial compliance FAQ covering notice requirements.
2026-03-10First IDOL investigation opened against Chicago-area employer for AI screening tool lacking required candidate disclosure.
🇺🇸California AB 2013Active
California, US · AB 2013 (2024)
medium risk
Scope: Training data transparency — AI systems trained on personal data must publish training data summaries
📅 In force Jan 1, 2026
Recent developments
2024-09-28Governor Newsom signed AB 2013. Requires developers of AI systems to post training data summaries on their websites.
2026-01-01Enforcement commenced. CalDOJ issued first compliance letters to AI developers lacking training data disclosures.
🇺🇸New York AI ActPending
New York, US · S 5152
high risk
Scope: Comprehensive AI regulation — high-risk AI in employment, housing, credit, healthcare; impact assessments required
📅 Pending — legislature 2026 session
Recent developments
2026-01-20S 5152 reintroduced in 2026 legislative session with expanded scope to include generative AI systems.
2026-04-08Senate Commerce Committee advanced S 5152 to floor. Enforcement date if passed: estimated Jan 2027.
🇪🇺EU AI Liability DirectiveProposed
European Union · COM/2022/496
medium risk
Scope: Civil liability for AI-caused harm — reversal of burden of proof, disclosure of AI system documentation
📅 Stalled — Council negotiations
Recent developments
2024-11-01Council negotiations stalled on burden-of-proof reversal provisions. No revised text published.
2026-02-01European Parliament renewed call for Directive adoption. Commission indicated revised proposal expected H2 2026.
🇨🇦Canada AIDAProposed
Canada · Bill C-27, Part 3
low risk
Scope: High-impact AI systems — mandatory impact assessments, transparency requirements, ISED enforcement
📅 Died on order paper — reintroduction expected
Recent developments
2024-03-28Parliament dissolved. Bill C-27 (including AIDA) died on the order paper without receiving Royal Assent.
2025-09-15Innovation Canada consultation on revised AIDA framework launched. Reintroduction expected Q1 2026 under new government.
🇨🇳China Generative AI RegulationsActive
China · CAC Order No. 15 (2023)
high risk
Scope: Generative AI services for Chinese users — security assessment, content labelling, user real-name registration
📅 In force Aug 15, 2023
Recent developments
2023-08-15Measures for the Management of Generative Artificial Intelligence Services entered into force.
2025-10-01CAC published revised draft expanding scope to AI-generated video and audio; public comment period closed.
🇧🇷Brazil AI BillPending
Brazil · PL 2338/2023
medium risk
Scope: Comprehensive AI governance — risk classification, transparency obligations, ANPD enforcement authority
📅 Pending — Senate approval
Recent developments
2024-12-11Brazilian Senate Committee on Science approved PL 2338/2023 with amendments. Full Senate floor vote pending.
2026-03-20Senate leadership confirmed floor vote scheduled for Q2 2026. If passed, ANPD to issue implementing regulations within 180 days.