The standing board of AI-regulation compliance dates. Updated weekly with each issue of AI Policy Wire. Movement since last update is flagged in bold.
European Union
| Instrument | What it requires | Who it binds | Compliance date | Status |
|---|
| Regulation (EU) 2024/1689, Art. 50 — transparency duties | Disclose AI interaction (chatbots); label AI-generated content | Providers and deployers | 2 August 2026 | Enforcement from that date |
| Art. 101 — GPAI enforcement powers | Commission may request documentation, evaluate models, order corrections, fine up to 3% of global turnover or EUR 15M | GPAI providers | 2 August 2026 | Exercisable from that date |
| GPAI model obligations (Arts. 53–55) | Documentation, copyright policy, training-data summary; systemic-risk duties | GPAI providers | Applied since 2 August 2025 | In force; enforcement from Aug 2026 |
| High-risk system obligations (Annex III) | Risk management, data governance, registration, human oversight | Providers and deployers of high-risk systems | December 2027 (per the amended timeline) | Not yet applicable |
| Prohibited practices (Art. 5) | Bans on manipulation, social scoring, most real-time biometric ID | All operators | Applied since 2 February 2025 | In force |
United States — states
| Instrument | What it requires | Who it binds | Compliance date | Status |
|---|
| Colorado ADMTA (SB 25B-004, replacing the delayed SB 24-205) | Consumer notice for algorithmic decision-making technologies; deployer duties on high-impact decisions | Deployers and developers operating in Colorado | 1 January 2027 | Enacted |
| Colorado companion AI laws (2025–26 sessions) | Sectoral duties incl. health and employment contexts | Sector-specific operators | Various, 2026–27 | Enacted |
| Other state AI bills (1,100+ introduced in 2025 sessions) | Varies — disclosure, impact assessment, sector rules | Varies | Rolling | Mostly proposed; enactments tracked in each issue |
United States — federal
| Instrument | What it requires | Who it binds | Compliance date | Status |
|---|
| Agency AI rules and executive actions | Sectoral guidance and rules (health, finance, employment, education) | Regulated sectors | Rolling | Tracked per agency |
Standards operators actually use
| Standard | What it is | Status |
|---|
| ISO/IEC 42001 | AI management-system standard; certifiable | Published; certification market forming |
| NIST AI Risk Management Framework | Voluntary US risk framework; frequent contractual reference | Current |
Independent regulatory intelligence — not legal advice. Matters with material exposure warrant counsel.