Selected enacted laws and notable measures, with status and effective dates as reported by primary or legislative trackers. Snapshot as of Nov 3, 2026. Not exhaustive and not legal advice.
How to read this snapshot
U.S. AI rules are mostly state-driven and topic-specific (employment, elections, frontier-model transparency, chatbots, NCII). Comprehensive federal AI legislation has not been enacted as of Nov 3, 2026. Executive Order 14365 (Dec 11, 2025) directs federal agencies concerning state AI laws the administration describes as burdensome, while stating that a future national standard should not override state child-safety rules or state government purchasing of AI.[18] Abroad, approaches range from binding risk-based statutes (EU) to promotion frameworks, voluntary codes, and pending bills. Court challenges have enjoined or narrowed some state election-deepfake statutes; NCSL summarizes those cases alongside the enacted count.[28]
Pending bills continue to move in many legislatures. Examples reported in 2025–2026 trackers include additional election-synthetic-media disclosure bills, chatbot safety proposals, and expansions of frontier-model reporting (for example, New York follow-on proposals building on the RAISE Act). Status changes quickly; the tables below prioritize enacted measures with dated effective dates, then flag representative international instruments.
Selected U.S. state measures (comprehensive / sector)
| State | Law / bill | Status (as of Nov 3, 2026) | Effective / key date | Topic |
|---|---|---|---|---|
| Texas | HB 149 (TRAIGA) | Enacted | Jan 1, 2026 | Government AI disclosure; banned uses; AG enforcement[1] |
| Texas | HB 3512 | Enacted | Sep 1, 2025 | Annual DIR-certified AI training for covered officials/staff[2] |
| California | SB 53 (Transparency in Frontier AI Act) | Enacted; in force | Jan 1, 2026 | Frontier / large frontier developer transparency, catastrophic-risk assessments, incident reporting[29] |
| California | AB 2013 (Training Data Transparency Act) | Enacted; in force | Jan 1, 2026 | Documentation of datasets used to train generative AI systems[30] |
| Colorado | SB 24-205 (original CAIA) | Repealed/replaced before enforcement | Never operated | Prior comprehensive high-risk AI duties[31] |
| Colorado | SB 26-189 (ADMT Act) | Enacted; rulemaking underway | Jan 1, 2027 | Automated decision-making in consequential decisions; consumer notice/correction rights[31] |
| Colorado | HB 26-1263 (Chatbot Safety Act) | Enacted | Jan 1, 2027 | Conversational AI age estimation, disclosures, teen safeguards, self-harm protocols[31] |
| Utah | AI Policy Act + SB 226 / HB 452 amendments | In force (Office sunset Jul 1, 2027) | Disclosure rules May 7, 2025 | Generative-AI consumer disclosures; mental-health chatbot rules; Office of AI Policy[32] |
| New York | RAISE Act (as amended Mar 27, 2026) | Enacted | Jan 1, 2027 | Frontier / large frontier developer transparency and safety reporting[33] |
| Illinois | HB 3773 (Human Rights Act AI amendment) | Enacted; in force | Jan 1, 2026 | Employer notice and anti-discrimination limits on AI in covered employment decisions[34] |
| Illinois | SB 315 (AI Safety Measures Act) | Enacted (tracker) | Jan 1, 2027 | Large frontier developer frameworks, audits, incident reporting[34] |
Election deepfakes and NCII (multi-state)
NCSL reports that, as of its Oct 13, 2026 update, 31 states have enacted laws regulating deepfakes in political messaging—mostly disclosure rules, with prohibition models in states including Texas, Minnesota, Maryland, and (as amended after litigation) California.[28] Federal TAKE IT DOWN (above) addresses non-consensual intimate imagery including certain AI forgeries. Many states also have separate intimate-image or digital-replica statutes; those are tracked in state codes and in federal-state comparisons rather than in the election table alone.
| State (examples) | Measure | Status notes (as of Nov 3, 2026) | Topic |
|---|---|---|---|
| Texas | Election Code §255.004 | In force; Class A misdemeanor for specified deceptive deep fake video within 30 days of an election | Election deepfakes[14] |
| California | Election deepfake / AI ad rules (amended after Kohls) | Prior version enjoined 2025; legislature amended; see NCSL case note | Election deepfakes / disclosures[28] |
| Minnesota | Minn. Stat. §609.771 | Enacted; federal challenges pending on merits per NCSL | Election deepfake prohibition[28] |
| Maryland | Md. Election Code §§2-110, 16-905 | Enacted 2026; year-round deceptive deepfake prohibition (NCSL) | Election deepfakes[28] |
| Multi-state | Various disclosure statutes | 27 states use disclosure approaches per NCSL Oct 13, 2026 summary | Political synthetic media labels[28] |
| Federal | TAKE IT DOWN Act | In force; platform removal duty since May 19, 2026 | NCII / digital forgeries[16] |
Selected international frameworks
| Jurisdiction | Instrument | Status (as of Nov 3, 2026) | Notes |
|---|---|---|---|
| European Union | AI Act + Digital Omnibus adjustments | In force; phased application | Risk-based duties; GPAI since Aug 2025; high-risk Annex III Dec 2, 2027; Annex I Aug 2, 2028[21] |
| United Kingdom | Sector-regulator / pro-innovation approach; AI Opportunities Action Plan | No comprehensive AI statute as of House of Commons Library briefings into 2026 | Existing regulators apply principles; frontier binding rules discussed but not enacted in the same form as the EU Act[35] |
| China | Interim Measures for Generative AI Services | In effect since Aug 15, 2023 | Filing, content, training-data, and security obligations for public generative services[36] |
| South Korea | AI Basic Act (Framework Act) | In effect Jan 22, 2026 | National AI governance, industry promotion, and trust/safety duties[37] |
| Japan | AI Promotion Act | Fully in force Sep 1, 2025 (Cabinet Office) | Promotion and risk-mitigation framework rather than EU-style product bans[38] |
| Canada | AIDA (Bill C-27 Part 3) | Not enacted; bill died on prorogation Jan 2025 | No dedicated federal AI statute as of late 2026 summaries; privacy and voluntary codes continue[39] |
| Brazil | PL 2338/2023 | Pending in Chamber after Senate approval Dec 2024 | Comprehensive AI bill not yet law as of 2026 tracker reports[40] |
| India | AI Governance Guidelines (Nov 2025); DPDP Act/Rules; MeitY advisories | Guidelines non-statutory; DPDP phased | No dedicated AI Act; risk-based guidance plus data-protection and platform advisories[41] |
| Singapore | Model AI Governance Framework / AI Verify | Voluntary tools in active use | Governance guidance and testing tooling rather than a single AI Act[42] |
| Australia | Existing law + voluntary guidance; 2026 policy announcements | No comprehensive AI statute as of mid-2026 parliamentary material | Government has discussed standards and possible later legislation[43] |
Multilateral efforts
- OECDAI Principles (2019, updated) remain a non-binding intergovernmental reference for trustworthy AI.[44]
- G7 HiroshimaHiroshima AI Process Code of Conduct (Oct 2023) is a voluntary code for advanced AI organizations, with OECD reporting support continuing into 2026.[45]
- Council of EuropeFramework Convention on AI, Human Rights, Democracy and the Rule of Law (2024) is the first global legally binding AI treaty open to non-members; entry into force depends on ratifications (EU ratified May 14, 2026 per reporting; treaty not yet in force as of mid-2026 summaries).[46]
- UNESCO / UNUNESCO’s Recommendation on the Ethics of AI (2021) is a global soft-law instrument. UN processes continue to discuss international AI governance without a single binding UN AI statute as of Nov 2026.
NCSL and national legislatures update databases frequently. Re-check primary bill text before drafting local policy.