In 2025, US state lawmakers introduced 1,208 AI-related bills - the first year all 50 states filed at least one - and enacted 145 into law (MultiState AI Legislation Tracker, 2026). By March 2026, 45 states had already introduced 1,561 more, surpassing the entire 2024 volume in under three months (MultiState, 2026). The federal government, by contrast, had passed exactly one AI-specific statute, and the Senate voted 99-1 in July 2025 to kill a proposed moratorium on state AI rules (Inside Privacy, 2025). Across the Atlantic, the EU AI Act’s next compliance wave lands on August 2, 2026, even as 78% of enterprises report taking no meaningful steps to prepare (Vision Compliance, 2026). This analysis consolidates data from Stanford HAI, the IAPP, NCSL, MultiState, Ballotpedia, Pew Research, the European Commission, and 13 other primary sources to map where AI regulation actually stands in 2026.
TL;DR
- US states introduced 1,208 AI bills in 2025 and enacted 145; all 50 states filed at least one (MultiState, 2026).
- By March 2026, 45 states had introduced 1,561 AI bills - already above the full 2024 total (MultiState, 2026).
- US federal AI regulations hit 59 in 2024, up from 25 in 2023, issued by 42 agencies (Stanford HAI, AI Index 2025).
- State AI laws climbed from 1 in 2016 to 49 in 2023 to 131 in 2024 (Stanford HAI, AI Index 2025).
- The Senate voted 99-1 to strike a 10-year state AI moratorium from the 2025 budget bill (Inside Privacy, 2025).
- The EU AI Act’s top penalty is 35 million euros or 7% of global turnover - above GDPR’s ceiling (European Commission, 2026).
- 78% of enterprises have taken no meaningful steps toward EU AI Act compliance (Vision Compliance, 2026).
- 47 countries now have active AI-specific legislation (Stanford HAI, AI Index 2026).
- The OECD tracks 900-plus national AI policies across 70-plus jurisdictions (OECD.AI, 2026).
- 47 states have enacted deepfake legislation since 2019, with 82% of those laws passed in 2024-2025 (Ballotpedia, 2025).
- The FCC ruled AI-generated voices in robocalls “artificial” under the TCPA - illegal without prior consent (FCC, 2024).
- 55% of US adults want more control over AI; 62% distrust federal regulators (Pew Research Center, 2025).
1. The US State Legislative Surge
State legislatures, not Congress, are where AI law is being written in 2026. The number of AI laws passed by US states jumped from a single law in 2016 to 131 in 2024 (Stanford HAI, AI Index 2025), and 2025 sustained that pace regardless of which tracker you consult. The most-legislated areas are generative AI, algorithmic accountability, hiring decisions, and nonconsensual deepfakes.
| Metric | Value | Source |
|---|---|---|
| AI-related state bills introduced (2025) | 1,208 | MultiState, 2026 |
| State AI bills enacted (2025) | 145 | MultiState, 2026 |
| State AI bills enacted (2025, NCSL count) | 131 | NCSL, 2025 |
| States enacting at least one AI law (2025) | 40 | NCSL, 2025 |
| Average AI laws enacted per active state (2025) | 2.6 | NCSL, 2025 |
| State AI bills introduced by March 2026 (45 states) | 1,561 | MultiState, 2026 |
| State AI laws passed: 2016 / 2023 / 2024 | 1 / 49 / 131 | Stanford HAI, AI Index 2025 |
Source: MultiState AI Legislation Tracker and Stanford HAI AI Index 2025, Policy and Governance.
The MultiState (145 enacted) and NCSL (131 enacted) totals both count 2025 accurately - they diverge because each defines “AI-related” differently, so treat them as a range rather than a contradiction.
2. Federal Action and the Preemption Fight
Washington has produced regulation without producing a law. The Senate voted 99-1 to strip a proposed 10-year moratorium on state AI enforcement from the 2025 budget reconciliation bill (Inside Privacy, 2025), leaving states as the primary regulators - a status quo the executive branch has since tried to contest. Federal rulemaking activity is real but scattered across dozens of agencies rather than consolidated in one statute.
| Metric | Value | Source |
|---|---|---|
| US AI-specific federal laws enacted (as of mid-2026) | 1 | IAPP, 2026 |
| Federal AI regulations introduced (2024, from 42 agencies, up from 25 in 2023) | 59 | Stanford HAI, AI Index 2025 |
| Senate vote to strip AI moratorium (July 2025) | 99-1 | Inside Privacy, 2025 |
| Legislators opposing renewed preemption (June 2026) | 203 from 42 states | Senator Mike Moore, 2026 |
| AI-related congressional witnesses: 2017 vs 2025 | 5 vs 102 | Stanford HAI, AI Index 2026 |
| Industry share of AI congressional witnesses (2025) | 37% | Stanford HAI, AI Index 2026 |
Congress reconsidered preemption twice more after the July 2025 vote - through the NDAA and a June 2026 draft bill - and rejected it each time, while a January 2025 executive order took a deregulatory stance and a December 2025 order sought to restrict state-level AI rulemaking (Stanford HAI, AI Index 2026).
3. The EU AI Act: Deadlines and Penalties
The EU AI Act is the most consequential comprehensive framework in force, and its penalty ceiling is deliberately steep. Prohibited-practice violations can cost up to 35 million euros or 7% of global annual turnover - higher than GDPR’s 20 million euro / 4% top tier (European Commission, 2026). The phased timeline means obligations keep landing through 2028, so “the AI Act is live” is true and incomplete at the same time.
| Metric | Value | Source |
|---|---|---|
| Entered into force | August 2024 | European Commission, 2024 |
| Prohibited-practice ban / GPAI obligations applied | Feb 2025 / Aug 2025 | European Commission, 2025 |
| Remaining provisions apply / GPAI enforcement begins | Aug 2, 2026 | European Commission, 2026 |
| High-risk obligations deferred (Digital Omnibus) | Dec 2, 2027 / Aug 2, 2028 | European Commission, 2026 |
| Max fine, prohibited practices | EUR 35M / 7% turnover | European Commission, 2026 |
| Max fine, high-risk and transparency breaches | EUR 15M / 3% turnover | European Commission, 2026 |
| GPAI Code of Practice published | July 10, 2025 | European Commission, 2025 |
Source: European Commission, General-Purpose AI Code of Practice.
The voluntary GPAI Code of Practice - organized into transparency, copyright, and safety chapters - was signed by Amazon, Anthropic, Google, IBM, Microsoft, OpenAI, Mistral AI, and Aleph Alpha, while Meta declined (European Commission, 2025). The copyright chapter is a live flashpoint; see our AI copyright lawsuits roundup for the litigation behind it.
4. Global AI Governance Beyond the US and EU
AI lawmaking is now a worldwide phenomenon, but enforcement lags adoption almost everywhere. 47 countries now have active AI-specific legislation, though only a fraction have established enforcement mechanisms (Stanford HAI, AI Index 2026). Asia moved fast in 2025: South Korea, Japan, and China all enacted frameworks within months of one another.
| Metric | Value | Source |
|---|---|---|
| Countries with active AI-specific legislation | 47 | Stanford HAI, AI Index 2026 |
| Global legislative mentions of AI (2024, 75 countries, +21.3% YoY) | 1,889 | Stanford HAI, AI Index 2025 |
| Growth in AI legislative mentions since 2016 | 9x+ | Stanford HAI, AI Index 2025 |
| National AI policies tracked (OECD) | 900+ across 70+ jurisdictions | OECD.AI, 2026 |
| China AI content labeling rules effective | Sept 1, 2025 | Cyberspace Administration of China, 2025 |
| Data localization measures worldwide: 2017 vs 2025 | 5 vs 102 | Stanford HAI, AI Index 2026 |
| Median trust to regulate AI: EU / US / China | 53% / 37% / 27% | Pew Research Center, 2025 |
Source: OECD.AI Policy Observatory and Stanford HAI AI Index 2025.
South Korea finalized its AI Framework Act in January 2025, Japan enacted a light-touch AI Promotion Act in May 2025, and Italy became the first EU member state to pass a national AI law in September 2025. For the technology driving all this legislation, see the broader generative AI adoption picture.
5. Voice, Likeness, and Deepfake Laws
Synthetic media is the single most legislated AI category at the state level, and voice is central to it. By July 2025, 47 states had enacted deepfake legislation since 2019 - only Alaska, Missouri, and Ohio had none (Ballotpedia, 2025). Consent is the through-line: laws increasingly require permission to clone a real person’s voice or likeness, and disclosure when synthetic media is published.
| Metric | Value | Source |
|---|---|---|
| States with deepfake laws enacted since 2019 | 47 | Ballotpedia, 2025 |
| Share of those laws passed in 2024-2025 | 82% | Ballotpedia, 2025 |
| Deepfake laws enacted (year to date, July 10, 2025) | 64 | Ballotpedia, 2025 |
| States covering sexually explicit deepfakes (2025, from 32) | 45 | Ballotpedia, 2025 |
| States covering political deepfakes (2025, from 21) | 28 | Ballotpedia, 2025 |
| First state voice-specific AI law (ELVIS Act, Tennessee) | March 2024 | Tennessee ELVIS Act, 2024 |
| FCC ruling: AI robocall voices “artificial” under TCPA | Feb 8, 2024 | FCC, 2024 |
| States with deepfake regulations by 2024 | 24 | Stanford HAI, AI Index 2025 |
At the federal level, the NO FAKES Act - which would create a federal right against unauthorized AI replicas of a person’s voice or likeness - advanced out of the Senate Judiciary Committee in June 2026 but has not passed either chamber (Holland and Knight, 2026). For deeper coverage, see our deepfakes in elections roundup and the voice-cloning legal cases recap.
6. Corporate Readiness and the Compliance Gap
The gap between regulation on the books and readiness inside companies is the defining tension of 2026. 78% of enterprises have taken no meaningful steps toward EU AI Act compliance (Vision Compliance, April 2026), even with the August 2026 deadline in view. Governance roles and policies are spreading, but the inventory and documentation regulators will actually ask for often does not exist yet.
| Metric | Value | Source |
|---|---|---|
| Enterprises unprepared for EU AI Act | 78% | Vision Compliance, 2026 |
| Organizations with no formal AI inventory | 83% | Vision Compliance, 2026 |
| Organizations with no designated AI compliance owner | 74% | Vision Compliance, 2026 |
| Organizations with no high-risk documentation process | 61% | Vision Compliance, 2026 |
| Firms citing GDPR as a top regulatory influence (from 65%) | 60% | Stanford HAI, AI Index 2026 |
| Firms citing ISO/IEC 42001 as an influence | 36% | Stanford HAI, AI Index 2026 |
| Firms with no responsible-AI policy (from 24%) | 11% | Stanford HAI, AI Index 2026 |
| Companies with mature governance for AI agents | ~1 in 5 | Deloitte, State of AI 2026 |
Maturity is inching up, not surging: McKinsey’s Global AI Trust Maturity Survey put the average responsible-AI score at 2.3 in 2026, up from 2.0 in 2025, with only about one-third of organizations reporting a maturity level of three or higher (McKinsey, 2026).
7. Public Opinion, Incidents, and the Governance Market
Voters want more oversight, documented harms are climbing, and a compliance-tooling industry is forming around the rules. 55% of US adults say they want more control over how AI is used in their lives, and 62% have little or no confidence in federal agencies to regulate AI effectively (Pew Research Center, 2025). The demand for regulation is bipartisan, and it is running ahead of the enforcement capacity that exists today.
| Metric | Value | Source |
|---|---|---|
| US adults wanting more control over AI | 55% | Pew Research Center, 2025 |
| US adults with little/no confidence in federal AI regulation | 62% | Pew Research Center, 2025 |
| US adults more worried government will not regulate enough | ~60% | Pew Research Center, 2025 |
| Global median more concerned than excited about AI | 34% | Pew Research Center, 2025 |
| US adults rating AI societal risks as high | 57% | Pew Research Center, 2025 |
| AI incidents recorded (2025, up from 233 in 2024) | 362 | AI Incident Database / Stanford HAI, 2026 |
| AI governance market: 2024 vs 2030 (forecast) | $0.75B vs $5.64B | Wissen Research, 2025 |
| AI governance market CAGR (2025-2030) | 40% | Wissen Research, 2025 |
Source: Pew Research Center, How the US Public and AI Experts View AI.
Market-sizing estimates diverge sharply and should be read as directional: MarketsandMarkets pegs the AI governance market at $0.89B in 2024 rising to $5.78B by 2029 (45.3% CAGR), a different base and endpoint than the Wissen Research figures above.
Summary: AI Regulation by the Numbers
| Metric | Value | Source |
|---|---|---|
| US state AI bills introduced (2025) | 1,208 | MultiState, 2026 |
| US state AI bills introduced by March 2026 | 1,561 | MultiState, 2026 |
| State AI laws passed (2024, vs 1 in 2016) | 131 | Stanford HAI, AI Index 2025 |
| US federal AI regulations (2024, vs 25 in 2023) | 59 | Stanford HAI, AI Index 2025 |
| US AI-specific federal laws enacted (mid-2026) | 1 | IAPP, 2026 |
| Senate vote to strip AI moratorium (2025) | 99-1 | Inside Privacy, 2025 |
| EU AI Act max fine (prohibited practices) | EUR 35M / 7% | European Commission, 2026 |
| Enterprises unprepared for EU AI Act | 78% | Vision Compliance, 2026 |
| Organizations with no formal AI inventory | 83% | Vision Compliance, 2026 |
| Countries with active AI-specific legislation | 47 | Stanford HAI, AI Index 2026 |
| China AI content labeling rules effective | Sept 1, 2025 | Cyberspace Administration of China, 2025 |
| States with deepfake laws since 2019 | 47 | Ballotpedia, 2025 |
| Share of state deepfake laws from 2024-2025 | 82% | Ballotpedia, 2025 |
| Deepfake laws enacted (YTD July 2025) | 64 | Ballotpedia, 2025 |
| First state voice-specific AI law (ELVIS Act) | March 2024 | Tennessee ELVIS Act, 2024 |
| US adults wanting more control over AI | 55% | Pew Research Center, 2025 |
| US adults distrusting federal AI regulation | 62% | Pew Research Center, 2025 |
| AI incidents recorded (2025) | 362 | AI Incident Database / Stanford HAI, 2026 |
| AI governance market (2030 forecast) | $5.64B | Wissen Research, 2025 |
Methodology and Sources
Data was gathered by aggregating figures from primary reports, government publications, named legislative trackers, and public survey datasets, prioritizing 2025-2026 releases and tracing trade-press mentions back to their original source. Every statistic above was verified against the cited report during research.
- Stanford HAI - Artificial Intelligence Index Report 2025, Chapter 6: Policy and Governance (link)
- Stanford HAI - Artificial Intelligence Index Report 2026, Chapter 8: Policy and Governance (link)
- IAPP - Global AI Law and Policy Tracker and US State AI Governance Legislation Tracker, 2026 (link)
- NCSL - Summary of Artificial Intelligence 2025 Legislation (link)
- MultiState - Artificial Intelligence (AI) Legislation Tracker, 2026 (link)
- Ballotpedia - State Deepfake Legislation Tracker, July 2025 (link)
- European Commission - EU AI Act text, penalty structure, and General-Purpose AI Code of Practice, 2024-2026 (link)
- Vision Compliance - 2026 EU AI Act Readiness Report, April 2026 (link)
- Pew Research Center - How the US Public and AI Experts View AI (April 2025) and Views of AI Around the World (October 2025) (link)
- FCC - Declaratory Ruling on AI-Generated Voices under the TCPA, February 2024 (link)
- OECD.AI Policy Observatory - live repository of national AI policies, 2026 (link)
- McKinsey - Global AI Trust Maturity Survey and The State of AI, 2025-2026
- Deloitte - State of AI in the Enterprise, 2026
- Cyberspace Administration of China - Measures for Labeling AI-Generated Synthetic Content, 2025 (via White and Case and Loeb and Loeb regulatory trackers)
- Inside Privacy (Covington) and Senator Mike Moore newsroom - US federal AI moratorium coverage, 2025-2026
- Holland and Knight - NO FAKES Act legislative status, June 2026
- AI Incident Database (Responsible AI Collaborative) - reported AI incident counts, via Stanford HAI AI Index 2026
- Wissen Research and MarketsandMarkets - AI governance market sizing, 2025-2026
Data watch: The Stanford HAI AI Index refreshes annually each spring (a 2027 edition covering 2026 data is expected around April 2027); NCSL, MultiState, IAPP, and Ballotpedia update their legislative trackers continuously through active sessions; Pew Research runs its AI opinion surveys annually; and the European Commission continues issuing AI Act guidance ahead of the August 2, 2026 deadline. Figures here will shift as 2026 legislative sessions close.
Last updated: July 10, 2026. We review and update this page quarterly as new data is published.