We spend most of our time here on AI companions — the apps people actually open and talk to. But those apps don't exist in a legal vacuum, and 2025 and 2026 turned out to be the years lawmakers finally noticed. Companion chatbots, AI therapy bots, and deepfakes are now written into statute in a fast-growing list of states. So we pulled the numbers on where AI regulation actually stands right now. Every figure is tied to its primary source with an access date. No fluff, just data.
The short version
- In 2025, state lawmakers in all 50 US states introduced AI-related bills for the first time — 1,208 bills total, with 145 enacted into law, per MultiState.
- NCSL's parallel count puts 2025 at 1,134 bills introduced and 131 enacted, with 40 states passing at least one AI law. Both are correct; they use different definitions of "AI-related."
- As of March 2026, lawmakers in 45 states had already introduced 1,561 AI-related bills, surpassing the entire 2024 total before most sessions had finished.
- The only federal AI-specific statute enacted in 2025 was the Take It Down Act, signed in May, covering nonconsensual intimate images including AI-generated deepfakes.
- A 99-to-1 Senate vote stripped a proposed 10-year moratorium on state AI laws from the "One Big Beautiful Bill Act" in 2025.
- 97% of Americans agree AI safety and security should be subject to rules and regulations (Gallup / Special Competitive Studies Project, September 2025).
AI legislation statistics in the US tell a story of states moving faster than any federal body can keep up with. In 2023, fewer than 200 AI-related bills were introduced across all US state legislatures combined. Two years later, every single state had introduced at least one. Washington has produced executive orders, task forces, and policy frameworks, but as of June 2026, only one AI-specific law has been enacted at the federal level. This article compiles verified data from MultiState, NCSL, Congress.gov, and primary polling sources to give an accurate picture of where AI regulation stands right now.
On this page
The State of AI Legislation in 2026
The clearest feature of AI lawmaking in 2025 and 2026 is the gap between state and federal action. States introduced thousands of bills, enacted hundreds of laws, and created binding compliance obligations that businesses now have to navigate. Congress debated dozens of AI bills without passing a comprehensive framework. The Trump administration pursued federal preemption through executive orders but failed to pass a moratorium through Congress. The result is a growing patchwork of state laws covering six main areas: deepfakes and synthetic media, AI in hiring and employment, healthcare AI, election integrity, children's safety, and transparency and disclosure.
Key AI Legislation Statistics You Should Know
- In 2023, fewer than 200 AI-related bills were introduced across all US state legislatures (MultiState, accessed June 2026).
- In 2024, 635 AI-related bills were introduced in 45 states — more than three times the 2023 volume — with 99 bills enacted (MultiState).
- In 2024, nonconsensual explicit deepfake laws were the single most widely enacted AI category, passing in 22 states (MultiState).
- Congress debated 107 AI-related bills in 2024 but enacted none as a comprehensive framework (MultiState).
- In 2025, 1,208 AI-related bills were introduced across all 50 states — the first year every state had at least one — with 145 bills enacted (MultiState).
- NCSL's parallel count for 2025: 1,134 bills introduced, 131 enacted, 40 states passed at least one AI law, an average of 2.6 AI laws per enacting state (NCSL data, via CCIA).
- The MultiState and NCSL figures differ because MultiState uses a broader definition of "AI-related" that includes task forces, budget items, and autonomous vehicle bills, while NCSL focuses on regulatory AI bills. Both are cited throughout this piece.
- As of March 2026, lawmakers in 45 states had already introduced 1,561 AI-related bills — already surpassing the entire 2024 total before most state sessions had finished (MultiState, March 2026).
- Only one AI-specific federal statute was enacted in all of 2025: the Take It Down Act (Sidley Data Matters, December 2025).
- The most active legislative areas in 2026 are generative AI regulation, algorithmic accountability, AI in hiring, and expanded deepfake protections (MultiState, March 2026).
State-by-State AI Bill Statistics
State AI lawmaking is not evenly distributed. A handful of states have enacted the most significant legislation, while others have stayed focused narrowly on deepfakes and fraud. Here's how the most active states break down.
Colorado
Colorado enacted SB24-205 in 2024 — the first comprehensive algorithmic discrimination law in the US. Its original February 2026 effective date was delayed to June 30, 2026, after a special session failed to agree on amendments. It remains the broadest US law covering high-risk AI systems used in consequential decisions.
California
California signed multiple AI bills in October 2025 covering AI safety, deepfake civil penalties, and children's online protections — the densest single batch of AI lawmaking in any state that year.
New York
New York's RAISE Act, signed in December 2025 with a chapter amendment signed March 27, 2026, applies to large frontier AI developers and requires safety plans, incident reporting within 72 hours, and related disclosures. It takes effect January 1, 2027.
Washington
Washington enacted three notable AI laws in 2026: SSB 5886 on AI-forged digital likenesses (effective mid-June 2026), HB 2225 regulating AI companion chatbots (effective January 1, 2027), and HB 1170 on AI content disclosure (effective February 2027). The companion-chatbot rules speak directly to questions we've raised about whether AI companions are safe.
Missouri
Missouri's SB 1019, which includes a prohibition on AI therapy chatbots, takes effect August 28, 2026 — among the most specific AI restrictions enacted by any state.
Other active states
- Oregon, Idaho, Iowa, Nebraska, Maine, and Georgia all enacted chatbot disclosure or safety laws in 2025 and 2026, one of the fastest-growing multi-state categories.
- Florida Governor DeSantis announced an "AI Bill of Rights" legislative package in late 2025 covering AI regulation and data centers.
- Of 44 states in the second year of a two-year session in 2026, 23 allow bills to carry over from 2025, so many bills that cleared committee last year are positioned to move quickly (MultiState, March 2026).
- 2025 state AI legislation shifted away from compliance mandates like impact assessments toward transparency measures — disclosure over documentation (Future of Privacy Forum, October 2025).
| State | Key enacted law | Effective date |
|---|---|---|
| Colorado | SB24-205 (algorithmic discrimination) | June 30, 2026 |
| New York | RAISE Act (frontier model safety) | January 1, 2027 |
| Washington | HB 2225 (AI companion chatbots) | January 1, 2027 |
| Missouri | SB 1019 (AI therapy chatbot ban) | August 28, 2026 |
| 22 states | Nonconsensual explicit deepfake laws | 2024 |
AI Legislation by Topic
Deepfakes and synthetic media are the most legislated AI categories at the state level. By the end of 2024, 22 states had enacted nonconsensual explicit deepfake laws, with more added through 2025 and 2026. The federal Take It Down Act, signed in May 2025, criminalizes nonconsensual distribution of intimate images, including AI-generated deepfakes, and imposes notice-and-removal obligations on platforms. At least half of US states had enacted legislation addressing generative AI and fabricated images by the end of 2025 (PBS NewsHour, citing NCSL). Many states have pivoted from outright bans to disclosure requirements because of First Amendment concerns emerging from early litigation. We tracked the broader scale of the problem in our deepfake statistics.
AI in hiring and employment is among the most active areas in 2026. Colorado's SB24-205 is the broadest US law in this space, covering high-risk AI systems used in employment decisions. New York City's Local Law 144, effective in 2023, requires bias audits of automated employment decision tools. High-risk AI approaches made up over 40% of all AI-related bills introduced in 2025 (Future of Privacy Forum, October 2025).
Healthcare AI saw targeted new activity. Missouri's prohibition on AI therapy chatbots is among the most specific restrictions enacted. Other states introduced transparency requirements for AI-assisted medical decisions and disclosure obligations for AI used in insurance underwriting. Four of the laws enacted in 2025 directly addressed healthcare AI (Future of Privacy Forum, October 2025).
Election integrity drew consistent bipartisan support. 87% of voters back disclosure requirements for generative AI in political ads, and 78% support banning deceptive AI-generated content targeting federal candidates (2024 YouGov poll, via Public Citizen). Montana enacted SB 25 specifically on deepfakes in elections in 2025.
Children's safety is one of the areas Trump's December 2025 executive order explicitly carved out from its preemption push, stating the national framework should not override state child safety laws (White House EO 14365). A bipartisan coalition of 42 state attorneys general wrote to major AI companies in December 2025, urging improved safeguards for children.
Transparency and disclosure — covering chatbot identification, synthetic content labeling, and automated decision notifications — represent the broadest and fastest-growing category. Oregon, Idaho, Iowa, Nebraska, Maine, and Georgia all enacted chatbot laws. It's the same disclosure question we run into across the AI chatbot statistics: how do you know when you're talking to a machine?
Federal AI Legislation Statistics
Federal AI lawmaking has trailed state activity significantly.
- Only one AI-specific federal statute was enacted in all of 2025: the Take It Down Act, signed in May, criminalizing nonconsensual distribution of intimate images and imposing notice-and-removal obligations on covered platforms.
- A proposed 10-year moratorium on state AI laws was stripped from the "One Big Beautiful Bill Act" by a 99-to-1 Senate vote in 2025.
- A similar preemption attempt through the National Defense Authorization Act also failed.
- On January 23, 2025, President Trump signed EO 14179, "Removing Barriers to American Leadership in Artificial Intelligence," revoking Biden's 2023 AI executive order.
- On December 11, 2025, Trump signed EO 14365 directing the DOJ to establish an AI Litigation Task Force to challenge state AI laws and linking BEAD broadband funding to AI policy alignment.
- On March 20, 2026, the White House released a National Policy Framework for Artificial Intelligence, urging Congress to replace state variation with a uniform federal standard — it is non-binding and creates no immediate compliance obligations.
- A bipartisan coalition of 36 state attorneys general wrote to Congressional leaders in November 2025, urging them to reject any federal moratorium on state AI laws.
- State AI laws remain operative unless and until Congress legislates. No binding federal AI framework exists as of June 2026.
Global Context
The EU AI Act (Regulation 2024/1689) is the world's first comprehensive AI legal framework. It entered into force on August 1, 2024, and phases in through 2027.
| EU AI Act milestone | Date | What it means |
|---|---|---|
| Prohibited practices enforceable | February 2, 2025 | Social scoring, biometric ID in public, and subliminal manipulation are banned |
| GPAI model obligations | August 2, 2025 | Transparency, copyright, and safety requirements for large model providers |
| High-risk AI systems | August 2, 2026 (Annex III delayed to Dec 2027) | Strict requirements for AI in hiring, healthcare, education, and law enforcement |
| Maximum fine | €35M or 7% of global turnover | Applies to the most serious violations |
- February 2, 2025: Prohibited AI practices became enforceable — social scoring, subliminal manipulation, real-time biometric identification in public, and exploitation of vulnerable groups. Maximum fines: €35 million or 7% of global annual turnover.
- August 2, 2025: General-purpose AI model obligations took effect, covering providers of models including GPT-4, Claude, and Gemini. The GPAI Code of Practice was published July 10, 2025.
- August 2, 2026: High-risk AI system obligations take effect for most categories. Under the EU Digital Omnibus provisional agreement of May 7, 2026, Annex III stand-alone high-risk obligations have been pushed to December 2, 2027.
- The Act applies to any organization developing or deploying AI systems affecting EU users, regardless of where that organization is headquartered.
- Italy enacted Law No. 132/2025 on October 10, 2025, establishing national-level fines up to €774,685 for AI violations.
- Brazil, Canada, the UK, and China all have active AI governance frameworks running in parallel with the EU Act.
Industry and Public Opinion Statistics
Support for AI regulation in the US is bipartisan, consistent across polling methodologies, and high. The only real variation is over who should regulate and how.
- 97% of Americans agree that AI safety and security should be subject to rules and regulations (Gallup / Special Competitive Studies Project, September 2025, nationally representative).
- 80% of US adults believe the government should maintain AI safety rules even if it slows AI development — supported by 88% of Democrats and 79% of Republicans and Independents (Gallup / SCSP, September 2025).
- 79% of Americans favor requiring AI programs to pass a government test before deployment, checking for regulatory violations, bias, or security vulnerabilities — supported by 84% of Republicans and 81% of Democrats (CISSM / University of Maryland, August 2025, n=1,202).
- 68% of Americans say the government should regulate AI to ensure public safety; 61% say the same for economic stability — both majorities hold across partisan groups (Reuters / Ipsos, August 2025, n=4,446).
- 71% of Americans want more regulation of the AI industry — a 15-point increase from a year earlier (YouGov, September 2024, via Public Citizen).
- Only 7% of Americans want no regulation of AI at all (AI Policy Institute, January 2025).
- By a greater than 2-to-1 margin, voters oppose federal preemption of state AI regulation (Echelon Insights / Common Sense Media, May 2025).
- Only 44% of US adults trust the US government a lot or somewhat to regulate AI effectively; 47% have little to no trust (Pew Research Center, June 2025, n=5,023).
- 59% of US adults have little or no confidence in US companies to develop and use AI responsibly (Pew Research Center, April 2025).
Recent Developments (2025 to 2026)
- January 23, 2025: Trump signs EO 14179, removing barriers to US AI leadership and revoking Biden's 2023 AI EO.
- February 2, 2025: EU AI Act prohibited practices become enforceable.
- May 2025: Federal Take It Down Act signed — the only AI-specific federal statute enacted in 2025.
- July 2025: Trump signs EO "Preventing Woke AI in the Federal Government," setting ideological neutrality requirements for government-purchased AI; the White House releases "Winning the AI Race: America's AI Action Plan."
- August 2, 2025: EU AI Act GPAI model obligations take effect.
- October 10, 2025: Italy enacts Law No. 132/2025, implementing national AI fines up to €774,685.
- December 2025: New York RAISE Act signed, takes effect January 1, 2027.
- December 11, 2025: Trump signs EO 14365, creating the AI Litigation Task Force and linking BEAD funding to state AI policy alignment.
- March 20, 2026: White House releases non-binding National Policy Framework for AI.
- May 7, 2026: EU Digital Omnibus provisional agreement delays the Annex III high-risk system deadline to December 2, 2027.
- June 2026: New York's law requiring disclosure of synthetic performers in advertising takes effect; Washington's SSB 5886, expanding personality-rights protections to AI-forged digital likenesses, takes effect.
Conclusion
As of June 2026, the US AI regulation landscape is a patchwork of enacted state laws, a handful of federal executive orders, one federal statute, and a preemption battle likely to produce years of litigation. States are moving faster than Congress. The public consistently backs regulation across partisan lines. The EU has a binding legal framework already partially in force. For any organization developing, deploying, or procuring AI systems — in the US or globally — compliance is not a future exercise. Multiple state laws have effective dates this year.
Data current as of June 2026. We update this piece quarterly; primary tracking sources include the MultiState AI Tracker, the NCSL AI Legislation Database, and the EU AI Act official page.