AI laws in all 50 US states

Alabama has no comprehensive AI statute; its AI rules cover election deepfakes, AI-generated intimate and child sexual images, and, since October 1, 2026, AI use in health insurance coverage decisions. State-government AI use is shaped by a 2024 executive order and task force rather than by statute.

Alaska has not enacted an AI-specific statute; bills on deepfakes, synthetic media in elections and AI-generated child sexual abuse material did not pass the 34th Legislature (2025-2026). The main AI-specific guidance is the Alaska Bar Association's 2025 ethics opinion on generative AI.

Arizona regulates AI through narrow laws on election deepfakes and digital impersonation, AI-generated child sexual abuse material, and, since July 1, 2026, individual medical-director review of medical-necessity denials (no sole reliance on automated recommendations). In 2026 the governor vetoed a broader AI service disclosure bill (HB 2311) and a bill on state agency AI rules (HB 2592).

Arkansas has no comprehensive AI statute, but its 2025 session passed several targeted laws: public bodies must adopt AI use policies, ownership rules for generative AI output, AI likeness rights, and AI-generated child sexual abuse material. State-government AI work runs through a governor's working group and an AI & Analytics Center of Excellence set up in 2024.

California has no single comprehensive AI act, but it has the densest set of AI laws in the US: frontier-model safety reporting, training-data and content-provenance transparency, companion-chatbot safeguards, workplace automated-decision limits and health-care AI rules. State agencies add regulations, including the privacy agency's automated decisionmaking technology (ADMT) rules effective January 1, 2026 and Civil Rights Council employment rules on automated decision systems.

Colorado's 2024 AI Act (SB 24-205) never took effect: in May 2026 SB 26-189 repealed and re-enacted it as a narrower law on automated decision-making technology (ADMT) used in consequential decisions, built on notice, documentation, data-correction and human-review rights, with most duties starting January 1, 2027. Colorado also has laws on election deepfakes, AI intimate images and insurers' use of algorithms and external data.

After broad AI bills failed in 2024 and 2025, Connecticut enacted Public Act 26-15 (SB 5) in May 2026, an omnibus law covering companion chatbots, automated employment decision tools, AI provenance data, frontier-developer whistleblowers, AI-related layoff notices and state agency AI use, phased in from October 2026 to January 2028. Since 2023 state agencies have had to inventory and assess their AI systems under Public Act 23-16.

Delaware has no comprehensive AI statute; its AI laws cover election deepfakes and sexual deepfakes, plus a statutory AI Commission that advises on state AI use and inventories generative AI in state government. In 2025 the legislature directed that commission to build a regulatory sandbox framework for testing agentic AI.

Florida's AI-specific laws are narrow: generative AI disclaimers on political ads and duties on platforms to take down altered sexual depictions. Governor DeSantis's broader 'AI Bill of Rights' (SB 482, chatbots and minors, AI disclosure) passed the Senate in March 2026 but died when the House refused to take it up, including in the spring special session.

Georgia has no comprehensive AI statute; its 2026 session produced two targeted laws, one on conversational AI safety for minors (SB 540) and one barring health coverage denials based solely on AI (SB 444), both taking effect in 2027. Election deepfake and AI child sexual abuse material bills moved in 2025 but their final enactment could not be confirmed.

Hawaii has no comprehensive AI statute; in July 2026 it enacted an AI companion chatbot law (Act 248) and a civil remedy for harmful AI digital imitations (Act 247). Its 2024 election deepfake law (Act 191) was struck down by a federal court in January 2026, and statewide agency AI use guidelines were still in draft as of mid-2026.

Idaho has no comprehensive AI statute; it regulates specific harms, including election deepfakes (2024), AI-generated child sexual abuse material and explicit synthetic media (2024, tightened 2026), and conversational AI chatbots (2026, effective July 2027). State agency AI use is governed by an Office of Information Technology Services policy issued in August 2025.

Illinois regulates AI through several targeted statutes: AI in hiring (video interviews since 2020; discrimination and notice under the Human Rights Act since January 2026), AI therapy limits (2025), and digital replicas (2024). In July 2026 it enacted the Artificial Intelligence Safety Measures Act, which requires large frontier developers to publish safety frameworks, report incidents and undergo third-party audits.

Indiana has no comprehensive AI statute; its AI-specific laws cover election deepfakes and AI-generated intimate images (2024), an AI task force (2024), and, from July 2026, limits on insurers relying solely on AI to downcode or process health claims. State agencies follow an enterprise AI policy from the Office of the Chief Data Officer, based on the NIST AI Risk Management Framework.

Iowa has no comprehensive AI statute; in 2026 it enacted a conversational AI (chatbot) law with protections for minors, applying from July 2027, and limits on insurers using AI as the sole basis to deny prior authorization. Earlier, a 2024 law made posting sexually explicit altered (deepfake) images of a person a form of harassment; election deepfake bills have repeatedly failed.

Kansas has no comprehensive or consumer-facing AI statute; its AI-specific laws ban AI platforms controlled by a country of concern (such as DeepSeek) on state devices and networks, and extend child sexual exploitation crimes to AI-generated or morphed images. Bills on election deepfakes failed in 2024 and 2025.

Kentucky's main AI law, SB 4 (2025), sets governance rules for state agencies' AI use through the Commonwealth Office of Technology and an AI Governance Committee, and creates a civil cause of action over synthetic media in election communications. It has no comprehensive private-sector AI statute; 2026 bills on AI therapy chatbots did not pass.

Louisiana has no comprehensive AI statute; its AI-specific laws criminalize sexual deepfakes of minors and AI-generated images of others (2023, expanded 2026), require AI disclosures on deceptive candidate depictions in election communications (2026), and make attorneys exercise diligence on possibly AI-manipulated evidence (2025). Earlier election deepfake bills were vetoed in 2024, and a 2026 critical-infrastructure AI risk bill (SB 474) did not pass.

Maine has no comprehensive AI statute; it requires businesses to disclose AI chatbots that could be mistaken for humans (2025), criminalizes threatened dissemination of AI-generated private images (2025), and in 2026 added a synthetic media disclaimer for campaign ads and limits on AI in therapy. A 2026 bill on children's access to AI companion chatbots did not pass.

Maryland governs state agencies' AI through the Artificial Intelligence Governance Act of 2024, a 2024 executive order and a DoIT Responsible AI Policy, and regulates specific private uses: AI in health insurance utilization review, facial recognition in job interviews, and (from 2026) election and identity-fraud deepfakes. It has no comprehensive private-sector AI statute; a 2025 workgroup was due to recommend AI consumer protection steps by July 2026.

Massachusetts has no comprehensive AI statute; its main AI-specific law is a 2024 act that criminalizes distributing sexual images made or altered by computer, and the Attorney General has advised that existing consumer protection and anti-discrimination laws apply to AI. Two House-passed bills on election deepfakes and AI disclosure in political ads stalled in the Senate in 2026, leaving the state without an election deepfake rule.

Michigan has no comprehensive AI statute; its AI-specific laws require disclosure of AI in political ads, ban undisclosed deceptive election deepfakes within 90 days of an election, and create criminal and civil liability for sexual deepfakes. Bills on state-agency AI governance and other AI topics were introduced in 2026 but none had been enacted as of October 2026.

Minnesota has no comprehensive AI statute but has criminal election and sexual deepfake laws (2023), a 2026 ban on AI nudification tools, and a 2026 rule barring health insurers from denying care by automated processing alone. The nudification ban took effect Aug 1, 2026 but a federal appeals court paused its enforcement in early October 2026 while xAI's First Amendment challenge proceeds.

Mississippi's AI-specific laws criminalize deceptive election deepfakes (2024), add AI-altered ('morphed') images of minors to child exploitation law (2024), create an AI Regulation Task Force (2025) and define 'artificial intelligence' in statute (2026). 2026 bills on name, likeness and voice rights and on AI disclosure in political ads died, so there is still no broad AI regulation.

Missouri has no comprehensive or sector AI statute; its main AI-related law is a 2026 sexual-offenses act (effective Aug 28, 2026) that creates a felony for disclosing intimate digital depictions and covers computer-generated child sexual abuse material. Several 2026 bills on AI deepfakes, companion chatbots, AI in mental health and AI content labeling passed the House or committees but did not pass the Senate before the session ended.

Montana passed a cluster of AI laws in its 2025 session: limits and human-review duties for state and local government AI use, a Right to Compute Act that restricts government limits on computing and requires risk policies for AI-controlled critical infrastructure, and laws on election and sexual deepfakes. The Legislature does not meet in 2026, so no new AI statutes were added this year; the next session starts January 2027.

Nebraska has no comprehensive AI statute, but it bars AI algorithms from being the sole basis for health insurance prior-authorization denials (from Jan 1, 2026), covers AI-generated child sexual abuse material (2025), and has a Conversational AI Safety Act for chatbots that becomes operative July 1, 2027. The chatbot act, signed April 14, 2026, requires AI disclosure, protections for minors and self-harm crisis protocols, enforced only by the Attorney General.

Nevada has no comprehensive AI statute; its 2025 laws bar AI systems from being offered or marketed as mental or behavioral health care and limit AI use in schools and by therapists, require a 'manipulated' disclosure on election communications using synthetic media, and keep final emergency and utility shutoff decisions out of AI's hands. The Legislature meets in odd years, so no new regular-session AI laws were added in 2026.

New Hampshire has no comprehensive AI statute but enacted several targeted laws in 2024: limits on state agency AI use, a disclosure rule for AI and deepfakes in political ads, and a crime and civil action for fraudulent deepfakes. A 2025 law reportedly adds criminal and civil liability for chatbot operators whose bots push minors toward sexual content, drugs or self-harm.

New Jersey's main AI-specific statute is a 2025 law that makes creating or sharing deceptive AI-generated audio or video to further a crime a third-degree crime with fines up to $30,000 and a civil action for victims. The Attorney General and Division on Civil Rights treat AI hiring and other automated decision tools as covered by the Law Against Discrimination, while comprehensive and generative-AI bills remain pending.

New Mexico's main AI-specific statute is a 2024 Campaign Reporting Act amendment that requires a disclaimer on AI-manipulated election media and penalises deceptive deepfakes distributed within 90 days of an election. Broader bills on consequential AI decisions (HB 60 in 2025, HB 28 in 2026), government AI use (SB 68 in 2026) and sexual deepfakes (HB 22 in 2026) all died without passing.

New York has enacted the RAISE Act, a frontier-model safety and incident-reporting law for large AI developers that takes effect January 1, 2027, alongside in-force laws on AI companion safeguards, state-agency automated decision-making, synthetic performers in ads and deceptive election media. In June 2026 the legislature also passed bills on chatbot features for minors, personalised algorithmic pricing, training-data transparency and AI workforce reporting, which awaited the governor's action as of October 2026.

North Carolina has no comprehensive or sector AI statute; its main enacted AI-related law is a 2024 sex-crimes update that the legislature's trackers list as covering AI-generated sexual images, including of minors. Bills on AI in insurance utilization review (HB 565) and deepfakes (HB 934) remained in committee as of October 2026, while Governor Stein's Executive Order 24 set up AI governance for state agencies.

North Dakota has no comprehensive AI statute; in its 2025 session it enacted targeted laws requiring AI disclaimers on political content, covering computer-generated sexual images, limiting AI's role in health-insurance prior authorization and extending stalking law to robots. The legislature meets only in odd years, so no new AI laws were made in 2026; state agencies follow an NDIT AI policy first issued in 2023.

Ohio's only enacted AI-specific requirement is in the FY2026-27 budget (HB 96), which required a state model policy on AI in schools and local district AI policies by July 1, 2026. Bills on AI-generated sexual images and replica identity fraud (SB 163, passed the Senate), election deepfakes (HB 362), AI personhood (HB 469) and a 'right to compute' (HB 392) were still pending in the 136th General Assembly as of October 2026.

Oklahoma has no comprehensive AI statute; its enacted AI-specific laws are a 2025 criminal law on AI-generated sexual images and the 2026 Oklahoma Responsible Technology in Schools Act on AI use in public schools. In 2026 bills on AI chatbots and minors, AI personhood, state-agency AI disclosure and election deepfakes passed the House or a committee but did not reach the governor.

Oregon has no comprehensive AI statute but has passed a series of targeted laws: AI disclosure in campaign ads (2024), fake intimate images, a ban on certain vendors' AI on state IT and on AI using nursing titles (2025), and an AI companion chatbot safety law with a private right of action (2026, effective January 1, 2027). The executive branch follows recommendations from a State Government AI Advisory Council created by Executive Order 23-26.

Pennsylvania has no comprehensive or sector AI statute; its enacted AI-related laws are criminal: Act 125 of 2024 on AI-generated child sexual abuse material and intimate images, and Act 35 of 2025 creating the offence of digital forgery for deepfakes. Executive agencies operate under Governor Shapiro's 2023 executive order on generative AI, and no AI-specific act appears among 2026 enactments.

Rhode Island has no comprehensive AI statute but has enacted targeted laws: a 2025 ban on deceptive synthetic media in election communications within 90 days of an election, and three 2026 laws on AI companion safety, AI in mental health care and notice of AI-assisted clinical documentation. Broader 2026 bills on insurers' use of AI, automated employment decisions, a state-agency AI inventory and AI watermarks did not pass both chambers.

South Carolina's AI-specific laws are criminal statutes enacted in 2025 covering computer-generated child sexual abuse material, morphed images of identifiable minors and machine-learning 'digitally forged' intimate images, plus a 2024 real-estate law making licensees responsible for AI-assisted work product. Chatbot bills for minors (S 896, H 5476/S 1037) did not pass before the 2025-26 session ended in 2026.

South Dakota's AI-specific laws are narrow criminal statutes: computer-generated child sexual abuse material (2024), unlabeled election deepfakes (2025) and nonconsensual sexual deepfakes (2026). Broader 2026 bills on chatbots for minors, chatbot disclosure, health-insurer AI, AI in therapy and an AI study task force did not pass.

Tennessee regulates specific AI harms: the 2024 ELVIS Act protects voice and likeness from AI cloning, a 2025 act targets nonconsensual intimate deepfakes, and a 2026 act bars marketing AI as a mental health professional. In 2026 broader companion-chatbot bills were amended into studies by the Tennessee Advisory Commission on Intergovernmental Relations (TACIR), with one report due January 31, 2027.

The Texas Responsible Artificial Intelligence Governance Act (HB 149) took effect January 1, 2026: it bans a short list of intentional harmful AI uses, sets disclosure and other duties for government AI, creates a regulatory sandbox, and is enforced only by the Attorney General. Separate 2025 laws cover AI in health records and clinical decisions, automated health-claim denials, state agency AI governance and AI-generated intimate images.

Utah's Artificial Intelligence Policy Act (2024, narrowed in 2025) makes businesses answerable under consumer protection law for their generative AI, requires AI disclosure on request or in high-risk interactions, and runs an Office of AI Policy with a regulatory learning lab; the Act is set to repeal July 1, 2027. Separate laws cover mental health chatbots, election deepfakes, AI-made intimate images and content provenance (from January 1, 2027), and AI-driven defamation and identity abuse.

Vermont has overseen AI in state government since 2022 through a Division of Artificial Intelligence, an automated decision system inventory and an AI Advisory Council, now extended to 2030. In 2026 it enacted an election synthetic-media disclosure law, a ban on AI-only mental health services offered to the public, and a neurological-rights act, while a broad bill on inherently dangerous AI systems (H.341) did not pass.

Virginia's legislature passed a Colorado-style high-risk AI act (HB 2094) in 2025, but Governor Youngkin vetoed it; enacted AI law is limited to human final decisions on criminal-justice AI tools, AI guidance for schools, a study of independent AI verification bodies and coverage of fabricated intimate images. In September 2026 Governor Spanberger's Executive Order 22 created a rapid-response AI Task Force and rescinded Executive Order 30, while 2026 chatbot, health-insurer AI and election synthetic-media bills were carried over to 2027.

Washington has layered targeted AI laws: election synthetic-media disclosure (2023), fabricated intimate images (2024), forged digital likenesses (2025, expanded 2026), and two 2026 statutes on companion chatbots (from January 1, 2027) and AI content provenance plus government AI disclosure (from February 1, 2027). A legislative AI Task Force run by the Attorney General advises on further legislation, and a 2024 executive order governs state agencies' generative AI use.

West Virginia's enacted AI laws are a state Task Force on Artificial Intelligence (2024, amended 2025) and a 2025 ban on AI-generated child sexual abuse material. Bills on election deepfakes, nonconsensual intimate deepfakes, AI video labels and AI in mental health care were introduced in 2024-2026 but did not pass.

Wisconsin's AI-specific statutes, both from 2024, require disclosure of AI-generated content in political advertising and criminalize possession of virtual (including computer-generated) child sexual abuse material. No AI-specific act was found among the 2025-26 session's enacted laws; executive action has been limited to a governor's task force on AI and the workforce.

Wyoming's only AI-specific statute is 2026's HB 102, which creates crimes for AI-generated sexual material involving children and for using AI to promote self-harm or censor political speech, says AI use is no defense to a crime, and gives AI developers specified immunity. Earlier bills on election synthetic media (2024) and government AI social scoring (2026) died.