AI Regulation Tracker
Every US state AI law, federal action, and EU AI Act milestone we track — updated daily from LegiScan, the Federal Register, the FTC, and EUR-Lex. Members also see per-regulation compliance notes and get deadline alerts.
51 of 51
| Regulation | Jurisdiction | Status | Impact | Effective | Source |
|---|---|---|---|---|---|
Illinois AI Safety Act Signed July 6, 2026, this frontier-model law targets developers with over $500M revenue training models above defined compute thresholds. It requires public transparency frameworks, pre-deployment reports, 72-hour safety-incident reporting, quarterly catastrophic-risk assessments to the AG, and — a first in the US — annual independent third-party safety audits, with penalties up to $3M per violation. frontier-modelstransparency | IL (state) | Enacted | Watch | 2027-01-01 | View |
EU AI Act Digital Omnibus The EU's simplification package amending the AI Act: it defers high-risk obligations from August 2, 2026 to December 2, 2027 for stand-alone (Annex III) systems and August 2, 2028 for product-embedded (Annex I) systems, extends SME simplifications to small mid-caps, and adds a new prohibition on AI generating CSAM and non-consensual intimate imagery. Parliament approved June 16, 2026 and the Council gave final approval June 29, 2026; publication in the Official Journal is imminent. generaltransparency | European Union | Passed legislature | Watch | — | View |
EU AI Act The world's first comprehensive AI law, applying to any company placing AI systems on the EU market or whose AI outputs are used in the EU — including US startups with EU users. It bans certain practices (social scoring, manipulative AI), imposes heavy obligations on 'high-risk' systems (hiring, credit, biometrics), transparency duties on chatbots and synthetic content, and a dedicated regime for general-purpose AI models. Penalties reach 7% of global turnover. generalfrontier-modelstransparencychatbotsdeepfakesbiometricsemployment | European Union | In effect | Action required | 2024-08-01 | View |
New York Kids Chatbot Safety Act Passed both chambers unanimously in June 2026 and now awaiting Governor Hochul's signature (she has until December 31, 2026). It would require age assurance and prohibit offering minors companion chatbots that personify themselves, simulate emotional relationships, engage in sexual content, promote self-harm, or use sycophancy to prolong engagement, with fines up to $25,000. chatbotsconsumer-protection | NY (state) | Passed legislature | Watch | — | View |
New York AI Employment Reporting Bill Passed the legislature in June 2026 and awaiting the Governor's decision. It would require businesses operating in New York with more than 50 employees (or publicly traded) to file annual reports with the Department of Labor by March 1 detailing how AI affected their workforce: jobs displaced, hours reduced, hires attributable to AI, positions left unfilled, plus oversight and data-protection practices. employmenttransparency | NY (state) | Passed legislature | Watch | — | View |
Great American AI Act A 269-page bipartisan discussion draft released June 4, 2026 by Reps. Obernolte (R-CA) and Trahan (D-MA). It would impose federal safety obligations on 'large frontier developers' ($500M+ revenue) and preempt state laws regulating AI development for three years — but explicitly not state laws about AI use or deployment. Not yet formally introduced. frontier-modelsgeneral | US Federal | Proposed | Watch | — | View |
FTC Cox Media Settlement May 2026 settlements totaling $930,000 (Cox Media Group $880K; MindSift and 1010 Digital Works $25K each) over marketing an 'Active Listening' AI ad service that supposedly targeted ads using conversations captured from smart devices. The service didn't use voice data at all — but the FTC also stated that even if it had worked, doing so without genuine consumer consent would violate Section 5. consumer-protectionbiometrics | US Federal | Enacted | Watch | — | View |
Colorado ADMT Act (revised Colorado AI Act) Repeals and replaces the Colorado AI Act with a narrower regime focused on 'covered automated decision-making technology' that processes personal data to materially influence consequential decisions (jobs, credit, housing, education, healthcare). It drops the original duty of care, impact assessments, and AG reporting, replacing them with consumer notice and appeal rights. Effective January 1, 2027. employmenttransparencyconsumer-protection | CO (state) | Enacted | Action required | 2027-01-01 | View |
Colorado AI Act (original) The first comprehensive US state AI law, targeting 'high-risk' AI systems used in consequential decisions (hiring, lending, housing, healthcare) with a duty of care against algorithmic discrimination, impact assessments, and risk management programs. It never took effect: its start date was pushed from February 1, 2026 to June 30, 2026, and it was then repealed and replaced by SB 26-189 before becoming operative. employmentconsumer-protectiontransparencygeneral | CO (state) | Superseded | Info | — | View |
Connecticut AI Act (pending) Senator Maroney's perennial comprehensive AI bill — covering high-risk AI in consequential decisions, transparency, and an AI regulatory sandbox — again failed to clear the House before Connecticut's 2026 session adjourned May 6, after passing the Senate in 2025. It remains the leading model bill for the multistate working group of AI legislators and is expected back in 2027. employmentconsumer-protectiontransparencygeneral | CT (state) | In committee | Watch | — | View |
Michigan Surveillance Pricing Bill Pending Michigan bill restricting 'surveillance pricing' — using AI and personal data to set individualized prices for consumers. Algorithmic/surveillance pricing is one of 2026's hottest state legislative themes, with similar bills in New York, California, and elsewhere. consumer-protection | MI (state) | In committee | Watch | — | View |
Michigan Kids Chatbot Bill Pending Michigan bill that would prohibit chatbot operators from offering products to minors unless the chatbot cannot encourage self-harm, suicidal ideation, violence, drug or alcohol use, or disordered eating. Part of the fast-moving 2026 wave of state kids-chatbot-safety bills. chatbotsconsumer-protection | MI (state) | Proposed | Watch | — | View |
Pennsylvania Companion Chatbot Bill Pending Pennsylvania bill requiring safety features in companion chatbot applications, particularly protocols for detecting and responding to suicidal ideation and self-harm. Modeled on the California SB 243 / New York Article 47 approach. chatbotsconsumer-protection | PA (state) | Proposed | Watch | — | View |
New Jersey AI Safety Test Bill Pending New Jersey bill directing the state Office of Information Technology to establish minimum requirements for an AI safety test applicable to AI technology sold, developed, deployed, or used in the state. Notably broad scope compared to peer bills that target only frontier developers. generalconsumer-protection | NJ (state) | Proposed | Watch | — | View |
Virginia AI IVO Law Signed April 13, 2026, making Virginia the first state to establish an Independent Verification Organization (IVO) framework — a voluntary, audit-based approach where AI systems can be certified by accredited third parties instead of prescriptive regulation. Positioned as a light-touch alternative to Colorado-style mandates. generaltransparency | VA (state) | Enacted | Info | 2026-07-01 | View |
Washington Provenance Law Signed in 2026 and effective February 1, 2027, this law requires covered generative-AI providers to embed provenance data in AI-created or materially altered video, image, and audio content so consumers can identify synthetic media. Washington joins California and Utah in mandating content provenance. transparencygenai-contentdeepfakes | WA (state) | Enacted | Action required | 2027-02-01 | View |
New York RAISE Act New York's frontier-model safety law, signed December 19, 2025 with a negotiated chapter amendment signed March 27, 2026. Large AI developers must publish safety and security protocols, avoid deploying models that create unreasonable risk of critical harm, and report safety incidents to the state within 72 hours. Takes effect January 1, 2027. frontier-modelstransparency | NY (state) | Enacted | Watch | 2027-01-01 | View |
Utah Provenance Law Enacted in Utah's 2026 session, this law addresses provenance of AI-generated content, requiring covered generative AI tools to support provenance data so consumers can tell when content is synthetic. Part of the 2026 wave of state provenance/watermarking laws alongside Washington's HB 1170 and California's SB 942. transparencygenai-contentdeepfakes | UT (state) | Enacted | Watch | 2026-05-06 | View |
Washington Companion Chatbot Law Signed March 24, 2026 and effective January 1, 2027. Companion-chatbot operators must conspicuously disclose the bot is not human (at session start, then every three hours for adults and every hour for minors), block conversations on suicide, self-harm, and eating disorders with referral to professionals, and face liability including a private right of action under Washington's Consumer Protection Act. chatbotsconsumer-protection | WA (state) | Enacted | Action required | 2027-01-01 | View |
Pennsylvania AI Disclosure Bill Pennsylvania Senate bill requiring disclosures and safeguards when businesses use AI in consumer interactions. It passed the Senate 49-1 on March 17, 2026 — unusually strong bipartisan support — and is now before the House Communications & Technology Committee. transparencychatbotsconsumer-protection | PA (state) | In committee | Watch | — | View |
AI Preemption Executive Order Signed December 11, 2025, this executive order attacks the state AI-law 'patchwork': it creates a DOJ AI Litigation Task Force to sue states over AI laws (Colorado's is named), directs Commerce to identify 'onerous' state laws and condition BEAD broadband funds on them, orders FTC/FCC preemption guidance, and calls for federal legislation preempting conflicting state AI laws (with carveouts for child safety, infrastructure, and state procurement). general | US Federal | In effect | Watch | 2025-12-11 | View |
CPPA ADMT Regulations Final CPPA rulemaking (approved by the Office of Administrative Law September 23, 2025) regulating businesses' use of automated decisionmaking technology for 'significant decisions' about consumers, including hiring and financial decisions. Businesses must give pre-use notice, honor opt-outs, and respond to access requests about ADMT logic, with full ADMT compliance required by January 1, 2027. Also mandates privacy risk assessments (from 2026) and phased cybersecurity audits. employmentconsumer-protectiontransparency | CA (state) | In effect | Action required | 2026-01-01 | View |
California Companion Chatbot Law (SB 243) The first state law regulating AI companion chatbots. Operators must disclose that users are talking to AI, maintain protocols to prevent and respond to suicidal-ideation and self-harm content, and give minors break reminders every three hours plus protections against sexual content. It carries a private right of action ($1,000 per violation or actual damages, plus attorney's fees). chatbotsconsumer-protectiontransparency | CA (state) | In effect | Action required | 2026-01-01 | View |
Texas TRAIGA Texas's AI law prohibits developing or deploying AI with intent to discriminate, manipulate behavior into self-harm or crime, or produce CSAM/unlawful deepfakes, and restricts government use of AI for social scoring and biometric identification. Unlike Colorado's approach, liability is intent-based rather than disparate-impact-based, and it includes a regulatory sandbox. Enforced exclusively by the Texas AG with a 60-day cure period. consumer-protectiontransparencybiometricsprocurement | TX (state) | In effect | Action required | 2026-01-01 | View |
California Training Data Transparency Act Requires developers of generative AI systems made available to Californians to post a public, high-level summary of the datasets used to train the system, including sources, whether data includes personal or copyrighted material, and collection timeframes. It applies to any GenAI system released on or after January 1, 2022, with no small-business exemption. transparency | CA (state) | In effect | Action required | 2026-01-01 | View |
California Frontier AI Act (SB 53) Requires 'large frontier developers' (over $500M annual revenue, training models above 10^26 FLOPs) to publish safety frameworks, report critical safety incidents to California's Office of Emergency Services, and protect whistleblowers. It is the first US frontier-model safety law. Most startups fall below its thresholds but should watch it as the template other states are copying. frontier-modelstransparency | CA (state) | In effect | Watch | 2026-01-01 | View |
Illinois AI Employment Law Amends the Illinois Human Rights Act to make it a civil rights violation for employers to use AI that discriminates against protected classes in recruitment, hiring, promotion, discipline, or discharge, or to use ZIP codes as a proxy for protected classes. Employers must also notify workers when AI is used for these decisions. employment | IL (state) | In effect | Action required | 2026-01-01 | View |
New York AI Companion Law Enacted in New York's FY2026 budget and effective November 5, 2025, this law requires operators of 'AI companions' to detect and respond to expressions of suicidal ideation, self-harm, or harm to others (referring users to crisis resources), and to tell users — at session start and every three hours — that the companion is not human. AG enforcement with penalties up to $15,000 per day. chatbotsconsumer-protection | NY (state) | In effect | Action required | 2025-11-05 | View |
California No Robo Bosses Act Would have required California employers to notify workers before using automated decision systems for employment decisions and barred relying solely on ADS for discipline or termination. Governor Newsom vetoed it in October 2025 as overly broad, but a revised version is expected, and the CPPA ADMT and Civil Rights Council rules already cover much of the same ground. employment | CA (state) | Vetoed | Watch | — | View |
California AB 853 Amends SB 942: delays its operative date to August 2, 2026, and extends provenance obligations beyond GenAI developers. Large online platforms (2M+ monthly users) must detect and surface provenance metadata in uploaded content starting January 1, 2027, and capture-device makers (cameras, phones) must offer provenance embedding starting January 1, 2028. transparencygenai-contentdeepfakes | CA (state) | Enacted | Watch | 2027-01-01 | View |
California AI Transparency Act (SB 942) Requires 'covered providers' of generative AI systems with over 1 million monthly users in California to embed latent (machine-readable) provenance disclosures in AI-generated images, video, and audio, offer users an optional visible AI disclosure, and provide a free public AI-detection tool. Originally operative January 1, 2026, AB 853 delayed it to August 2, 2026, aligning with the EU AI Act's Article 50 timeline. transparencygenai-contentdeepfakes | CA (state) | Delayed | Action required | 2026-08-02 | View |
California Employment ADS Regulations FEHA regulations clarifying that using automated-decision systems (including AI resume screeners, video-interview analysis, and personality tests) can constitute unlawful employment discrimination if they disparately impact protected groups. Employers and their agents (including vendors) are liable, and anti-bias testing is relevant evidence of good faith. employment | CA (state) | In effect | Action required | 2025-10-01 | View |
Montana Right to Compute Act The first 'right to compute' law: it restricts Montana state and local government from limiting private ownership or use of computational resources (including AI development) unless restrictions are narrowly tailored to a compelling interest. It also requires risk-management policies for AI controlling critical infrastructure. general | MT (state) | In effect | Info | 2025-10-01 | View |
California AB 1018 A broad Colorado-style bill imposing impact assessments, disclosures, and human-review rights on developers and deployers of automated decision systems used in consequential decisions (employment, housing, credit, healthcare). It stalled in September 2025 (moved to the Senate inactive file) but remains a live two-year bill in the 2026 session and is the most consequential pending California AI bill for SMBs. employmentconsumer-protectiontransparency | CA (state) | In committee | Watch | — | View |
FTC Companion Chatbot Study In September 2025 the FTC issued 6(b) orders to Alphabet, Character.AI, Instagram, Meta, OpenAI, Snap, and xAI, compelling detailed information on how companion chatbots are tested, monetized, and safeguarded for children and teens, including COPPA compliance. A 6(b) study isn't enforcement, but it typically previews the FTC's next enforcement priorities and possible rulemaking. chatbotsconsumer-protection | US Federal | Enacted | Watch | — | View |
Colorado AI Act Delay Bill Passed in an August 2025 special session, this bill delayed the effective date of the Colorado AI Act (SB 24-205) from February 1, 2026 to June 30, 2026, buying the legislature time to rework the law. That rework became SB 26-189, which replaced the original act entirely. general | CO (state) | Enacted | Info | 2025-08-28 | View |
Illinois AI Therapy Ban Prohibits offering AI-powered therapy or psychotherapy services to Illinois residents unless conducted by a licensed professional; licensed clinicians may use AI only for administrative or supplementary support, not therapeutic communication or treatment decisions. Enforced by the Illinois Department of Financial and Professional Regulation with penalties up to $10,000 per violation. healthcarechatbots | IL (state) | In effect | Action required | 2025-08-01 | View |
America's AI Action Plan The administration's July 23, 2025 AI policy blueprint: 90+ actions across accelerating innovation, building AI infrastructure, and international AI diplomacy. For businesses it signaled deregulation, federal pushback on state AI laws (later operationalized in the December 2025 EO), procurement preferences for 'unbiased' LLMs, and streamlined permitting for data centers. generalprocurement | US Federal | In effect | Info | 2025-07-23 | View |
Utah Mental Health Chatbot Law Regulates AI chatbots that provide mental-health support to Utah users: operators must clearly disclose the bot is AI (before first use and on request), may not advertise products within a session without disclosure, and may not sell or share users' individually identifiable health information. Creates an affirmative defense for operators who follow best practices and file policies with the state. healthcarechatbotsconsumer-protection | UT (state) | In effect | Action required | 2025-05-07 | View |
Utah AI Policy Act The first state GenAI consumer-protection statute: businesses can't blame the AI when generative AI output violates consumer-protection law, and must disclose AI use when asked (proactively, for regulated occupations like health and finance). 2025 amendments (SB 226, SB 332) narrowed the disclosure duty to high-risk interactions and extended the act's sunset, and HB 452 added specific rules for mental-health chatbots. consumer-protectiontransparencychatbots | UT (state) | In effect | Action required | 2024-05-01 | View |
FTC Workado Settlement April 2025 FTC settlement with Workado over marketing its AI content detector as '98% accurate' when independent testing on general-purpose content showed accuracy near 53%. The order requires competent and reliable evidence for accuracy claims and notice to customers. consumer-protectiongenai-content | US Federal | Enacted | Info | — | View |
New Jersey Deepfake Law Signed April 2, 2025 and effective immediately, this law makes producing or distributing deceptive deepfake media for unlawful purposes a third-degree crime and creates civil liability for victims. It covers non-consensual intimate imagery, fraud, and election-related deception using AI-generated media. deepfakesgenai-content | NJ (state) | In effect | Watch | 2025-04-02 | View |
FTC Cleo AI Settlement March 2025 FTC action against AI-branded fintech Cleo AI, which paid $17 million over allegations it deceived consumers about the amount and speed of 'instant' cash advances and made subscription cancellation difficult. Shows AI branding doesn't change core FTC rules on money claims and dark patterns. consumer-protection | US Federal | Enacted | Info | — | View |
Virginia High-Risk AI Act (vetoed) A Colorado-style high-risk AI bill passed by Virginia's legislature in February 2025 and vetoed by Governor Youngkin on March 24, 2025, who cited burdens on startups and small business. Most 2026 Virginia AI bills covering similar ground were tabled to 2027, so Virginia currently has no comprehensive AI statute. employmentconsumer-protectiongeneral | VA (state) | Vetoed | Info | — | View |
FTC accessiBe Settlement January 2025 FTC order requiring accessiBe to pay $1 million over claims its AI widget could make any website fully WCAG-compliant, which the FTC said was false, along with undisclosed paid reviews. A warning shot for 'AI fixes compliance' product marketing. consumer-protection | US Federal | Enacted | Info | — | View |
California Health AI Disclosure Law Requires California health facilities, clinics, and physician practices that use generative AI to produce patient-facing clinical communications to include a disclaimer that the message was AI-generated and instructions for reaching a human provider. Communications reviewed by a licensed provider are exempt. healthcaretransparencychatbots | CA (state) | In effect | Action required | 2025-01-01 | View |
FTC Evolv Settlement November 2024 FTC settlement with Evolv Technologies over claims that its AI-powered security scanners could reliably detect weapons in schools; the FTC alleged the marketing overstated real-world accuracy. The order restricts unsupported accuracy claims and lets certain school customers cancel contracts. consumer-protection | US Federal | Enacted | Info | — | View |
Operation AI Comply The FTC's September 2024 enforcement sweep against companies using AI hype to deceive consumers, including DoNotPay (the 'robot lawyer' that couldn't substantiate lawyer-equivalence claims, $193K settlement), Rytr (an AI review generator shut down for enabling fake reviews), and three 'AI-powered' business-opportunity schemes. The sweep established that existing FTC Act authority fully applies to AI claims, and follow-on cases have continued into 2026. consumer-protection | US Federal | Enacted | Watch | — | View |
Tennessee ELVIS Act The first state law protecting voice against AI cloning: it adds 'voice' to Tennessee's right-of-publicity statute and creates liability for distributing AI tools whose primary purpose is producing unauthorized replicas of a person's voice or likeness. Aimed at music-industry deepfakes but applies to any commercial voice cloning. deepfakesgenai-content | TN (state) | In effect | Watch | 2024-07-01 | View |
FTC Impersonation Rule Final FTC rule (effective April 1, 2024) prohibiting impersonation of government agencies and businesses, giving the FTC power to seek civil penalties and consumer redress directly — a key tool against AI-driven impersonation scams. A proposed extension covering impersonation of individuals (deepfaked people) and means-and-instrumentalities liability for AI tool providers was noticed but has not been finalized. deepfakesconsumer-protection | US Federal | In effect | Watch | 2024-04-01 | View |
NYC AEDT Law New York City's pioneering law on AI hiring tools: employers and employment agencies using automated employment decision tools to screen NYC candidates must obtain an independent bias audit annually, publish a summary of audit results, and notify candidates at least 10 business days before use. Enforcement began July 5, 2023. employmenttransparency | NY (state) | In effect | Action required | 2023-07-05 | View |
Upcoming deadlines
2026-08-02
EU AI Act: Article 50 transparency duties and GPAI enforcement begin
Chatbots must disclose they are AI, synthetic media must be marked machine-readably, and deepfakes must be labeled for EU users. The Commission's enforcement powers for GPAI obligations and member-state penalty regimes also kick in.
2026-08-02
California AI Transparency Act operative (as delayed)
Covered GenAI providers (1M+ monthly CA users) must embed latent provenance disclosures in AI-generated media, offer visible disclosures, and provide a free public AI detection tool. Delayed from January 1, 2026 by AB 853.
2027-01-01
California large online platform provenance duties begin
Online platforms with 2M+ monthly users must detect and surface provenance metadata in content distributed on their services.
2027-01-01
CCPA ADMT full compliance deadline
CCPA-covered businesses using automated decisionmaking technology for significant decisions must have pre-use notices, opt-out mechanisms, and access-request handling in place; CPPA enforcement expected after this date.
2027-01-01
Colorado ADMT Act effective
Colorado's revised AI law takes effect: deployers of covered automated decision-making technology must provide pre-use notices, 30-day adverse-decision notices, and honor data access/correction and human-review requests.
2027-01-01
New York RAISE Act effective
Large frontier AI developers must publish safety and security protocols and report safety incidents to New York within 72 hours, as finalized by the March 2026 chapter amendment.
2027-01-01
Illinois AI Safety Measures Act effective
Covered frontier developers ($500M+ revenue) must publish transparency frameworks, file pre-deployment reports, report safety incidents within 72 hours, and undergo annual independent third-party audits.
2027-01-01
Washington companion chatbot law effective
Companion chatbot operators must implement not-a-human disclosures (every three hours for adults, hourly for minors), crisis topic-blocking with referrals, and face a private right of action under Washington's CPA.
2027-02-01
Washington AI content provenance law effective
Covered GenAI providers must include provenance data in AI-created or materially altered video, image, and audio content available to Washington users.
2027-08-02
EU AI Act: compliance deadline for pre-existing GPAI models
General-purpose AI models placed on the EU market before August 2, 2025 must be brought into full compliance with GPAI obligations by this date.
Latest developments
Proposed Information Collection; ATUS Artificial Intelligence (AI) Questions
The Department of Labor is adding AI-related questions to its American Time Use Survey to collect data on how Americans interact with AI tools; SMBs should stay aware of emerging labor/productivity data that may inform future AI workplace regulations.
Illinois enacts AI Safety Measures Act (SB 315)
Governor Pritzker signed SB 315 on July 6, 2026 — a frontier-model safety law requiring large developers ($500M+ revenue) to publish transparency frameworks, report incidents within 72 hours, and undergo annual independent third-party safety audits, a US first. Effective January 1, 2027.
EU Council gives final approval to Digital Omnibus AI amendments
The Council approved the Digital Omnibus on June 29, 2026, deferring EU AI Act high-risk obligations to December 2, 2027 (Annex III) and August 2, 2028 (Annex I). Article 50 transparency duties — chatbot disclosure and synthetic-media marking — still apply August 2, 2026.
New York legislature passes kids companion-chatbot ban unanimously
S9051B passed both chambers (137-0, 60-0) on June 13, 2026 and awaits Governor Hochul's signature. It would require age assurance and prohibit companion chatbots for minors that simulate relationships, engage in sexual content, or use engagement-prolonging sycophancy.
Don’t check this page — we’ll tell you.
Free weekly digest of what changed, every Monday morning.