There is no single US AI law — there is federal law applied to AI, state AI laws with different triggers, state privacy laws with profiling rights, and sector regulators, and most of it is already in force. Pick where you have users, customers, employees, or job applicants and your industry, answer each law's own screening questions, and get a deterministic answer per law: in scope or not, which obligations apply to your role, each deadline, the statutory citation, and who enforces it. Nothing you enter leaves your browser unless you ask for the emailed report.
The US AI laws this check screens
Texas — Responsible AI Governance Act (TRAIGA, HB 149)
In force since 1 January 2026. Reaches anyone who advertises or does business in Texas, produces a product used by Texans, or develops or deploys AI in Texas. Its prohibitions are intent-based (manipulation toward self-harm or crime, unlawful discrimination, sexual deepfakes of minors, government social scoring), government agencies must disclose AI interaction, and health-care providers must disclose AI use in care. The Attorney General enforces exclusively, with a 60-day cure period, civil penalties of $10,000–$12,000 per curable violation and $80,000–$200,000 per uncurable violation, and no private right of action. The AG's complaint portal has been live since 1 September 2026.
California — seven AI laws with separate gates
AB 2013 training-data disclosure (since 1 January 2026, no size threshold) for anyone who develops or fine-tunes generative AI offered to Californians; the AI Transparency Act (SB 942, since 2 August 2026) for generative image, audio, or video systems with more than 1,000,000 monthly users, at $5,000 per violation; the FEHA automated-decision-system regulations (since 1 October 2025) for employers with five or more employees; the CCPA automated-decisionmaking rules (compliance from 1 January 2027) for businesses over the CCPA thresholds; SB 243 companion-chatbot duties (since 1 January 2026) with a private right of action for the greater of actual damages or $1,000 per violation; the B.O.T. Act bot-disclosure rule; and SB 53 frontier-model duties above 10^26 operations. Roughly 30 more AI bills sit on the Governor's desk until 30 September 2026, including SB 1000, which would remove the 1,000,000-user threshold.
Colorado — Automated Decision-Making Technology law (SB 26-189)
Applies from 1 January 2027. It repealed and replaced the 2024 Colorado AI Act, which never took effect, and swaps the duty-of-care and impact-assessment regime for disclosure and record-keeping duties on developers and deployers of ADMT that materially influences consequential decisions about Colorado residents in employment, education, housing, lending, insurance, health care, and government services. The Attorney General enforces as a deceptive trade practice, up to $20,000 per violation, with a 60-day cure only where the AG deems a cure possible; there is no private right of action. The AG's implementing rules are proposed, with a hearing on 26 October 2026.
Connecticut — Public Act 26-15 (as amended by Public Act 26-100)
Phased: from 1 October 2026, using an AI hiring tool is no defense to a discrimination claim, large generative-AI providers must embed provenance data, and frontier developers owe whistleblower protections; from 1 January 2027, AI-companion operators owe disclosure, crisis-protocol, and minor-safeguard duties; from 1 October 2027, employers must give pre-decision notices for automated employment-decision technology. The Attorney General enforces under CUTPA with no private right of action.
Illinois — HB 3773 (AI in employment)
In force since 1 January 2026 as an amendment to the Illinois Human Rights Act. Employers with Illinois employees or applicants must notify workers when AI is used in hiring, promotion, discipline, or similar decisions, must not use AI that has a discriminatory effect, and may not use zip codes as a proxy for protected classes. There is no regulator fine: enforcement runs through a charge with the Department of Human Rights and the usual civil-rights remedies. The department's proposed notice rules were published in May 2026 and then postponed.
New York City — Local Law 144 (automated employment decision tools)
Enforced since 5 July 2023. Employers and employment agencies hiring or promoting for NYC-based roles may not use an automated employment decision tool unless it has had an independent bias audit within the past year, a summary is published, and NYC candidates get at least ten business days' notice. The Department of Consumer and Worker Protection enforces at up to $500 for a first violation and $500–$1,500 for each subsequent one, with every day of use and every missed notice counted separately.
Utah — Artificial Intelligence Policy Act
In force since 1 May 2024 and narrowed on 7 May 2025 (SB 226): any business using generative AI with Utah consumers must disclose that fact when clearly asked, and regulated professionals must disclose prominently in high-risk interactions such as health, financial, or legal advice. A separate 2025 chapter regulates mental-health chatbots. The Division of Consumer Protection enforces at up to $2,500 per violation, and a chatbot that identifies itself as AI throughout the interaction is shielded.
Multi-state chatbot, companion, and AI-therapy laws
Already in force: Maine's chatbot disclosure duty (since 24 September 2025), New York's AI-companion safeguards (since 5 November 2025), Hawaii's conversational-AI law (since 14 July 2026), and AI-therapy restrictions in Nevada, Illinois, Utah, Rhode Island, Tennessee, Vermont, Maine, Colorado, and Missouri. A 2027 wave follows: Colorado, Washington, Oregon, and Rhode Island from 1 January 2027, then Nebraska, Idaho, Georgia, and Iowa from 1 July 2027 — disclosure cadences, crisis protocols, minor safeguards, and, in Washington and Oregon, private rights of action.
Frequently asked questions
Is there a federal AI law that overrides state AI laws?
Not as of September 2026. No federal statute preempts state AI laws; the December 2025 executive order directs the Justice Department to challenge some of them in court, but the laws remain in force and enforceable until a court says otherwise.
Does the Colorado AI Act apply to my company?
The original 2024 act was repealed before taking effect. Its replacement, SB 26-189, applies from 1 January 2027 to developers and deployers of automated decision-making technology that materially influences consequential decisions — employment, education, housing, lending, insurance, health care, or government services — about Colorado residents. Insurers, HIPAA entities, and FDA-regulated products have exemptions.
Is Texas TRAIGA in effect?
Yes, since 1 January 2026. It reaches anyone doing business in Texas or whose product is used by Texans. Most private companies owe only intent-based prohibitions and cure-readiness; government agencies and health-care providers owe AI-interaction disclosures. The Attorney General enforces with a 60-day cure period and no private right of action.
Which states regulate AI in hiring?
Illinois (HB 3773, since 1 January 2026), New York City (Local Law 144, since 2023), and California (FEHA automated-decision-system regulations, since 1 October 2025) are in force now. Connecticut's no-defense rule applies from 1 October 2026 and its pre-decision notices to deployments from 1 October 2027; Colorado's ADMT law reaches employment decisions from 1 January 2027.
Do I have to tell users my chatbot is AI?
In Maine, yes, wherever a reasonable consumer could think they are talking to a human (since 24 September 2025). New York requires it for AI companions, Hawaii for public conversational AI, Utah on a clear request, California for companion chatbots, and Texas for government and health-care uses. From 2027 Colorado, Washington, Oregon, Rhode Island, Nebraska, Idaho, Georgia, and Iowa add disclosure cadences and crisis protocols.
Does the California AI Transparency Act apply to small companies?
Not today — SB 942 covers generative image, audio, or video systems with more than 1,000,000 monthly users, plus anyone who licenses such a system. But AB 2013 training-data disclosure has no size threshold, SB 243 companion duties have none either, and SB 1000, awaiting the Governor's signature until 30 September 2026, would remove the 1,000,000-user threshold entirely.
What if my company is not based in any of these states?
Location of the company rarely matters. These laws attach to users, customers, employees, or applicants in the state, to products used there, or to advertising into the state. If you cannot rule a state out, the check assumes it reaches you and shows what would follow.
Is the result legal advice?
No. It is a deterministic classification against versioned regulation packs, each verified against the enacted text and dated. Confirm obligations with counsel, especially where the result flags a statutory exception for review.
