Tracking of artificial-intelligence laws state by state in the United States and of the tension between state and federal regulation. In the absence of a comprehensive federal law, states have created a patchwork —Colorado, California, Texas, Utah, Illinois— that companies must map jurisdiction by jurisdiction. But that patchwork is under pressure: in December 2025 a federal executive order sought to preempt 'excessive' state laws, and in April 2026 a federal court paused Colorado's law weeks before it took effect. The tracker measures the gap between the state law passed and the law actually applicable. Each record documents a rule or milestone with its status and source.
Statistical readings derived from the attributes of each recorded case. All figures come from the documented events; amounts are computed only over cases with a sum expressed in the indicated currency, without converting between currencies.
The United States has no federal artificial-intelligence law, so fifty states legislate on their own and create a patchwork that companies must map one by one. But that patchwork is under fire: a federal executive order seeks to preempt it and a court has just frozen the most comprehensive law, Colorado's. This tracker measures the gap between legislating and being able to enforce.
While the European Union applies a single AI Act to its twenty-seven members, the United States does the opposite: it has no comprehensive federal law, and in its absence each state regulates its own way. The result is a patchwork of laws with different approaches and dates that any company operating across several states must map jurisdiction by jurisdiction. This tracker orders that patchwork and, above all, measures something the law count hides: which are actually in force and which are not.
Because in 2026 the distinction is critical. California (frontier-AI transparency), Texas (responsible governance) and Illinois (employment discrimination) have laws in force since 1 January. But the most ambitious of all —the Colorado AI Act, the first comprehensive state law on high-risk AI— has had a troubled history: its effective date was delayed from February to June, and in April a federal court froze it weeks before the new date. Passing a law and being able to enforce it are, here too, two different things.
Of the eight laws and milestones tracked, five are in force, but two —including Colorado's flagship law— are judicially paused. The difference between 'state X regulates AI' and 'that law can be enforced today' is exactly the information a multi-state compliance team needs, and which the headline hides.
What makes the US case particularly unstable in 2026 is not just the fragmentation, but a new tension between two levels of government pulling in opposite directions. In December 2025, the White House signed an executive order proposing a uniform federal framework and seeking to preempt state laws deemed 'excessive' —citing by name, as an example, precisely Colorado's—. A few months later, a federal court froze that same law. The wind's direction is clear: federal deregulatory pressure against ambitious state legislation.
For a company, this creates the worst possible combination: obligations that exist on paper but whose real applicability depends on litigation and on a political tug-of-war between Washington and the state capitals. The tracker takes no side in that fight; it documents its milestones with their status and source, so that whoever must comply knows, at each moment, which rule is alive, which is frozen and which hangs by a judicial thread. That traceability is the asset.
Utah's law requires disclosure when a consumer interacts with an AI system, applicable across all industries. It is an early, light-touch transparency model, predating the comprehensive 2026 laws.
Texas's Responsible Artificial Intelligence Governance Act took effect on 1 January 2026. It is one of the most prominent state AI laws, with a focus on responsible use and the prohibition of certain uses, in a state with a light-touch regulatory tradition.
On 11 December 2025, President Trump signed the executive order 'Ensuring a National Policy Framework for Artificial Intelligence', proposing a uniform federal framework that preempts state laws deemed inconsistent. It cites the Colorado AI Act by name as an example of 'excessive State regulation'. It marks the start of the federal-state tension.
The United States has no comprehensive federal AI law. Federal governance relies on agency enforcement of existing laws (FTC under Section 5), executive orders and voluntary standards like the NIST AI RMF. Congress continues to debate broader legislation. That absence is what has pushed states to legislate.
In the early months of 2026, Connecticut, Washington, Oregon, Idaho, Nebraska, Maryland, Vermont, Georgia and Illinois advanced or enacted significant AI laws, according to industry compilations that put at around 2,000 the AI-related bills introduced across the 50 states. The fragmented environment —diverging 'high-risk system' definitions, different disclosure requirements and inconsistent enforcement mechanisms— complicates multi-state compliance. Several lawsuits, including a certified class action against Workday's AI resume screener, established that both vendors and employers can face liability for discriminatory AI outcomes.
Illinois's law (Public Act 103-0804) amends the state Human Rights Act to prohibit employers from using AI that discriminates against protected classes. It is part of the wave of state laws focused on employment and algorithmic hiring.
Governor Jared Polis signed SB 189, 'Automated Decision-Making Technology,' on May 14, 2026, repealing and replacing the Colorado AI Act (SB 24-205) mere weeks before its June 30, 2026 effective date. The new law adopts a lighter-touch regime that eliminates certain obligations for developers and deployers of AI systems; as a result, the 2024 Colorado AI Act will not take effect. The new rule applies from January 1, 2027 and retains broad applicability in consequential decisions, albeit with more streamlined compliance obligations.
On 27 April 2026, a federal court paused enforcement of the Colorado AI Act weeks before its expected effective date (30 June 2026), preventing the state from initiating enforcement actions while the litigation is pending. It is neither a repeal nor a permanent delay, but a freeze that leaves companies in legal uncertainty.
The first comprehensive US state law targeting 'high-risk' AI systems: it requires developers and deployers to use reasonable care to prevent algorithmic discrimination, conduct impact assessments and give consumer notices. Fines of up to $20,000 per violation, no private right of action. Its effective date was delayed from Feb to June 2026, and on 27 April 2026 a federal court paused its enforcement. It is the only state law cited by name in the December 2025 federal executive order.
California's Transparency in Frontier Artificial Intelligence Act took effect on 1 January 2026, alongside several transparency and employment laws. California has led AI regulation with a disclosure-based approach, complemented by the AI Transparency Act (SB 942) on labeling AI-generated content.
Each record documents a state AI law or a milestone in the federal-state tension, with its application status (in force, passed not in force, judicially paused, delayed, under review) and its source. The decisive field is the status: a passed law is not an applicable law, and this tracker separates the two. The sanction amounts are each rule's legal maximums. Legislative or judicial outcomes are not predicted. Coverage prioritizes comprehensive laws and federal preemption milestones for their impact on multi-state compliance.