Three of the largest states now regulate the most powerful artificial intelligence systems in nearly the same way, and the people behind those laws have started acting together. California's law took effect in January, and New York's and Illinois's take effect on January 1, 2027. The lawmakers who wrote them issued a joint appeal to AI companies in September, and New York's law lets its regulator share companies' confidential safety reports with other governments. No formal three-state agreement has been announced, and the laws mostly require disclosure and reporting rather than limits on what companies can build. The Trump administration, for its part, wants one national standard and has ordered the Justice Department to challenge state AI laws it considers burdensome.
Three laws, one template
California's Transparency in Frontier Artificial Intelligence Act, known as SB 53, came first. New York amended its RAISE Act this year to bring it closer to California's, and Illinois Gov. JB Pritzker signed the Artificial Intelligence Safety Measures Act on July 6. All three cover "frontier" models, those trained with more than 1026 computing operations, and put most of their duties on developers with more than $500 million in annual revenue. Those companies must publish safety frameworks, file assessments of catastrophic risk with a state agency, and report serious safety incidents: within 15 days in California, and within 72 hours in New York and Illinois, or 24 hours when the danger is imminent. Illinois alone requires annual independent audits. The laws reach any developer whose models are available to people in those states, which in practice means every major AI company.
Industry has not uniformly resisted. OpenAI, which in March 2025 urged the federal government to preempt state AI laws, endorsed the Illinois bill as "one of the strongest frontier AI safety laws in the country" and said California, New York and Illinois "are beginning to create a de facto national framework."
Where the coordination is
On September 4, the lead authors of the three laws, including New York Assemblymember Alex Bores and Sen. Andrew Gounardes, Illinois Rep. Daniel Didech and Sen. Mary Edly-Allen, and California Sen. Scott Wiener, jointly called on AI labs to "immediately establish a Mutually Agreed Pacing Framework (MAP Framework) that is jointly negotiated and independently verifiable by third parties," to slow the race toward more capable systems "until progress on alignment and safety have demonstrably surpassed" it. They wrote that this was "not a replacement for robust actions at the state, federal, and/or international levels."
The other channel is New York's Department of Financial Services, which will enforce the RAISE Act through a new office and opens registration for large developers in November. Unlike the California and Illinois laws, New York's lets the department share companies' incident reports and internal risk assessments with "other governmental entities." A Lawfare analysis argues that phrase has no geographic limit and covers agencies in other states, and that the department could share the reports confidentially through the licensing system state financial regulators already use. Crypto Briefing reports that the department's superintendent, Kaitlin Asrow, has spoken publicly about interstate coordination on AI oversight.
The push from Washington
An executive order Trump signed on December 11, 2025, "Ensuring a National Policy Framework for Artificial Intelligence," directs the Justice Department to challenge state AI laws in court and calls some state rules "legally deceptive." The law firm Paul Hastings listed California's SB 53 among the laws potentially vulnerable under it, while noting that the order "does not, and cannot, overturn existing state law"; only Congress or the courts can. Congress has been reluctant: in July 2025 the Senate voted 99 to 1 to strip a proposed 10-year ban on state AI laws from the budget bill. A bipartisan House draft, the Great American AI Act from Reps. Jay Obernolte and Lori Trahan, would write much of the three states' framework into federal law and block new state rules on AI model development for three years.
Sources: the lawmakers' joint statement, September 4, 2026; Lawfare; Crowell & Moring via Mondaq, July 10, 2026; Crypto Briefing, October 5, 2026; Paul Hastings; the Future of Privacy Forum.
