Leaving AI Regulation to the States Could Strangle AI
Without federal preemption, a regulatory thicket of state AI laws threatens to slow the technology's development.
Artificial intelligence is advancing rapidly in the United States, but one thing threatens the industry's growth: state regulations. In July 2025, 38 states passed about 100 AI measures, according to the National Conference of State Legislatures. To counteract this push from states, President Donald Trump is calling on Congress to implement a federal preemption of state AI regulations. However, it appears increasingly unlikely that such a law will pass.
Before Thanksgiving, Trump took to Truth Social to call for one federal standard instead of dozens of local laws to keep America ahead of China in the global AI race. To this end, the president called on congressional Republicans to work federal preemption into the National Defense Authorization Act (NDAA) or to pass a separate bill so that "nobody will ever be able to compete with America."
A few days later, an executive order entitled "Eliminating State Law Obstruction of National AI Policy" was leaked. Public Citizen, a progressive consumer rights nonprofit, reports that the order would have directed the Justice Department to establish a state AI law task force charged with challenging local laws that violate the Constitution or federal preemption. Like the failed AI moratorium in an early Senate version of the One Big Beautiful Bill Act (OBBBA) earlier this year, the order directs the Commerce Department to withhold federal broadband funding from states with laws that run afoul of the president's AI Action Plan.
After pushback from GOP lawmakers decrying the rumored order as violating the spirit of federalism, the White House withheld the order from publication. However, Trump has continued advocating for language to be included in the NDAA that would nullify technology-specific state and municipal AI regulations through federal preemption. Such a provision is opposed by Reps. Marjorie Taylor Greene (R–Ga.), Thomas Massie (R–Ky.), and Chip Roy (R–Texas), who regard it as federal overreach that compromises federalism instead of a prudent regulation of interstate commerce. Politico reported Tuesday that House Majority Leader Steve Scalise said the NDAA "wasn't the best place for [preemption language] to fit."
In the Senate, meanwhile, opposition from Sens. Marsha Blackburn (R–Tenn.)—who previously blew up the OBBBA moratorium—and Josh Hawley (R–Mo.)—who introduced four bills this year strictly regulating AI technology and its applications—seems to indicate that, at this point, a standalone preemption won't meet the 60-vote threshold needed to clear the filibuster.
Although concerns about federal overreach are understandable, there are also strong reasons to oppose state-level regulation. Kristian Stout, the director of innovation policy at the International Center for Law and Economics, argues that a patchwork of state AI regulations "threatens to impose substantial costs while failing to achieve the consumer protection goals that motivate such regulation." While policymakers who fear Big Tech believe regulations will rein in these companies, they should pause to consider why Anthropic is actively calling for regulation. Incumbent firms with large legal departments can navigate a byzantine maze of state-specific AI laws; the compliance burden throttles innovation—harming the American economy and national security—by effectively foreclosing the market to capital-poor startups.
Trump's recently announced Genesis Mission, touted as a Manhattan Project for AI, would "launch coordinated funding opportunities…to incentivize private-sector participation in AI-driven scientific research." American AI doesn't need industrial policy, but a nationwide regulatory framework that provides nascent and incumbent AI developers with the clarity required for continued private investment and development. Unfortunately, it is unlikely to pass in the current Congress.