House lawmakers on Thursday released a long-awaited 269-page discussion draft of a national artificial intelligence framework that would override state AI safety laws in California, New York, and Illinois for at least three years.
The bill was introduced by Rep. Jay Obernolte (R-CA) and Rep. Lori Trahan (D-MA), the two House members who have led bipartisan negotiations on federal AI regulation for months.
The draft calls for federal preemption of state laws that target AI model developers, a provision directly aimed at blocking regulations that California, New York, and Illinois have enacted to require top AI companies to disclose information about new models and identify potential security risks. Under the draft, those state laws would be superseded by a federal standard for three years, with a sunset provision that would allow states to re-engage afterward.
Supporters argue a national framework is necessary to prevent a fragmented regulatory environment from slowing American AI development. Critics on both sides have raised questions about whether the federal standard goes far enough, or in some cases too far, in reining in the industry.
The release of the draft follows a March 2026 White House policy framework that called on Congress to preempt state laws interfering with a “minimally burdensome” national standard. That White House document also included provisions on protecting children online, intellectual property, and free speech in AI-generated content.
One of the central unresolved questions in the draft involves whether a federal vetting regime for advanced AI models should be mandatory or voluntary. That debate has complicated negotiations and remained a sticking point as of late May.
The legislation is framed as a discussion draft, meaning it is intended to serve as a starting point for congressional debate rather than an immediate path to a floor vote. Both Obernolte and Trahan have described the process as collaborative and emphasized they do not expect the final product to look exactly like the document released Thursday.
The move puts Congress in a direct confrontation with Democratic-led states that have moved aggressively to regulate AI companies. California Gov. Gavin Newsom signed several AI-related bills in 2025, though he vetoed the high-profile SB 1047 the year prior. New York and Illinois have since passed their own versions of AI accountability legislation targeting large model developers.
