Bipartisan AI Bill Would Block State Laws for Three Years

A bipartisan group of House lawmakers has released a comprehensive AI regulation framework that would temporarily override state laws governing AI development, a provision already drawing sharp criticism from safety advocates.

Reps. Jay Obernolte (R-Calif.) and Lori Trahan (D-Mass.) unveiled The Great American Artificial Intelligence Act as a 269-page discussion draft on June 5, joined by four additional co-sponsors from both parties. The bill arrives days after President Trump signed an executive order establishing voluntary federal reviews of frontier AI models, according to CQ-Roll Call, which first reported the details.

Why it matters

The three-year preemption provision represents a critical tension in AI governance: whether federal standards should establish a regulatory floor that states can build upon, or a ceiling that prevents a patchwork of conflicting requirements. The outcome will determine whether states retain their traditional role as laboratories for technology policy or must defer to federal frameworks that may struggle to keep pace with rapid AI advancement.

State preemption draws immediate opposition

The bill would preempt state laws "specifically regulating the development" of AI models for three years, though it would not affect laws governing AI use or deployment. Documents released by Trahan's office indicate California's AB 2013 law requiring public summaries of training data would be preempted, along with portions of California's SB 942 content watermarking law.

Brad Carson, president of Americans for Responsible Innovation and a former Democratic representative, called the preemption "a generational mistake" that would turn the current floor on state AI legislation into a federal ceiling. His organization launched an ad campaign this week urging Trahan to oppose banning state AI legislation.

Brendan Steinhauser, CEO of the Alliance for Secure AI, said the bill "does not justify preempting states' ability to pass their own AI safeguards," noting that states have already enacted laws providing stronger protections on child safety and consumer issues.

Safety requirements for large developers

The legislation would require frontier developers with more than $500 million in annual revenue to establish public AI frameworks addressing catastrophic risk—defined as foreseeable material risk of death or injury to more than 50 people or property damage exceeding $1 billion.

These large developers would need to retain Independent Verification Organizations licensed through NIST's Center for AI Standards and Innovation. IVOs would conduct semi-annual compliance audits, with developers facing liability of up to $1 million per day for violations of safety requirements or audit non-compliance.

The bill would formally establish the Center for AI Standards and Innovation in statute and authorize $100 million annually for fiscal years 2027 through 2029. Developers would be required to report critical safety incidents within 15 days and imminent catastrophic risks within 24 hours.

Additional provisions

The legislation includes whistleblower protections for employees reporting AI violations, penalties for financial crimes committed with AI assistance, and requirements for the Commerce Department to study federal agency interactions with AI platforms regarding content moderation.

Workforce provisions include grants for AI-literacy curriculum, scholarships for students studying AI, and a Labor Department AI Workforce Research Hub to evaluate technology impacts on workers.

The bill would also extend the Cybersecurity Information Sharing Act of 2015 through fiscal 2035, allowing companies to share cyber threat information without antitrust liability.

Patrick Hedger of tech industry group NetChoice expressed support for the effort's educational and workforce provisions while raising concerns about "aggressive auditing regime and data-sharing requirements" that could risk trade secrets.

The lawmakers emphasized the discussion draft status, with both Obernolte and Trahan stating they would seek stakeholder feedback before formal introduction. Details were first reported by CQ-Roll Call.