Reps. Jay Obernolte (R-Calif.) and Lori Trahan (D-Mass.) released a 269-page discussion draft of the Great American AI Act on June 4, the first comprehensive federal AI bill to attach binding obligations on frontier labs to a preemption of state AI development laws. The trade is explicit: a national floor for model-level governance, paid for with three years of quiet on the state-legislative front.

That preemption is the live wire. Per Roll Call, the draft blocks states from “specifically regulating the development” of AI models for three years while leaving use-and-deployment laws untouched. A document from Trahan’s office names California’s AB 2013 and Colorado’s SB26-189 as casualties. The framing is narrower than last year’s failed 10-year moratorium, which the Senate killed 99-1, but the political memory is fresh. The House Democratic Commission on AI opposed the draft within hours.

The bill is organized into four titles, per Tech Policy Press: Frontier AI Governance, Workforce, Cybersecurity, and Research & International Cooperation. DLA Piper’s read of the operative sections is that “large frontier developers,” defined as labs with $500M+ in annual revenue that have trained a frontier model, would face binding disclosure under Section 111, third-party audits under Section 112, and whistleblower protections. The Cybersecurity Information Sharing Act of 2015 gets extended through fiscal 2035.

Reps. Scott Franklin (R-Fla.), Suhas Subramanyam (D-Va.), Erin Houchin (R-Ind.), and Scott Peters (D-Calif.) joined the rollout, giving the draft a bipartisan surface area its predecessor never had. Energy and Commerce is the likely committee of jurisdiction.

Critics aren’t buying the trade. Alliance for Secure AI CEO Brendan Steinhauser said the bill “does not justify preempting states’ ability to pass their own AI safeguards.” That’s the bet the sponsors are making with a public comment window: that binding rules on the labs are worth more than Sacramento and Denver combined.

Sources

Sources