Reps. Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a 269-page discussion draft of the Great American AI Act on June 4, freezing state AI development laws for three years in exchange for a federal audit, disclosure, and incident-reporting regime aimed at large frontier developers. It’s the most serious bipartisan attempt at a comprehensive federal AI framework to date, and the preemption clock starts the moment it becomes law.

The structural trade is legible. Developers with more than $500 million in prior-year gross revenue would’ve to publish frontier AI frameworks and obtain third-party audits, per Roll Call, and report critical safety incidents to the Center for AI Standards and Innovation within 15 days, dropping to 24 hours for imminent risks of death or serious injury. Civil penalties cap at $1 million per violation, with each day counted separately. State attorneys general can opt in to receive safety reports. CAISI gets codified inside Commerce with a $100 million annual authorization for FY2027–2029, FedScoop reports.

That cap is where the bargain gets contested. The Future of Privacy Forum notes the $1 million ceiling sits below California’s SB 53 and well under New York’s RAISE Act and SB 315, which reach $3 million for repeat violations. Trade a tougher state regime for a softer federal floor, and the math gets uncomfortable fast.

Reaction split immediately. The House Democratic Commission on AI came out against the draft within hours. Brendan Steinhauser, CEO of the Alliance for Secure AI, otherwise praised the bill but warned: “A national AI standard should protect at least as much as it preempts.”

Cosponsors Scott Franklin (R-FL), Suhas Subramanyam (D-VA), Erin Houchin (R-IN), and Scott Peters (D-CA) signed on at release. Both lead sponsors sit on Energy and Commerce. No introduction date is set, which is itself the point: the draft exists to absorb three years of state-law momentum into a federal frame before the states finish writing their own.

Sources

Sources