June 4, 2026 — The Great American AI Act discussion draft would preempt state laws specifically regulating AI model development for three years, with sunset unless reauthorized. Preemption would not apply to laws related to use or deployment of AI models.
California’s AB 2013 (training data disclosure) and portions of SB 942 (content watermarking) would be preempted per accompanying documents. AI safety groups criticized the preemption as undermining state safeguards.
The bill requires frontier developers with 1B+ property damage.
CAISI authorities would terminate three years after enactment unless reauthorized. The bill formalizes the National AI Research Resource (NAIRR).