Bipartisan negotiations in the U.S. Senate over artificial intelligence (AI) safety have resurfaced disagreements, signaling potential obstacles for legislation that remains the most likely path for action this year.
The dispute centers on whether federal legislation should preempt state-level AI safety laws. Senate Commerce Committee lead Democrat Maria Cantwell of Washington posted on X on Wednesday expressing concern: she welcomed the urgency to act but warned that the answer is not a weak federal standard that could become a backdoor to eliminating stronger state protections.
The draft bill under negotiation is associated with Commerce Chair Ted Cruz (R-Texas), Senator Amy Klobuchar (D-Minn.), and Majority Leader John Thune. Sources say the current language of the proposal still contains provisions that would preempt state AI safety laws.
A source familiar with the matter told Semafor that Cantwell continues to push for a stricter version of the bill that would impose legal duties on AI companies to manage catastrophic risks. That source said Cantwell, Anthropic, and several safety organizations are "refusing to play ball"—unwilling to accept a bipartisan compromise that would give the federal government authority over AI models while allegedly preserving continued innovation and due process in the courts.
Other sources familiar with the negotiations told Semafor that major AI labs are still working in good faith with congressional negotiators.
These renewed disagreements highlight that bipartisan talks are not seamless and that preemption—the question of whether federal law will override stronger state rules—remains a central sticking point as lawmakers attempt to finalize the bill.
Why this matters
The core issue is whether a federal floor would set minimum standards while allowing states to maintain tougher protections, or whether the federal law would displace state-level rules. The outcome will affect how legal responsibility is assigned to AI companies, the regulatory landscape across states, and the balance between promoting innovation and enforcing safety measures.
(This article draws on reporting by Semafor; the original piece was by Ashley Gold.)
