Two Lawmakers at Odds Over Federal AI Vetting Requirements
Representatives Kat Trahan and Jay Obernolte are deadlocked on a core question: should the federal government require AI companies to submit to safety reviews, or let them choose?
Trahan, a Massachusetts Democrat, wants mandatory data-sharing and disclosure requirements for advanced AI developers. Obernolte, a California Republican, favors a voluntary approach where companies decide what information to report to the government, according to three people familiar with the closed-door talks.
The disagreement reflects a broader split in Washington over how to regulate AI. The two lawmakers have been negotiating a federal framework that would preempt conflicting state AI laws - a proposal that has drawn fire from both safety advocates and fellow Democrats.
Mythos Accelerates the Debate
The negotiations gained urgency after Anthropic released Mythos, an AI model that can identify cybersecurity vulnerabilities human hackers cannot find. The capability triggered internal discussions at the White House about whether new AI systems should require government approval before deployment.
President Trump is considering an executive order to vet advanced AI models for security risks. Inside the administration, some officials back a hands-off approach while others push for mandatory vetting or pre-clearance requirements - mirroring the Trahan-Obernolte divide.
State Regulation vs. Federal Preemption
AI companies have spent months arguing that conflicting state laws create an unmanageable regulatory patchwork. The industry wants federal rules to override state restrictions.
The proposed preemption is narrow on paper - it would apply only to laws directly regulating advanced AI development. But safety advocates worry companies will use it as a legal weapon to block state laws on children's safety or privacy, claiming those rules would force changes to how they build models.
"It will be a litigation magnet," said one AI policy advocate with knowledge of the talks.
Political Pressure Mounts on Trahan
Trahan faced swift backlash for negotiating with Obernolte. Massachusetts state legislators sent her a warning letter. A coalition of safety advocates and Massachusetts voters launched a petition opposing any deal that would weaken state AI safeguards.
When asked about the political risk, Trahan said she sees merit in discussions about national security and cybersecurity threats in a "post-Mythos world."
The talks represent Congress's latest attempt to establish federal AI rules after multiple failed efforts to reach consensus. Spokespeople for both lawmakers declined to comment on the negotiations.
For more on how government agencies are approaching AI regulation, see AI for Government.
Your membership also unlocks:
