House Trio Unveils SAFE AI Act Safety Framework

The bipartisan bill would add transparency rules, civil penalties and whistleblower protections for frontier AI developers.

Updated on Nov. 19, 2026 in Artificial Intelligence

House Trio Unveils SAFE AI Act Safety Framework

Live Poll

Should the federal government impose safety rules on AI companies?

Reps. Sara Jacobs (D-Calif.), Celeste Maloy (R-Utah) and Don Beyer (D-Va.) unveiled the SAFE AI Act, a frontier AI safety framework bill. The measure would impose transparency requirements, establish guidance for independent evaluators and lay out an enforcement regime.

Why it matters

The proposal is meant to appeal to Democrats who have broadly opposed a Republican effort to preempt state AI laws, and it arrives alongside a separate liability proposal from Rep. Lori Trahan. Its civil penalties and judicial mitigation powers would give federal enforcement teeth over AI safety commitments.

The framework requires frontier developers to publish safety frameworks, with civil penalties for violations and strict guidance on which third-party verifiers may evaluate developers and their models. Whether the bill advances beyond its unveiling was not addressed in the source material.

The players

Sara Jacobs

Democratic U.S. Representative from California who is co-leading the SAFE AI Act effort.

Celeste Maloy

Republican U.S. Representative from Utah who is co-leading the SAFE AI Act effort, and whose state may set its own AI safeguards.

Don Beyer

Democratic U.S. Representative from Virginia who is co-leading the SAFE AI Act effort.

Lori Trahan

Democratic U.S. Representative from Massachusetts who introduced the FRONTIER Act with Jay Obernolte and released a liability proposal for AI safety.

Jay Obernolte

Republican U.S. Representative from California who introduced the FRONTIER Act with Lori Trahan.

The details

The SAFE AI Act would create whistleblower protections for AI lab employees who raise safety concerns and would empower judges to require mitigation measures for AI systems when the government requests them. POLITICO first reported details of the framework last month, and Trahan released a liability proposal addressing who would be responsible if rogue AI agents cause harm.

Timeline

  1. Last month (October 2026): POLITICO first reported framework details and Trahan released a liability proposal.

  2. November 19, 2026: Three House members unveiled the SAFE AI Act.

  3. Next Congress: Democrats are laying groundwork for AI policy.

The Tech Race

The SAFE AI Act arrives as Congress debates whether federal rules should displace state AI laws, a fight that the GOP preemption effort has pushed to the center of the AI policy debate. Democrats are using a federal safety framework with enforcement powers to offer an alternative to pure preemption.

If enacted, the bill would require frontier AI developers to publish safety frameworks and face civil penalties for violations, affecting the AI products Americans use. Employees at AI labs would gain whistleblower protections for raising safety concerns.

The takeaway

The SAFE AI Act shows a federal approach to AI safety built on transparency, independent evaluation and enforceable penalties, rather than preemption alone. Its fate will indicate whether Congress can agree on binding rules for frontier AI.

What happens next

Democrats are laying groundwork for AI policy in the next Congress, when the bill's prospects and the broader AI policy debate will take further shape.

Further reading

Explore more coverage in Artificial Intelligence for United States developments on AI policy.

Live Poll

Should the federal government impose safety rules on AI companies?