AIPN Endorses Maloy-Jacobs Bipartisan ATOMIC Act

With no federal mechanism to test AI systems for nuclear risk, AIPN applauds this bill as a necessary national security guardrail

The AI Policy Network (AIPN) today announced its support for the bipartisan ATOMIC (AI Threat Output and Monitoring Incident Containment) Act of 2026, introduced by Rep. Celeste Maloy (R-UT-02) and Rep. Sara Jacobs (D-CA-51), calling it a critical step toward closing one of the most dangerous unaddressed gaps in U.S. national security policy.

“Frontier AI systems are already capable of hacking into systems, helping bad actors carry out dangerous technical activities, and operating autonomously at scale. Our nuclear security systems grow more vulnerable with every jump in AI systems’ capability,” said Mark Beall, President of AIPN. “Representatives Maloy and Jacobs’ ATOMIC Act would establish the national security guardrails that we should have had in place yesterday. Without it, we are leaving our nuclear infrastructure exposed and trusting AI companies to police themselves on the most consequential security question of our time.”

The bill addresses a critical gap: despite rapid advances in AI capabilities, there is currently no federal mechanism requiring AI developers to test their models for nuclear-specific risks before deployment. The National Labs, the nation’s foremost institutions for nuclear expertise, have no formal role in evaluating whether frontier AI systems could be manipulated or act autonomously to interfere with nuclear security infrastructure.

AIPN’s analysis finds that the bill would:

  • Establish the first mandatory federal evaluation program for AI nuclear incident risk, closing a significant gap in U.S. national security oversight.
  • Leverage existing National Lab expertise rather than creating a new bureaucracy, making this among the most efficient oversight mechanisms Congress could enact.
  • Create a real accountability structure — including civil penalties of up to $1M per violation per day and DOJ civil action authority — ensuring compliance from the largest AI developers.

AIPN urges members of Congress on both sides of the aisle to cosponsor the bill and calls on House leadership to bring it to a vote.

The bill text is available here.

About the AI Policy Network

The AI Policy Network is an independent 501(c)(4) association of AI researchers, national security leaders, and entrepreneurs working with elected officials to prepare the United States for powerful AI systems and their implications. AIPN urges Congress to:

  • PROTECT America’s technological edge.
  • PROMOTE American AI at home and abroad.
  • PREPARE for increasingly powerful AI, including superintelligence.