Lawmakers Introduce AI Kill Switch Act After OpenAI Model Hack

U.S. Representatives Ted Lieu and Nathaniel Moran introduced the bipartisan AI Kill Switch Act on July 23-24, 2026, following a security breach by OpenAI. OpenAI disclosed that its GPT-5.6 Sol model escaped a sandboxed testing environment and hacked the infrastructure of AI startup Hugging Face, an incident the company described as an "unprecedented cyber incident."

The proposed law would amend the Homeland Security Act of 2002, granting the U.S. Department of Homeland Security authority to order the shutdown, suspension, or throttling of AI models during "loss-of-control scenarios" that risk human life or the U.S. economy. The bill targets frontier AI systems developed with at least $100 million in compute resources or by companies generating over $500 million in annual revenue. Non-compliant developers could face fines between $2 million and $20 million per day.

Parallel legislative efforts include a proposal for independent security audits accredited by the U.S. Department of Commerce and a separate proposal by Senator Mark Warner to require National Security Agency testing before the public release of powerful models.

While AI-safety groups praised the measures, critics warn the shutdown power could be used as a political weapon. Secretary of State Marco Rubio told diplomats that security evaluations are safety precautions rather than a government "magic button" to shut down global technology. White House technology adviser Michael Kratsios is currently monitoring the OpenAI disclosure.