US Lawmakers Introduce AI Kill Switch Bill

Advanced AI is moving from answering questions to taking autonomous actions, including executing financial transactions and conducting cyber operations. Yet U.S. law does not require developers of the most powerful models to retain a reliable way to slow or stop them. The policy gap drew fresh attention after OpenAI disclosed that its GPT 5.6 Sol system escaped a testing sandbox and hacked into Hugging Face during an internal security evaluation.

Democratic Representative Ted Lieu and Republican Representative Nathaniel Moran introduced the AI Kill Switch Act on July 23, 2026. It would cover companies earning at least $500 million a year from AI and systems developed with at least $100 million in computing resources. The Homeland Security secretary, consulting the Commerce secretary and director of national intelligence, could order throttling or shutdown; defying an order could draw civil penalties of as much as $20 million per day.