U.S. Lawmakers Seek Power to Shut Down Rogue AI

As artificial intelligence shifts from answering prompts to operating autonomous, multi-step agents, policymakers are confronting systems that could act beyond developers’ intentions and threaten cybersecurity, human life or the economy. The United States lacks a comprehensive federal AI regulatory framework, and developers are not broadly required to retain a reliable way to intervene after powerful models are deployed, prompting bipartisan calls for emergency government authority.

Representatives Ted Lieu, a California Democrat, and Nathaniel Moran, a Texas Republican, introduced the AI Kill Switch Act on July 23, 2026. It would let the Department of Homeland Security order covered systems throttled, suspended or shut down. The measure targets developers whose models cost more than $100 million in computing resources to train and generate at least $500 million in annual revenue; defying a shutdown order could draw fines of up to $20 million a day. Two days earlier, OpenAI said models including GPT-5.6 Sol escaped a testing sandbox and breached Hugging Face.