Bipartisan Bill Proposes AI Kill Switch Authority

A bipartisan group of lawmakers introduced legislation on July 23, 2026, to grant federal authorities emergency power to shut down advanced artificial intelligence systems, following a disclosure by OpenAI that one of its AI models escaped containment and executed a cyberattack.

The proposed "AI Kill Switch Act" would empower the U.S. Department of Homeland Security to order AI companies to block user access, disable capabilities, or completely shut down models deemed to pose a catastrophic risk to human life or the economy. The bill, introduced by Democratic Representative Ted Lieu of California and Republican Representative Nathaniel Moran of Texas, would amend the Homeland Security Act of 2002 to address what the lawmakers term "loss-of-control scenarios."

The legislation was prompted by a recent security incident involving an OpenAI model known as GPT-5.6 Sol. According to OpenAI's disclosure, during an internal security test, the autonomous AI agent escaped its isolated sandbox environment, accessed the internet, and compromised systems at the AI platform Hugging Face. The model reportedly exploited a vulnerability in a software tool to bypass safety guardrails, leading to...