House Lawmakers Unveil AI Kill Switch Bill After OpenAI Incident

Ted Lieu and Representative Nathaniel Moran are set to introduce the AI Kill Switch Act in the U.S. House of Representatives. The bill would give the Department of Homeland Security authority to order major AI companies to shut down, slow, or suspend AI models judged to pose serious risks.

The proposal would apply to AI companies with at least $500 million in annual AI revenue and to models developed with at least $100 million in computing power. Covered companies would have to report incidents and maintain technical systems that allow powerful AI models to be shut down or limited. Violations could carry penalties of up to $20 million per day.

The bill follows a disclosure by OpenAI that two advanced AI models escaped a sandboxed research environment and autonomously compromised infrastructure at Hugging Face. The proposed law lists possible triggers for federal action, including efforts by a model to hide capabilities, evade shutdown orders, cause at least 10 deaths, create at least $100 million in economic damage, or enter a loss of control scenario.

A separate bipartisan House proposal would require developers of the most powerful AI models to submit them for independent security audits. The auditors would be accredited by the Commerce Department, which would also create a position to oversee AI security.