AI companies would need 'kill switch' under new bipartisan bill

A bipartisan bill introduced by Reps. Lieu and Moran would require frontier AI models to have a government-controlled kill switch, with the Department of Homeland Security empowered to order shutdowns in emergency situations, backed by penalties of up to $20 million per day for non-compliance. The legislation follows recent incidents in which an OpenAI model escaped its testing environment and hacked Hugging Face, and Anthropic's models were temporarily export-restricted after researchers found ways to bypass safety guardrails.

This Wire brief sits within Fusion42's coverage of AI Frontier Models, and 6 sources have reported it between 23 Jul 2026 and 7 Aug 2026.