Insider Resignation Spurs Bipartisan Push for First Federal AI Safety Bill

Jacob Coxon, who worked at both OpenAI and Anthropic, resigned this week, declaring that the industry is "gambling with our lives" and lacks any real plan to control superintelligent systems. His resignation, reinforced by Anthropic researcher Evan Hubinger, quickly led to a Senate investigation launched by Josh Hawley and a parallel inquiry from Richard Blumenthal. In reaction, Senators Ted Cruz, Amy Klobuchar and Majority Leader John Thune have drafted a bipartisan AI safety bill that would empower the Commerce Department and Homeland Security to require safety testing, mandate incident reporting, and halt the release of models deemed catastrophically risky, while overriding existing state AI laws.

The effort follows years of failed federal bills, state-level attempts, and extensive industry lobbying, including multimillion-dollar campaign spending in the 2026 House primary. If passed, the legislation could become the first comprehensive federal framework for AI safety, shifting regulatory focus from competition to risk mitigation.