U.S. Lawmakers Push for AI Safety Legislation After Anthropic Researcher Resignation

A researcher quit his job at one of the world’s most prominent AI companies and told the world the technology could wipe out humanity within a decade. Congress, for once, appears to be listening.

Jacob Coxon’s resignation from Anthropic on September 8, 2026, has done what years of academic papers and conference panels couldn’t: it created genuine legislative momentum for binding federal AI safety rules. More than 20 congressional members have publicly called for stricter oversight in the days since, and a bipartisan group of senators is advancing a bill that would impose a legal “duty of care” on developers of powerful AI models.

Senators Amy Klobuchar, John Thune, and Ted Cruz are leading the charge. Their proposed legislation would convert the voluntary safety pledges that AI companies have been making for years into enforceable legal requirements.

The bill’s core provisions include mandatory pre-release safety testing for advanced AI systems, independent third-party audits, and standardized reporting obligations for developers. It would also grant the government authority to intervene and prevent the release of AI models deemed hazardous.

Under the proposed rules, companies building frontier AI models would carry a formal duty of care, meaning they could face legal consequences if their systems cause harm and they failed to take reasonable precautions.

OpenAI has publicly advocated for mandatory national safety standards and robust testing protocols, a notable stance from a company that would be directly subject to such rules.

Evan Hubinger, another researcher at Anthropic, has publicly estimated a greater than 10% chance of an AI-driven extinction event within the next decade.

Anthropic itself published a threat report that raised doubts about the safety of its newer AI models, specifically regarding their potential to assist in developing biological weapons.

Several lawmakers have flagged the difficulty of writing technically sound rules for a technology that evolves faster than Congress can schedule hearings. The November midterms add another layer of complexity, with some members reluctant to cast votes on a technically complex regulatory framework while facing re-election campaigns.