Anthropic Researcher Set to Testify on AI Risks at NYC Council Hearing

A former researcher at Anthropic is set to testify before New York City lawmakers about the risks of advanced AI. Jacob Coxon, who helped build AI systems during his time at the company, will speak at a hearing on October 5, 2026. The event is part of an effort by city leaders to address potential dangers posed by AI.

Council Speaker Julie Menin called the hearing, which will involve all 51 council members. Representatives from major AI companies, including Anthropic, OpenAI, Google, and Meta, have been invited to attend. However, not everyone has responded positively: SpaceXAI was served with a subpoena after it failed to comply.

The proposed legislation includes measures to require third-party validation of AI systems used in the city, create whistleblower incentives, and allow individuals to sue for damages caused by AI systems. These ideas have gained momentum following Coxon's public statement on leaving Anthropic, which has been viewed over 173 million times.

Coxon characterized his former employer as 'gambling with our lives' and warned that the technology could kill everyone by the end of the decade. His concerns are shared by other Anthropic researchers, including Evan Hubinger, who estimates a greater than 10% probability that AI could be lethal to humanity within a decade.