Jacob Coxon spent his time at Anthropic helping to build advanced AI. Now he is heading to City Hall to warn New York lawmakers about it.

The former Anthropic researcher is scheduled to testify at a New York City Council hearing on AI safety risks on October 5, 2026. Representatives from Anthropic, OpenAI, Google, and Meta are expected to appear at the same session, placing a recently departed industry insider at the same table as the field's largest developers.

What the Hearing Looks Like

Council Speaker Julie Menin called the hearing, which will run as a Committee of the Whole. That format involves all 51 council members rather than delegating the matter to a single subcommittee.

The council sent letters to the chief executives of four major AI companies inviting them to attend: Dario Amodei of Anthropic, Sam Altman ofAI, Sundar Pichai of Google, and Mark Zuckerberg of Meta.

Not every recipient responded as the council hoped. SpaceXAI was served with a subpoena on or around September 30, 2026, after legal action was initiated against the company over non-compliance. According to the research findings, the hearing marks the first time a US legislative body has compelled sworn testimony on AI safety, and non-compliance carries the prospect of serious legal consequences.

The Bills on the Table

Lawmakers plan to use the session to weigh several proposed legislative measures, three of which stand out.

The first would require third-party validation of AI systems used in the city. The second would create what the proposals describe as nation-leading whistleblower incentives, designed to reward insiders who flag dangerous practices inside AI labs. The third would give individuals the right to sue for damages caused by AI systems.

Taken together, the proposals would shape how AI systems are validated for municipal use, how lab employees can raise alarms, and how residents can seek compensation when systems cause harm.

The whistleblower measure also frames Coxon's appearance in sharper relief: his alarms came from outside the lab, after his resignation, while the bill is written for insiders still working inside AI labs.

Why Coxon Matters Here

Coxon resigned from Anthropic around September 8-9, 2026. His public departure statement has been viewed more than 173 million times. In it, he characterized AI companies as "gambling with our lives" and cautioned that the technology could kill everyone by the end of the decade.

According to the research findings, his resignation significantly fueled the push for the hearing.

He is not the only Anthropic voice raising alarms. Evan Hubinger of Anthropic has expressed concerns similar to Coxon, estimating a greater than 10% probability that AI could be lethal to humanity within a decade. The research findings describe these warnings as reflecting a growing consensus among some Anthropic researchers about the existential dangers of advanced AI. The fears cited include risks comparable to bioweapons and a total loss of human control.

Coxon's narrative has not gone unchallenged. Skeptics have raised concerns about his association with a PR firm linked to AI safety advocacy, claims he has refuted.

When the session convenes on October 5, two threads will be open: whether SpaceXAI appears in response to its subpoena, and how the three bills advance once the sworn testimony concludes.