Jacob Coxon Warns NYC Council About AI Control as Lawmakers Weigh Safeguards

New York City lawmakers heard a warning from inside the AI industry on October 5, 2026. Former Anthropic researcher Jacob Coxon testified that the race to develop AI carries risks the industry does not fully understand, as the City Council considered imposing its own safeguards.

Coxon’s testimony, reported by Bloomberg News, centered on a basic question: whether the people building AI know how to control it. He told the Council that they do not.

Representatives from leading AI developers also appeared at the hearing. Their participation put Coxon’s warning before lawmakers in a proceeding that included the industry he was criticizing. The Council’s stated task was to weigh whether New York City should establish protections around a business whose reach extends well beyond the city.

We do not know how to control any AI system yet.

Jacob Coxon, testifying before the New York City Council

Coxon left Anthropic in September, according to Reuters. His departure came with a much more sweeping public warning: he said the “people building AI earnestly believe that it could kill us all by the end of the decade.” He also accused Anthropic and OpenAI of “gambling with our lives.” Those statements express Coxon’s view of the risks and the companies’ conduct.

The route to the hearing involved an invitation from Council Speaker Julie Menin. Bloomberg reported on October 4, citing people familiar with the matter, that Coxon would appear at her request. Reuters relayed that account and said Menin had urged AI whistleblowers to testify while lawmakers considered a package of safeguard bills.

Coxon was arguing that the industry does not fully understand the risks it is creating. The Council was examining whether those concerns warrant protections established through city legislation.