NYC Council Hears Warnings of AI Extinction Risk as Whistleblowers Testify

The New York City Council convened a rare Committee of the Whole hearing on Monday to question executives from OpenAI, Anthropic, Google, and Meta on AI safety. Speaker Julie Menin led the session, which also considered a legislative package including mandatory third-party audits and "kill switches" for AI systems deployed in the city.

Former Anthropic researcher Jacob Coxon testified voluntarily that humanity is "more likely than not" to lose control of advanced AI, potentially ending in extinction. He compared current safety practices to "duct tape that will fall off later" and urged a slowdown in frontier model development.

“On the current path, I think it is more likely than not that humanity loses control to these AIs, and it could end in human extinction.”

Former OpenAI researcher Daniel Kokotajlo, testifying under subpoena, warned the field resembles psychology more than engineering because systems are grown rather than designed. He said the industry's ability to detect misalignment is poor and set to worsen.

Company representatives refused to quantify the probability of catastrophic harm. OpenAI's Morgan Dwyer stated she did not know the risk level but argued no percentage was acceptable. The firms also declined to commit that failed safety tests would automatically block releases or to accept blanket legal liability for harms.

Elon Musk's xAI ignored a subpoena to appear. Menin said the Council would pursue legal action to compel testimony from the company, which responded to the subpoena only with a letter expressing willingness to cooperate.