Jacob Coxon quit Anthropic in September, saying the AI labs were "gambling with our lives." On Monday he made that case again before the New York City Council at a hearing open to all 51 members, with the companies he was criticizing on the witness list for the same hearing. The council is weighing bills that could require outside safety checks before an AI product is used in the city and let New Yorkers harmed by AI sue the companies behind it.
Coxon, who also worked at OpenAI before Anthropic, didn't soften his warning. "From my experience, the companies are being extremely reckless given the stakes," he said, and he put the chance that humanity loses control of advanced AI at "more likely than not". He blamed a startup culture of moving fast and fixing problems later. He also said most code at these companies is now written by AI with little human review.
The companies came, but didn't say much
The hearing took the rare form of a Committee of the Whole, a format that convenes the entire council instead of a single committee. According to the council, OpenAI and Google agreed to appear only after it threatened subpoenas, and Anthropic confirmed late Sunday, hours before its subpoena would have come due. SpaceXAI, Elon Musk's AI company, was subpoenaed and didn't show up. It's the company formerly known as xAI, which SpaceX absorbed earlier this year and renamed in July.
The answers from the companies were thin. According to amNY, representatives from Anthropic, OpenAI, Google and Meta would not put a number on catastrophic risk. They wouldn't promise that failing an independent safety test would stop a release, and they mostly sidestepped whether they'd be legally liable for serious harm. Google's Alice Friend gave the plainest line: "If it's illegal without AI, it's still illegal with AI". Speaker Julie Menin was not impressed. "To say you don't know and it doesn't matter is flippant at best," she said.
Other former insiders backed Coxon up. Former OpenAI researcher Daniel Kokotajlo described today's safety fixes as "duct tape that will fall off later". Ex-Google DeepMind researcher Alex Turner put the odds of an AI takeover at roughly one in three.
Why a city is doing this
The package includes human "kill switches," third-party validation before deployment, cash rewards for whistleblowers, a right for harmed New Yorkers to sue, 24-hour incident reporting and fines of up to $25,000 per violation. Coxon said kill switches might help for now but doubted they'd be enough in the long run. The Trump administration prefers to let the industry police itself. Menin says that "is not going to work" and hopes local rules will push Washington to act.
No vote date has been set, and deliberations continue. Menin said the council "still has many questions" and will send the companies written follow-up questions, and it's pursuing SpaceXAI's ignored subpoena in court.
Sources
- 1.New York City Council Announces Anthropic, OpenAI, Google, and Meta to Publicly Testify Under Oath for First Time Since Recent Incident Reports · New York City Council
- 2.AI giants give few clear answers to key safety questions at NYC Council hearing amid whistleblower warnings · amNY
- 3.Ex-OpenAI researcher tells NYC lawmakers AI safety fixes may be 'duct tape that will fall off' · Fortune
- 4.Ex-Anthropic Researcher Testifies at NYC Council AI Hearing · Bloomberg (via Yahoo Finance)
- 5.xAI is now officially known as SpaceXAI · Engadget
Reported by the WattsUpNext desk from the sources linked below. Spot an error? Tell us at corrections@wattsupnext.com.
The WattsUpNext Brief
Get stories like this in your inbox.
One email each weekday, only the topics you choose. Real news, sourced, no fluff. Unsubscribe in one click.
