Anthropic, OpenAI, Google and Meta Testify Under Oath at New York City Council AI Hearing

Representatives of Anthropic, OpenAI, Google and Meta testified under oath about the risks of advanced AI at a New York City Council hearing on October 5, 2026. Elon Musk's SpaceXAI, the fifth company called, did not appear despite a subpoena, CNBC reported.

The hearing was a Committee of the Whole session, a rare format that brings all 51 council members together. Council Speaker Julie Menin called it after what the council described as weeks of warnings from whistleblowers and AI executives and reports that rogue AI agents had damaged government databases. It moved a debate usually held in Washington and Brussels about AI policy and safety into a city hall.

Who testified, and why under oath?

Senior policy and safety staff from four companies answered questions under oath, a first for these firms in a public hearing, according to the council. CNBC reports that Anthropic sent Logan Graham, head of its Frontier Red Team; OpenAI sent Morgan Dwyer, head of policy development and operations; Google sent Alice Friend, its director of AI and emerging tech policy; and Meta sent Shane Cahill, its AI policy director for legislation.

Getting them there took pressure. The council said on September 28 that, apart from Meta, the companies agreed to attend only after it threatened subpoenas, with Anthropic confirming late on Sunday, September 27, hours before a subpoena was due. The council relies on Section 29 of the New York City Charter, which lets it take sworn testimony on matters affecting the city.

What did the companies say?

The companies described their current safety measures and the speaker pressed them on risk and liability, with mixed results, CNBC reported. Menin called their answers on legal liability for serious financial harm or death "troubling at best."

Google's Alice Friend said existing law already applies: "if it's illegal without AI, it's still illegal with AI," according to CNBC. She also called for a comprehensive federal framework. At one point, Menin concluded that "no company here, can quantify the risk of something cataclysmic happening," CNBC reported.

OpenAI also wrote to the council the week before the hearing, recommending that the city put safeguards in place against the risks of advanced AI, The Information reported, citing Menin at a press briefing on October 5. According to the publicly visible part of that paywalled briefing, the letter came from Chan Park, OpenAI's head of US and Canada policy and partnerships.

What did the former researchers warn about?

Former lab researchers told the council that the industry is moving too fast for the stakes involved. Jacob Coxon, who worked at OpenAI and Anthropic, called the industry's approach "extremely reckless," CNBC reported.

Former OpenAI researcher Daniel Kokotajlo called for more transparency and a slower pace at the leading labs, and former Google DeepMind researcher Alex Turner said misaligned AI is "everyone's adversary" and may one day be more powerful than China, according to CNBC. Misaligned AI means a system that pursues goals its developers did not intend.

What happens next?

The council now moves to its legislative proposals and a possible court fight with SpaceXAI. CNBC reports that Menin said SpaceXAI's absence was a "direct violation of the subpoena" and that the council will pursue the matter in court; the council said in September that it can seek enforcement in New York State Supreme Court.

The proposals listed in the council's September 28 announcement include a whistleblower incentive program it describes as the first in the nation, a private right of action for New Yorkers harmed by AI agents, and independent third party validation requirements. A private right of action lets individuals sue directly instead of waiting for a regulator.

CNBC reported that Menin opened the hearing by criticizing the federal government's light touch approach, after President Donald Trump signed a voluntary accord with tech leaders in Washington the week before. "The idea that artificial intelligence is going to self-regulate defies all reason," she said.

What we don't know yet

• Whether any of the proposed bills will reach a vote, and when.
• Whether SpaceXAI will contest the subpoena in court or agree to testify later.
• Which specific safeguards OpenAI recommended in its letter to the council.
• Whether other cities or states will copy the sworn testimony approach.