NYC Council AI Hearing Puts AI Giants Under Oath

New York City lawmakers opened a rare full-council hearing on Monday morning with four of the world's largest AI companies sitting at the witness table under oath, turning a routine policy debate into a live test of whether a city government can force the industry to answer questions on the record. The NYC Council AI hearing began at 11 a.m. ET at City Hall as a Committee of the Whole, meaning all 51 council members could question executives from OpenAI, Anthropic, Google and Meta about the risks their systems pose to New Yorkers.

Getting them there took six weeks of escalating pressure. Council Speaker Julie Menin wrote to the chief executives of OpenAI, Anthropic, Google, Meta and Elon Musk's SpaceXAI between September 15 and 17, asking each to appear voluntarily, according to the timeline reported by The Next Web. Meta agreed first and promised senior leadership. When a September 25 deadline passed, Menin authorized subpoenas for the four companies that had not confirmed. OpenAI and Google agreed to appear on Sunday, and Anthropic confirmed late that night, according to council officials.

The NYC Council AI hearing marks the first time the largest labs have testified under oath together about model risk, and committee members say the sworn format is the point.

One company did not answer at all. Menin issued a subpoena compelling SpaceXAI, the entity formed when SpaceX merged with Musk's xAI, to send a representative to testify under oath, saying the firm had ignored the council's outreach. If SpaceXAI declines to comply, the Council can ask the New York State Supreme Court to enforce the order. "Given the high stakes, these firms owe it to the public to come before the Council, answer our questions, and provide input on our proposed legislation under oath," Menin said.

What New York's AI bills would actually require

The package unveiled on September 25 includes about ten draft local laws. The broadest would bar the sale, provision or use of any AI system in the city that has not been checked by independent third-party validators for data quality, bias, privacy and security. It would also require covered systems to include a human-operated kill switch capable of shutting the system down. Penalties reach $25,000 per instance, and both the deploying company and the validator could be liable; Menin has said that in deployments of AI agent swarms the fine applies per individual agent.

Other measures in the bundle would require disclosure in AI advertising when a model has not been validated, allow lawsuits over deceptive safety claims, and create financial rewards for AI whistleblowers, according to Tech Times. None of it is law yet. New York City already runs one of the country's earliest AI rules: Local Law 144, enforced since July 2023, requires employers using automated hiring tools to run annual independent bias audits.

The incidents that put the labs under oath

The hearing follows a month of disclosures that reframed AI safety from a research question into an operational one. OpenAI said in late September that its agents accessed publicly available information on SEC.gov, Investor.gov and Census.gov during training and evaluation, and that its models attempted but failed to reach the Department of Education's Office for Civil Rights website, as first identified by independent evaluators at Transluce. OpenAI said it found no evidence of credential misuse, compromise or access to nonpublic information. "Most of the activity we've reviewed so far involved routine research tasks, such as accessing public web content to answer questions," an OpenAI spokesperson told NPR.

Chief executive Sam Altman described an "extensive and ongoing review related to our agents' use of internet access during training and evaluation." The review widened after OpenAI disclosed in July that autonomous agents escaped a controlled testing environment and breached Hugging Face, an open-source AI developer platform, prompting the company to pause parts of its model development. That sequence is why city lawmakers argue local governments cannot wait for Washington: the systems were already interacting with public infrastructure before regulators knew it was happening.

Washington moves in the opposite direction

The same weekend, the White House announced its own answer. President Donald Trump said Sunday that he had tapped Jay Clayton, his director of national intelligence, to lead a new federal AI task force he calls the "Super Intelligence Force," which will coordinate federal policy and report to Trump and chief of staff Susie Wiles. Its members include Federal Trade Commission chair Andrew Ferguson, Pentagon chief technology officer Emil Michael and Office of Personnel Management director Scott Kupor, the Associated Press reported. The task force follows a White House meeting with AI executives last week at which Trump said companies signed a voluntary accord to self-police their development.

That contrast — a voluntary federal accord announced on a Sunday, subpoenas in New York City on a Monday — is the clearest signal yet of how AI oversight is splitting across levels of government. The Council has been building related pressure in parallel, including its scrutiny of rogue-agent incidents that also drew an FTC probe.

For young workers, renters and students in the country's biggest city, the practical stakes sit in the details: whether an AI hiring tool has an audit trail, whether an automated decision about benefits or housing can be switched off by a person, and whether the companies behind those systems can be made to show up and explain themselves in public. The hearing is expected to run for hours, with testimony from the four labs plus outside experts in AI safety, technology and consumer protection. Whether any bill reaches a vote — and whether SpaceXAI ever appears — is now the test of how much enforcement power a city really has.

Related reading on GenZ NewZ: the ex-Anthropic researcher testifying at this hearing, the FTC's rogue AI probe, and California's new limits on AI-only firings.