NYC AI Oversight Hearing Puts OpenAI, Google, Meta Under Oath

NEW YORK — Artificial intelligence just had its first day in court — sort of. On Monday, October 5, 2026, the New York City Council convened the first NYC AI oversight hearing, compelling sworn testimony about AI safety and putting representatives from OpenAI, Google, Anthropic and Meta on the record, under oath, about what their systems could do to the public and what they are doing to keep people safe.

The hearing was convened as a Committee of the Whole — meaning all 51 council members sat together in a rare show of institutional force — and was organized by Council Speaker Julie Menin. Her argument was blunt: the federal government has left AI companies to regulate themselves, so the city would step in. According to reporting on the proceedings, Menin told the companies’ representatives that New York would act where Washington had not.

Who actually showed up matters. OpenAI sent Morgan Dwyer, its head of policy development and operations. Google sent Alice Friend. Anthropic sent Logan Graham, who leads its Frontier Red Team — the group that stress-tests models for dangerous behavior. Meta sent Shayne C. Cahill. Three of the four companies only agreed to appear after Menin threatened them with subpoenas, reported by Tech Times and the Associated Press. The fifth invitee, SpaceXAI — the company Elon Musk formed through SpaceX’s February 2026 acquisition of xAI — was formally subpoenaed and then failed to appear at all.

The exchange that stole the hearing

The most dramatic moment came when Menin asked each representative to quantify the worst-case risk their technology poses. How likely, she wanted to know, is a catastrophic AI failure? Dwyer answered that he didn’t know the odds — and didn’t think the number was the point: “whether it’s 1% or 10% or 20% chance that something catastrophic will go wrong. None of these levels is remotely acceptable.” He added that OpenAI should not train models it cannot make an extremely strong case for keeping under human control.

Menin called the response “flippant at best.” The sharper warnings came from the people who used to build these systems. Jacob Coxon, a former Anthropic researcher who resigned in September 2026, testified that on the industry’s current path, it is “more likely than not that humanity loses control to these AIs and it could end in human extinction.” A second outside expert, former Google DeepMind researcher Alex Turner, warned that misaligned AI could become a bigger threat than China.

The companies’ representatives pushed back gently. Dwyer said OpenAI shares the council’s commitment to safe development and that its benefits should be widely shared. Anthropic’s Graham said the company welcomes “smart regulation” and that state and local governments have a role in building frameworks. “AI is powerful enough that getting this wrong has real consequences, for individuals, institutions, public trust,” he said. “We have to work relentlessly to do the science of mitigating the risks.”

Ten bills that could rewrite the rules

The NYC AI oversight hearing wasn’t just theater — it was the opening move for legislation. The council is weighing a package of ten AI bills, eight of which are set for formal introduction on October 8, according to the council’s legislative records. The proposals are strikingly concrete. One bill from Menin would make it illegal to sell or deploy an AI model in New York City unless an outside validator has reviewed it and a human can shut it down — a literal kill switch. Others would let people sue an AI company when someone else misuses its model to cause harm in the city, pay whistleblowers a share of the money the city recovers from violators, require chatbots to disclose that users are talking to a machine, and ban unauthorized AI-generated depictions of public officials.

New York is using an unusual legal lever. Under Section 29 of the City Charter, the council can compel testimony and documents from any person or organization doing business in the five boroughs, with enforcement through the state Supreme Court. That is how a city council ended up doing what Congress has not: two years of AI oversight bills on Capitol Hill have produced zero binding measures, while a single city hearing produced sworn testimony and a legislative package in a week.

Why this matters far beyond New York

For Gen Z, this isn’t abstract. The systems under discussion are the chatbots millions of young people use for homework, job applications and creative work — and the same technology behind deepfakes, algorithmic hiring filters and the automation reshaping entry-level jobs. A kill-switch rule or a chatbot-disclosure law passed in New York has a way of becoming the template other cities copy, the way the city’s food-labeling and gig-worker rules once did.

Governor Kathy Hochul has separately moved to impose AI reporting and transparency standards at the state level, and officials from Mayor Zohran Mamdani’s administration attended the hearing — a sign the bills could get executive backing. The next milestone is October 8, when the bills are formally introduced. After that comes the harder part: turning sworn testimony into enforceable law — and finding out whether the industry’s promises survive contact with it.

This story sits on the AI News beat — see also earlier coverage of ChatGPT’s invisible text watermark rolling out in Europe, another front in the fight over AI transparency. For background on how the hearing was compelled, see Tech Times’ reporting on the subpoena fight and Bushwick Daily’s account of the testimony.