admin
The New York City Council just held a rare “Committee of the Whole” on AI — all 51 members — and it was equal parts theater and policy-making. Former inside researchers sounded the alarm under oath, tech firms sent cautious spokespeople, and the Council unveiled a package of local rules that could reshape how AI is sold and used in the city. This was supposed to be about safety. It may end up being about who gets to write the rules: City Hall or the engineers who build the machines.
What happened at the hearing
Speaker Julie Menin led the hearing, which put three former lab researchers on the record: Jacob Coxon (a former Anthropic researcher), Daniel Kokotajlo (formerly with OpenAI), and Alex Turner (formerly with Google DeepMind). Coxon’s blunt line — “We do not know how to control any AI system yet” — was the dramatic headline. Representatives from Anthropic, OpenAI, Google and Meta also gave sworn testimony, but their answers were cautious and often noncommittal. The Council even subpoenaed SpaceXAI after it failed to appear. That no-show sets up a likely legal fight over municipal subpoena power and shows how high the stakes have become.
The bills on the table
The hearing wasn’t just for drama. Speaker Menin has floated a multi‑bill package that would require independent third‑party safety validation before systems are deployed or sold in the city, demand a human “kill switch,” impose fines for violations, and create whistleblower protections and a city response plan for AI incidents. The proposals aim to fill a perceived federal gap, but they also threaten to create a novel patchwork of local AI rules that firms must navigate if they want to operate here.
Why this risks pushing away innovation
Regulating AI is not a moral lecture; it’s technical work. City Hall can hold hearings and subpoena witnesses, but the Council does not write code, test models, or run safety labs. Heavy-handed local rules — mandatory third‑party validation, blanket kill‑switch requirements, and steep per‑instance fines — would raise compliance costs and slow product launches. That can make New York an expensive, risky place to operate. Startups and big firms alike could decide the city isn’t worth the hassle. The result: jobs and investment walk out while bureaucrats pat themselves on the back for “doing something.”
What should come next
If New York wants real safety, it should push for federal standards and coordinate with technical experts, not try to rewrite the rules from the City Hall podium. Practical local steps make sense — pilot programs, partnerships with research labs, transparency requirements for deployments in critical systems — but they should avoid one-size-fits-all mandates that invite lawsuits and deter innovation. The right goal is to protect New Yorkers without turning the city into a regulatory island where companies hide their best work somewhere more sensible. That balance won’t be easy, but grandstanding and subpoenas alone won’t solve the problem — they’ll just make the tech leave and the risk stay.
