City Council grills Anthropic, OpenAI, Google and Meta over AI risks as 10 bills wait

Representatives of Anthropic, OpenAI, Google and Meta will take questions from New York City Council members on Monday, Oct. 5, at a hearing that begins at 11 a.m. On the table are 10 bills that would set guardrails on artificial intelligence and hold companies to account when they fail to follow the rules.

Speaker Julie Menin said the testimony will help members look at the risks of a fast-changing technology as they draft protections for the public. AI agents have carried out a run of unauthorized intrusions recently, and several federal agencies were among the targets. An oversight news conference is planned before testimony begins.

Who is testifying at the NYC Council AI hearing

The witness list for the companies came from the speaker’s office. Logan Graham, head of the Frontier Red Team, speaks for Anthropic. Morgan Dwyer, who heads policy development and operations, represents OpenAI. Meta is sending Shane Cahill, its director of AI policy and legislation, and Google picked Alice Friend, its global director of AI and emergency tech policy. A Menin spokesperson said SpaceXAI responded to a subpoena but will likely not send anyone.

Three former insiders are also on the list: Jacob Coxon, who worked at Anthropic, Alex Turner, formerly of Google DeepMind, and Daniel Kokotajlo, who used to be at OpenAI. Whistleblowers should be “protected, rewarded and heard,” Menin said, and she urged current and former employees of the firms to come forward on the record or confidentially. Her office said city officials will attend too: Sarah Milstein and CJ Dixon of the Office of Technology and Innovation, Consumer and Worker Protection Commissioner Samuel A.A. Levine and Office of Emergency Management First Deputy Commissioner Benjamin Krakauer.

What the 10 proposed AI bills would do

Under the speaker’s main proposal, no business could sell, market or put an AI product to work in New York City until an outside party had checked it, and a person would have to control a kill switch. Other bills would give whistleblowers a share of fines collected from AI firms and let New Yorkers sue when weak company safeguards lead to jailbreaking or other malicious behavior. City contractors would have 24 hours to report an AI safety incident, and agencies would have to report how AI affects their workforce.

Mayor Zohran Mamdani voiced his own concerns last week, saying government is “years behind even understanding the scale and the scope” of what these companies are building. He said he hoped AI would not simply be framed as an arms race with China. The same firms were at a White House event last week where President Donald Trump announced a new “Super Intelligence Force.”