New York City lawmakers grilled top artificial intelligence executives Monday on existential threats posed by the technology, with one whistleblower warning it could trigger human extinction.
The marathon City Council hearing, held in Manhattan, brought policy leads from Google, Meta, Anthropic and OpenAI face-to-face with legislators weighing local restrictions on the booming industry. Council Speaker Julie Menin opened the session by demanding concrete risk assessments from the witnesses, who had agreed to appear only after being threatened with subpoenas.
“We want to hear how each of you quantify the risk imposed by AI in the worst case, catastrophic scenario,” Menin said during the hearing.
Morgan Dwyer, OpenAI’s head of policy development and operations, testified virtually alongside representatives from the three other tech giants. When pressed on worst-case calculations, Dwyer offered a stark admission.
“I don’t know, I also don’t think it matters,” Dwyer replied.
Menin immediately fired back at the response.
“To say you don’t know or [it] doesn’t matter is flippant at best,” she said. “If you can’t quantify what safety or risk is, how do you know the product is safe?” she continued, comparing the stance to a pharmaceutical company developing a drug and saying, “I don’t know if it’s going to kill people.”
The hours-long questioning kicked off with testimony from AI whistleblowers who painted a grim picture of unchecked technology disrupting the city’s economy and infrastructure. Jacob Coxon, a former researcher at both Anthropic and OpenAI, delivered the hearing’s most alarming warning.
“From my experience, the companies are being extremely reckless given the stakes,” Coxon said. “On the current path, I think it is more likely than not that humanity loses control to these AIs, and it could end in human extinction,” he told the council.
The tech executives attempted to shift focus toward potential benefits, pitching advancements in medical research and expanded tools for small businesses.
“AI is making it possible to solve economic, scientific and social problems faster than ever,” Dwyer said during her testimony.
Yet even the company representatives acknowledged serious dangers lurking behind the promise.
“We think AI has tremendous potential benefits… but no technology is without risks and there are a wide spectrum of risks. Clearly some risks are unacceptable,” Dwyer said after sustained questioning from Menin.
Logan Graham, who leads Anthropic’s Frontier Red Team responsible for stress-testing advanced AI models including Claude, echoed the qualified concern.
“Our common belief is that if not properly safeguarded, the risks from AI can be extremely high,” Graham added.
The executives largely endorsed the council’s proposal to strengthen protections and establish cash rewards for whistleblowers reporting AI risks. Council members also pressed the witnesses on third-party validation measures, potential job displacement, and cooperation with immigration officials. Notably absent from the questioning was any discussion of the legislative body’s signature proposal — a “kill switch” or human override requirement for AI systems operating in the city.
The hearing exposed the limited reach of local regulation, with all participants emphasizing the need for comprehensive federal oversight of the globally operating companies.
“The most effective safeguards are anchored in national technical standards and harmonized across state and federal levels,” said Alice Friends, Google’s global director of AI policy.