Alex Turner warns NYC Council about misaligned AI risks
Former Google DeepMind researcher Alex Turner warned New York City lawmakers that AI development could create a domestic adversary.
By Maya Okafor · Markets Writer
· 3 min read
Former Google DeepMind researcher Alex Turner used an October 5 New York City Council hearing to issue a stark Alex Turner AI warning: he said the race to develop advanced systems could create an adversary in the United States. For readers following AI policy, the hearing also put specific city-level safeguards, rather than only broad safety concerns, on the table.
Turner made the comment after Council Speaker Julie Menin asked how the city should weigh development of frontier AI models against technological competition with China, CNBC reported. Turner said China was not the only potential adversary and warned that, in his view, misaligned AI could eventually become more powerful than China. His remarks were a warning about a possible outcome, not evidence that such an outcome has occurred.
The Committee of the Whole meeting was held at City Hall on October 5 at 11 a.m. The official agenda identified its oversight subject as “Examining the Risks Posed by Artificial Intelligence.” The Council had said the body, which includes all 51 members, was scheduled to consider AI risks, company safeguards and possible legislative responses.
What did Alex Turner tell the NYC Council about AI?
Turner was one of several former AI-lab researchers who pressed lawmakers to treat the pace of development as a safety concern. Jacob Coxon, a former Anthropic researcher, called the industry’s approach “extremely reckless” and said he believed people were more likely than not to lose control of AI systems on the current path, according to CNBC. That was Coxon’s assessment, not a confirmed forecast.
Daniel Kokotajlo, a former OpenAI researcher, called for greater transparency and for companies to slow development of leading models, CNBC reported.
Representatives of Anthropic, OpenAI, Google and Meta testified under oath, according to CNBC. The outlet identified them as Anthropic Frontier Red Team head Logan Graham, OpenAI policy development and operations head Morgan Dwyer, Google AI and emerging-technology policy director Alice Friend, and Meta legislative AI policy director Shane Cahill.
Friend supported a comprehensive federal framework for AI risks, citing national-security concerns, CNBC reported. Menin said the responses from company representatives did not directly quantify the risk of a catastrophic event.
What AI measures was the Council considering?
The Council’s proposals had not become law. Its September announcement included a bill that would require third-party validation before an AI system could be marketed, sold or deployed in New York City, along with a human shutdown capability. The proposal set penalties of $25,000 per specified instance of noncompliance or falsified validation.
- A whistleblower-incentive proposal that could award individuals a share of penalties recovered from AI companies that violate applicable laws.
- A proposal creating a private right of action in certain cases involving foreseeable harms from malicious use or the bypassing of safety controls.
- A measure requiring covered city contractors or agencies to report AI safety incidents to the Office of Cyber Command within 24 hours, with public disclosure of reported incidents within 24 hours.
The official meeting record lists the agenda as final, but says the minutes remain draft and provides no completed legislative results. That means the available record does not establish that the proposals were enacted or that the hearing produced final Council action.
This story draws on original reporting from CNBC.
