Updated:
Former Anthropic researcher Jacob Coxon is expected to testify Monday, October 5, at a New York City Council hearing on artificial-intelligence safeguards, Bloomberg News reported Sunday, citing people familiar with the matter. The hearing will bring together lawmakers, AI-company representatives and outside critics as the Council considers proposed rules for AI systems used or offered in the city.
Coxon is appearing at the request of Council Speaker Julie Menin, according to the report. He left Anthropic last month and has warned publicly about risks from increasingly capable AI. His appearance would place a former industry researcher’s concerns alongside testimony from representatives of Anthropic, OpenAI, Google and Meta, who are also expected, according to reporting on the hearing.
A rare Council-wide examination
The October 5 proceeding is a Committee of the Whole hearing, a format that convenes all 51 members of the City Council. Menin announced the hearing in September, saying lawmakers would examine risks to New Yorkers, safeguards used by leading AI companies and possible city-level legislative or regulatory steps.
Company attendance follows a push by Menin’s office. Bloomberg previously reported that OpenAI, Anthropic and Alphabet’s Google agreed to send executives after being warned that they could face subpoenas; Meta had committed to sending a senior executive earlier. The companies’ scheduled participation does not mean they endorse the Council’s proposals, and the record available ahead of Monday did not establish which individuals would appear.
Reuters, in a report carried by AOL, said Coxon would testify at Menin’s request as the Council weighs a package of bills. Reuters said it could not immediately verify Bloomberg’s report. A subsequent Investing.com report, carrying Reuters material, also identified former Google DeepMind researcher Alex Turner and former OpenAI researcher Daniel Kokotajlo among those expected to testify.
Proposals include audits and whistleblower incentives
The Council’s September 25 announcement described a slate of proposals that would impose new obligations on some AI businesses and create additional routes for public oversight. One bill sponsored by Menin would require third-party validation of AI systems marketed, offered for sale or deployed in New York City, including checks related to data quality, bias, privacy and security. It would also require a human override, or “kill switch,” and set a proposed $25,000 penalty for each instance of covered deployment without proper validation or with falsified validation.
Other proposals would let whistleblowers receive a share of penalties recovered from AI companies for violations, and create a private right of action in certain cases where people are harmed through malicious use or the circumvention of safety controls. A separate proposal would require reporting of AI safety incidents involving certain city contracts. These are proposed measures for Council consideration, not enacted requirements.
The proposals build on city rules already adopted. The Council says laws passed in 2025 created an Office of Algorithmic Accountability to review city agencies’ use of AI and algorithmic systems, set standards for government use and require a public inventory of covered systems. The new package would reach further into private-sector AI activity, though the scope and legal design of any final legislation remain to be determined through the legislative process.
Coxon’s warnings enter the policy debate
Coxon’s public departure from Anthropic helped bring AI safety concerns into the Council’s debate. The Council’s September announcement noted his resignation and warnings about potential catastrophic risks. Reuters reported that Coxon had criticized Anthropic and OpenAI for what he viewed as excessive risk-taking in the race to develop more advanced systems; those criticisms are his claims, not findings established by the Council.
The hearing follows other high-profile discussion of AI risk. The Associated Press reported that Anthropic Chief Executive Dario Amodei and OpenAI Chief Executive Sam Altman had recently addressed the pace of development and the need for safeguards, while emphasizing differing approaches to managing risks. Such company statements form part of the wider debate, but Monday’s hearing is focused on what New York City can examine or regulate within its authority.
What happens next
The hearing is scheduled for October 5, with all Council members convening to consider the proposed safeguards and hear from witnesses. Testimony could inform whether bills are revised, advanced or debated further; the Council’s announcement did not set a vote or specify a timetable for passage. Any measure would still need to proceed through the city’s legislative process.
It also remains unclear what specific subjects Coxon and the other outside witnesses will address, what questions lawmakers will prioritize, or whether testimony will prompt changes to the proposals. The proceeding will give lawmakers a public forum to assess competing views, but it will not itself establish that the risks described by witnesses have occurred or settle how far the city can regulate AI companies operating across jurisdictions.
