AI Whistleblower and Tech Executives Face New York City Council Hearing Over Safety and Oversight

New York City lawmakers are bringing a former Anthropic researcher and executives from major artificial intelligence companies together for a public hearing on AI safety, oversight, and recent incident reports. Jacob Coxon, a former Anthropic researcher who has warned that advanced AI could become too powerful for people to control, is scheduled to testify alongside representatives of Anthropic, OpenAI, Google, and Meta.

The hearing places an industry whistleblower and the companies building large AI systems on the same public stage. The New York City Council said the companies will testify publicly under oath, marking the first such appearance since recent incident reports, according to the council’s announcement.

Coxon’s testimony centers on the risks created by the competition to develop increasingly capable AI systems. He has warned that the industry does not fully understand the dangers associated with that race, including the possibility that advanced systems could outstrip human control.

The concerns bring a safety-focused perspective into a hearing otherwise dominated by companies with commercial and technical interests in AI development. Coverage identifies Coxon as a former Anthropic researcher and describes his warnings as involving safeguards, control mechanisms, and the need for stronger oversight. The New York Post’s headline also points to debate over “kill switches” and regulation, though detailed testimony was not available in the published excerpts.

Anthropic, OpenAI, Google, and Meta are appearing before the council as lawmakers examine how AI companies manage safety risks and respond to incidents involving their systems. The lineup covers several of the most influential developers and platforms in the field, giving the hearing significance beyond a single company’s policies.

Executives face questions about how companies test models, monitor harmful behavior, respond when systems fail, and communicate risks to the public. The companies’ participation also gives city officials a direct forum for examining whether voluntary safeguards are adequate as AI products spread through workplaces, schools, government services, and consumer applications.

The hearing reflects New York City’s effort to shape AI oversight even as national and international rules remain unsettled. City lawmakers are weighing whether local safeguards are needed for a fast-growing industry whose products increasingly affect residents, employers, and public institutions.

The companies’ attendance followed pressure from the council, with one report stating that OpenAI, Google, and Anthropic agreed to participate after threats of subpoenas. Public testimony under oath raises the political stakes because statements made before lawmakers can influence later regulatory proposals and public scrutiny. The council’s approach also signals that local governments intend to question major AI firms directly rather than wait for federal action.

The hearing links technical safety concerns with practical questions about accountability. Coxon’s warnings address the limits of current understanding, while the company appearances focus attention on the systems, policies, and safeguards that developers put in place before releasing powerful models.

That combination matters because AI risks are no longer confined to research laboratories. Model failures can affect consumers, businesses, public agencies, and workers, while decisions about testing and deployment remain concentrated among a small group of companies. The hearing gives lawmakers a chance to compare industry assurances with whistleblower concerns in a public setting, where disagreements over safety standards and enforcement are easier to examine.

The immediate outcome will depend on the witnesses’ answers and on whether council members pursue new city safeguards. The hearing can produce commitments, expose gaps in company practices, or provide evidence for future legislation, but the published announcements do not establish a specific policy package.

The appearance also creates a public record that can be used in later debates over AI procurement, consumer protection, workplace rules, and emergency responses. For the companies, the session is a test of how clearly they can explain safety decisions under direct questioning. For lawmakers, it is an early measure of whether city-level oversight can keep pace with the technology’s rapid deployment.