NYC Lawmakers Grill AI Firms Over Risks and Regulation

On October 5, representatives from OpenAI, Anthropic, Google, and Meta faced tough questions from New York City lawmakers during a rare hearing on AI safety. Former AI researchers, including Jacob Coxon and Daniel Kokotajlo, warned that rapidly advancing systems could outpace human control, raising concerns about potential extinction risks. Coxon, a former researcher at Anthropic and OpenAI, criticized the industry’s startup-style culture, arguing that it is poorly suited for developing superintelligent AI.

The hearing highlighted a central issue in the AI safety debate: companies may acknowledge severe risks but struggle to agree on how to measure or mitigate them. Lawmakers pressed OpenAI’s Morgan Dwyer and Google’s Alice Friend for specific answers about catastrophic scenarios, with mixed responses. The Council is now considering new AI regulations, including mandatory third-party validation, human-controlled kill switches, and whistleblower protections.

Meanwhile, Elon Musk’s SpaceXAI faced legal pressure after failing to appear despite a subpoena. The hearing also addressed broader concerns like privacy, deepfakes, and harmful content, with UK MP Jess Asato testifying about her lawsuit against xAI over alleged privacy violations. As governments debate whether voluntary safeguards are sufficient, New York City is pushing for stronger oversight of private-sector AI systems.