HearingsHauled before the NYC Council, AI companies dodged whether they would be liable if a model goes rogue

The New York City Council held an AI oversight hearing on October 5. Executives from Anthropic, OpenAI and Google, which agreed to attend after subpoena warnings, and Meta, which committed earlier, testified virtually. SpaceX did not show despite a subpoena, and legal action is pending.

Former employees were blunt. Ex-Anthropic staffer Jacob Coxon said that on the current path humanity is more likely than not to lose control to AI, and ex-OpenAI researcher Daniel Kokotajlo warned companies could use it to become dictators or oligarchs. When Speaker Julie Menin asked whether a company bears legal responsibility if its frontier model causes serious financial harm, the executives avoided a direct answer. The council is weighing 10 bills, covering AI depictions of public officials, human shutdown capability and consent for training data.