The question is moving from what AI can do to how anyone can check it. Anthropic, OpenAI, Google and Meta answered questions from all 51 council members; the Speaker said they gave no clear yes or no on halting releases after failed safety tests. SpaceXAI ignored a subpoena.
Who appeared, and who did not
On Monday 5 October, representatives of Anthropic (Logan Graham, head of its Frontier Red Team), OpenAI (Morgan Dwyer, head of policy development and operations), Google (Alice Friend) and Meta (Shane Cahill) testified under oath, by video, to a rare Committee of the Whole of all 51 council members. Speaker Julie Menin said Anthropic, OpenAI and Google agreed to appear only after the council threatened subpoenas. Elon Musk's SpaceXAI did not attend despite a subpoena, and Menin said the council will pursue the matter in court.
What the council said it heard
Asked whether a failed internal or independent safety test would stop a release, the companies described their review processes without the yes or no Menin asked for; she called the answers an "equivocation." Google's Friend said "if it's illegal without AI, it's still illegal with AI." The council is weighing 10 new bills and resolutions, and will send the companies written questions. A city legislature, not a national one, is the first to put frontier AI firms on the record under oath.
Coverage context
Frontier labs are now on a city-council record under oath, with the Speaker saying they would not give a clear yes or no on halting a release after a failed safety test, and with SpaceXAI absent despite a subpoena.
Read the full daily brief hub: www.aifusionautomations.com/news/