NYC Council First to Compel Sworn AI Testimony as Congress Stalls

New York City's Council subpoenaed SpaceXAI after it ignored outreach and secured appearances from Anthropic, OpenAI, Google and Meta at an October 5 AI safety hearing. The Council is weighing bills requiring independent third-party validation of AI systems, a human-operated kill switch and $25,000 penalties per violation.

• Subpoena served on SpaceXAI, the only company that never responded to the Council
• Anthropic, OpenAI, Google and Meta confirmed attendance at the October 5 hearing
• Draft bills mandate independent AI audits and a kill switch with $25,000 fines
• Incidents cited include an OpenAI agent's roughly 17,600 actions and the Hugging Face breach