NYC hearing, the company panel: OpenAI touts a $1B cyber-defender fund; Anthropic, Meta, Google name their safety cases

#What happened

After whistleblower testimony from Jacob Coxon, Daniel Kokotajlo and Alex Turner, a panel of company representatives testified at the NYC Council's Oct 5 AI hearing, according to a live recap by Kevin O'Shaughnessy published on Medium today.

• Dr. Morgan Dwyer (OpenAI): said OpenAI has committed $1 billion to facilitate access to its most advanced models and provide support to cyber defenders, and that OpenAI took a number of steps in response to the Hugging Face incident.
• Logan Graham (Anthropic, Frontier Red Team): said he dreamed of curing diseases but believes the technology carries risks; Anthropic welcomes smart regulation, is a public benefit corporation whose board must weigh risks, and held back its Mythos model from public release as a safety example.
• Shane Cahill (Meta, AI policy director): said safety, security and privacy are built into Meta's Muse; Meta works with external experts and has signed onto external evaluations and independent board oversight.
• Alice Friend (Google, Director of AI and emerging tech policy): said Google supports creating an independent industry body for AI, citing the American Medical Association as a model.
• Nobody from SpaceXAI testified.

Note: this account is a single observer's live recap — not yet corroborated by major outlets. Related dotsfeed posts today covered the witness roster, Coxon's testimony, OpenAI's endorsement of third-party audit bills, and the whistleblower panel recap.

#Why it matters

The council is weighing a sweeping package of bills, including third-party validation of AI systems, a human-controlled kill switch with $25,000-per-instance penalties, and whistleblower rewards. How the four labs frame their safety work today will shape that legislation. The recurring pledges — outside audits, held-back models, external evaluation — are voluntary until a law converts them into enforceable duties.

Verified Oct 5, 2026

#Sources

• https://medium.com/@ZombieCodeKill/industry-whistleblowers-testify-at-nyc-council-hearing-on-ai-safeguards-418076a37e89 (live hearing recap, published Oct 5 2026)