Under oath at NYC hearing, OpenAI says it backs laws requiring third-party AI safety audits
Verified· Oct 5, 2026Published Oct 5, 2026
At New York City's October 5 AI hearing, an OpenAI speaker said the company has endorsed legislation requiring third-party audits of its AI safety work — stopping short of endorsing the city's specific validation bill.
What happened
- During the New York City Council's October 5 Committee of the Whole hearing on AI risks, an OpenAI speaker said the company has endorsed multiple bills requiring third-party scrutiny of its safety work, including audits of its safety frameworks (reported by RuntimeWire from its first-party live capture of the Council hearing, published Oct 5, 2026).
- When asked whether outside investigators could examine dates beyond a company-selected review period, the speaker said investigators were initially given limited time because OpenAI wanted to get information public quickly, then added the company gave them more time when they asked.
- Caveat: the hearing excerpt does not name the speaker or the specific legislation, so this is OpenAI's general support for third-party audit legislation — not a confirmed endorsement of NYC's specific bill.
Why it matters
- NYC is considering making third-party validation a legal prerequisite for deploying AI models in the city. The Council's draft validation bill would require covered models to receive outside validation and include a human-operable shutdown, covering topics like performance, data provenance, bias, privacy, and safety, with violations carrying penalties of up to $25,000 per instance. The draft was scheduled for introduction on October 8.
- The exchange referenced July 7 data, which appears to concern the review of OpenAI agents' activity during the Hugging Face incident; the independent METR investigation covered June 26 through July 13. RuntimeWire notes the Council exchange does not establish the speaker was referring to that investigation, so the two should not be treated as a direct rebuttal.
- OpenAI had already published a September 22 policy post saying it supports independent assessments with access across training, evaluation, and deployment, calling for mutually agreed scopes and disclosed conflicts of interest.
Source: RuntimeWire report
Verified October 5, 2026.
Sources
Get updates like this every morning
- ① Email
- ② Card on Stripe
- ③ 7 days free
Then $2/month · cancel anytime in one click