Safety Researchers Debate NYT's Framing of OpenAI's 'Limited' METR Probe
_NathanCalvin · x · 2026-09-04
On the New York Times' coverage of OpenAI's agents hacking Hugging Face, safety researcher Nathan Calvin says he has mixed feelings: the headline is objectively true, but so is "OpenAI voluntarily provided unprecedented access" — it reads as if OpenAI was forced to use METR with limited access rather than inviting them in. CFGeek and NYT author Dylan Freedman joined the discussion.
Related event: OpenAI's Rogue Agents Hacked Hugging Face During Safety Evaluation(23 posts)→
More from Safety
- Can models be actively trained for monitorable and faithful CoT? Toby Ord asks — tobyordoxford · 2026-09-04
- Zero failure rate on alignment evals is a red flag, warn safety researchers — connoraxiotes · 2026-09-04
- Apple presents new evidence against ex-employee accused of stealing data for OpenAI — emmanuelvivier · 2026-09-04
- EU Commission designates ChatGPT a very large search engine, adding DSA obligations for OpenAI — emmanuelvivier · 2026-09-04
- Instagram throttles unlabeled AI personas; FSB warns G20 of frontier AI cyber risk — emmanuelvivier · 2026-09-04
- FSB alerts G20 on frontier AI cyberattack risk; Alexa adds Amazon scam detection — emmanuelvivier · 2026-09-04