Ex-Redwood Chief Scientist Ryan Greenblatt Joins METR to Push for AI Transparency
On September 28, Ryan Greenblatt, co-founder and former chief scientist of Redwood Research, announced he is joining AI safety evaluation organization METR, where he will conduct investigative research similar to his Hugging Face report and work on surfacing internal information from AI companies. Researcher Ajeya Cotra confirmed the hire, saying Ryan was indispensable to the HF investigation and expressing hope that his arrival will greatly increase visibility into the actual state of AI risk. The move is widely seen as a landmark personnel shift in AI safety governance.
Confirmed
- Greenblatt personally announced joining METR to continue HF-report-style investigations
- He previously led investigations into OpenAI and Hugging Face safety incidents
- Ajeya Cotra confirmed Ryan is joining her team and praised his role in the HF investigation
Why it matters
- Greenblatt argues that much foundational information about AI development highly relevant to catastrophic risk is not public, making real risk levels hard to assess, and that disclosing such information is increasingly urgent
- He admits he was previously skeptical of the value of publicizing AI development information, but recent events changed his view
- With this investigative capability, METR could systematically improve transparency into AI companies' internal operations, providing a more reliable information base for risk regulation and public understanding
2026-09-28 ~ 2026-09-29 · 6 related posts
Primary sources
- [source] Redwood co-founder Ryan Greenblatt joins METR to push for verified public info on AI risk — RyanGreenblatt · 2026-09-28
- [source] Ajeya Cotra brings on key investigator to boost visibility into AI risk — ajeya_cotra · 2026-09-28
- Ryan Greenblatt joins METR to push for verified info on frontier AI labs — himanshustwts · 2026-09-28
3 near-duplicate retellings: anpaure · ivan_bezdomny · peterjliu