Anthropic has granted external evaluator METR "employee-like access" to its facilities and systems, marking the first implementation of a September pledge by Dario Amodei that was also supported by rivals Sam Altman and Elon Musk. The move aims to increase transparency and safety oversight as AI models become more powerful, though critics warn of potential conflicts of interest and argue for greater government regulation.
Background
The AI safety debate has intensified as models grow more capable, with leading labs facing pressure from governments and the public to implement stronger safeguards and independent oversight mechanisms.
- Source
- Ars Technica
- Published
- Oct 19, 2026 at 12:25 AM
- Score
- 7.0 / 10