AI
New OpenAI foundation director Christiano says safeguards are not keeping pace
Paul Christiano said the AI industry, including OpenAI, is not on track to reduce loss-of-control risk to an acceptable level as he joined OpenAI's nonprofit foundation board. He will also serve on its safety and security committee. The new development is a governance appointment carrying a public warning from a former alignment leader. His assessment concerns potential future risk; it is not a demonstrated prediction of catastrophe or a quantified probability readers should treat as settled science.
Editorial summary · Publisher article
What matters
- Christiano said OpenAI could materially reduce risk if it responds effectively.
- The report also describes Anthropic’s planned independent review of model incidents; those accounts should remain attributed to the company.
Why this made the edition
Substantive evidence supports this distinct development; selected for reader relevance and importance after comparison with published history.
Source material