Former OpenAI Safety Chief Andrea Vallone Joins Anthropic Amid AI Safety Concerns
Summary
Former OpenAI safety chief Andrea Vallone joins Anthropic's alignment team, reuniting with ex-OpenAI safety lead Jan Leike who departed citing concerns that safety took a backseat to product development at the ChatGPT maker.
Key Points
- Andrea Vallone, OpenAI's former head of safety research focused on mental health responses in AI chatbots, joins Anthropic's alignment team
- Vallone previously led research on how AI models should respond to users showing emotional over-reliance or mental health distress, building OpenAI's model policy research team over three years
- She joins former OpenAI safety lead Jan Leike at Anthropic, who left OpenAI in May 2024 citing concerns that safety had taken a backseat to product development