John Schulman, co-founder of OpenAI, announced he is leaving to join rival Anthropic, the AI startup backed by Amazon.
He follows out of the door senior figures such as Jan Leike and Ilya Sutskever.
Schulman's move follows OpenAI's reported recent disbandment of its 'superalignment' team, which aimed to ensure control over advanced AI systems.
Fellow co-founder and chief scientist Ilya Sutskever announced in May that he was also leaving the AI startup.
At OpenAI, Schulman co-led the post-training team refining AI models for ChatGPT and developer interfaces. He also headed alignment science and served on a safety and security committee.
His new employer, Anthropic, launched a new AI model, Claude 3.5, in June that it claimed outperforms competitor models such as OpenAI's GPT-4o or Google's Gemini a wide range of tasks.