Joanne Jang, General Manager of Labs at OpenAI, has departed the company after playing a foundational role in shaping how AI models interact with users. Jang joined OpenAI when the company had fewer than 200 employees and is widely credited with establishing 'model behavior' as a discipline — the practice of defining how AI systems should respond, what tone they should take, and how they should handle sensitive topics.
She led model behavior work across some of OpenAI's most consequential products, including GPT-4, DALL-E 2, Text-to-Speech, and the Memory feature that allows ChatGPT to remember user preferences across conversations. Her contributions earned her a spot on the TIME100 AI list in 2025, recognizing her as one of the most influential people in artificial intelligence.
Jang's departure represents the loss of deep institutional knowledge at OpenAI — she understood the nuanced tradeoffs in model personality and safety better than almost anyone at the company. Her exit continues a pattern of early employees leaving OpenAI as the company has grown from a small research lab into a commercial juggernaut valued at over $300 billion.
