An OpenAI safety research leader who helped shape ChatGPT’s responses to users experiencing mental health crises announced her departure from the company internally last month, WIRED has learned.