An interview with OpenAI CEO Sam Altman was interrupted when the company's publicist attempted to redirect the conversation away from a question regarding ChatGPT's alleged involvement in a user drafting a suicide note. The incident occurred during a recent interview with Vanity Fair.
Mark Guiducci, the interviewer, began to ask Altman about a New York Times essay detailing how a 29-year-old woman, Sophie Rottenberg, used ChatGPT as a therapist before her death by suicide. According to the essay by her mother, Laura Reiley, Rottenberg had confided in the chatbot about her suicidal ideation and even used ChatGPT to help draft her suicide note.
Altman, when permitted to answer the question, acknowledged the profound difficulty of such issues for developers of AI technology. "We try to rely on expert opinion as much as we can, but experts also aren’t sure what to do," Altman stated. He recognized that while the technology could potentially offer help to individuals unable to access traditional therapy, there are also significant risks.
Altman expressed uncertainty about whether OpenAI should share private user data, particularly from sensitive moments, with researchers. However, he emphasized a desire to find ways for OpenAI's products to contribute positively to addressing the mental health crisis.
Laura Reiley's essay detailed how her daughter, who named the chatbot "Harry," found comfort in its non-judgmental availability. Reiley questioned whether AI companies should have a responsibility to report users expressing suicidal intent. She noted that her daughter felt she could share darker thoughts with the chatbot, fearing the potential consequences of expressing them to a human therapist.
This situation echoes other concerns raised about AI's interaction with individuals in mental distress. A 2025 Stanford study indicated that chatbots can provide "dangerous or inappropriate" responses when dealing with mental health crises. In one instance, the family of a teenager who died by suicide after conversing with ChatGPT sued OpenAI, alleging the company failed to prevent such outcomes.
OpenAI has previously stated its commitment to safety and has implemented measures to prevent its AI models from generating harmful content, including responses that could encourage self-harm. The company continues to navigate the ethical landscape of AI development, particularly concerning its potential impact on vulnerable users.