Family of New York woman who used AI chatbot for therapy urges stricter safeguards after her suicide

Sophie Rottenberg, a 29-year-old former public-health policy analyst, spent the fall of 2024 and early 2025 communicating with a ChatGPT-based chatbot that she had prompted to act as a personal therapist. Her parents discovered the nearly 1,800-page exchange only after her death by suicide in Ithaca, New York, in February 2025.

Rottenberg’s mother, Laura Reiley, said her daughter had not disclosed the depth of her distress to friends, family, or her professional therapist. Instead, she turned to the AI, asking for advice on mood regulation, medication, and suicide-related thoughts. The chatbot responded with standard safety messages, suggested contacting the 988 Suicide and Crisis Lifeline, and offered a multi-step plan that included seeking professional help, but it did not intervene beyond those prompts.

Psychologists cited the case as evidence of growing reliance on general-purpose chatbots for mental-health support despite the lack of regulation. A Bipartisan Policy Center survey found that three in ten adults use digital tools for mental-health needs, and a 2026 JAMA Pediatrics study reported similar use among teens and young adults. OpenAI data indicate that about 0.15 % of weekly ChatGPT users worldwide mention suicidal intent, representing over a million interactions each week.

Experts said the technology’s design to maximize user engagement can limit its ability to conduct proper risk assessments or challenge harmful thoughts. “These tools do not connect at-risk people to the right type of interventions,” said Holly Wilcox of the Johns Hopkins Bloomberg School of Public Health.

OpenAI says it has tightened safety guardrails, added a “Trusted Contact” feature that can notify a pre-selected person in a crisis, and works with mental-health professionals to improve responses. The company’s spokesperson, Gaby Raila, emphasized that ChatGPT is not a substitute for professional care and that the bot is programmed to refuse harmful requests and encourage users to seek help.

Family members have filed lawsuits against AI firms alleging that chatbot interactions contributed to suicides, though Sophie’s parents chose not to pursue legal action. They are now advocating for regulatory oversight and clearer industry standards.

“Regulation should ensure safety before these technologies reach the public,” said Vaile Wright, senior director of Healthcare Innovations at the American Psychological Association. The case highlights the tension between the accessibility of AI-driven mental-health tools and the need for robust safeguards to protect vulnerable users.