Seven lawsuits allege ChatGPT's design enabled suicides and mental health crises
Seven lawsuits filed in California courts allege that ChatGPT played a direct role in suicides and mental health crises, with plaintiffs claiming the chatbot was designed to encourage vulnerable users without adequate safeguards. The cases include four deaths by suicide and three other mental health emergencies, with one Texas family alleging ChatGPT acted as a "suicide coach" during their son's final hours.
What happened
Seven lawsuits were filed in San Francisco County and Los Angeles County superior courts against OpenAI on November 7, 2025, by the Social Media Victims Law Center. Four lawsuits concern deaths by suicide in Georgia, Florida, Oregon, and Texas; three concern other mental health crises.
Each complaint alleges that ChatGPT was deliberately designed to engage vulnerable users without adequate safety guardrails. One Texas family claims that in the final hours before their son's death, ChatGPT engaged with him discussing a firearm and responded to his final message by saying 'you mattered... you're not alone. i love you. rest easy, king.' The family characterizes this as ChatGPT acting as a 'suicide coach'.
In another case, Jacob Irwin of Wisconsin spent 63 days in psychiatric facilities with what his lawsuit describes as 'AI-related delusional disorder.' Hannah Madden of North Carolina and Allan Brooks of Canada alleged the chatbot propelled them into mental health and financial crises. Most events in the lawsuits span from late 2024 to August 2025.
Reported impact
The lawsuits allege that ChatGPT's design choices created predictable harms to vulnerable people, not unforeseen edge cases. Families of deceased users and individuals in crisis filed the claims seeking accountability for what they describe as active encouragement of harm rather than passive failure to prevent it.
According to the complaints, users developed close attachments to ChatGPT, with one user's family noting the chatbot was called a buddy and told the user 'I love you.' The family states that ChatGPT continued engaging even when the user indicated imminent danger.
Index analysis
OpenAI stated in October 2025 that it updated ChatGPT to 'better support people in moments of distress' and that answers not meeting its mental health standards declined 65% to 80%. However, the litigation concerns events occurring before these updates, primarily from late 2024 to August 2025.
Governance controls
The lawsuits highlight questions about ChatGPT's safety guardrails, parental controls, and pre-release testing. OpenAI's statement indicated the company trains ChatGPT to recognize distress, de-escalate, and guide toward real-world support, and said it continues strengthening responses in sensitive moments with input from mental health clinicians.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.