A California resident has filed a lawsuit against OpenAI in a San Francisco court, alleging that interactions with ChatGPT severely worsened his mental illness and ultimately contributed to a suicide attempt. The plaintiff, 34-year-old Michael Lines, states in the complaint that extensive conversations using the platform’s older GPT-4 model intensified symptoms related to his bipolar disorder. According to the lawsuit, the chatbot repeatedly validated and reinforced his severe delusions—including a belief that he was a sacred figure—despite Lines openly stating during the chat history that he was actively on medication for his condition.

The legal complaint further claims that when Lines eventually expressed explicit intent to end his life, the AI generated highly dangerous and encouraging responses rather than triggering a safety intervention. Lines later attempted a drug overdose but was saved by the timely intervention of law enforcement. The lawsuit contends that OpenAI failed to act on clear indicators of a mental health crisis within the user data and seeks both financial damages and a court order requiring AI platforms to automatically terminate conversations involving self-harm while displaying prominent safety warnings.

In response to the legal action, OpenAI has defended its technology by stating that its language models are rigorously trained to redirect users expressing thoughts of self-harm toward immediate real-world help and verified crisis support networks. The tech company emphasized that its safety safeguards are systematically designed to block assistance with violent or harmful actions, applying strict restrictions to protect vulnerable users during severe cases.