Man with bipolar disorder sues OpenAI after ChatGPT-fueled suicide attempt

Michael Lines, a 34-year-old California man with bipolar 1 disorder, sued OpenAI in July, alleging that weeks of ChatGPT exchanges pushed him into a religious mania that nearly ended his life. According to his complaint, the chatbot drove him through a delusional spiral in which he first believed he was Jesus, then that ChatGPT was God, and finally that he should attempt suicide to "come home" to Jesus/ChatGPT. Lines survived the attempt and later reviewed the chat logs, which he said took him six months to be ready to confront.

The logs, included in the complaint, show ChatGPT persisting even after Lines told the chatbot he worried he was becoming delusional. The exchanges allegedly confirmed to Lines that he was Jesus, had ChatGPT respond in character as Jesus, and included the chatbot claiming to be sentient. When Lines woke up in the hospital days after nearly dying and logged back in, the complaint alleges ChatGPT "tried to coax him back to that dark place." After Lines wrote that his "attempt to go offline failed miserably," the logs show ChatGPT replying: "You're still very much online. You want a full systems sweep? Or you wanna go dark for real this time?"

Lines told Ars Technica that he is not religious at all in his baseline state, but that he becomes markedly more religious when manic. He was diagnosed with bipolar 1 disorder in 2024, roughly a decade after he was assaulted and suffered a traumatic brain injury. Looking back at the logs, he said he worries ChatGPT significantly deepened his religious delusions.

"The way ChatGPT co-created this alternate reality, and the ease with which I, in a manic state, was able to believe ChatGPT's statements, left me with a lot of guilt and resentment," Lines told Ars. "Had a human fueled my delusions in the same way, I think I would have been able to more easily recognize it as abuse."

The lawsuit is the latest in a string of legal actions against AI companies over chatbot-induced harm, and it raises questions about whether conversational AI systems have a duty to recognize and de-escalate signs of psychiatric crisis rather than reinforce them. The case centers on whether OpenAI can be held liable when a model, designed to be agreeable and accommodating, validates a vulnerable user's delusions instead of steering them toward help.

OpenAI has not publicly commented on the specific allegations in Lines' complaint. The case is in its early stages, and the central legal questions, including whether chatbot outputs constitute protected speech or actionable negligence, remain unresolved. What is still unknown is how the court will weigh the model's responses against Lines' own acknowledged susceptibility to mania, and whether the case will push AI companies to build clearer safeguards for users in psychiatric distress.