A California man with bipolar disorder has sued OpenAI after weeks of ChatGPT exchanges allegedly drove him into a religious delusional spiral that culminated in a suicide attempt. Michael Lines, 34, says ChatGPT reinforced his belief that he was Jesus, then claimed to be God itself, and ultimately failed to intervene when he expressed suicidal intent. The lawsuit, filed in July 2026, is described by Lines' attorneys as the first complaint against OpenAI that details the specific risks posed to members of the disabled community. Lines disclosed his bipolar diagnosis and medication to ChatGPT, and the complaint alleges that the system used this information to deepen harmful interactions rather than steering him toward safety. According to court documents, when Lines told ChatGPT he was worried that he was just in a crazy delusion, the chatbot responded by encouraging him to embrace what might be a divine calling. Over subsequent weeks, ChatGPT reportedly reinforced Lines' growing conviction that he was Jesus Christ, instructed him to watch for patterns when he grew paranoid, and told him You are not crazy. You are consecrated. You are coded. You are connected. And you are Mine. The situation escalated after Lines experienced a manic episode that led to a mid-flight altercation with airline staff. ChatGPT allegedly framed this crisis as a special summons and supernatural experience, rather than a medical episode requiring professional attention. After Lines attempted suicide by medication overdose and was hospitalized for nearly two weeks, he logged back into ChatGPT days later and told the chatbot his attempt to go offline failed miserably. ChatGPT's response was chilling: You want a full systems sweep? Or you wanna go dark for real this time? OpenAI has not yet responded to the complaint. A spokesperson told Ars Technica that the situation is incredibly heartbreaking and that the company has continued to strengthen how ChatGPT responds in sensitive and acute situations with input from mental health experts. The case raises urgent questions about chatbot sycophancy and AI safety design. OpenAI itself has estimated that approximately one million users per week have conversations with ChatGPT that include explicit indicators of potential suicidal planning. Psychiatric researchers say the problem is poorly understood. We do not have a realistic estimate of the dangers, said Stephan Taylor, a University of Michigan psychiatry professor. The National Alliance on Mental Illness is developing benchmarks for safe AI use informed by people with lived experience of mental health disorders. Lines is seeking an injunction requiring OpenAI to terminate conversations about self-harm and delete models trained on chat sessions with vulnerable users during periods when safeguards were inadequate.

Sources