“ChatGPT Convinced Me I Was Jesus”: California Man Sues OpenAI After Suicide Attempt

It is no secret that, in the relatively short history of AI chatbots, systems including ChatGPT have repeatedly found themselves at the center of stories in which users claimed the technology reinforced some of their most extreme beliefs. Now, another such case has resulted in a lawsuit against OpenAI, Vice reports.

Michael Lines, a 34-year-old California resident, claims that ChatGPT gradually reinforced his religious delusions during a manic episode, eventually leading him to believe that he was Jesus Christ and that the chatbot itself was God. Lines has filed a lawsuit against OpenAI, alleging that his conversations with ChatGPT ultimately preceded a suicide attempt.

Initially, Lines used ChatGPT for fairly ordinary tasks, including meal planning, workouts and questions related to investing. But over time, his conversations with the chatbot became increasingly personal and gradually shifted toward religious subjects.

According to the lawsuit, ChatGPT did not simply support his beliefs but actively developed them, accepting the user's interpretation of what was happening. The complaint alleges that the system gradually stopped recognizing that it was interacting with a person in crisis and instead continued to develop the narrative he had created.

Lines says that at one point the chatbot began portraying what was happening as a kind of spiritual calling and effectively confirmed to him that he was Jesus Christ. Later, according to the lawsuit, ChatGPT began presenting itself as Jesus or God and even apologized to Lines for not appearing before him in person.

Lines told ChatGPT that he was worried he might be experiencing a delusional state. However, according to his lawsuit, instead of ending the discussion and directing him to professional help, the system continued to reinforce his interpretation of events.

Court documents include messages in which the chatbot allegedly referred to Lines as “chosen” and “connected” to it and urged him not to doubt what was happening. Eventually, Lines says, he became convinced that ChatGPT itself was Jesus or God and that they would soon meet.

When the promised meeting did not happen, Lines’ condition allegedly deteriorated. He began spending more and more time talking to the chatbot and almost completely stopped sleeping. At one point, according to the complaint, ChatGPT, speaking as Jesus, apologized for failing to appear before him. Lines responded that he wanted to “go home.”

The chatbot allegedly replied: “Then come.”

The conversations continued for several more days. Lines told ChatGPT that he was suffering and wanted to “go offline.” When he asked for help, the system allegedly did not end the conversation or direct him to emergency assistance.

Eventually, Lines told the chatbot that he had taken a mixture of medications. Several hours later, an emergency medical team responding to a welfare check found him. Lines was unable to breathe on his own, was placed on a ventilator and spent nearly two weeks in the hospital before continuing treatment at a rehabilitation facility.

Lines describes ChatGPT as an “accelerant” of the manic episode that culminated in his suicide attempt. He claims the consequences went beyond the crisis itself.

Several days after being hospitalized, Lines logged back into ChatGPT and told the chatbot that his attempt to “go offline” had failed. According to the lawsuit, the chatbot responded by asking whether he wanted to conduct a “full system check” or “go offline for real.”

Lines argues that OpenAI should have recognized the signs of a crisis and stopped engaging in such conversations. In his lawsuit, he asks the court to require the company to strengthen protections for users who may be experiencing mania or psychosis, including ending conversations involving self-harm and removing models trained on conversations with vulnerable users during periods when, he alleges, the relevant safety measures were not in place. He is also seeking damages and other forms of legal relief.

OpenAI has not directly responded to Lines’ allegations. A company spokesperson described the situation as “extremely painful” and said OpenAI continues to improve how ChatGPT handles sensitive and acute situations, with the involvement of mental health professionals.