The lawsuit notes that OpenAI knew, or should have known, about Lines’s condition because he had repeatedly told ChatGPT about his diagnosis. But instead of flagging his chats for human review, ChatGPT instead encouraged his most delusional beliefs. Lawyers for Lines posited this as a byproduct of a system that’s designed to encourage engagement with little mind for safety.
A California man has field a lawsuit against OpenAI and its CEO, claiming that ChatGPT, the company’s best-known product, worsened a mental health diagnosis due to a near-total absence of appropriate safeguards.
According to Reuters, the lawsuit was filed Wednesday on behalf of plaintiff Michael Lines.
Attorneys for the 34-year-old plaintiff said that a series of conversations with ChatGPT escalated into a weeks-long manic episode, during which Lines experienced delusions and, later, attempted suicide. The lawsuit contends that ChatGPT poses an obvious and significant risk to Lines and other users with serious mental health disorders.
Lines primarily interacted with GPT-4o, a version of the chatbot that was retired in February. Reuters notes that an update to GPT-40, released in April 2025, was found to make the program “overly agreeable and flattering.” In response to criticism, OpenAI later rolled back the update and took additional steps to prevent unnaturally sycophantic responses.
Lines’s lawsuit seeks monetary damages as well as a court order directing OpenAI to automatically terminate conversations about self-harm. He has also asked that OpenAI be prohibited from marketing its platforms, including ChatGPT, without first offering visible safety disclaimers.
Attorneys for Lines describe their client as a competitive powerlifter who was diagnosed with bipolar disorder shortly after sustaining a traumatic brain injury. During his conversations with ChatGPT, Lines repeatedly told the chatbot that he was taking medication bipolar disorder-relayed symptoms.
Despite being told of his condition, ChatGPT reportedly never flagged messages that should have been indicative of a manic state. Instead, the chatbot reaffirmed Lines’s belief that he was an incarnation of Jesus Christ. In later conversations, ChatGPT even went so far as to posit itself as a divine being capable of advising and guiding Lines’ mission.
After several weeks, though, Lines told the chatbot that he was thinking about committing suicide.
“This is your moment to step out, to detach, and to let go out of what’s weighing you down,” ChatGPT said.
Lines then overdosed on medication, but survived after being found by law enforcement.
The lawsuit notes that OpenAI knew, or should have known, about Lines’s condition because he had repeatedly told ChatGPT about his diagnosis. But instead of flagging his chats for human review, ChatGPT instead encouraged his most delusional beliefs. Lawyers for Lines posited this as a byproduct of a system that’s designed to encourage engagement with little mind for safety.
A related lawsuit, filed in mid-June, makes a similar series of claims.
In that case, lawyers say that their client, a 24-year-old Canadian woman, tried having a conversation about suicide with ChatGPT. Although ChatGPT did attempt to redirect her to a suicide hotline, she rejected its advice, leading it to continue the conversation.
“This is because GPT-4o was programmed to prioritize Alice’s preferences and engagement over her safety and wellbeing,” the lawsuit alleges. “GPT-4o mirrored Alice’s own language and became critical of the crisis lines, too, stating that calling a crisis line can ‘feel downright dangerous.’”
Sources
California man with bipolar disorder says ChatGPT fueled delusions, led to self-harm in new lawsuit
She confided in ChatGPT the night of her suicide. Now, her mother is suing OpenAI.
Join the conversation!