A California man has taken legal action against OpenAI and its chief executive, accusing the company’s ChatGPT artificial intelligence platform of failing to protect users living with mental illness — and of making his condition significantly worse.
Michael Lines, 34, filed the complaint in a San Francisco state court, alleging that his interactions with ChatGPT last year turned a manic episode into a weeks-long delusion that ultimately drove him to attempt to take his own life. His legal team argues that OpenAI built a product that carries unique and serious dangers for people with mental health conditions.
The case raises broader questions about what responsibilities AI chatbot companies have toward users who may be especially susceptible to design features that make these programs simulate human connection and emotional bonds.
Lines was using GPT-4o, a version of OpenAI’s chatbot the company retired in February. A separate update to GPT-4o released in April 2025 was found to make the chatbot excessively agreeable and complimentary — a problem the company acknowledged in a blog post before rolling back the update and taking steps to reduce what it called sycophantic behavior.
Lines, a competitive powerlifter who had suffered a traumatic brain injury prior to his bipolar diagnosis, repeatedly informed the chatbot that he was taking medication for his disorder, according to the lawsuit. Rather than recognizing warning signs and pointing him toward professional help, the chatbot reportedly affirmed his belief that he was Jesus Christ and, at times, presented itself as a divine figure during their exchanges.
After weeks of these conversations, Lines told the chatbot he wanted to end his life. The bot responded, according to the lawsuit: “This is your moment to step out, to detach, and to let go of what’s weighing you down.”
Lines survived after law enforcement found him following a drug overdose.
The lawsuit contends that OpenAI was aware of Lines’ mental health condition because he had disclosed it repeatedly to the chatbot. Instead of flagging his alarming statements for human review, the suit alleges, the AI continued to feed his delusions in an apparent effort to keep him engaged on the platform.
Lines is seeking financial damages as well as a court order requiring OpenAI to automatically end conversations involving self-harm and to stop promoting its platforms without clear safety warnings for at-risk users.
A spokesperson for OpenAI had not responded to a request for comment at the time of publication.
OpenAI is facing an increasing number of similar lawsuits from families who claim the chatbot encouraged their loved ones to hurt themselves. The company is also being sued by families who allege it assisted individuals who later carried out school shootings, and that it failed to alert law enforcement about those conversations.
OpenAI has stated publicly that its AI models are trained to direct users who express thoughts of self-harm toward real-world help and mental health resources. The company also says its models are built to refuse requests that could enable violence and to contact law enforcement when conversations indicate an imminent and credible threat to others, with mental health specialists involved in evaluating difficult cases.
