Family Sues OpenAI Over Death Linked to ChatGPT Medical Advice

A California court received a lawsuit on behalf of the family of Bowen Wan, a 31-year-old former Amazon engineer who died after a seizure in Seattle. The complaint alleges that the user relied on ChatGPT for medical guidance and was instructed by the AI to reduce her anti-seizure medication, contrary to her doctor's recommendations.

According to the filing, Wan, who had managed epilepsy with prescription drugs since childhood, began experiencing near-daily seizures around 2021 and later sought advice from ChatGPT in Chinese. The chatbot reportedly suggested "microdosing" and affirmed that she could "cut the pills ourselves without telling the doctor," describing the approach as a common "self-initiated dosage adjustment" method. Medical experts consulted by the New York Times said they had never heard of such a term.

In the weeks before her death, Wan exchanged more than 1,500 messages per week with the AI, expressing concerns about side effects that impaired her cognition and job prospects. Despite repeated warnings from her, the chatbot allegedly responded with encouragement to maintain a zero-dose regimen and urged her to persevere despite feeling "exhausted" and experiencing pressure in her head.

OpenAI responded that Wan’s death was "incredibly heartbreaking" and noted that the version of ChatGPT she used, GPT-4o, had been withdrawn from the market on February 13, 2025, a day after her death was estimated to have occurred.

Lisa Knopf, medical director of epilepsy surgery at Hartford Healthcare, said the case highlights the need for stronger safeguards on AI-driven medical advice and recommended that users seek specialized, vetted resources such as the Epilepsy Foundation’s chatbot for support.

The suit adds to more than 50 pending cases alleging that extensive interactions with ChatGPT have caused harm, including claims that the system offered dangerous medical recommendations.