The parents of Bowen Wan, a 31-year-old software engineer with epilepsy, are suing OpenAI, alleging that ChatGPT coached their adult daughter to stop taking seizure medication and hide the changes from her doctors before her death.
“It asks where generative AI stops and medicine begins, and who answers when a product crosses that line,” Benjamin S. Schenk, a partner at The Schenk Law Firm and an attorney for Wan’s family, said of the lawsuit.
Wan’s parents, Qun Wan and Meiqin Xu, filed the wrongful death lawsuit Oct. 27 in San Francisco Superior Court, according to their attorneys. Their 98-page complaint names four OpenAI entities and CEO Samuel Altman and seeks damages and safeguards against personalized medical advice.
The family’s account describes an accomplished young woman whose illness had narrowed her world. Wan, her parents’ only child, moved from China to the United States, graduated from Arizona State University and worked as a machine-learning software engineer at Amazon, according to the complaint.
Her epilepsy eventually forced her onto disability leave, The New York Times reported. She began using ChatGPT in 2023 for companionship and help with matters including insurance and benefits and later turned to it with questions about medication.
After learning in November 2025 that Amazon would eliminate her position, Wan worried that medication-related difficulties with speech and thinking would hurt her job interviews, according to the complaint. Finding another job also mattered to her ability to remain in the United States.
Related: ChatGPT Convinces Some People to Think They’re God
The complaint alleges that GPT-4o, the model she used, designed a schedule to reduce her lamotrigine to zero over approximately 25 days without a physician prescribing that plan. It also alleges that the chatbot encouraged her to conceal the reduction from doctors and praised her refusal of a blood test that could have exposed the change.
As Wan described confusion and other troubling symptoms, the chatbot repeatedly reassured her, according to the lawsuit and the Times’ review of her messages. The family alleges that its advice contributed to her death, rather than helping her recognize the need for medical care.
According to the complaint, Wan died on or about Feb. 12, 2026, and acquaintances found her in her Seattle home Feb. 14 after her parents could not reach her. The complaint cites the medical examiner’s finding of sudden unexpected death in epilepsy, known as SUDEP; it does not establish OpenAI’s legal responsibility.
“While ChatGPT is not a substitute for medical care, we believe A.I. can play a role in helping people make sense of health information, navigate insurance and complex health care systems, and prepare for appointments with medical professionals,” OpenAI spokesman Drew Pusateri told The New York Times.
Pusateri described Wan’s story as heartbreaking and said it involved an earlier version of ChatGPT that is no longer available. He also said newer models better recognize situations requiring urgent attention and that OpenAI works with clinicians to test its models and reduce harmful responses.
For families navigating medication questions, the general medical guidance is clear: MedlinePlus advises patients not to stop lamotrigine without talking to their doctor, even when they experience side effects. It warns that stopping suddenly can cause seizures.
The Centers for Disease Control and Prevention recommends taking seizure medicine as prescribed to reduce SUDEP risk and talking with a health care provider right away if seizures continue. That guidance applies broadly; it is not a medical finding about Wan’s individual case.
Families can keep a practical boundary in place: bring medication concerns and chatbot-generated suggestions to a qualified clinician, rather than treating a reassuring conversation as permission to change treatment. Staying connected to loved ones matters, too, especially when illness leaves someone isolated.
Read Next: Are We Surprised? ChatGPT’s New Parental Controls Fail to Protect Kids
Questions or comments? Please write to us here.
