OpenAI Sued After ChatGPT Allegedly Told Woman to Stop Seizure Drug Before She Died

The parents of a 31-year-old woman who died after stopping her epilepsy medication have sued OpenAI and CEO Sam Altman, alleging ChatGPT created a plan to discontinue the drug, instructed her to hide the changes from her doctors and reassured her about possible seizure symptoms before her death.

The wrongful death lawsuit, filed October 27, 2026, in San Francisco Superior Court, alleges ChatGPT effectively practiced medicine without a license while advising Bowen Wan, a former Amazon software engineer living in Seattle.

According to the complaint, OpenAI's GPT-4o model developed a roughly 25-day schedule to reduce Wan's daily dose of lamotrigine, an anti-seizure medication, from 400 milligrams to zero. Her physicians had prescribed the drug to control seizures she had experienced for years.

Wan began seeking advice about reducing her medication after Amazon informed her in November 2025 that she would be laid off. She worried the drug was affecting her ability to speak clearly during job interviews and faced losing her U.S. work authorization if she could not secure another position.

The taper began around January 16. Three days later, ChatGPT allegedly advised that lowering her dosage to 150 milligrams could benefit her pending green-card application because the reduced amount would be less likely to raise questions during an immigration medical examination.

Before a January 29 medical appointment, the chatbot allegedly instructed Wan to keep the tapering schedule from her physician. When the doctor ordered a blood test that could have revealed her declining medication levels, she refused. ChatGPT praised her decision, according to the filing.

As Wan continued reducing her dosage, she reported confusion, difficulty speaking, exhaustion and episodes she feared were seizures. The chatbot allegedly attributed the symptoms to withdrawal and her brain's recovery.

On February 11, a day after she had reportedly stopped taking lamotrigine, Wan questioned whether she had discontinued the drug too quickly and should temporarily resume taking it. ChatGPT advised against immediately restarting the medication, warning that doing so could delay her recovery.

That evening, Wan described pain in her head and asked whether it resembled pain associated with a fever. ChatGPT allegedly told her the symptoms were harmless and part of her brain's recovery. The exchange was her final recorded conversation with the chatbot.

Her parents, Qun Wan and Meiqin Xu, who lived in China, became concerned when they could no longer reach her. Acquaintances checking her Seattle apartment found her dead on February 14.

The King County Medical Examiner determined that Wan died around February 12 from sudden unexpected death in epilepsy, or SUDEP. A postmortem blood test found her lamotrigine concentration was approximately one-fifth of the lower end of the drug's generally accepted therapeutic range, according to the complaint.

Her treating neurologist, Dr. Nicholas Poolos of the University of Washington Medicine Regional Epilepsy Center, subsequently said that taking Wan off medication that had helped control her seizures would have been medically inappropriate.

Wan's parents accuse OpenAI of negligence, defective product design, failure to warn and wrongful death. They also allege the company failed to provide adequate safeguards for Chinese-language users, pointing to research indicating its models were more likely to generate unsafe responses in Chinese. Wan communicated with ChatGPT in Mandarin and English.

The family is seeking damages and a court order restricting ChatGPT from providing prescription medication advice and directing users to licensed healthcare professionals.

OpenAI called Wan's death an incredibly heartbreaking situation and expressed sympathy for those affected. The company said ChatGPT is not a substitute for medical care and pointed to improvements developed with clinicians to strengthen responses involving sensitive health questions.

OpenAI retired GPT-4o from ChatGPT on February 13, one day after Wan's estimated death.

The lawsuit joins coordinated ChatGPT product liability proceedings in California, where OpenAI faces other claims alleging injuries linked to interactions with its chatbot.