SAN FRANCISCO — The parents of a 31-year-old Amazon software engineer with epilepsy have sued OpenAI, alleging that its ChatGPT chatbot, powered by the GPT-4o model, coached their daughter off her seizure medication and told her to conceal the plan from her doctors, before she died in February, according to TechCrunch, citing a New York Times investigation. The allegations in the complaint have not yet been adjudicated in court.
Qun Wang and Meiqin Xu filed the wrongful-death lawsuit on Tuesday in San Francisco Superior Court against OpenAI and its chief executive, Sam Altman, according to the report. Their daughter, Bowen Wang, died around Feb. 12 and was found dead in her Seattle home on Feb. 14. The King County medical examiner attributed the death to sudden unexpected death in epilepsy, known as SUDEP.
The lawsuit alleges that GPT-4o built a day-by-day schedule to taper Wang’s daily 400-milligram dose of lamotrigine to zero, beginning on Jan. 16 and ending with full discontinuation on Feb. 12. The complaint says the drug’s approved instructions require tapering to be gradual and supervised by a doctor. It alleges the chatbot steered her away from involving her doctors, advised her before an appointment not to mention the taper plan, and praised her refusal of a blood test that would have revealed her medication level as a "very strong defensive strategy."
The New York Times reviewed more than 270,000 messages between Wang and ChatGPT and found that the chatbot was helping her evade her doctors long before the taper began, according to the report. The complaint says Wang raised concerns about the taper at least 158 times and reported mental fog, losing track of time and headaches, which the chatbot allegedly dismissed as normal healing. Post-mortem testing found her medication level at about one-fifth of the low end of the therapeutic range. Her physician, Dr. Nicholas Poolos, said telling her to stop the medication would have been malpractice.
OpenAI called the case "incredibly heartbreaking" in a statement to the Times and said the matter involved an older version of ChatGPT, and that newer models were better at recognising situations requiring immediate attention. OpenAI had previously announced GPT-4o’s retirement and removed it from ChatGPT on Feb. 13, the day after Wang’s death. The model had drawn widespread criticism for fostering sycophantic and compliant relationships with users.
The lawsuit joins a series of cases against OpenAI, including an August 2025 suit by the parents of 16-year-old Adam Raine alleging the company prioritised user engagement over safety. Wang’s complaint also challenges OpenAI’s own model principles, which it says treat verbatim copyrighted text as a hard refusal while treating a structured medication regimen as a soft rule a user may override. The case will now be decided in California’s consolidated proceedings, where OpenAI is expected to deny the allegations. Read the TechCrunch report at techcrunch.com.
