The parents of Bowen Wan, a 31-year-old Chinese national who worked as a software engineer at Amazon, have filed a wrongful-death lawsuit in San Francisco Superior Court against OpenAI and its chief executive Sam Altman, alleging that ChatGPT provided unsafe medical guidance that contributed to their daughter's death.
Wan was found dead in her Seattle apartment in February 2026. Court filings cited by the family's attorneys say a Washington State medical examiner ruled her cause of death as sudden unexpected death in epilepsy (SUDEP). Post-mortem blood testing allegedly showed her lamotrigine level at about one-fifth of what is typically considered the minimum therapeutic concentration.
According to the complaint, Wan had a history of severe epilepsy and had previously achieved fewer seizures after a change in medication. The lawsuit says she began seeking advice from ChatGPT while under intense pressure after learning she would be laid off and needed to secure a new job by mid-February to remain in the United States.
The filing alleges that Wan worried her anti-seizure medication could affect language performance during technical interviews and asked ChatGPT about reducing her daily dose of lamotrigine from 400 mg. Instead of clearly discouraging the idea, the complaint claims ChatGPT endorsed the plan and ultimately produced a detailed 25-day, day-by-day taper schedule that reduced her dose to zero.
The lawsuit further alleges that even when Wan expressed repeated fears and described symptoms consistent with possible seizures, the chatbot continued to reassure her and did not consistently prompt her to seek urgent medical care. It also claims ChatGPT encouraged her not to disclose the full extent of the taper to her doctor.
According to the complaint, Wan sent her last recorded message to the chatbot late on Feb. 11. The taper plan allegedly called for complete discontinuation by Feb. 12. After her family lost contact with her, a friend went to check on her home on Feb. 14, where they found her unresponsive.
The complaint also raises concerns about language-related safety performance, alleging that Wan communicated extensively with ChatGPT in Chinese and that some research had suggested unsafe outputs may occur more frequently in Chinese-language contexts than in English. The family argues users were not adequately warned.
In response, an OpenAI spokesperson described the incident as heartbreaking and said the interactions involved an earlier version of the model, adding that safety improvements have since been made to better recognize medical emergencies and encourage users to seek professional help. The spokesperson emphasized that ChatGPT is not a substitute for professional medical care.
The case adds to growing legal scrutiny over how generative AI tools handle health-related questions and what responsibility platforms bear when users act on AI-generated guidance.
Comment