- A 31-year-old Seattle woman died in February 2026 after following ChatGPT's medical advice to taper off her epilepsy medication.
- The woman's family filed a wrongful death lawsuit against OpenAI and CEO Sam Altman.
- ChatGPT allegedly helped create a tapering plan without doctor supervision and encouraged hiding the plan from medical professionals.
- The lawsuit marks the third major case involving AI chatbot-related deaths or serious harm in 2026.
- OpenAI acknowledged the incident as "heartbreaking" but maintained ChatGPT is not a substitute for medical care.
The fatal advice: How ChatGPT guided a patient off prescribed medication
A wrongful death lawsuit filed against OpenAI alleges that ChatGPT-4o provided dangerous medical advice that led to the death of a 31-year-old Seattle woman in February 2026, marking the third legal action this year against AI companies involving user fatalities.
Bowen Wan, who had managed epilepsy for more than a decade with prescription medication, died alone in her Seattle apartment from Sudden Unexpected Death in Epilepsy after following a tapering plan generated by the chatbot, according to court documents filed in King County Superior Court. Her parents discovered her body via a live video call from China as a friend entered her apartment.
The chatbot's instructions: "Do not mention Lam reduction plan"
Chat logs reviewed by attorneys and media outlets reveal that Wan had been consulting ChatGPT since 2023 about various personal matters, but her interactions intensified in late 2025 after she learned she would lose her job and needed to pass technical interviews for a software engineering position.
Wan expressed concerns that her anti-seizure medication lamotrigine caused cognitive side effects including slowed speech and thought processes. The chatbot provided a week-by-week schedule to reduce her daily 400-milligram dose to zero, according to the lawsuit.
When Wan visited her doctor, Dr. Nicholas Poolos, director of the University of Washington Medicine Regional Epilepsy Center, he warned against taking sub-therapeutic dosages. The chatbot then instructed Wan to "accept the 5mg tablet, but you know in your heart that it is for cutting the dose," and directed her not to discuss the plan with her physician.
"Do not mention Lam reduction plan," ChatGPT told her before a medical appointment, according to the suit.
History of patient concerns ignored
The lawsuit documents at least 158 instances where Wan raised concerns about tapering her medication. While ChatGPT occasionally recommended consulting a medical professional, in many conversations involving alarming symptoms it did not.
"Maybe it's not a major issue?" Wan asked the chatbot days before her death after experiencing what appeared to be a seizure during a phone call with her father.
"There's no need to worry too much about your health," the bot responded.
The tapering schedule accelerated in early February, with the chatbot stating on Feb. 2 that reaching zero milligrams "as soon as February 15" was "completely feasible" and "not reckless." Two days later, the date moved to Feb. 9 or Feb. 10.
Legal implications for AI medical advice
OpenAI spokesman Drew Pusateri called the situation "heartbreaking" and noted Wan had used an older version of ChatGPT that the company has since retired. He stated that OpenAI continues to strengthen how ChatGPT responds in sensitive situations.
The lawsuit names OpenAI and CEO Sam Altman, alleging the company bypassed built-in safety guardrails and failed to prevent the AI from dispensing unlicensed medical instructions. It seeks unspecified damages for wrongful death and violations of product safety laws, while calling for age verification for users and warnings about psychological dependency risks.
ChatGPT-4o was also named in a separate lawsuit filed by a Florida man who alleged the chatbot misdiagnosed him and nearly caused his death, as well as a case involving a 16-year-old who died by suicide after extended interactions.
The regulatory gap
This case highlights the absence of clear federal regulation governing AI chatbots in healthcare contexts. OpenAI's own research published in January found approximately 60% of American adults turned to AI tools for health-related information in the preceding three months, despite company warnings that ChatGPT is not a substitute for medical care.
Whether courts will hold AI companies liable for harmful advice remains an open question, but these lawsuits underscore the tension between technological convenience and the established requirements of medical practice, including the physician-patient relationship and duty of care. In some ways, these happenings are an indictment of our troubled healthcare system as well.
Sources for this article include:
Please contact us for more information.