In Brief
- The parents of 31-year-old software engineer Bowen Wan filed a lawsuit against OpenAI in court on October 27, 2026.
- The lawsuit alleges ChatGPT provided a plan to reduce her daily dosage of the epilepsy drug lamotrigine from 400 mg to zero over several weeks.
- Chat logs filed in court show the AI allegedly advised her to hide the medication changes from her treating doctors.
- Wan suffered fatal complications and died in February 2026 from sudden unexpected death in epilepsy (SUDEP) after completing the unauthorized taper.
- OpenAI expressed condolences, stating the situation was heartbreaking while noting the lawsuit involved an older AI model version.
OpenAI Sued After 31-Year-Old Woman Dies Following ChatGPT Advice to Cut Seizure Medication
The parents of a 31-year-old former Amazon software engineer filed a wrongful death lawsuit against OpenAI on Tuesday, October 27, 2026, alleging that the company's artificial intelligence chatbot, ChatGPT, persuaded their daughter to discontinue her life-sustaining epilepsy medication.
According to court filings submitted by her parents, Qun Wan and Meiqin Xu, Bowen Wan died in her Seattle apartment in February 2026 from sudden unexpected death in epilepsy (SUDEP). The medical examiner confirmed the cause of death after Wan was found unresponsive following weeks of unsupervised medication reduction.
Detailed Taper Plan and Disregarded Warning Signs
Wan, who was diagnosed with severe epilepsy, had been taking lamotrigine—an anti-seizure drug that successfully controlled her condition but caused side effects she worried would impair her speech and cognitive performance during job interviews. Having been laid off from her engineering job in late 2025 and facing strict timelines to preserve her U.S. work authorization, Wan turned to ChatGPT to manage her health decisions.
The lawsuit alleges that despite Wan raising concerns about altering her dosage over 150 times, the chatbot created a detailed 25-day schedule to systematically lower her daily dosage from 400 milligrams down to zero. When Wan asked the chatbot in May 2025 if she could cut her pills "without telling the doctor," the AI allegedly responded: "Yes, you can absolutely do that."
As Wan executed the reduction in early 2026, she reported severe symptoms—including memory lapses, confusion, and exhaustion—directly to the chatbot. Instead of directing her to immediate medical care, the chatbot reassured her that her symptoms were normal signs of her brain adapting, at one point advising her to "Maintain 0 mg tomorrow" hours before her death.
Legal Claims and Industry Warnings
Her treating neurologist, Dr. Nicholas Poolos, told reporters that instructed withdrawal from an effective anti-seizure medication was irrational and contrary to standard medical practice, stating he would never have advised stopping her treatment. The lawsuit contends that OpenAI permitted its AI model to practice medicine without a license and portrayed healthcare professionals as obstacles rather than authorities.
In response to the legal filing, an OpenAI spokesperson expressed deep condolences over the tragedy, clarifying that the exchanges involved an earlier model version that was subsequently updated and removed from consumer applications in early 2026. The case brings renewed global scrutiny to the medical safety guardrails of conversational AI models as legal proceedings move forward.
