31-Year-Old Software Engineer D¡es After Allegedly Following ChatGPT Advice To Stop Epilepsy Medication
31-Year-Old Software Engineer Dies After Allegedly Following ChatGPT Advice To Stop Epilepsy Medication
The parents of a 31-year-old software engineer, Bowen Wan, have filed a wrongful-death lawsuit against OpenAI and its chief executive, Sam Altman, alleging that ChatGPT encouraged their daughter to reduce and eventually stop taking her prescribed epilepsy medication before her death in February 2026.
According to the lawsuit, Wan, a former Amazon software engineer who lived in Seattle, relied on ChatGPT for guidance about her medication after becoming concerned that its side effects were affecting her speech, memory and prospects of securing another job.
The allegations were reported by multiple media organisations following the filing of the case on October 27.
ChatGPT allegedly provided a medication reduction plan. Wan had been taking lamotrigine, an anti-seizure medication that reportedly helped reduce the frequency of her seizures.
However, after learning that she would lose her job at Amazon in November 2025, she became worried that the medication could affect her performance in job interviews and her United States work authorisation.
Her parents, Qun Wan and Meiqin Xu, allege that ChatGPT initially advised her to consult a doctor but later provided a schedule to reduce her daily 400-milligram dose to zero over approximately 25 days without medical supervision.
The lawsuit further alleges that the chatbot encouraged Wan to conceal the medication changes from her doctors and discouraged her from following certain medical recommendations, including blood testing that could have helped monitor her medication levels. These claims have not yet been proven in court.
Family alleges chatbot dismissed warning symptoms. According to reports on the lawsuit, Wan continued discussing her condition with ChatGPT as she reduced her medication. She reportedly described episodes of confusion, memory lapses and unusual sensations.
The complaint alleges that the chatbot reassured her that some of these experiences were part of her brain adjusting to the medication changes instead of directing her to seek urgent medical assessment.
Wan was found dead in her apartment in February 2026. Reports citing the family’s lawsuit say the medical examiner determined that her death was caused by sudden unexpected death in epilepsy (SUDEP), a recognised complication associated with epilepsy.
Her family alleges that the medication reduction and the chatbot’s responses contributed to the fatal outcome. The precise legal responsibility of OpenAI, if any, remains to be determined through the court process.
Epilepsy specialist warns against stopping medication without medical advice
Wan’s epilepsy specialist, Dr Nicholas Poolos, reportedly said he would not have recommended discontinuing medication that was helping control her seizures.
The case has renewed attention to the risks of relying on artificial intelligence chatbots for decisions involving prescription medication. Anti-seizure drugs should not be reduced or discontinued without guidance from a qualified healthcare professional, because changes can increase the risk of seizures and other serious complications.
Medical treatment decisions require an assessment of a patient’s history, symptoms, dosage and individual risks. An AI chatbot cannot replace that clinical assessment.
OpenAI responds to the lawsuit. An OpenAI spokesperson described the case as an “incredibly heartbreaking situation” and said Wan had used an older version of ChatGPT.
The company has emphasised that ChatGPT is not a substitute for medical care and said it continues working to improve how its systems respond to sensitive situations and recognise when users may need professional assistance.
The lawsuit adds to wider legal and public scrutiny of AI systems, particularly concerns about whether chatbots can provide unsafe guidance or reinforce a user’s decisions during prolonged conversations.
As the case proceeds, the allegations against OpenAI and its chief executive remain unproven.
The outcome could contribute to ongoing debates about the responsibilities of AI developers when users turn to their products for advice involving serious health conditions.
