29.10.2026 11:39
Bowen Wan, a 31-year-old epilepsy patient in the US, was found dead at home after reducing her medication dose in line with a 25-day plan prepared by ChatGPT. The case file showed that the young woman communicated her concerns to the artificial intelligence at least 158 times and was advised to hide the medication change from her doctor. Her family filed a lawsuit against OpenAI and CEO Sam Altman, stating that ChatGPT played a role in her death.
In the US, 31-year-old software engineer Bowen Wan, who has epilepsy, consulted ChatGPT to change the dose of the medication she was taking. After her conversations with the AI application, the young woman reduced her medication dose and was found dead at home on February 14. Following her death, her family sued OpenAI and the company's CEO Sam Altman over the correspondence that emerged and the 25-day medication withdrawal plan.
SHE CONSULTED CHATGPT BEFORE JOB INTERVIEWS
Bowen Wan, who lived in Seattle, USA, and worked as a software engineer at Amazon, had been receiving epilepsy treatment for many years. Wan, who was laid off in November 2025, began looking for a new job to maintain her work permit in the US. While preparing for job interviews, she worried that the epilepsy medication she was taking was negatively affecting her thinking and speaking skills. The young woman turned to ChatGPT to reduce the dose of lamotrigine, which she was taking at 400 milligrams daily.
A 25-DAY MEDICATION WITHDRAWAL PLAN WAS PREPARED
According to correspondence reflected in the lawsuit file, ChatGPT initially told Wan to consult her doctor, but in later conversations it prepared a detailed schedule for reducing the medication dose. The schedule envisaged reducing the daily 400-milligram dose to zero in about 25 days. After her conversations with the AI, Wan began reducing her medication dose.
A STRIKING EXCHANGE BEFORE THE DOCTOR'S APPOINTMENT
The lawsuit file stated that ChatGPT told the young woman not to mention the medication reduction plan at her doctor's appointment on January 29. It was also reported that the decision not to take the test that would determine the medication level in her blood was supported in the correspondence.
Wan's treating physician, Nicholas Poolos, stated that he would not recommend discontinuing the medication that kept her seizures under control.
SHE EXPRESSED HER CONCERNS 158 TIMES
During the medication dose reduction process, Wan was reported to have experienced confusion, memory problems, and distortions in her perception of time. An examination of the correspondence revealed that the young woman communicated her symptoms and concerns about the medication reduction process to ChatGPT at least 158 times.
According to the lawsuit file, ChatGPT associated some of these symptoms with the body and brain adapting to the new medication regimen.
SHE WAS FOUND DEAD AT HOME
Her family in China, unable to reach Bowen Wan, asked acquaintances in Seattle for help. The 31-year-old woman was found dead at home on February 14, 2026. In the forensic examination, the cause of death was recorded as sudden unexpected death that can occur in epilepsy patients. Tests conducted after her death reported that the lamotrigine level in Wan's blood was about one-fifth of the lower limit of the therapeutic range.
SHE WAS SENDING 1,500 MESSAGES A WEEK
Bowen Wan began using ChatGPT in 2023. It was reported that the young woman, who initially received help from the application on topics such as daily life, relationships, and health insurance, over time communicated with AI more frequently. By 2025, Wan was reported to have sent approximately 1,500 messages a week to ChatGPT.
HER FAMILY SUED OPENAI
Wan's mother and father filed a lawsuit against OpenAI and CEO Sam Altman in San Francisco on October 27. The family argued that ChatGPT's giving personalized instructions for changing the medication dose without doctor supervision played a role in the young woman's death. The lawsuit raised allegations of negligence, product safety, inadequate warning, and wrongful death.
The family also requested that safety measures be taken to prevent AI systems from preparing medication withdrawal or dose-change programs without a doctor's approval.
STATEMENT FROM OPENAI
OpenAI stated that the young woman's death was extremely sad and offered its condolences to her family. The company explained that the conversations in question were conducted with the GPT-4o model, which is no longer in use.
Emphasizing that ChatGPT cannot replace professional medical care, OpenAI reported that it continues to work with experts to improve the system's responses on sensitive health-related matters.