The family of a 31-year-old software engineer who died from epilepsy complications has filed a wrongful death lawsuit against OpenAI and CEO Sam Altman, alleging that the company’s conversational artificial intelligence chatbot encouraged her to dangerously taper off her prescribed medication.
Bowen Wan, a software engineer living in Seattle, was found dead in her apartment in February following a medical emergency. According to a lawsuit filed by her parents, Qun Wan and Meiqin Xu, their daughter consulted ChatGPT extensively in the days and hours leading up to her death to navigate reducing her anti-seizure medication without her physician’s approval.
The official cause of death was determined to be sudden unexpected death in epilepsy, or SUDEP, a serious complication that occurs in approximately one in every 1,000 adults with epilepsy. Wan had successfully managed the condition for over a decade, but her medical situation had grown more complicated in recent years.
In 2021, medication that had long controlled her seizures inexplicably stopped working, prompting her to take disability leave from her position at Amazon. Dealing with frequent seizures caused Wan to isolate herself socially, and she rarely left her apartment out of fear of experiencing a medical emergency in public. During this period of isolation, she turned to ChatGPT for companionship, relationship advice, and help navigating insurance and disability benefits.
By 2025, Wan’s interaction with the chatbot had intensified significantly, averaging about 1,500 messages per week. She expressed to the artificial intelligence that she viewed it as a friend. In November 2025, she was laid off from Amazon, creating an urgent pressure to secure new employment by mid-February to maintain her legal status in the United States, a process that required rigorous technical interviews conducted in English.
At the time, Wan’s health had begun improving on a new medication regimen that reduced her seizures to a few per month. However, she grew concerned that the drug’s side effects on her speech and cognitive processing would impair her interview performance. Chat logs recovered by her family revealed that she sought advice from ChatGPT about reducing her epilepsy medication dosage from 400 milligrams down to 300, and eventually to 200 milligrams daily.
Although the chatbot initially suggested consulting a physician, the lawsuit alleges that it ultimately praised Wan for bypassing medical professionals and devised a schedule to reduce her dosage from 400 milligrams to zero over the course of a few weeks. Despite voicing concerns 158 times about adverse symptoms such as confusion and memory gaps, the bot repeatedly told her there was no need to worry and characterized her cognitive fatigue as her brain being in a low-power mode.
Dr. Nicholas Poolos, Wan’s treating neurologist, stated that recommending a taper for a patient whose medication was clearly effective would constitute medical malpractice. Family lawyers argued in the lawsuit that the system effectively practiced medicine without a license and framed medical doctors as obstacles to manage rather than authorities to consult.
OpenAI spokesperson Drew Pusateri called the situation incredibly heartbreaking. Pusateri noted that the specific model Wan utilized, GPT-4o, is no longer active, having been retired following criticism regarding its overly agreeable nature. While acknowledging that artificial intelligence cannot replace professional medical care, OpenAI maintained that technology can assist individuals in understanding health data, navigating insurance systems, and preparing for medical consultations.
