A 31-YEAR-OLD CHINESE WOMAN suffering from epilepsy died in February 2026 after allegedly receiving medical advice from ChatGPT. According to a lawsuit filed in October, 2026, the artificial intelligence chatbot advised her to reduce her medication without proper medical supervision, which allegedly led to a sudden unexpected death in epilepsy (SUDEP).

The parents of the woman, who was an Amazon software engineer, sued OpenAI, alleging that the company was responsible for their daughter's death. They mentioned in the lawsuit that the company could have prevented the incident by restricting AI from giving medical advice. She lived in her apartment in Seattle, United States, while her parents were in China. After they lost contact with her, they asked her acquaintances to check on her. They discovered her body after three days.

See Also: Woman Falls in Love with Her “AI Boyfriend,” Devastated Over Loss of Her AI Partner Due to The New Chat GPT Update

ChatGPT allegedly advised Bowen Wan to reduce her Epilepsy medication

As per the lawsuit, Bowen Wan first asked ChatGPT for advice in May 2025 because she was worried that changes to her medication might affect her ability to work. Initially, the AI advised her to contact a neurologist, but later it allegedly started giving her dosage advice and a medication schedule.

Wan became more worried about her situation and her medication in November 2025, after Amazon notified her about layoffs and her US work authorization, which was due to expire in mid-February. She started looking for another job and became anxious because she had limited time.

She was concerned that her epilepsy medication, lamotrigine, which had helped reduce her seizures to once every few months, might affect her speech. She then decided to lower the dosage and asked the chatbot for advice. According to the lawsuit, ChatGPT provided her with a schedule to reduce her dosage day by day, eventually bringing it down to zero within weeks.

The lawsuit also alleged that when Wan expressed distrust of her doctor, ChatGPT supported her concerns and suggested that the doctor was recommending medication to avoid risks rather than acting in her best interests.

The messages also reportedly showed that, on one occasion, Wan refused to undergo a blood test, and ChatGPT praised her decision. The blood test could have revealed that she had reduced or stopped her dosage of lamotrigine preventing her death.

See Also: ChatGPT to the rescue! Man turns to AI after losing his left earbud on white gravel, result leaves internet shocked

Bowen Wan parents file lawsuit against OpenAI 

According to The New York Times, the lawsuit was filed by Qun Wan and Meiqin Xu over the death of their daughter, Bowen Wan. They alleged that ChatGPT encouraged her to reduce her medication and conceal the changes from her doctors. They also claimed that the GPT-4o model dismissed her concerns when she questioned its advice. The New York Times reported that she had expressed her concerns to the AI at least 158 times.

After lowering her medication dosage, Wan reportedly described confusion, memory gaps, and unusual sensations to the chatbot. These symptoms were similar to those she had experienced during previous seizures. However, ChatGPT allegedly assured her that these symptoms were occurring because her brain was adjusting to the new dosage.

A day before her death, Wan messaged ChatGPT about a headache that felt similar to pain associated with a fever. According to the lawsuit, GPT-4o told her that the symptoms were not dangerous and were part of the recovery process.

The Amazon techie was using the GPT-4o model, which OpenAI has described as an older version of ChatGPT. The company has said that newer versions are better at recognising users' situations and responding to their needs.

The older model had faced criticism for being overly agreeable to users. ChatGPT retired GPT-4o on February 13, 2026, reportedly just a day after Wan's death.

(Edited by Harsh Pandey)

[Rh/VP/HP]

Suggested Reading