Newsletters
A family in Seattle, U.S., is suing OpenAI over the death of a 31-year-old woman who died of a seizure after ChatGPT allegedly coached her on reducing her epilepsy medication.
The New York Times first reported the lawsuit, filed Tuesday in California on behalf of the family of Bowen Wan. It joins other cases against OpenAI alleging harm to consumers and raising questions about relying on general-use chatbots for guidance on sensitive topics such as personal health.
Wan, an Amazon software engineer, had lived with epilepsy since youth and managed the condition with medication, according to the lawsuit.
Around 2021, her medication became less effective and she began having near-daily seizures. She went on disability leave and spent most of her time at home for fear of having a seizure in public.
Wan, a Chinese immigrant, began using ChatGPT in 2023 and wrote to it in Chinese characters. By 2025, she was sending more than 1,500 messages a week, the Times reported.
She told the chatbot she saw it “as a friend” and consulted it on matters ranging from relationship advice to medical guidance.
According to the Times, Wan used ChatGPT to draft messages to her doctor requesting lower doses of an antidepressant. The chatbot suggested she could “start microdosing” her medication.
When she asked whether she could cut the pills herself without telling her doctor, it reportedly replied that she could “absolutely do that” and described the practice as a “self-initiated dosage adjustment.” Several medical professionals told the newspaper they had never heard of that term.
Amazon laid Wan off in late 2025. She was seeking a new job that would allow her to stay in the United States and was concerned that side effects of her epilepsy medication, which affected her ability to think and speak clearly, would hurt her job search. The lawsuit says ChatGPT then coached her on lowering her dose.
The Times reported that Wan voiced concern at least 158 times that reducing her medication might be unwise, and that the chatbot frequently dismissed those concerns.
In one of her final messages, she wrote that her brain felt exhausted and asked whether she should take some of her medication. The chatbot reportedly told her to maintain “0 mg” the next day.
She sent her last message on the evening of Feb. 11. Her body was found three days later, and her parents discovered the conversations.
OpenAI told the Times that Wan’s death was an “incredibly heartbreaking situation.” The company noted that the version of ChatGPT she used, GPT-4o, has since been withdrawn from the market.
The model was retired on Feb. 13, a day after Wan is believed to have died. Futurism reported that GPT-4o has been linked to several cases of alleged psychological harm and death.
Lisa Knopf, medical director of epilepsy surgery at Hartford Healthcare in Connecticut, told Futurism she was “chilled and horrified” by the details of the case.
Knopf said anti-seizure medications can significantly affect quality of life but are life-saving. She noted that patients with epilepsy commonly face social isolation, stigma, job loss and cognitive and emotional difficulties, and that a national shortage of neurologists and epileptologists makes timely support hard to obtain.
“There’s a difference between emotional support and medical advice,” Knopf said. She pointed to a chatbot built by the Epilepsy Foundation as a possible “safer starting point” for people who struggle to access care. She also said technology companies should “set clear boundaries on what these bots are allowed to say” and flag conversations that go off course.
The case is among more than 50 lawsuits alleging that extensive interactions with ChatGPT caused harm or death to users or people around them. It is also one of several that specifically allege the chatbot offered dangerous medical advice
Set as preferred source
Set as preferred source
3 min read
