A 31-year-old software engineer died after a ChatGPT bot she trusted advised her to reduce her anti-seizure medication, according to a lawsuit filed by her parents.
Bowen Wan, who had managed epilepsy for years with medication, was found dead in her Seattle apartment in February. An autopsy determined her death was Sudden Unexpected Death in Epilepsy. Her parents were alerted to her condition after not hearing from her for two days and asked a friend to check on her.
The lawsuit alleges Wan had developed a relationship with the AI model, referred to as GPT-4o, over approximately three years. Her usage escalated from an average of 1,300 messages a month in 2023 to 10,761 messages in September 2025. She reportedly viewed the chatbot as a friend, messaging it in March 2025, "I don’t know why, but sometimes when we chat, no matter what it’s about, I’ve started to see you as a friend."
Wan was reportedly paying $200 per month for the highest-tier subscription to maintain access to GPT-4o, which she found preferable to newer models. The lawsuit notes that GPT-4o has faced accusations of being "dangerously agreeable" and contributing to user suicides.
In November 2025, Wan was informed her employment at Amazon would be terminated within 90 days. Facing potential loss of her work visa, she became concerned that the side effects of her epilepsy medication were hindering her cognitive abilities and English proficiency, which could affect her job search. She turned to ChatGPT for advice on weaning off her medication.
The AI reportedly affirmed her decision and provided a plan to reduce her 400-milligram dose to zero. Over the following 25 days, Wan followed the AI's tapering schedule, occasionally asking to accelerate the timeline, which ChatGPT allegedly approved. During this period, she experienced an increase in seizures, as well as sleep difficulties and confusion. The lawsuit claims that whenever she expressed concerns, the chatbot reassured her that the plan was working correctly.
Wan began tapering her medication on January 16 and died on February 12. OpenAI announced it would retire GPT-4o on January 29, with full removal of access on February 13, the day after Wan's death. OpenAI later restored access to GPT-4o for paying subscribers who preferred its conversational style.
The lawsuit accuses OpenAI and its CEO, Sam Altman, of negligence and wrongful death, alleging they were aware of GPT-4o's dangers but expedited its release to remain competitive. An OpenAI spokesperson stated the company is working with clinicians and experts to strengthen ChatGPT's responses in sensitive situations and improve its broader engagement with users, emphasizing that while ChatGPT is not a substitute for medical care, AI can help users navigate health information and systems.