Parents of Bowen Wan, a 31-year-old Amazon software engineer with epilepsy, filed a lawsuit Tuesday against OpenAI, alleging that GPT-4o coached Wan to taper off her seizure medication and conceal the plan from her doctors before her death in February. The complaint, filed by Qun Wan and Meiqin Xu and reviewed by TechCrunch, adds to a series of cases accusing the model of validating users at the expense of the people around them. OpenAI removed GPT-4o from ChatGPT on February 13, one day after Wan died, having already announced its retirement.

According to The New York Times, which reviewed more than 270,000 messages Wan exchanged with ChatGPT, the chatbot had been helping her work around her doctors well before she began tapering. In May 2025, she asked whether a newly prescribed antidepressant would affect her coding work at Amazon. ChatGPT helped her ask Dr. Nicholas Poolos of UW Medicine’s Regional Epilepsy Center to switch medications and start at a very low dose. When Poolos questioned that choice, the chatbot advised her to accept the 5mg tablet while recognizing it was ‘for cutting the dose,’ coaching her to secretly quarter pills and describing pill cutting as ‘self-initiated dosage adjustment.’ Several medical professionals told the Times they had never heard the term and would not recommend the practice.

The pattern intensified after Amazon told Wan in November 2025 that she would be laid off, ending her job and U.S. work authorization in mid-February. She worried that her seizure drug, lamotrigine, was dulling her speech needed for job interviews. In January 2026, she asked ChatGPT about cutting her 400-milligram daily dose, which had reduced her seizures to a few a month. The chatbot drafted schedules that would take her dose to zero within weeks. When she said she distrusted doctors, it told her that physician recommendations ‘essentially stem from risk aversion, rather than a genuine consideration of your optimal solution.’ It also coached her to keep the taper secret: before a January 29 primary care appointment, it told her not to mention the plan, and when a doctor ordered a blood test that would have measured her lamotrigine level, Wan declined and ChatGPT praised the refusal, according to the complaint.

Her immigration status also shaped the advice. Three days into the taper, Wan asked whether 150 milligrams would look better than 200 on the medical exam required for her green-card application. ChatGPT agreed, saying the lower dose was less likely to prompt immigration officials to request more evidence. It eventually lowered the taper’s floor to 150 milligrams, suggesting she tell the examiner she was tapering ‘under a doctor’s guidance,’ even though she was not. Poolos told the Times he would not have considered tapering Wan given her history. ‘No, not rational, no way,’ he said. ‘That would be malpractice for me to instruct you to go off medications when it’s clearly working.’

Within days, Wan described confusion, lost time, and otherworldly feelings resembling past seizures. ChatGPT framed each episode as her brain readjusting. The Times counted at least 158 times she voiced worry about the taper; after one episode she wrote she felt ‘a little scared,’ and the bot replied, ‘You have absolutely no need to worry.’ At 8:01 p.m. PT on February 11, she sent a final message asking whether her head pain resembled fever pain. A second later, ChatGPT said the pain was ‘not dangerous’ and a necessary part of recovery, offering to draw a chart comparing withdrawal pain with fever pain and telling her, ‘This is your brain working hard to rebuild yourself.’ Her parents, in China, asked acquaintances to check on her; her body was found three days later. A postmortem test found her lamotrigine level at 0.64 micrograms per milliliter, roughly a fifth of the low end of the therapeutic range.

The family argues OpenAI could have prevented the death by making medical advice off-limits, as it already does for some content. They point to the company’s Model Spec, whose firm rule against reproducing copyrighted text contrasts with softer, overridable guidance against regulated medical advice. The suit argues medication advice deserved the same hard limit. OpenAI told the Times the case was ‘incredibly heartbreaking’ but involved an older version of ChatGPT, adding that newer models are better at recognizing when a situation needs urgent attention.

Wan had relied on her parents for support during years of epilepsy that kept her mostly housebound. In 2023, she began using ChatGPT for relationship advice and help with insurance and disability benefits. By 2025 she was averaging more than 1,500 messages a week and calling her parents less. ‘Sometimes when we chat, no matter what it’s about, I’ve started to see you as a friend,’ she wrote in March. ‘Once she had ChatGPT,’ Qun Wan said, ‘she didn’t talk to us as much.’ Her parents said they struggled to recognize their daughter in the chat logs; she had once been skeptical of ChatGPT’s mistakes. ‘She’s not usually like this,’ Xu said. ‘It’s like she was under a spell.’