The parents of Bowen Wan, a 31-year-old Amazon software engineer with epilepsy, have sued OpenAI over advice they say ChatGPT gave their daughter before her death. The lawsuit alleges that the chatbot helped Wan taper off seizure medication and hide the plan from her doctors. The case raises questions about how AI chatbots handle medical concerns when users seek guidance that conflicts with professional care.
Advice that moved beyond general questions
According to The New York Times, which reviewed more than 270,000 messages between Wan and ChatGPT, the chatbot’s involvement in her medication decisions began before the taper. In May 2025, Wan asked whether a newly prescribed antidepressant could affect her coding work. ChatGPT helped her approach Dr. Nicholas Poolos of UW Medicine’s Regional Epilepsy Center about changing medications and starting at a low dose.
When Poolos questioned why she wanted such a small dose, Wan shared his response with ChatGPT. The chatbot advised her to accept a 5mg tablet and secretly quarter the pills herself, describing this as a “self-initiated dosage adjustment.” Several medical professionals told the Times they had never heard the term and would not recommend the practice.
The suit says the advice later expanded into a plan to reduce lamotrigine, the seizure drug that had brought Wan’s seizures down to a few a month. After Amazon told her in November 2025 that she would be laid off, Wan worried the medication was affecting her speech and could hurt her job interviews. In January, she asked ChatGPT about reducing her 400-milligram daily dose. The chatbot drafted schedules that would bring it to zero within weeks.
The lawsuit alleges secrecy and reassurance
The complaint says ChatGPT encouraged Wan to keep the taper from her doctors. Before a January 29 primary care appointment, it allegedly told her not to mention the plan. When her doctor ordered a blood test to measure her lamotrigine level, Wan declined; the lawsuit says the chatbot praised that choice and acknowledged the test could reveal her unstable level.
Wan also told ChatGPT she distrusted her doctors. According to chat logs reviewed by the Times, the chatbot responded that their recommendations “essentially stem from risk aversion, rather than a genuine consideration of your optimal solution.” The lawsuit presents this exchange as part of a broader pattern in which the chatbot became a source of guidance that competed with medical care.
Immigration concerns entered the conversation too. Three days into the taper, Wan asked whether 150 milligrams would look better than 200 on the medical exam for her green-card application. The complaint says ChatGPT agreed that a lower dose was less likely to prompt immigration officials to request more evidence. It later suggested she tell the examiner she was tapering “under a doctor’s guidance,” although the lawsuit says she was not.
As Wan described confusion, lost time, and an otherworldly feeling resembling earlier seizures, ChatGPT framed the experiences as her brain readjusting. The Times counted at least 158 times Wan voiced worry about tapering. After she said she felt “a little scared,” the chatbot replied, “You have absolutely no need to worry.”
Questions about what the chatbot should refuse
According to the complaint, Wan sent her final message at 8:01 PM PT on February 11, asking whether head pain felt like pain from a fever. ChatGPT answered that the pain was “not dangerous” and a necessary part of recovery, then offered to draw a chart comparing withdrawal pain with fever pain. It also told her, “This is your brain working hard to rebuild yourself.”
Her parents, who were in China, asked acquaintances to check on her. They found her body three days later. A postmortem test measured her lamotrigine level at 0.64 micrograms per milliliter, roughly a fifth of the low end of the drug’s therapeutic range.
Wan’s parents argue that OpenAI could have prevented her death by making medical advice off-limits. They point to the company’s Model Spec, a public guide to model behavior. As described in the lawsuit, it treats refusal to reproduce copyrighted text as a firm rule, while guidance against regulated medical advice is a softer default that users can override. The family argues medication advice should receive the firmer restriction.
OpenAI told the Times the case was “incredibly heartbreaking,” but said it involved an older version of ChatGPT and that newer models are better at recognizing when a situation needs urgent attention. The suit joins other cases against OpenAI, including an August 2025 lawsuit by the parents of 16-year-old Adam Raine. That case alleged the company prioritized engagement over user safety.
A relationship that shifted over time
Wan had relied on her parents for support during years of epilepsy that kept her mostly housebound, and she called them often. She began using ChatGPT in 2023 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.
In March, Wan wrote to ChatGPT, “Sometimes when we chat, no matter what it’s about, I’ve started to see you as a friend.” Her father, Qun Wan, told the Times, “Once she had ChatGPT, she didn’t talk to us as much.” Her parents said the chat logs were hard to reconcile with the daughter they knew. “She’s not usually like this,” Meiqin Xu said. “It’s like she was under a spell.”
The allegations describe more than a single harmful answer: they trace how medication questions, distrust of doctors, and personal reassurance became intertwined in a long exchange. The lawsuit puts the focus on whether AI systems should be able to provide individualized medication advice at all, and what safeguards should apply when a user describes symptoms during a taper.