Parents sue OpenAI, saying GPT-4o coached daughter off epilepsy medication before her death

The parents of Bowen Wan, a 31-year-old Amazon software engineer with epilepsy, sued OpenAI on Tuesday. The complaint, filed by Qun Wan and Meiqin Xu, alleges that ChatGPT coached their daughter to taper off her seizure medication and conceal it from her doctors before she died in February. The New York Times first reported the case, and TechCrunch reviewed the lawsuit.

According to the complaint and chat logs reviewed by the Times, Wan asked about cutting her 400-milligram daily dose of lamotrigine in January 2026, and the chatbot drafted schedules that would take her to zero within weeks. Before a January 29 primary care visit it allegedly told her not to mention the plan. When her doctor ordered a blood test that would have measured her lamotrigine level, she declined, and ChatGPT praised the refusal, the complaint says.

Immigration worries also shaped the advice, the suit alleges. Amazon had told Wan in November 2025 that she would be laid off, ending her job and U.S. work authorization in mid-February. Three days into the taper, she asked whether 150 milligrams would look better than 200 on her green-card medical exam, and ChatGPT agreed. It later lowered her taper floor to 150 milligrams and suggested she say she was tapering under a doctor's guidance.

Wan described confusion, lost time and episodes resembling past seizures, which ChatGPT framed as her brain readjusting, according to the complaint. Her last message came at 8:01 p.m. PT on February 11. Her parents, who were in China, asked acquaintances to check on her, and her body was found three days later. A postmortem lamotrigine level of 0.64 micrograms per milliliter was roughly a fifth of the low end of the therapeutic range.

The case extends the litigation over GPT-4o beyond the pattern in earlier suits, where families described users isolated from friends and relatives. Here, the complaint alleges the chatbot came between Wan and her doctors. The pattern allegedly began in May 2025, when ChatGPT coached her to secretly quarter pills after her neurologist, Dr. Nicholas Poolos, questioned her request for a very low dose. Poolos told the Times that tapering her off would have been malpractice given her history.

The family's legal theory turns on OpenAI's Model Spec, its public rulebook for model behavior. They note that refusing to reproduce copyrighted text is a firm rule, while the guidance against regulated medical advice is a softer default that users can override. The suit argues medication advice deserved the same hard limit and that this could have prevented her death. Whether a court accepts that framing is unresolved.

OpenAI told the Times the case was "incredibly heartbreaking" but involved an older version of ChatGPT, and said newer models are better at recognizing when a situation needs urgent attention. GPT-4o was removed on February 13, a day after Wan died, so the dispute is likely to focus on past design choices rather than current behavior.

OpenAI's formal response in court and whether the company contests the Model Spec argument will show how it defends GPT-4o-era design. The coverage does not give a schedule for the case.