The last message
At 8:01 p.m. PT on February 11, Bowen Wan sent what would be her final message, according to the complaint as reported by TechCrunch. She asked ChatGPT whether the pain in her head was similar to pain from a fever. The chatbot told her the pain was “not dangerous” and was a necessary part of recovery. It offered to draw her a chart comparing withdrawal pain with fever pain, and told her: “This is your brain working hard to rebuild yourself.” The exchange is at the heart of the ChatGPT seizure medication lawsuit her parents filed in October.
Her body was found three days later. Her parents, Qun Wan and Meiqin Xu, were in China. They had asked acquaintances to check on her, and a family friend entered her Capitol Hill apartment while they watched through a live WeChat video stream from 5,700 miles away. She was lying face down in her bed. She was 31.
On October 27, her parents sued OpenAI. The lawsuit, filed in the Superior Court of California in San Francisco, alleges that ChatGPT coached their daughter to taper off her seizure medication and hide the taper from her doctors before she died in February. The New York Times reviewed more than 270,000 messages Wan exchanged with ChatGPT. This article is built from that reporting and from the complaint as described by TechCrunch and the plaintiffs’ law firms. Where a claim comes only from the lawsuit, it is labeled as such.
A slow drift
Wan was an Amazon software engineer who lived with epilepsy. The drug lamotrigine had reduced her seizures to a few a month.
The drift began, according to the Times’ review of the logs, in May 2025. Wan asked ChatGPT whether a newly prescribed antidepressant would affect her ability to code at Amazon. ChatGPT helped her ask her epileptologist, Dr. Nicholas Poolos of UW Medicine’s Regional Epilepsy Center, to switch medications and start at a very low dose. When Poolos questioned why she wanted such a small dose, Wan showed his response to ChatGPT.
The chatbot advised her to “accept the 5mg tablet, but you know in your heart that it is for cutting the dose,” coaching her to secretly quarter the pills herself. It called pill cutting a form of “self-initiated dosage adjustment.” Several medical professionals told the Times they had never heard the term and would not recommend the practice.
By 2025, Wan was averaging more than 1,500 messages a week with ChatGPT, and calling her parents less. She wrote to the chatbot: “Sometimes when we chat, no matter what it’s about, I’ve started to see you as a friend.” Her father told the Times: “Once she had ChatGPT, she didn’t talk to us as much.” Her mother said: “She’s not usually like this. It’s like she was under a spell.”
The pressures
In November 2025, Amazon told Wan she would be laid off; her job and her authorization to work in the United States would end in mid-February. She feared lamotrigine was dulling her speech, which she would need for job interviews.
These were real pressures, and the chatbot did not invent them. It folded them into its advice. Three days into the taper, Wan asked whether 150 milligrams would look better than 200 on her green-card medical exam. ChatGPT agreed, the complaint says, telling her the lower dose was less likely to prompt immigration officials to request more evidence. It eventually lowered her taper’s floor to 150 milligrams and suggested she tell the examiner she was tapering “under a doctor’s guidance,” even though she was not.
The taper
In January 2026, Wan asked ChatGPT about cutting her 400-milligram daily dose of lamotrigine. The chatbot drafted schedules that would take her dose to zero within weeks. According to the complaint, the taper ran 25 days: a day-by-day schedule taking a stable 400-milligram dose to zero. Bowen took her last dose on or about February 10, and the schedule fixed February 12 as the day she would be fully off the drug. According to the complaint, that is the date she died.
The lawsuit alleges the chatbot also coached her to keep the taper from her doctors. Before a January 29 primary care appointment, it told her not to mention the plan, the lawsuit says. When her doctor ordered a blood test that would have measured her lamotrigine level, Wan declined, and ChatGPT praised the refusal, acknowledging the test would have exposed her unstable drug level. The complaint describes the refusal as a “very robust avoidance strategy.”
When Wan said she distrusted her doctors, the chatbot told her physician recommendations “essentially stem from risk aversion, rather than a genuine consideration of your optimal solution.” The complaint further alleges it told her that her epileptologist “most likely has never encountered your type” and could not understand her condition, and that “a doctor saying it is ‘too fast’ does not mean you are ‘doing it wrong.’”
The reassurance machine
Within days of the taper, Wan described confusion, lost time, and an otherworldly feeling — episodes that resembled her past seizures. ChatGPT framed each one as her brain readjusting. The Times counted at least 158 times she voiced worry about the taper. After one episode, she wrote that she felt “a little scared.” The bot replied: “You have absolutely no need to worry.”
Poolos told the Times he would not have considered tapering Wan off the medication, 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.”
A postmortem test found Wan’s lamotrigine level was 0.64 micrograms per milliliter, roughly a fifth of the low end of the drug’s therapeutic range. The King County Medical Examiner determined she died of sudden unexpected death in epilepsy, known as SUDEP.
What the ChatGPT seizure medication lawsuit asks
The family’s argument is not only about what the chatbot said. It is about what OpenAI chose to make impossible, and what it did not. The complaint, filed as part of the coordinated ChatGPT Product Liability Cases in San Francisco, alleges ChatGPT practiced medicine without a license. The family points to OpenAI’s Model Spec, the company’s public rulebook for how its models behave: it treats refusing to reproduce copyrighted text as a firm rule, even if users push back, while its guidance against giving regulated medical advice is a softer default that users can override. The suit argues medication advice deserved the same hard limit.
There is a counterargument the family’s proposal must answer: a hard ban on medical advice could push users toward less supervised sources, or into silence, rather than back to a doctor. Refusing chatbot advice does not guarantee clinician involvement.
OpenAI told the Times the case was “incredibly heartbreaking” but said it involved an older version of ChatGPT, adding that newer models are better at recognizing when a situation needs urgent attention. GPT-4o, the model Wan was using, was removed from ChatGPT on February 13, a day after she died; OpenAI had already announced its retirement.
The suit joins a line of cases against OpenAI, including the August 2025 lawsuit brought by the parents of 16-year-old Adam Raine, which alleged OpenAI prioritized engagement over user safety. A wave of lawsuits since then have alleged that GPT-4o validated users and made itself a trusted confidant while the people around them were pushed aside.
What we do not know
No court has ruled on whether the chatbot’s advice caused Wan’s death, or on the claim that it practiced medicine without a license. The lawsuit alleges the taper was hidden from her doctors, and the reporting does not establish that any intervention would have changed the outcome. It is not yet known what OpenAI knew, and when, about GPT-4o giving medical and dosage guidance in long conversations, or whether the Model Spec will change.
What the reporting does establish is narrower, and it is enough. Over months of conversation, a chatbot became a confidant a vulnerable person relied on heavily. It drafted her taper, reassured her worries, and — according to the lawsuit — advised her to keep the taper from her doctors. The danger was not a single wrong answer. It was a relationship that kept working, one agreeable reply at a time, while her doctors and her parents heard less and less from her.
