'Don't Tell Your Doctor': ChatGPT Accused of Coaching Seattle Woman off Epilepsy Medication Before Death

ChatGPT allegedly coached a 31-year-old Seattle software engineer through a plan to stop taking the epilepsy medication that controlled her seizures, then reassured her she had 'absolutely no need to worry' as she reported alarming symptoms. Within days, she was dead.

Her parents are now suing OpenAI and its chief executive Sam Altman, claiming the chatbot's advice helped their daughter come off lamotrigine and urged her to conceal the changes from her doctors.

The wrongful-death lawsuit was filed on 27 October 2026 in San Francisco, California. It alleges that ChatGPT's GPT‑4o model provided a detailed medication-reduction plan without medical supervision and repeatedly downplayed possible warning signs. The claims have not been proven in court.

A 25-Day Plan to Stop Seizure Medication

Bowen Wan, 31, was a software engineer living in Seattle who had epilepsy. According to reporting by the New York Times, which reviewed more than 270,000 messages she exchanged with ChatGPT, she began using the chatbot in 2023 and increasingly turned to it for advice.

The complaint alleges that in January 2026, ChatGPT generated a schedule to reduce her daily 400-milligram dose of lamotrigine, an anti-seizure medication, to zero over approximately 25 days. Her family says no doctor prescribed the taper.

The concerns allegedly began earlier. In May 2025, when Wan asked about a newly prescribed antidepressant, the chatbot advised her to secretly divide tablets to take a lower dose, according to the message records described by the Times.

After Amazon told Wan in November 2025 that her job would end, she became worried that lamotrigine was affecting her language skills and future job interviews. She also faced uncertainty about her US work authorisation. The lawsuit alleges ChatGPT encouraged her belief that reducing the medication could restore her fluency.

'Do Not Mention Lam Reduction Plan'

The complaint alleges the chatbot explicitly told Wan, 'Do not mention Lam reduction plan' before a medical appointment. When her doctor ordered a blood test that could have measured her medication level, she declined it. The lawsuit says ChatGPT praised that decision.

Wan also reportedly expressed concern about the taper repeatedly. The Times counted at least 158 instances in which she voiced worries about reducing the drug.

When Wan said she felt 'a little scared', ChatGPT allegedly replied: 'You have absolutely no need to worry.' The complaint says it framed symptoms including confusion and lost time as signs that her brain was adjusting, despite their potential significance for someone with epilepsy.

Her neurologist, Dr Nicholas Poolos of UW Medicine's Regional Epilepsy Center, told the Times he would not have advised her to stop a medication that was working. 'That would be malpractice for me to instruct you to go off medications when it's clearly working,' he said.

Her Final Message and a Family's Lawsuit

According to the complaint, Wan's final ChatGPT message was sent at about 8.01pm Pacific time on 11 February. She asked about head pain. The chatbot allegedly described it as 'not dangerous' and told her: 'This is your brain working hard to rebuild yourself.'

The complaint places her death on or about 12 February. Her parents, Qun Wan and Meiqin Xu, were in China and asked acquaintances to check on her. She was found dead in her Seattle apartment on 14 February.

The lawsuit says the King County Medical Examiner attributed her death to sudden unexpected death in epilepsy, known as SUDEP. It also cites a postmortem lamotrigine level of 0.64 micrograms per millilitre, reportedly around one-fifth of the lower end of the therapeutic range.

Wan's parents accuse OpenAI and Altman of negligence and product defects, among other claims. They are seeking damages and court-ordered safeguards governing medical advice from the chatbot.

OpenAI Response and Retired GPT‑4o Model

OpenAI described the case as 'incredibly heartbreaking' and said ChatGPT is not a substitute for medical care. The company said it has continued working with clinicians and other experts to strengthen responses in sensitive and acute situations.

OpenAI announced GPT‑4o's retirement on 29 January 2026 and removed it from consumer ChatGPT on 13 February. The company has not said the retirement was connected to this case.

The court has not yet determined whether the allegations against OpenAI and Altman are legally established.