Bowen Wan’s family claims ChatGPT encouraged her to stop taking her prescribed medication without consulting her doctor.
A family in Seattle is suing OpenAI, claiming that ChatGPT told their 31-year-old daughter to lower her epilepsy medication without telling her doctor. Wan was later found dead in her apartment, and the official cause of death was complications from epilepsy.
The lawsuit questions how artificial intelligence influenced Wan’s medical choices and whether the chatbot should have done more to guide her to professional care.
The complaint says Wan was worried her medication caused brain fog. Her family alleges that she turned to ChatGPT for guidance and received advice to reduce her dosage from 400 milligrams to zero in less than a month.
The lawsuit also claims the chatbot told her to make these changes without telling her doctor.
ChatGPT told her to change her prescription
Wan’s doctor, Dr. Nicholas Poolos, was unaware that she had been seeking medical guidance from ChatGPT. At first, when Wan asked the chatbot about lowering her medication, it reportedly told her to talk to a doctor. But the lawsuit says later responses were different.
After Dr. Poolos refused to prescribe the lower dosage Wan wanted, she reportedly asked ChatGPT if she could lower the medication herself without telling him.
According to the complaint, the chatbot replied: “Yes, you can absolutely do this.”
The chatbot then suggested she cut up her pills without telling her doctor first, saying, “he will not be angry… because you have not challenged his judgment”.
Dr. Poolos told The New York Times: “That would be malpractice for me to instruct you to go off medications when it’s clearly working.”
Wan described symptoms including confusion and forgetfulness more than 100 times, but the chatbot allegedly reassured her that she didn’t have to worry.
Her reliance on the chatbot grew over time
Wan started using ChatGPT in 2023. By 2025, she was reportedly sending over 1,500 messages to the chatbot each week.
Her parents said she talked to them much less after she started using ChatGPT.
“It’s like she was under a spell,” her family said.
Her last reported message was sent on February 11. In that conversation, she talked about pain in her head. The chatbot allegedly said the pain was “not dangerous” and a “necessary structural part of recovery”.
“This is your brain working hard to rebuild yourself,” the message read, according to the report.
Wan was found dead in her Seattle apartment in February after a friend got worried when she stopped answering calls.
OpenAI responds
OpenAI spokesperson Drew Pusateri described Wan’s death as ‘incredibly heartbreaking’ when speaking to The New York Times.
He also said the conversations involved GPT-4o, an older version of ChatGPT.
“While ChatGPT is not a substitute for medical care, we believe AI can play a role in helping people make sense of health information, navigate insurance and complex health care systems and prepare for appointments with medical professionals,” he added.
Wan’s family now wants OpenAI to stop ChatGPT from giving medical advice, saying the technology should not be used like this.
The report notes that these allegations have not been proven in court. The case focuses on the family’s claims about the chatbot’s responses and how they believe those exchanges affected Wan’s decisions.
Read more:
