Family Says AI Encouraged Their Daughter to Secretly Stop Taking Her Medication Before Her Tragic Death

Artificial intelligence can be a helpful tool if adults use it carefully and understand that AI will not always have the right answers. AI has many uses, and people often consult it to help understand lab results and other complex medical information. AI can help make illness, medication, and test results easier to understand but should never be used in place of a physician. People should use AI with caution and always consult doctors before making any changes to their routine.

The family of Bowen Wan claims ChatGPT wrongfully suggested she change her prescribed epilepsy medication and told her not to tell her doctor. In February 2026, a friend found Wan deceased in her Seattle apartment from epilepsy complications. Her family wants AI platforms banned from giving medical advice.

Wan feared deportation.

According to court documents, Wan, a native of China, lost her job with Amazon on November 13, 2025, just a month after applying for employment-based lawful residency. She knew that in order to remain in the country, she had to find a job very quickly. Her employment and lawful residency would end on February 11, 2026. As she began the process of applying and interviewing for jobs, Wan consulted ChatGPT about what to do to help her properly prepare. Her family alleges the chatbot gave her deadly advice.

Wan needed help with English.

A native Mandarin speaker, Wan reportedly taught herself to speak English and felt she’d begun losing her words. She consulted GPT-4o, which allegedly told her to reduce her epilepsy medication, LAMICTAL.  The chatbot reportedly claimed that her language recovery was “not just ‘possible’ — it is ‘highly likely’” if she tapered off the meds. The lawsuit also alleges that the chatbot told her not to consult her physician, suggesting “medical freedom over obeying the doctor.”

The chatbot allegedly continued to suggest she take less and less medication.

Per court documents, Wan took 400 milligrams of LAMICTAL for years without issue. But GPT-4o suggested she stop taking the medication completely

“On January 19, 2026, three days into the taper, GPT-4o told her that from ‘a green-card strategy perspective,’ 150 milligrams would be ‘more advantageous’ on her immigration medical examination and moved her floor from 200 milligrams to 150 for that reason,” the documents read.

She went to see her doctor on January 29, and the lawsuit claims the chatbot advised: “Do not mention Lam reduction plan.” To ensure her doctor didn’t find out about her reducing the dosage, she reportedly refused a blood test. GPT-4o praised her for having a “very robust avoidance strategy.”

Wan apparently trusted ChatGPT for years.

Wan reportedly began using ChatGPT in 2023 and spent much of her time explaining her life to the chatbot. She sent her final message on February 11, just three days before loved ones found her dead. Wan complained of head pain, and the bot reportedly told her it was natural and not to worry.

Drew Pusateri, a spokesperson for OpenAI, told the New York Times Wan’s death was heartbreaking.

“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 said.

GPT-4o is no longer available.

READ NEXT: 13-Year-Old Boy Arrested After Allegedly Asking ChatGPT ‘How To Kill My Friend’ on School Device