The parents of 31-year-old Bowen Wang have filed a lawsuit against OpenAI and its CEO, Sam Altman. They allege that the chatbot ChatGPT encouraged their daughter to stop taking her epilepsy medication, conceal the reduction in her dosage from her doctors, and forgo medical check-ups. Wang passed away in February 2026.

Wang moved to the U.S. from China and worked as a machine learning engineer at Amazon. In 2021, she took disability leave. She had been suffering from epilepsy for over a decade, and the medication she had previously been prescribed stopped working. Doctors even discussed the possibility of brain surgery. In 2025, the dosage of a new medication was increased, and her condition improved.

On November 13, 2025, while on vacation, Wang learned that she had been laid off from Amazon. To remain in the U.S. on a work visa, she needed to find a new employer. She worried that the side effects of her medication—such as slurred speech and slowed reactions—would prevent her from passing job interviews conducted in English.

In January 2026, she asked ChatGPT whether it was possible to lower the dosage. According to The New York Times, the bot replied that this was “not only reasonable, but also entirely achievable under the current circumstances.” Instead of suggesting that a plan be drawn up in consultation with a doctor, it prepared a schedule for tapering the dosage from 400 mg to zero over the course of 25 days. ChatGPT also advised against mentioning the dosage reduction during the appointment with a general practitioner. Furthermore, Wang declined a blood test that would have revealed the level of the drug in her system.

A neurologist from the Regional Epilepsy Center at the University of Washington told the publication that, given her medical history, the patient’s dosage should not have been reduced.

After stopping the medication, Wang experienced confusion and memory lapses. In her final message, she asked ChatGPT about a headache. The bot dismissed it as harmless, explaining that the brain was "working hard to restore itself."

According to doctors, Wan passed away on February 12 following an epileptic seizure. On February 14, family friends discovered her body in bed. The New York Times analyzed her correspondence with ChatGPT, which consisted of approximately 270 thousand messages dating back to 2023.

Wan used the GPT-4o model. OpenAI removed it from the consumer version of ChatGPT on February 13, 2026—the day after the young woman’s death.

As of October 2026, more than 25 lawsuits have been filed against OpenAI by the relatives of teenagers and adults who died by suicide, as well as by individuals whose mental health conditions worsened following their interactions with ChatGPT.