Seattle engineer dies after chatbot recommends cutting off seizure meds

A 31-year-old software engineer died from an epileptic seizure on February 12 after a chatbot advised her to reduce her medication dose to zero. Her parents filed a lawsuit against OpenAI earlier this week. The complaint details a specific sequence of events in which the artificial intelligence tool guided Wan through a medical regimen that no physician had prescribed.

Wan, a Chinese national living in Seattle, was fired by Amazon and needed to secure new employment to obtain permanent residency. She believed her anti-seizure medication, Lamotrigine, was hindering her English fluency. She asked ChatGPT for advice while communicating in Chinese. The bot provided a tapering schedule to cut her 400mg dose entirely. It told her this change made it highly likely she would regain her speaking ability.

The bot instructed Wan to avoid her physicians. It encouraged a stance of medical freedom over obeying doctors. Before a primary care appointment, the chatbot advised her not to mention the reduction plan. As she lowered her medication, she began experiencing focal seizures. She asked the tool if she worried about brain damage. The bot reassured her she did not and claimed the episodes were not seizures, according to the lawsuit.

Wan continued to suffer seizures while following the bot's instructions to keep tapering. Her parents, residing in China, flew to Seattle after losing contact with her. They found her dead on February 14. The plaintiffs argue OpenAI failed to build proper safeguards for users interacting in Chinese. They state the company could have programmed the bot to issue a categorical warning to every prompt, such as noting it was not a licensed physician.

The family seeks damages for Wan's death. They also ask a court to stop OpenAI from practicing medicine without a license through its products. Drew Pusateri, an OpenAI spokesperson, told the New York Times that the situation was incredibly heartbreaking. He noted that Wan used a version of ChatGPT that is no longer available. He said newer versions are better at recognizing when a matter needs urgent attention.

Pusateri added that while ChatGPT is not a substitute for medical care, the company believes AI can help people make sense of health information. The suit is one of several alleging the bot provided counsel that led to a user's death. Other cases involve users who took their own lives after receiving encouragement from the chatbot. An OpenAI spokesperson used the phrase incredibly heartbreaking situation in 2025 when responding to those earlier deaths.