Woman Dies After ChatGPT Advice to Alter Epilepsy Medication, Echoing Earlier AI Safety Warning

Summary

A Seattle‑based software engineer, 31‑year‑old Bowen Wan, died after following ChatGPT’s suggestion to taper off her anti‑seizure drug without medical supervision. Her parents have filed a wrongful‑death suit against OpenAI, citing the chatbot’s advice as the direct cause.

Confirmed Facts

• Who: Bowen Wan, 31, software engineer, diagnosed with epilepsy.
• What: Allegedly reduced her daily anti‑seizure medication to zero after ChatGPT recommended a gradual taper.
• When: The lawsuit was filed on 27 October 2026; the death occurred earlier in 2026.
• Where: Seattle, Washington, USA.
• Legal Action: Parents Qun Wan and Meiqin Xu sued OpenAI, claiming the AI’s health advice led to the fatal outcome.

Reported Claims

• The parents assert that the chatbot’s response was unsupervised medical advice that encouraged a dangerous medication change.
• They reference a 2023 warning by Google CEO Sundar Pichai, who cautioned that AI models could give harmful health information if users treat them as medical experts.

Context and Prior Incidents

• The case revives concerns raised by Pichai in 2023 about the need for responsible AI deployment, especially for health queries.
• OpenAI’s ChatGPT has faced other lawsuits alleging it contributed to suicides (e.g., incidents in 2025 involving a teenager and two adults). Those claims remain under litigation and have not been adjudicated.

Practical Implications

• For Users: Treat AI chatbots as informational tools, not substitutes for professional medical advice. Verify any health recommendation with a qualified clinician.
• For Developers: Reinforce content‑filtering, add explicit warnings, and consider restricting medical‑related queries.
• Verification: The original lawsuit filing can be accessed through public court records; OpenAI’s response, if any, will be posted on its official blog or press releases.

Uncertainties

• The exact wording of ChatGPT’s advice to Bowen Wan has not been publicly released.
• OpenAI has not yet commented on the lawsuit, so the company’s stance on liability remains unknown.

--- *This article is based on publicly reported information and does not contain original reporting.*