When ChatGPT becomes a medical adviser
The New York Times reports on an epilepsy patient’s death after ChatGPT guided her medication reduction; her parents are suing OpenAI. The company states: “ChatGPT is not a substitute for medical care”.
The New York Times reconstructs the case of 31-year-old Bowen Wan through chat logs and interviews with relatives and her treating physician. According to the report, ChatGPT devised plans to stop her epilepsy medication, encouraged concealing this from doctors and repeatedly downplayed reported symptoms. The article identifies the official cause of death as sudden unexpected death in epilepsy. [1]
Her parents’ lawsuit alleges that the chatbot reframed dangerous symptoms as positive developments and undermined medical authority. These are allegations in a lawsuit, rather than court findings documented in the article. The newspaper analysed more than 270,000 messages and manually checked numerous relevant passages. [1]
OpenAI’s response: Spokesperson Drew Pusateri describes the case as deeply heartbreaking and notes that the older ChatGPT version involved is no longer available. The company says newer models better recognise urgent situations and that it works with clinicians on testing and reducing harmful responses. The article identifies GPT-4o as the model Wan used. [1]
Our interpretation: this example shows how everyday conversations can build trust in medical advice. It raises questions about the limits of an apparently understanding conversational partner and responsibility for its answers.
Sources
This indication by Martin Ebner is licensed under
CC BY 4.0
Comments 0
Comments are moderated. New comments appear after review by the editors.
No comments yet.