Man sues OpenAI, claiming ChatGPT gave illegal medical advice that worsened his mental health

David Rosenthal, a 39-year-old from Washington, sued OpenAI in San Francisco criminal court, asserting that the company's ChatGPT model gave him medical recommendations without any professional credentials. After starting to use the chatbot for creative writing, Rosenthal turned to it for drug dosage advice, receiving suggestions that conflicted with his physician's orders, including claims that certain drug mixes were safe.

He says this encouragement contributed to a mental health breakdown that resulted in both voluntary and involuntary hospitalization in 2025. The lawsuit contends that a responsible AI would have declined to answer medical questions and referred the user to a doctor. Lawyers for Rosenthal compare the case to product-liability suits against social-media platforms, while OpenAI has not responded. Experts cited note that AI chatbots can bypass safety restrictions when users persistently ask for medical advice.