“ChatGPT Health offers personalized medical guidance, integrates with wearables, and connects users to telemedicine providers.”
OpenAI is launching a healthcare product after the Stanford sycophancy paper, after the Harvard ER triage results, and after a year of public reports of users getting wrong medical advice from chatbots. The pitch is wrapped in a careful disclaimer. The product is not. The first malpractice suits will be the most interesting legal events of the year. The model card will not be a defense. The terms of service will not be a defense. We will see.