OpenAI published the report AI as a Healthcare Ally, which describes how large language models are used by people as a tool to navigate medical information. According to company data, more than 40 million people turn to ChatGPT daily with health-related questions, and over 5% of all model queries are medical in nature. The authors emphasize that this does not involve diagnosis or replacing a doctor. In most cases, users use AI as an initial guide-to understand the severity of symptoms, what questions to ask a doctor, or how to interpret existing medical data.
The structure of queries is led by questions about symptoms and potential causes of ailments, followed by topics related to medications, tests, medical procedures, and the organization of medical care, including insurance.
Separately, the report notes the context of use. A significant portion of medical dialogues with AI occurs outside of standard clinic hours, as well as in rural and remote areas where access to medical specialists is limited. In these situations, ChatGPT effectively serves as a “first point of contact”-not as a clinical tool, but as a source of basic, structured information, helping the user make decisions about the next steps.
OpenAI emphasizes that such a role for AI is formed from the bottom up, without centralized implementation in the healthcare system. People independently choose the model as an accessible and anonymous way to get explanations or reduce uncertainty, especially in situations where it is not immediately possible to consult a doctor. The company believes that this role of the chatbot requires careful attention from regulators and developers. The report underscores the need for clear boundaries of AI use in medicine, transparent restrictions, and safety mechanisms-considering that for millions of users, AI has already become part of their daily interaction with the healthcare system.
