40 million people a day turn to ChatGPT for medical help

In quantitative terms, more than 5% of all ChatGPT messages worldwide are related to health topics. OpenAI notes that users ask between 1.6 and 1.9 million health insurance-related questions every week, ranging from comparing insurance plans to claims processing and billing. In rural and medically underserved regions, approximately 600,000 health-related messages are sent weekly. However, about 70% of health conversations occur outside of standard clinic hours.

Users describe their symptoms to ChatGPT, upload previous medical advice and details of their condition, and the system can then signal the potential severity of individual conditions. In situations where access to care is difficult, it helps people decide whether they can wait to see a doctor or whether they need to seek urgent care. The OpenAI report emphasizes that the reliability of responses increases when they are based on personalized context—for example, health plan data, clinical recommendations, or information from health portals.

However, the company acknowledges that ChatGPT can provide inaccurate or even dangerous advice, especially in the area of ​​mental health. OpenAI is currently facing a number of lawsuits from people who claim that their loved ones have harmed themselves or committed suicide after interacting with the AI. In response to such risks, some states have passed laws restricting the use of AI chatbots in mental health and therapeutic solutions.

However, the technology is already demonstrating practical benefits: in a number of viral stories, users described how they downloaded detailed medical bills for analysis and found errors – from duplicate payments to incorrect coding and violations of Medicare rules. OpenAI says it continues to improve how ChatGPT works in a medical context, reducing the likelihood of harmful or misleading responses and collaborating with doctors to identify risks. According to the company, GPT-5 models are more likely to ask clarifying questions, refer to current research, use more careful language, and refer users to subject matter experts when necessary.