More than 40 million Americans now turn to ChatGPT daily for health information, according to a report from OpenAI shared exclusively with Axios. The findings highlight how artificial intelligence is increasingly filling gaps in access, affordability, and understanding within the U.S. health care system.
Axios reports that patients are using ChatGPT as an “ally” to help decode medical bills, identify potential overcharges, appeal insurance denials, and seek guidance when access to doctors is limited. In some cases, users rely on the tool to self-diagnose symptoms or manage aspects of their care. Analysis of anonymized ChatGPT interactions and a December 2025 survey conducted using the AI-powered research tool Knit found that more than half of respondents used AI to check or explore symptoms, while others used it to understand medical terminology, ask questions at any time of day, or learn about treatment options.
According to OpenAI data cited by Axios, more than 5% of all ChatGPT messages globally are related to health care. Users submit between 1.6 million and 1.9 million health insurance-related questions each week, ranging from plan comparisons to billing disputes. In underserved rural communities alone, nearly 600,000 health care-related messages are sent weekly, and seven in 10 health-related conversations occur outside normal clinic hours.
Axios notes that ChatGPT can help patients assess whether symptoms are severe enough to require emergency care or whether they can wait for a scheduled appointment, particularly when traditional care is unavailable. OpenAI says reliability improves when responses are grounded in patient-specific context such as insurance documents, clinical instructions, or health portal data.
However, the report also underscores significant risks. ChatGPT can provide incorrect or potentially dangerous advice, especially in mental health-related conversations. Axios points out that OpenAI is facing multiple lawsuits from individuals who claim loved ones were harmed after interacting with the technology, while several U.S. states have enacted laws restricting AI chatbots from offering mental health or therapeutic decision-making.
Behind the scenes, OpenAI says it is working with clinicians to reduce harmful or misleading responses and improve safety in health-related interactions. According to Axios, newer GPT-5 models are more likely to ask follow-up questions, use cautious language, browse for up-to-date research, and direct users toward professional medical evaluation when appropriate.
As Axios observes, the potential expiration of enhanced Affordable Care Act subsidies could further accelerate reliance on AI tools, particularly among uninsured and underinsured Americans. The key questions ahead center on accuracy, legal responsibility, and data access as AI becomes an increasingly common source of medical guidance without a doctor directly involved.
