Overview
- OpenAI reports that health queries now exceed 5% of all ChatGPT messages, with roughly one in four users submitting a health-related prompt each week.
- Usage concentrates when care is harder to reach, with about 70% of health conversations occurring outside clinic hours and roughly 580,000 to 600,000 weekly messages originating from rural or hospital‑desert areas.
- Consumers are leaning on the tool for the financial side of care, sending an estimated 1.6 million to 1.9 million weekly questions on insurance comparisons, billing, claims and potential overcharges.
- Independent studies and active lawsuits highlight risks, including reports of dangerously inaccurate advice on medical and mental health topics and error rates near 13% for some leading models in testing.
- OpenAI says newer GPT‑5 behavior emphasizes follow‑up questions, web research and directing users to professional evaluation, and the company plans a healthcare policy blueprint as state rules tighten and a recent federal order seeks a national framework.