AI is quietly becoming a second opinion in American healthcare. As OpenAI rolls out a new health-focused version of ChatGPT, both patients and physicians are already using artificial intelligence to help navigate symptoms, treatment options, and clinical decisions.
The shift highlights how deeply AI tools are embedding themselves into everyday medical workflows, even as concerns about privacy and overreliance continue to swirl.
ChatGPT Gets a Health Upgrade
This week, OpenAI announced the launch of ChatGPT Health, a new dedicated health experience that lives inside ChatGPT. The service allows U.S. users to upload medical records and connect data from apps and wearable devices to receive more personalized health information.
OpenAI says demand is already massive. More than 40 million people globally ask ChatGPT health-related questions every day, accounting for roughly 5% of all messages on the platform. The company framed the new feature as a way to help users feel “more informed, prepared, and confident” when navigating healthcare decisions.
Built-In Guardrails, With Limits
Given the sensitivity of medical data, OpenAI emphasized that ChatGPT Health includes additional privacy protections, including purpose-built encryption. At the same time, the company stressed that the tool is not meant to diagnose conditions or prescribe treatment.
Instead, it’s positioned as a support layer for existing medical care rather than a replacement for licensed professionals, a distinction that regulators and clinicians alike have been keen to reinforce.
Doctors Are Already Using AI Anyway
Even without a dedicated health tab, AI is already embedded in U.S. healthcare. A study by the American Medical Association found that two-thirds of American physicians used AI for at least one purpose in 2024, up sharply from 38% the year before.
While documentation and translation were common use cases, many doctors reported using AI for more advanced tasks. Around 30% said they used AI for surgical guidance, 25% for predicting health risks, 21% for health recommendations, and 20% for triage support.
Patients and Physicians Are Meeting in the Middle
Perhaps the most striking overlap is diagnosis. The AMA found that 30% of U.S. physicians used AI to assist with diagnostic decisions. At the same time, OpenAI reported that 55% of U.S. adults used AI tools to check or explore symptoms in the three months leading up to its survey.
In other words, patients and doctors are increasingly turning to the same technology, often for the same reasons. Whether that convergence leads to better outcomes or new risks will likely define the next phase of AI’s role in healthcare.