OpenAI plugged ChatGPT into Epic EHR on September 1. Every headline called it a historic partnership. There is no partnership. Here is what actually happened.
The integration runs on Epic’s free public FHIR APIs, the same ones any developer can use. A health system admin registers an app, adds credentials, publishes it. Clinicians sign in with their existing Epic logins. ChatGPT sees exactly what the doctor can see. Nothing more.
Epic’s official response on launch day: OpenAI’s app is ‘one of thousands’ that connect through free APIs, and health systems choose which to enable. Translation: we didn’t invite them, we can’t stop them, liability is yours.
OpenAI published physician safety ratings: 99.1% of responses rated safe across 4,363 evaluations. Do the math. That’s roughly 39 unsafe responses in a controlled setting. At real health system scale, 0.9% is a patient safety committee agenda item, not a rounding error.
Subscribe to www.onhealthcare.tech for free and paid articles, podcasts, and more.
