OpenAI Gives Doctors ChatGPT to Decode Medical Records

OpenAI is plugging ChatGPT into Epic, one of the largest medical record systems in the world. Epic holds the health data of over 325 million patients. With this new integration, doctors can ask an AI to pull up appointment notes, lab results, medications, and specialist reports — then summarize them in simple language. In some hospitals, ChatGPT appears right inside the patient's chart, so caregivers can review a patient's history and prepare for an upcoming visit without flipping through dozens of screens. The goal is to save doctors time and help them spot important changes at a glance.

OpenAI says this AI is on a short leash. It is "read-only," meaning it can look at medical records but cannot write anything back or change a chart. The company also added a plug-in that lets healthcare workers search public sources like ClinicalTrials.gov, drug databases, and insurance coverage policies. That could help staff quickly check whether a patient qualifies for a drug trial or understand why a medicine was approved. But these advanced tools are only for hospitals that have signed strict privacy contracts.

The big safety question remains. OpenAI tested the system with 4,300 physician responses across 27 real-world tasks like pre-visit reviews and medication checks. They said 99.1% of responses were safe — but that still means dozens could be wrong. And a few recent lawsuits are dragging ChatGPT into court: a Florida pastor says it gave him a near-fatal recommendation, and a family blames ChatGPT for wrong medication dosage advice. OpenAI insists its health tool is not meant to diagnose or treat anyone.

Still, this marks a huge step in making AI a normal part of hospital work. For you, it could mean your doctor walks into the exam room already knowing your full history — including things you forgot to mention. But it also means a computer has a hand in your care. For now, your doctor is still the one making the call, not the algorithm.