OpenAI Links ChatGPT to Epic EHRs, Adds Health Data Plugin

OpenAI has connected ChatGPT for Healthcare to Epic electronic health records and introduced a plugin linking clinicians to nine public healthcare data sources, according to a Sept. 4, 2026 report from Healthcare IT News.

What the Epic integration does

Clinicians can pull patient information into ChatGPT while preparing for visits, reviewing a patient's history, or tracking changes in a patient's condition over time, Healthcare IT News reported, citing OpenAI. According to the report, the tool can pull together details from visit notes, lab results, medications and specialist records, and it links back to the original chart entries that informed its output. Some health systems can embed ChatGPT within their Epic screens so staff do not have to leave the record to use it, the report said.

OpenAI told Healthcare IT News that UCSF Health is piloting the EHR integration. Access is limited to ChatGPT for Healthcare customers through workspace administrators and is not available to individual accounts, per the report.

The public data plugin

OpenAI also launched a Healthcare Public Data plugin linking ChatGPT to nine government and public healthcare data sources, including ClinicalTrials.gov, CMS Coverage, RxNorm, DailyMed and PubMed, Healthcare IT News reported. The publication said the tool can help compare records and identifiers across sources, aiding researchers looking for open trials and eligibility criteria, and letting pharmacy staff check drug labeling through DailyMed. Eligible U.S. users of ChatGPT for Clinicians can access the plugin, according to the report.

Context: broader AI governance discussions

The Epic announcement came the same week Healthcare IT News published a roundup on AI in virtual care and ran separate HIMSS TV segments on AI governance. In one, Infinitus CEO Ankit Jain discussed safeguards needed as AI agents gain more autonomy, so unintended or unauthorized actions do not affect patient care. In another, carried by Healthcare Finance News, HIMSS chief scientific research officer Anne Snowdon described the difficulty healthcare leaders face in locating and securing AI built into medical devices, from sensors to smartphones.

MarketScale analysis: Health IT teams evaluating this integration should confirm what patient data fields are exposed, whether prompts and outputs generate logs that integrate with existing Epic audit and security review processes, and which of the nine plugin data sources matter most for their specialty mix. Teams should also ask how source updates or format changes would be communicated before expanding use beyond a pilot like the one underway at UCSF Health.