techcrunch.com web signal

OpenAI's ChatGPT Health Plugs Into Epic's Patient Charts

By Published Updated
5 sources tracking this story
OpenAI Healthcare Enterprise AI ai-business

TL;DR

  • OpenAI signed a BAA for read-only Epic access across 325 million patient records; the AI cannot write back to the chart in any configuration.
  • Seven health systems are launch partners with UCSF Health as the cited pilot, but clinical validation data remains proprietary and unpublished.
  • The 99.1% physician safety rating derives from 4,363 internal evaluations across 27 use cases and has not been submitted for independent peer review.

OpenAI has wired ChatGPT into Epic, the electronic health record system that holds data for more than 325 million patients.

Clinicians using the new ChatGPT Health integration can pull appointment notes, lab results, medications and specialist documentation into a chat window without leaving the patient chart, TechCrunch reported. The "integration allows only read-only access to health records, and AI doesn't write anything back."

Alongside it, OpenAI shipped a Healthcare Public Data plug-in that reaches into ClinicalTrials.gov, CMS Coverage, RxNorm, DailyMed and PubMed for trial eligibility, medication identifiers and coverage policies. Organizations that sign Business Associate Agreements can now use ChatGPT Work, Codex, apps and connectors in HIPAA-compliant workflows.

OpenAI puts a safety number on the launch. In an evaluation of 4,300 responses from physicians across 27 clinical use cases, the company says 99.1% were safe, and for each of the five connected data sources more than 93% of responses were rated "good" or better on accuracy. It also puts weekly ChatGPT health-related queries at around 300 million.

UCSF Health is the featured pilot customer. Its president and CEO, Suresh Gunasekaran, said the technology has the potential to reduce time spent synthesizing data and give clinicians more time with patients.

The launch lands in a busy stretch for healthcare AI, coming shortly after the FDA signaled it may weigh clinician-style tests for generative AI medical devices. It's the latest in nearly 350 OpenAI stories we've tracked over the past 90 days.

What others are reporting

Coverage cluster as of 24h after publish

  1. TechTarget SearchHealthIT Read →

    Covers UCSF Health pilot specifics and the nine-dataset public data plugin as a research synthesis layer, framing this as a broader ecosystem play beyond EHR summarization.

    Patient context, medical evidence, public healthcare data, and organizational knowledge often live in different places.
  2. HealthSystemCIO Read →

    Uniquely frames the story as a CIO portfolio management problem: three vendors billing for chart summarization, with Epic's own governance stance toward this integration unconfirmed.

    Health systems that sort the overlap on purpose will spend less than those that discover it later, once three products are live.
  3. Unite.AI Read →

    Names all seven launch partners, distinguishes enterprise versus individual clinician access tiers, and quotes UCSF Health CEO on time-savings rationale.

    By bringing relevant information together more quickly and comprehensively, the technology has the potential to reduce time spent synthesizing data.
  4. PYMNTS Read →

    Maps the three-phase healthcare rollout timeline across 2026, surfacing OpenAI's deliberate consumer-to-clinician-to-EHR sequencing as a market-entry pattern.

    Clinicians can pull authorized information from appointment notes, laboratory results, medications and specialist documentation.

Shared on Bluesky by 2 AI experts

Originally reported by techcrunch.com

Read the original article →

Original headline: OpenAI Wires ChatGPT Into Epic EHR for Clinicians, Adds Healthcare Public Data Plug-In