OpenAI announced on September 1, 2026 that ChatGPT Health now integrates with Epic's electronic health record system, software that holds data for more than 325 million patients, allowing clinicians to import patient data and ask questions of it in natural language. The company also launched a Healthcare Public Data plug-in that pulls from major clinical reference sources, and opened compliant workspace features for healthcare organizations that operate under a Business Associate Agreement.
The integration, reported by TechCrunch, Fierce Healthcare, and Becker's Hospital Review, marks OpenAI's deepest move yet into clinical workflows, following last month's rollout of ChatGPT Health to all US consumers, a launch the company accompanied with the disclosure that people direct around 300 million health-related queries to ChatGPT every week. For more context on this story, see our ongoing breaking AI news.
What clinicians can do with the Epic integration
With the integration in place, clinicians can access information such as appointment notes, laboratory results, medications, and specialist documentation, and use ChatGPT to summarize them. OpenAI says the assistant can help review patient history, identify changes over time, and prepare for upcoming appointments. In some deployments, ChatGPT sits directly inside EHR workflows, letting clinicians run pre-visit reviews and build clinical timelines without leaving the patient chart.
Access is read-only: OpenAI specified that the AI does not write anything back into the health record. That design choice will matter to health-system CIOs, for whom write access by a third-party AI touches a much more sensitive set of safety and regulatory questions than summarization and retrieval.
Public data sources and compliant workspaces
Alongside the Epic integration, OpenAI launched a Healthcare Public Data plug-in that can fetch information from ClinicalTrials.gov, CMS Coverage, RxNorm, DailyMed, and PubMed. The company says the plug-in can help healthcare workers synthesize trial eligibility criteria, medication identifiers, coverage policy versions, and provider records.
OpenAI is also allowing organizations with a Business Associate Agreement, the contract required under HIPAA when vendors handle protected health information, to use ChatGPT Work, Codex, apps, and connectors in their workspaces for compliant workflows. That effectively extends the product from an assistant for individual clinicians to a platform health systems can deploy across teams.
The physician feedback OpenAI collected maps closely onto the work the integration is built for. The 27 clinical use cases span pre-visit review, clinical timelines, medication review, and handoff summaries, tasks that consume clinical time without requiring diagnosis. Fierce Healthcare's report on the announcement paired the Epic integration with the new public health data connections, giving clinicians a way to check trial eligibility criteria or coverage policies without leaving the chart.
Safety record and the liability question
OpenAI says it collected more than 4,300 physician responses across 27 clinical use cases, including pre-visit review, clinical timelines, medication review, and handoff summaries, and found that 99.1 percent of responses were rated safe. The company continues to maintain that AI is not suitable for diagnosis or treatment, and the integration is positioned around information synthesis and preparation rather than clinical judgment.
The safety margin, however, is not the same as zero risk, and the legal environment is active. TechCrunch notes that a Florida-based pastor sued OpenAI shortly before this rollout, alleging ChatGPT gave him a near-fatal recommendation, and the company was sued in May by family members of a user who blamed the chatbot for wrongful dosage advice. Embedding the model inside the record system for 325 million patients raises the stakes of any failure, even as read-only access and clinician-in-the-loop use are designed to contain it.
Epic's scale makes this the biggest EHR-AI tie-up yet
Epic is the dominant EHR vendor in US hospitals, and its decision to host ChatGPT inside clinical workflows gives OpenAI distribution that no pilot program could match. Competing assistants have pursued similar ground: Microsoft's Nuance DAX and Epic's own in-house AI features are already embedded in documentation pipelines, and Google has health integrations through its cloud unit. What distinguishes the OpenAI deployment is consumer-grade conversational ability arriving in the exam-room workflow, with a usage base already habituated by 300 million weekly health questions on the consumer side.
The pace is also notable. ChatGPT Health reached all US consumers only last month, and within roughly four weeks the product has gained read-only EHR access through Epic, a multi-source clinical reference plug-in, and BAA-gated workspace features. The sequence, consumer adoption first, regulated-industry deployment second, mirrors the playbook OpenAI used to push ChatGPT into other enterprise functions, and it will test whether health systems move at the same speed the consumer market did.
The coming test is adoption in regulated environments. If health systems embrace pre-visit summaries and timeline building at scale, the integration could meaningfully cut documentation time; if incident reports accumulate, expect regulators and plaintiffs' lawyers to probe the 0.9 percent of responses that did not meet OpenAI's safety bar. Either way, the boundary OpenAI has drawn, read-only access, no diagnosis, clinicians reviewing everything, is now the template other AI vendors will be measured against.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
