OpenAI announced it is integrating ChatGPT Health with Epic‘s electronic health record (EHR) system, which holds data for over 325 million patients. Clinicians can now import patient data and use ChatGPT to ask questions about appointment notes, laboratory results, medications, specialist documentation, and more, with the ability to summarize them easily. In some deployments, ChatGPT will be embedded directly within EHR workflows, allowing clinicians to access pre-visit review and build clinical timelines without leaving a patient chart. The integration is strictly read-only: AI does not write any information back into the health records. Additionally, OpenAI is launching a Healthcare Public Data plugin that can pull information from ClinicalTrials.gov, CMS Coverage, RxNorm, DailyMed, and PubMed, helping healthcare workers synthesize data for trial eligibility criteria, medication identifiers, coverage policy versions, and provider records. Organizations with a Business Associate Agreement can now use ChatGPT Work, Codex, apps, and connectors in their workspace for compliant workflows.

The health-focused rollout follows OpenAI‘s launch of ChatGPT for health to all U.S. consumers last month, noting that people ask 300 million health-related queries to ChatGPT every week. Despite deep integrations with medical systems, OpenAI maintains that AI is not suitable for diagnosis or treatment. The company collected over 4,300 responses from physicians across 27 clinical use cases—including pre-visit review, clinical timelines, medication review, and handoff summaries—and found that 99.1% of responses were safe. However, the article notes that even a few unsafe answers can lead to harmful results. A Florida-based pastor recently sued OpenAI, alleging that ChatGPT gave a near-fatal recommendation. In May, the company was also sued by family members of a user who blamed ChatGPT for giving wrongful advice related to dosage.

The integration of ChatGPT with Epic marks a significant step in bringing AI into clinical workflows, but the associated safety concerns and ongoing lawsuits underscore the risks. OpenAI‘s read-only design and disclaimer about diagnosis are intended as safeguards, but the article highlights that harmful outcomes remain possible.