OpenAI Plugs ChatGPT Into Epic's Health Records for 325 Million Patients

OpenAI on September 1, 2026, integrated ChatGPT with Epic Systems' electronic health records, enabling authorized clinicians in ChatGPT for Healthcare and HIPAA-enabled Enterprise workspaces to review read-only patient data from Epic, with UCSF Health as the first pilot partner. Epic's EHR holds data for over 325 million patients, but the integration does not grant ChatGPT access to all those charts; it connects to records clinicians already have permission to see. OpenAI also launched a Healthcare Public Data plugin that searches nine official sources including PubMed, ClinicalTrials.gov, DailyMed, RxNorm, and CMS Coverage.

OpenAI's Epic connection is current, but the cleaner story is narrower than the headline suggests: authorized clinicians can review Epic chart data in ChatGPT for Healthcare, starting with UCSF Health as a pilot partner. OpenAI introduced the Epic integration on September 1, 2026, and the important word is authorized. This isn't ChatGPT roaming through hospital records. In approved ChatGPT for Healthcare and HIPAA-enabled Enterprise workspaces, clinicians can connect to the patient information they already have permission to see in Epic, then ask ChatGPT to pull together notes, lab results, medications, conditions and encounters. It is read-only. That matters. OpenAI's help documentation says the Epic plugin can't update medical records, place orders, message patients or override existing chart permissions. If you're a patient, that distinction is not legal decoration. It is the line between a tool that helps a clinician read a chart faster and a tool that starts acting inside the record. TechCrunch reported that Epic's EHR system holds data for more than 325 million patients. That figure explains why this landed with force, but it should be read carefully. The UCSF Health pilot doesn't mean ChatGPT suddenly has access to 325 million charts. It means OpenAI has connected its healthcare workspace to one of the biggest record systems in U.S. medicine, with San Francisco-based UCSF Health serving as the first named pilot partner. According to OpenAI's announcement, the integration works in two ways. A clinician can bring authorized Epic context into ChatGPT to review a patient's history before an appointment, or, in supported deployments, use ChatGPT directly inside an EHR layout without leaving the patient chart. The company also launched a separate Healthcare Public Data plugin that searches nine official sources, including PubMed, ClinicalTrials.gov, DailyMed, RxNorm and CMS Coverage. OpenAI Is Buying So Many Mac Minis and Studios That Apple Can't Keep Up https://startupfortune.com/openai-is-buying-so-many-mac-minis-and-studios-that-apple-cant-keep-up/ OpenAI has quietly bought tens of thousands of Apple Mac minis and Mac Studios to train its AI agents, and Anthropic is renting more through AWS for similar work. The buying spree has left Apple's own supply chain scrambling, with Tim Cook warning the shortage will last months even after a hardware refresh. - OpenAI buying Mac minis for AI model training https://startupfortune.com/openai-is-buying-so-many-mac-minis-and-studios-that-apple-cant-keep-up/ - Apple Mac Studio shortage due to OpenAI purchases https://startupfortune.com/openai-is-buying-so-many-mac-minis-and-studios-that-apple-cant-keep-up/ Epic Is The Gate Epic Systems is not just another software vendor in this story. It is the daily screen for a huge share of American hospitals, and once a health system builds around it, leaving is painful. That is why this is more than a chatbot feature. If OpenAI wants to be useful inside clinical work, it has to meet doctors where the chart already lives. Epic letting OpenAI into that workflow, even under tight rules, is the bigger signal. OpenAI has been preparing the ground for months. In July, Fierce Healthcare reported that OpenAI made Health in ChatGPT broadly available to U.S. users, letting people connect medical records and wellness data for more personal health conversations. The company said at the time that more than 300 million people use ChatGPT for health-related questions each week. That is a consumer habit. The Epic integration is the clinical version of the same bet. OpenAI is also trying to show hospitals that this isn't a toy. Its September announcement included a quote from Suresh Gunasekaran, UCSF Health's president and CEO, saying the pilot is meant to help clinical teams understand what has changed across a complex patient record and spend less time synthesizing data. That is the practical sell. Doctors don't need another tab. They need five scattered notes and three abnormal labs turned into something they can check quickly before walking into the room. The Risk Is Still Real Security is the whole pitch because it has to be. OpenAI says healthcare organizations using eligible products can operate under a Business Associate Agreement, and its business documentation points to SAML single sign-on, SCIM user management, role-based access controls, audit logs, data residency options and enterprise key management. Those controls help. They don't make the worry disappear. Patient records are not normal business documents. A medication list, an oncology note or a behavioral health diagnosis carries consequences if it leaks or ends up somewhere it shouldn't. Read-only access lowers the risk of direct clinical harm, but it still sends sensitive context into an AI workflow. Hospitals will have to prove to their own lawyers, regulators, clinicians and patients that the gain is worth it. Microsoft is already in the room. Nuance DAX, now part of Microsoft, is used for ambient clinical documentation, and Epic has worked with Microsoft on AI charting. Fierce Healthcare's coverage of the OpenAI rollout put the announcement in that wider fight over which AI tools become part of the EHR itself. That is exactly the right frame. The question is not whether AI reaches the chart. It already has. The question is whose system gets trusted with the closest view of the patient. OpenAI, Anthropic and Over 100 Firms Warn AI Cyberattacks Are Months Away https://startupfortune.com/openai-anthropic-and-over-100-firms-warn-ai-cyberattacks-are-months-away/ OpenAI, Anthropic, Microsoft and more than 100 other companies, including CrowdStrike, Visa and Capital One, signed an August 27, 2026 letter warning that AI-powered cyberattacks on hospitals and water systems are months, not years, away. The same labs sounding the alarm are already selling the defensive AI tools meant to stop it. - AI powered cyberattacks on critical infrastructure timeline https://startupfortune.com/openai-anthropic-and-over-100-firms-warn-ai-cyberattacks-are-months-away/ - when will AI cyberattacks happen to hospitals https://startupfortune.com/openai-anthropic-and-over-100-firms-warn-ai-cyberattacks-are-months-away/ UCSF Health didn't sign up to test a novelty. It signed up because complex charts waste real clinician time, and because summarizing what changed since the last visit is exactly the kind of task a model can help with if the guardrails hold. The test now is simple: whether doctors actually use it when the exam room is full and the schedule is already running late. Also read: AfterQuery Becomes Y Combinator's Fastest Unicorn Ever at $3.2 Billion https://startupfortune.com/afterquery-becomes-y-combinators-fastest-unicorn-ever-at-32-billion/ • Ilya Sutskever Warns Neoclouds Lack Security to Stop a Rogue AI Takeover https://startupfortune.com/ilya-sutskever-warns-neoclouds-lack-security-to-stop-a-rogue-ai-takeover/ • SoftBank's SB Energy Files for IPO While Admitting It Needs OpenAI to Pay Up https://startupfortune.com/softbanks-sb-energy-files-for-ipo-while-admitting-it-needs-openai-to-pay-up/