Sheba Medical Center, Israel’s largest hospital and a fixture on Newsweek’s list of the world’s best hospitals, has signed a strategic partnership with OpenAI that makes it the first medical institution outside the United States to gain access to the company’s dedicated suite of healthcare AI models. The agreement, finalized in late July 2026 through Sheba’s ARC innovation arm, was signed with Nate Gross, OpenAI’s Head of Health, and marks a significant expansion of OpenAI’s push into clinical medicine beyond its home market.
What the Partnership Covers
Under the deal, physicians and researchers at Sheba’s Tel Hashomer campus near Tel Aviv will get secure access to OpenAI’s healthcare-specific models, including ChatGPT Health, a clinician-facing tool introduced in January 2026. The system is designed to let doctors rapidly search millions of peer-reviewed studies, clinical guidelines and public health datasets to support decisions at the bedside, rather than relying on memory or manual literature review during time-pressured rounds. Administrative uses are also part of the rollout: OpenAI’s models will help reduce documentation burden, a persistent complaint among physicians who report spending nearly as much time on notes as on patients.
Tailoring a Global Model to a Local Health System
A distinguishing feature of the Sheba deployment is that the hospital’s own treatment protocols will be integrated into OpenAI’s models, allowing recommendations to reflect Israel’s specific clinical guidelines, drug formularies and patient population rather than a generic, U.S.-centric baseline. Sheba’s ARC innovation center, founded in 2019 by Prof. Yitshak Kreiss and Dr. Eyal Zimlichman, has spent years building partnerships with global technology firms and startups, positioning the hospital as a test bed for emerging medical technology before wider rollout. A Sheba delegation is scheduled to visit OpenAI’s headquarters in August 2026 to continue building out the collaboration.
Background: A Crowded Race for Clinical AI
The partnership arrives amid intensifying competition among major AI labs to establish footholds in hospitals. Anthropic has been expanding its presence in biology and health research ahead of a planned public offering, while a consortium of a dozen U.S. hospital systems recently teamed with imaging AI company Aidoc on diagnostic workflows. OpenAI’s approach with Sheba differs in scope: rather than a narrow diagnostic tool, it offers a broad clinical decision-support and documentation platform meant to touch many parts of a physician’s day, from literature review to note-taking.
Two Perspectives on the Rollout
Supporters argue that giving clinicians instant access to synthesized medical literature could meaningfully cut the time between a patient presenting with an unusual case and a doctor identifying the right diagnostic path, particularly valuable in a hospital system that treats complex trauma and rare disease cases. OpenAI, in a statement, said healthcare is “one of the fields where artificial intelligence can have the greatest impact,” and framed the Sheba deal as a template for other health systems globally.
Skeptics, including some bioethicists and hospital administrators elsewhere, caution that generative AI tools still hallucinate citations and can present outdated guidance with unwarranted confidence. Physicians will be expected to independently verify recommendations before acting on them, and Sheba has said the tools are intended to augment, not replace, clinical judgment. Israeli privacy regulators have also signaled they will monitor how patient data interacts with the models, given cross-border data flows to a U.S.-based company.
What Comes Next
Sheba plans to roll out ChatGPT Health access incrementally across departments over the coming months, starting with internal medicine and oncology units before expanding to emergency and surgical services. If the deployment proves successful, OpenAI executives have indicated interest in replicating the model with hospital systems in Europe and Asia, using Sheba’s rollout as a proof of concept. For patients, the near-term impact may be subtle — faster answers behind the scenes — but hospital leadership says the long-term goal is a fully AI-augmented hospital where administrative friction is minimized and more clinician time is freed for direct patient care.
Sheba treats roughly 1.7 million patient visits a year across its Tel Hashomer campus, including a large volume of complex trauma, rehabilitation and rare-disease referrals from across Israel and neighboring countries, giving the partnership an unusually broad testing ground for how generative AI performs across specialties rather than a single department. Hospital administrators say early internal testing focused heavily on oncology and infectious disease cases, where rapidly synthesizing recent literature has the clearest potential to change a treatment decision within the same clinical encounter. Training for staff is expected to roll out alongside the technology itself, with ARC innovation leaders emphasizing that physicians will need explicit guidance on when to trust a model’s output and when to escalate to a human specialist or literature search of their own.
