Sheba Medical Center becomes OpenAI’s first international hospital partner

Sheba Medical Center in Israel has established a partnership with OpenAI, marking the first time the artificial intelligence company has collaborated with a hospital outside of the United States. The agreement is designed to integrate advanced AI tools into the medical center’s operations, with the broader objective of transforming the institution into an AI-powered facility by 2030. Ayelet Akselrod-Ballin, the head and chief technology officer of Sheba’s AI Center, stated that the initiative aims to democratise access to AI tools for all staff members, including clinicians and other professionals working across the campus.

The partnership originated from internal discussions within Sheba’s engineering leadership, who approached OpenAI to address specific operational needs. Akselrod-Ballin explained that the vision extends beyond merely introducing technology; it involves educating staff and providing them with the ability to use self-service tools. Under the agreement, clinicians across the network will gain access to ChatGPT for Healthcare on both desktop and mobile devices. Implementation across clinical, research, and operational teams is expected to begin in the coming months. Additionally, the partnership grants Sheba early access to OpenAI’s latest healthcare models before they are made broadly available to the public.

The longer-term strategy includes enabling employees to build their own AI agents and utilise these tools for research purposes. The potential applications are not limited to direct patient care. Akselrod-Ballin identified surgical, regulatory, and financial workflows as key areas where similar AI approaches could be applied. This expansion aligns with Sheba’s existing portfolio of AI projects, which are already in deployment across the medical center.

One such project is SmartER, an ambient AI system developed with the clinical documentation startup ScribeMD. This tool operates in the emergency department by listening to conversations between patients and physicians, transcribing them, and generating structured clinical summaries for review. Another initiative, Project K, was developed in collaboration with Microsoft Israel and KPMG. It utilises an AI avatar to manage patient intake and provides an automated, AI-reliant process for triage. A third system, Beyonder, screens admissions and routes stable patients to Sheba Beyond, a virtual hospital, while flagging complex cases for in-person care.

The rollout of these technologies occurs against a backdrop of global clinician shortages. According to the OECD’s 2025 profile, Israel has 3.5 practicing doctors per 1,000 people, which is below the OECD average of 3.9. Other nations face similar challenges, with Japan reporting 2.6 doctors per 1,000 people. The World Health Organization projects a shortage of nearly one million health workers in the European Region by 2030, while the Association of American Medical Colleges estimates the United States could face a deficit of up to 86,000 physicians by 2036. Akselrod-Ballin described this situation as a crisis in healthcare, noting that AI can help bridge the gap by reducing the time clinicians spend on administrative tasks and information retrieval.

Akselrod-Ballin, who has extensive experience in AI for radiology, distinguishes generative AI from earlier radiology-specific models. She noted that while previous AI applications focused on image interpretation, generative AI addresses the basic level of reducing time wasted on searching electronic health records and summarising text. This type of administrative work does not require the specialised knowledge of nurses or radiologists. Sheba’s internal AI Center, which comprises more than 50 staff members, continues to develop technologies through its innovation arm, ARC. The center has worked on projects such as SmartER and Project K, and its ecosystem includes AI-focused companies like AISAP, which uses AI to support point-of-care ultrasound.

As Sheba builds these capabilities, it is also working to establish the necessary regulatory guardrails, monitoring, and governance structures. Akselrod-Ballin emphasised that the hospital is building these policies as it goes, leveraging the advantage of having an AI Center embedded directly within the institution. The AI Center maps recurring problems across the medical center to identify reusable infrastructure and components. For example, solutions developed for one of Sheba’s five emergency departments can often be adapted for other departments. The aim is to create a generalisable agentic infrastructure where different agents work on various tasks, allowing for the reuse of components across different workflows. This approach may extend to clinical research, where agents could handle operational and administrative tasks, potentially making clinical trials significantly faster.