Israel’s Health AI Sandbox Lets Machines Take Over Clinical Tasks — and Rewrite the Rules

Israel’s Ministry of Health and the Israel Innovation Authority have opened the country’s first-ever regulatory sandbox specifically for artificial intelligence in healthcare — and it’s not designed for the meek. The program explicitly targets AI systems that can fully automate tasks currently performed by healthcare professionals, redistribute work among care teams, and even create new professional roles. The submission deadline for the operational sandbox is March 15, 2026, and the initiative is already accepting proposals from Israeli companies that can pilot with clinical sites in the country.

Branded as Israel’s first Health AI Regulatory Sandbox , the initiative is jointly led by the Ministry of Health (MoH) and the Israel Innovation Authority (IIA) , operating under the IIA’s Regulatory Pilot Program (Pilot Fund) . The sandbox has two phases: an exploratory Call for Information issued in 2025 to map potential participants, and now a formal Call for Proposals that makes the sandbox operational. The earlier call, with deadlines in June and July 2025, was a market-sounding exercise. The current call, published in late 2025/early 2026, is the real deal: companies can apply with AI solutions that are already in development, pilot, or implementation stages, as long as they commit to working with a dedicated regulatory team.

The focus areas, confirmed by both official MoH documents and a LinkedIn post by senior participant Ran Milman, include:

Eligible entities range from startups and technology companies to industrial corporations, healthcare organizations, and research institutes. All solutions must be AI-based and capable of piloting in Israeli clinical sites. Participants must submit a concept document (up to three pages) detailing their AI solution, regulatory challenges, and potential impact.

Israel currently has no AI-specific statute for digital health. Medical AI/ML products are regulated under the Medical Devices Law (2012) , the National Health Insurance Law (1995) , Patients’ Rights Law (1996) , and the Protection of Privacy Law (1981) , among others. These frameworks were never designed for AI systems that can autonomously make clinical decisions or redistribute professional authority. The sandbox is a deliberate attempt to design appropriate regulation based on real-world pilots rather than theoretical rules.

An opinion piece in the Jerusalem Post described the initiative as a “bold move” to balance innovation with risk mitigation. The sandbox provides a supervised testing environment where advanced AI can operate in real clinical settings while the regulators — MoH and IIA — observe, collect data, and shape future rules. This is a pragmatic approach in a country that leads in AI startups but has lagged behind in formal regulation for AI in healthcare.

For Israeli companies building high-autonomy medical AI, the sandbox offers a rare combination of financial assistance and structured regulatory engagement. The IIA’s Pilot Fund provides direct funding, and the MoH’s Digital Health Division offers case-specific guidance on navigating existing laws — including privacy, data security, and medical device approvals. Companies that participate will not only test their products but also help write the rulebook for future medical AI.

The program is also open to industrial corporations, not just startups. This signals that the government expects large players — possibly hospital networks, pharma companies, or tech conglomerates — to bring mature AI solutions into the sandbox. The requirement to pilot with clinical sites in Israel means companies must have local partnerships or operations.

The sandbox will run through 2026 and likely beyond, with the MoH reserving the right to open additional calls. The most immediate milestone is the March 15, 2026 submission deadline for the operational sandbox. The first cohort of selected projects will be announced in mid-2026, and their results will start informing regulatory updates. If the sandbox succeeds, Israel could emerge as a global reference for how to regulate autonomous medical AI — not through static laws, but through iterative, evidence-based rulemaking. The rest of the world will be watching closely.