Israeli Health Ministry Blocks AI Tools in Hospitals Amid Rising Cybersecurity and Clinical Risks
Last week, Israel's Ministry of Health proactively blocked access to external AI tools such as ChatGPT, Gemini, and Claude on computers connected to government hospital networks. This decision follows a directive issued in March amid increased cybersecurity threats to the healthcare system during Operation Breaking Dawn. The ministry has now tightened restrictions by intentionally blocking these AI tools, aiming to reduce regulatory liability for negligent use. However, doctors can still access AI applications via personal devices like smartphones and laptops, limiting the effectiveness of the ban.
A physician interviewed by ynet explained that due to many blocked websites in hospitals, he routinely uses his personal laptop to access AI tools, which he finds beneficial for medical research and administrative tasks, though not for clinical diagnosis. He noted AI significantly reduces time spent on presentations and research grant applications.
Professor Shahar Shalev of the Technion and Rambam Medical Center highlighted that the ban addresses only one risk vector: preventing cyberattacks via hospital networks. It does not mitigate risks of sensitive patient data leakage, as doctors can still input information through uncontrolled channels, nor does it reduce clinical risks from potentially inaccurate AI-generated medical advice. Shalev emphasized that blocking AI access merely shifts responsibility from the health system to individual physicians without managing the underlying risks.
AI use in medicine has surged globally, driven by staff shortages and workload. A 2025 American Medical Association survey showed AI adoption among doctors nearly doubled from 38% in 2023 to 66% in 2024, with growing enthusiasm despite ongoing concerns. Israeli hospitals have some internal AI tools integrated into their networks, but these currently assist only with simpler tasks like drafting emails or summarizing articles, not complex clinical decision-making.
Shalev argued the Ministry of Health should focus on defining clinical standards for AI use rather than outright bans. He called for regulatory frameworks that specify permissible data sharing and usage practices, enabling safe, monitored AI integration. He warned that AI is transforming medical cognition and that resistance through blocking is futile.
The ministry’s move sparked debate on social media, with medical professionals noting widespread informal sharing of sensitive patient data via personal devices and messaging apps. Professor Ido Wolf of Ichilov Hospital criticized the policy, suggesting secure, organizational AI chatbots would be preferable but require investment. ChatGPT itself commented that a total ban is extreme and supports regulated, cautious AI use to balance benefits like time savings and diagnostic aid against risks such as misinformation and privacy breaches.
