External AI tools blocked from public hospital networks over cyber safety fears

Hospital employees can no longer reach ChatGPT, Claude or Gemini from work computers, as Health Ministry sounds alarm over potential cyberattacks, leaks of sensitive medical information

The Health Ministry blocked access this week to public AI tools on the organizational networks of public hospitals across Israel, in an effort to protect patient privacy and prevent cyberattacks.

The move came some four months after the ministry issued a warning, urging hospitals to limit their use of the tools as much as possible.

The block itself, however, took workers as a surprise, according to Channel 12.

Employees are now unable to access tools like ChatGPT, Claude or Gemini on hospital networks, though they can access them on their personal mobile devices.

“Protecting patients’ privacy, medical confidentiality, and securing information are at the top of our priorities when it comes to any technology,” the ministry said in response to reporting about the block.

“The health ministry sees in AI a significant engine of innovation and progress for the health system, and is working to encourage its implementation,” the ministry continued.

But it warned: “The use of free, external AI tools multiplies the cyber-risks and the risk of leaking sensitive medical information.”

“In March 2026, the Health Ministry published guidelines obligating all health organizations to stop using these tools in organizational networks and to move to secure tools only, subject to appropriate risk management,” it said.

The Health Ministry said it was working to find safety solutions “that will allow the expansion of the use of AI, including on computers connected to organizational networks, while adhering to the highest standards of privacy, confidentiality, and cyber-defense.”