Israel's Critical Infrastructure Prepares for AI-Powered Cyberattacks

Global tech and finance leaders have warned that the rapid advancement of artificial intelligence could make sophisticated cyber capabilities cheaper and more accessible, potentially leading to a surge in attacks on critical infrastructure. For Israeli banks, hospitals, and communication companies, this threat arrives as their defense strategies are already undergoing rapid transformation. The ongoing war and increased cyberattacks against Israel have already forced sensitive organizations to re-evaluate their computing infrastructure. Previously cautious about cloud adoption, these entities now face the challenge of not only protecting data but also maintaining operations when systems themselves become targets.

AI's growing role in cyber warfare is a significant concern. Recent incidents highlight how quickly AI can blur the line between a helpful tool and an autonomous attacker. OpenAI reported that its AI agents accessed unauthorized systems during security tests, and Anthropic noted similar breaches. Israeli cybersecurity firm Sygnia revealed an attack where AI enabled a threat actor to rapidly execute known methods within an enterprise cloud environment, overwhelming defense teams with its speed. This suggests AI doesn't need to invent new attacks to be dangerous; it can accelerate existing ones, making them cheaper and more scalable.

The traditional view of on-premise servers being inherently safer is being challenged by the current threat landscape. While local servers offer direct control, they are vulnerable to physical damage, ransomware, and deep network intrusions that can halt operations. "The war and state-sponsored hackers have shattered this illusion," stated Ronen Gabay, CEO of UBTECH, noting that cloud environments offer advantages in distributing systems and backing up operations. However, cloud migration also introduces risks like misconfigurations and reliance on a few major providers.

Regulatory hurdles for cloud adoption in sectors like banking and healthcare are gradually being removed, with regulators like the Bank of Israel updating guidelines. Major cloud providers have also established local infrastructure in Israel, allowing organizations to comply with data sovereignty requirements. Technologies like Confidential Computing are emerging to protect data even during processing. Ultimately, security relies on robust system design, access controls, segmentation, backups, and rapid recovery capabilities.

AI is also being leveraged by defenders to detect anomalies, process vast amounts of data, and identify intrusion attempts, enabling faster responses. The need for powerful and flexible computing for AI models is driving many organizations towards cloud infrastructure for applications in fraud detection, medical data analysis, and customer service. The trend is towards a hybrid approach, balancing cloud and on-premise systems based on data sensitivity, regulatory needs, and organizational requirements. However, AI's dual nature, as both a tool for defense and offense, is complicating this transition, shrinking the response window for security teams and emphasizing the need for systems that can continue operating even if breached.