Cybersecurity

How AI is transforming cyber warfare and making critical infrastructure vulnerable targets

Nadar Zafrir, CEO of Check Point Software Technologies, an Israeli cybersecurity firm, speaking exclusively to Il Sole 24 Ore: “Critical infrastructure is just the tip of the iceberg, and governments can slow down the spread of AI models, but they cannot stop it.”

by Gianni Rusconi

11 July 2026

Aggiungi Il Sole 24 Ore
ai preferiti su Google

Check Point Engage 2026 - Nadav Zafrir Ceo

3' min read

Translated by AI

Versione italiana

3' min read

Translated by AI

Versione italiana

The wars in Ukraine and the Middle East are also changing the face of cyber security, but the real turning point has (probably) yet to come. This is the view of Nadav Zafrir, CEO of Check Point Software Technologies and former commander of Unit 8200 of Israeli military intelligence, who believes that the true potential of cyberwarfare will be accelerated by the spread of artificial intelligence. In this interview with *Il Sole 24 Ore*, he explains why critical infrastructure will become the main target of attacks and what role governments and businesses might play in the race for AI.

Have recent conflicts changed the face of cyber warfare?

I believe we have not yet witnessed anything truly dramatic: cyber warfare has existed practically since the dawn of the Internet, and with the world’s progressive digitalisation, the attack surface has expanded and, consequently, the opportunities for attack have increased. Cybercrime continues to grow because it is a phenomenon that is difficult to combat; alongside financially motivated initiatives, there are the activities of nation states, ranging from industrial espionage to more destructive ends. Cyberspace knows no borders and is one of the tools in the arsenal of any conflict, used primarily to gather intelligence; yet the most profound change is yet to come.

Why does he claim that the worst is yet to come?

Because the democratisation of technology and artificial intelligence will make the world far more vulnerable. Critical infrastructure, essential public services and the healthcare sector are set to become increasingly exposed targets. We have already seen this with a number of attacks on hospitals, but I believe that what we are seeing today is merely the tip of the iceberg. The most advanced AI models are becoming increasingly effective at identifying vulnerabilities, and this makes it much more difficult to protect industrial and IoT environments, Operational Technology (OT) and SCADA systems, connected networks and PLC controllers.

What has changed compared to the past?

The distinction between military and civilian targets is becoming increasingly blurred because everyone uses the same digital infrastructure. We have entered an era in which the vulnerability of critical systems – which cannot be updated as quickly as traditional IT systems – is on the rise. That is why I believe that cybersecurity needs to adopt a new approach.

How?

In recent years, we have invested heavily in our ability to detect attacks and respond to incidents, but this is no longer enough. We need a revival of prevention: the primary objective of security must once again be to prevent an attack from succeeding. At the same time, we must automate our countermeasures, thinking in terms of hours, minutes and, in some cases, milliseconds. Attackers now operate at the speed of machines, and defences must keep pace. It is no longer simply a matter of protecting networks, data centres or devices: the aim is no longer merely to provide cybersecurity, but to ensure that AI-driven transformation is secure in every business and public sector organisation.

The United States has lifted certain restrictions on access to Anthropic’s most advanced models. What role can governments play?

It is understandable that there should be a certain level of control over cutting-edge artificial intelligence models and the ways in which they are made available, but we must be realistic: governments can slow down this process of innovation, but they cannot stop it. If one company does not develop a model, another will: this is not a race confined to the West; it is a global phenomenon.

What is the new frontier in defence?

Agents. More and more organisations will be entrusting operational and decision-making tasks to these autonomous tools, and this introduces a completely new attack surface. In the financial sector, for example, we are seeing the emergence of autonomous agents that will act as banking advisers, opening up a completely new attack surface. This is why we are developing Small Language Models capable of understanding the agent’s behaviour and protecting it in real time, preventing it from being manipulated. The challenge is not merely to integrate increasingly powerful models, but to build a layer of security capable of operating during the execution phase, at the very moment the AI makes a decision.

Brand connect

I prossimi eventi

Tutti gli eventi

Notizie e approfondimenti sugli avvenimenti politici, economici e finanziari.

How AI is transforming cyber warfare and making critical infrastructure vulnerable targets

Nadar Zafrir, CEO of Check Point Software Technologies, an Israeli cybersecurity firm, speaking exclusively to *Il Sole 24 Ore*: “Critical infrastructure is just the tip of the iceberg, and governments can slow down the spread of AI models, but they cannot stop it.”

Brand connect

I prossimi eventi

Newsletter

Nadar Zafrir, CEO of Check Point Software Technologies, an Israeli cybersecurity firm, speaking exclusively to Il Sole 24 Ore: “Critical infrastructure is just the tip of the iceberg, and governments can slow down the spread of AI models, but they cannot stop it.”