Dutch intelligence agencies: ‘AI amplifies existing threats to national security’

Dutch intelligence agencies have published an extensive analysis of the impact artificial intelligence (AI) can have on national security.

According to Dutch intelligence agencies like the General Intelligence and Security Service (AIVD), the Military Intelligence and Security Service (MIVD), and the National Coordinator for Counterterrorism and Security (NCTV), AI is a comprehensive technology that will fundamentally change our society. It has the potential to make our lives easier in many ways and can help solve numerous societal problems.

However, the development of AI also has an intended and unintended impact on our national security. The intelligence agencies argue that AI will most likely reinforce existing threats to our national security.

AI in itself isn’t dangerous. However, it allows malicious threat actors to do things that used to require more technical knowledge, time, and money more easily, convincingly, and cost-efficiently.

“It can, therefore, lead to more or better-executed sabotage, disinformation, hacks, radicalization, or theft. Major powers are investing in the development of AI because of its military, digital, and economic potential. As a result, AI can contribute to a shift in the balance of power in the world,” the AIVD, MIVD, and NCTV conclude.

At the moment, Western military technology is superior to that of countries like China and Russia. According to the intelligence agencies, this technological superiority contributes to the current military balance of power and, thus, to the territorial security of the Netherlands, its allies, and the international legal order. But if other countries are better able to deploy AI in the military domain, this balance of power may shift in their favor.

The intelligence agencies claim that although AI’s influence on cyberattacks isn’t clearly present, AI can have a facilitating role in existing cyberattacks. For example, threat actors could use large language models (LLMs) to automatically detect and select potential targets and collect information about them. LLMs could also assist threat actors in composing phishing emails and messages and executing cyberattacks they wouldn’t be able to do without the help of AI, the report says.

The Dutch intelligence agencies say there are two developments to which AI can contribute to national security threats. First, AI applications currently require a lot of high-level expertise and financial resources. These will become more widely and easily available over time, which means threat actors will be able to abuse AI applications more easily in the future.

Second, AI applications are likely to become part of our economy and society. This will provide economic and societal benefits on the one hand, but it will also create new vulnerabilities that malicious actors will try to exploit on the other.

“The impact of AI on national security in the coming years is likely to be determined primarily by malicious actors who deliberately use AI applications to achieve their own goals at the expense of Dutch national security interests,” the intelligence agencies conclude.