Artificial Intelligence (AI) is rapidly transforming global technological landscapes. The transformative technology introduces new levels of disruption and reorders traditional power balances. At the same time, it offers possibilities for increased productivity, efficiency and automation. AI’s far-reaching implications span the societal, economic, political and military domains, with direct consequences for national security.

This report assesses how AI developments may threaten national security interests, while recognising that responses often require both Dutch and EU-level action. In line with the 2023 Dutch National Security Strategy, we understand national security as the protection of the Netherlands’ six national security interests:[2] safeguarding its territory (territorial security); protecting its people (physical safety); the functioning of critical societal and economic sectors (economic security); the resilience of its living environment (ecological security); the democratic rule of law and social cohesion (social and political stability); and the country’s capacity to act autonomously in international affairs and uphold the international legal order and stability.

The current AI boom and increasing availability of AI software and open-source models is facilitating the proliferation and potential democratisation of AI technologies. The AI models that are most used by consumers – especially Large Language Models (LLMs) – are ‘proprietary models’, as they belong to a few (mostly) American companies. A shift towards open-source models – of varying quality, safety and bias – could help reduce Europe’s digital dependencies and thereby contribute to addressing national security concerns. After all, both state and non-state actors increasingly have access to powerful tools capable of shaping public discourse, executing cyber operations and deploying autonomous systems.

In this volatile context, foresight exercises, including scenario building, can aid in imagining and preparing for a range of so-called ‘Plausible Tomorrows’, each with a unique set of challenges and opportunities. Against this backdrop, this report asks: How might AI developments threaten or strengthen the Netherlands’ national security interests, and what actions can Dutch and EU policymakers take to mitigate risks and seize opportunities? To address this question, we employ a foresight approach that uses scenario building to explore ‘plausible tomorrows of AI’s impact on national security.[3]

The report proceeds as follows: it first introduces the analytical framework and key trends shaping AI developments; then presents three Plausible Tomorrows that explore possible futures; and finally distils policy implications and recommendations for Dutch and EU decision-makers.

On the Global Chessboard: Competing for AI Power

While Generative AI, for instance LLMs, has captured public attention since the launch of the ChatGPT chatbot at the end of 2022, AI’s disruptive potential goes well beyond natural language processing. Progress in computer vision and autonomous robotics, for instance, supports new capabilities in surveillance, battlefield automation and decision-making systems. The impact is currently most visible in battlefield operations, in theatres like Ukraine or the Middle East, but also in the United States (US), where big-data analytics and integration company Palantir promises to take surveillance a step further by integrating data from multiple federal agencies,[4] and in China, which is emerging in pole position in the development and commercialisation of humanoid robots.[5]

Amid these trends, leading powers the United States and China have adopted proactive strategies to secure their positions. The Trump administration’s AI Action Plan of July 2025 adopts a largely hands-off, pro-innovation approach to governing AI and calls for the US to ‘win the race’ in this rapidly evolving technology and related industries.[6] The US and China lead on frontier models (mostly LLMs) and applications (with the US still leading) as well as in AI-related research and development and innovation (R&D&I, where China has taken the lead).[7] New players in the Middle East and the Gulf, such as Israel, Saudi Arabia and the United Arab Emirates, are entering agreements with companies like Google and chip design giant Nvidia to get a foot in the door.[8]

The European Union (EU) and leading European countries are dealing with limited access to capital and computing power, as well as regulatory dilemmas and industrial gaps. As frontrunners, the United Kingdom and France do hold key AI software (such as British Synthesia) and a leading LLM foundational model (French Mistral’s Le Chat). The Netherlands also holds relative AI software strengths, including on the infrastructure side (Axelera AI and Nebius AI) and in sectors such as AgriTech and HealthTech (for example, Nedap and Philips). Even when taken together, however, European countries lag behind the US and China.[9]

Set against this context, the EU positioned itself as the regulatory standard-setter for AI: the world’s first binding regulation for ethical AI, the AI Act, has been in effect since August 2025. More recently, the EU and its member states have also been aiming to enhance the bloc’s competitiveness in the field. Recent and upcoming EU initiatives like InvestAI, the AI Continent Invest Plan and the Cloud and AI Developments Act include investments in computational power (hereafter ‘compute’), access to finance for European start-up and scale-up companies, and commercialisation of R&D&I. The success of such initiatives and the ingenuity of European citizens and businesses will be indispensable in a future where AI systems are expected to play a central role in military deterrence, economic competitiveness and potentially also for work automation in aging societies.

Plausible Tomorrows Guiding Action Today

The nature of AI-induced disruptions, as well as asymmetric access to advanced AI technologies, create real and present challenges for the Dutch national security apparatus, today and tomorrow. Starting with a baseline assessment of present-day AI technologies and applications based on expert interviews, this report will introduce three scenarios or ‘Plausible Tomorrows’, intended to help with imagining potential AI-induced national security risks. An analysis of the national security implications for the Netherlands and the EU follows the description of each plausible tomorrow.