Description

Given the rapid advances both in capabilities and availability of open-source AI systems, it is becoming easier for malicious actors to abuse AI. Self-navigating drone swarms, for instance, are commonplace weapons used by not only state actors such as Russia and Ukraine, but also by terrorists, activists and journalists. The complexity of police and intelligence work is sky-high: following the proliferation of AI, the internet has been poisoned with disinformation, making truth-finding nearly impossible, and reducing trust and increasing polarisation throughout society.

The upcoming elections for the European Parliament in 2029 are being manipulated on a massive scale, by various actors spreading ‘deepfake’ videos of politicians delivering fake speeches or behaving badly. Meta can still tell some deepfake images from real ones, and as the owner of Facebook, they are better positioned to analyse the origins of posts on their platform. However, Meta has been arbitrary in its willingness to lend a helping hand, loosely trending about the fluctuations in President Trump’s mood, who is serving a third term following the Second – ‘AI-induced’ – Capitol Siege in 2028.

Against the backdrop of all this uncertainty and the enhanced capacity to cause chaos from the widespread availability of open-source AI, a wave of drones, flying in perfect formation under the command of an AI navigation algorithm that constantly re-routes them to evade jamming, strikes its target. This time, the drones hit five large container ships at the Port of Rotterdam. As the drones’ Cobra 5[17] payloads detonate on impact, windows shatter and dockworkers shudder. Minutes later, under the sinister sound of sirens, the port lights up. The ships glow red beneath a thick layer of black smoke: hundreds of Chinese humanoid robots incinerated at the dock. ‘Another attack’, sighs a tired police researcher, ‘but who’s behind this one?’

In Denmark, meanwhile, Danish conservative Rasmussen is pleased and somewhat relieved. ‘This is sure to get me the missing votes!’, he hopes. Later that day, the Danish parliament will vote on extensive legislation aiming to increase surveillance. Massive monitoring operations using advanced AI-enabled cameras will keep everyone safe and sound, promises Rasmus. In his pocket, he carries a letter signed by a group of 1,984 industry advocates.

Scenario Analysis 1: National Security Implications for the Netherlands in the EU Context

In this Plausible Tomorrow, AI is essentially open and widely accessible through open-source models. The Netherlands and the EU have access to the latest AI technologies, just like any other country as well as companies, NGOs, journalists, activists, criminals, extremists and terrorists. Overall, non-state actors are rapidly gaining information and military capabilities, narrowing the gap with nation-states. This wide availability makes it very easy to spread disinformation or coordinate online troll armies, and to enact physical attacks such as the one envisioned above. The Rotterdam drone strike is just the latest case in a new era of ‘semi-autonomous terrorism’.

While AI contributes to detecting cyber threats and vulnerabilities, it equally empowers more users to carry out cyberattacks. AI-generated images and videos are nearly indistinguishable from real content, further weakening traditional media, which were already struggling under the pressure of AI-driven competition. Deepfakes and disinformation are no longer sporadic but systemic – ‘truth decay’ is now a structural condition. This is more than just a media problem: it corrodes the foundations of Dutch democratic governance, and social cohesion is breaking down fast and evermore difficult to avert. Citizens feel disconnected from public discourse, overwhelmed by distrust and a growing sense of fear towards AI systems, speculating about espionage via consumer apps like TikTok or Instagram and feeling a lack of confidence in institutional communication. Even essential sectors like healthcare become compromised. Actors may unknowingly adopt Trojan Horse AI tools, exposing systems to surveillance and cyberattacks. AI-generated content fuels conspiracy-driven civic unrest, for instance around the legitimacy of democratic elections and immigration.

Economically, societies become more vulnerable. The scenario imagines hybrid threats that disrupt physical and digital infrastructure, inevitably impacting the economy. There are also opportunities: in this scenario, the Netherlands makes use of open-source AI to expand its lead in sectors like AgriTech and HealthTech, and to support developments in new high-tech niches where the country has a strong position, like quantum technologies. Other countries apply AI to their existing strengths and new fields too: notably, AI allows the Global South to bridge the knowledge gap.

In the military domain, states struggle to regulate the proliferation of autonomous weapons. There are no clear standards or control mechanisms. The threshold for pre-emptive and preventive strikes lowers, while lethal autonomous weapon systems spread beyond traditional armies. The primary threat is no longer dominance by superpowers, but rather the uncontrollable scale and diffusion of destructive technology. Beyond direct attacks, adversaries may also target the systems that monitor and control Dutch airspace. AI-enabled drone swarms could deliberately jam, spoof or saturate air-traffic control radars and communication channels, creating unsafe skies and jeopardising commercial aviation. This risk links closely to surveillance, as when monitoring systems themselves are manipulated or overloaded, the ability to distinguish genuine threats from noise is eroded. In effect, AI turns the very infrastructure designed to ensure safety into a vulnerability, compounding both territorial and societal security risks.

Besides immediate cyber-, information and military threats, widespread open-source proliferation also poses ecological risks: the uncontrolled spread of energy- and water-intensive AI models accelerates pressure on Dutch and European sustainability targets, with data centres and chip production directly impacting the living environment. Moreover, the absence of effective international rules to govern such diffusion undermines the Netherlands’ capacity to uphold the international legal order, as norms on autonomous weapons, cyber operations and digital sovereignty remain fragmented or ignored.

Many of these challenges already exist today – but the capabilities will be far more accessible to a wider range of actors. For the Netherlands, this Plausible Tomorrow presents a strategic paradox: while AI boosts innovation in key sectors, it also erodes national resilience through asymmetric vulnerabilities.