In the heart of Brussels, discussions about Artificial Intelligence (AI) often revolve around data protection, ethics, and the celebrated AI Act. However, there is a sector that remains deliberately in the shadows, an "ungoverned space" that will determine the continent's survival in the 21st century: the military application of AI. A recent analysis by the CEPR highlights a harsh truth: strategic autonomy is not a product that can simply be bought off the shelf from American or Chinese tech giants. It requires a domestic infrastructure that Europe is still struggling to build.
The Regulatory Vacuum and the Defense Exception
While the European Union promotes itself as the world's regulatory superpower, the AI Act explicitly excludes applications concerning national security and defense. This exception creates a paradox: while citizens are protected from algorithmic discrimination in banking or employment, systems that will make life-and-death decisions on the battlefield are being developed in a legal vacuum. The lack of a common European framework for military AI means that each member state is following its own path, undermining interoperability and collective defensive strength.
Military AI is not just about the "killer robots" that dominate popular imagination. It is about processing vast amounts of data from satellites, optimizing supply chains during crises, and enhancing cyber defense. As CEPR researchers note, the failure to govern this space is not merely a bureaucratic oversight but a geopolitical vulnerability. Without clear rules of engagement and ethical boundaries, Europe risks being dragged into an arms race where the rules are dictated by others.
The Illusion of Purchased Autonomy
One of the most critical points of current geopolitical reality is the dependence on external technology. Many European states are rushing to equip themselves with AI systems from the US, believing they are filling the gap. However, the "black box" nature of these algorithms means Europe lacks full control over its tools. If the code, training data, and hardware (chips) belong to third powers, then "autonomy" is just a word in official press releases.
- Dependence on Nvidia and Intel for hardware makes European defense a hostage to global supply chains.
- The lack of European cloud infrastructure for military data exposes sensitive information to foreign jurisdiction.
- The "brain drain" of European scientists to Silicon Valley weakens the capacity to develop domestic solutions.
The strategic autonomy that "cannot be bought" refers precisely to this: the need for an integrated ecosystem. From education and research to semiconductor production and the creation of European champion companies in defense tech—such as Germany's Helsing or France's Mistral (in its civilian iteration)—the road is long and requires capital that Europe hesitates to invest.
The War in Ukraine as an Accelerator
The war in Ukraine has served as a violent laboratory for military AI. We have witnessed the first mass use of drones with autonomous targeting capabilities and the use of image analysis software to detect enemy movements in real-time. This experience has shown that the speed of algorithmic decision-making now exceeds human capabilities. For Europe, the lesson is clear: delay in adopting and regulating AI is equivalent to disarmament.
"Technological sovereignty is no longer an option, but a prerequisite for peace. If we do not define the parameters of our military AI, we will be forced to live in a world defined by Silicon Valley or Beijing."
In conclusion, Europe's "ungoverned space" in military AI must be closed. This requires a bold shift from mere regulation to active development. Europe must decide whether it will remain an ethical observer dependent on the protection of others or become a sovereign player capable of defending its values with its own means. Autonomy cannot be bought; it is earned through investment, innovation, and political will.
