June 2025
In just a few years, generative artificial intelligence surpassed laboratory experiments and became the central point of global security. Up to mid-2025, it has become a part of the state’s infrastructural power. It no longer functions only as a technological resource. It has become a mechanism of political influence and a new form of encryption of power in the digital environment.
Conflicts around access to models and data increasingly resemble classic geopolitical fights. States protect their systems as strategic nodes. We are entering the era of algorithm politics, in which the control over data and servers defines the rank on the global scale.
In this economy, a tri-layered hierarchy is formed. At the top, there are the US and China. The United States relies on private giants and coordination with state security structures. China is building its digital sovereignty through centralized models and state institutions. Each model becomes both a technological and a political product.
The mid-level consists of the EU, India, Japan, and South Korea. They are attempting to balance between the two poles. The European Union is attempting through the AI Act to become a regulatory center, but its influence depends on the development of the domestic infrastructure. Without their own models, regulation remains an empty form.
At the bottom of the pyramid are smaller countries. The Balkan countries are in this category. They do not possess capacities to develop their own systems. They use someone else’s technology. They often serve as testing polygons. Instead of energy dependence, an algorithmic form of dependence.
Sovereignty nowadays is less and less measured through territory. It is measured through control over data flow. Generative models work on big data – from public texts to biometric forms. Whoever holds these flows also holds the formation of social narrative.
Here, the deepest layer of power is invisible. Generative artificial intelligence influences the definition of truth. Its algorithms have soft, but still strong effects. Countries thus often create regimes of control over national models. Security standards and ethics serve as content filters. In practice, it is a new version of information monitoring.
Access to models is becoming a strategic issue. Data is being transformed into geopolitical currency. The one who manages them also dictates the rhythm of public discourse.
Artificial intelligence is also present in diplomacy. Technologically developed countries offer models as instruments of influence. Aid, loans, and transfer of technology – it all comes with political conditions. China is promoting the Digital Silk Road 2.0 and is spreading its infrastructure in Africa, Central Asia, and the Middle East. The United States responded by forming the democratic AI Alliance, with assistance from the G7 and the EU. Behind ethical principles lies the logic of limiting Chinese influence.
Digital diplomacy is becoming the new battlefield. Small countries must choose models. With this choice, they are also choosing the geopolitical framework as well.
The European Union is attempting to present itself as the third pillar, but is facing structural limitations. Its rules run late behind innovations. The majority of European companies rely on American and Chinese infrastructure. And so, the EU becomes a regulator without control. Attempts to strengthen digital sovereignty – such as the AI4EU fund – depend on a complex coordination of member states.
In the Balkans, an additional paradox emerges. States are obligated to apply European standards, but they do not possess their own infrastructure. They depend on someone else’s algorithms. This issue of digital sovereignty will become the key geopolitical topic in the years to come.
Generative artificial intelligence is no longer just a technological challenge. It is the new infrastructure of power. The control over models functions as a model version of the control over energy sources. The countries that don’t protect their own digital space are risking losing their political autonomy. In such dynamics, generative intelligence is becoming the division line: standing between the ones who construct order and the ones who adapt to someone else’s rules.
Author: Aleksandar Stanković