The First Decade of AI Policymaking in Europe: Trends, Gaps, and Strategic Transformations

The 27-Piece Puzzle of European AI Policymaking

• Between 2018 and 2020, 19 EU Member States adopted long-term national AI strategies. However, 18 of those countries revised or updated their strategies within just four to five years.
• The wave of reforms introduced in 2024–2025 indicates that, given rapid technological evolution, AI policymaking is increasingly shifting toward a “rolling-strategy” model.
• European experience underscores the importance of flexible, continuously updated AI strategies, particularly for countries like Georgia, which have yet to develop their first national AI strategy.

Between 2017 and 2021, Europe experienced a notable wave of „AI Strategization“. During this period, 24 of the 27 EU Member States developed and adopted national AI strategies. Although these documents varied in structure and duration, they all shared a common objective: to maximize the potential of AI technologies.

In the modern world, artificial intelligence has become an integral part of human activity. There are few fields in which its integration has not produced transformative change. As Henry Kissinger notes in The Age of AI, artificial intelligence is not merely a technological advancement, but a systemic shift, one that will fundamentally reshape long-standing perceptions and understandings of the world (Kissinger, Schmidt, & Huttenlocher, 2021).

The evolution of AI technologies has also influenced systems of state governance. Today, there is no country that does not rely on AI to perform even basic administrative or operational tasks, whether analyzing public opinion, deploying “robot judges,” using drones to ease traffic jams, or operating software for unmanned military systems. As artificial intelligence becomes increasingly central to public administration and national security, governments have recognized the need to develop coherent national strategies and action plans to guide its adoption and regulation.

The European Human-Rights-Centric Model and Global Competition

Today, there are three main actors that shape the global development of artificial intelligence.

• The United States: A “corporate-centric” model driven by private sector innovation and investment.
• China: A “state-centric” model where the government plays a central role in directing AI development.
• Europe: A “human-rights-centric” model that prioritizes ethical considerations and the protection of individual rights.

European states are trying to balance between the unhindered development of artificial intelligence on the one hand and the protection of human rights on the other. The EU’s aim is to ensure that the rapid advancement of AI – now functioning as a “factory” for processing the digital data of internet users – does not come at the cost of the excessive or unauthorized use of citizens’ personal information. In response, regulatory frameworks such as the GDPR and the EU AI Act have been introduced.

Compared to the Chinese and American models, the development of the European AI model is further complicated by the fact that the European Union is not a single state but a union of 27 equal members, each with its own approaches and attitudes toward AI development. Although the European Commission adopted the EU Coordinated Plan on AI in 2018, followed by a more detailed and operationalized version in 2021 – these documents remain non-binding. They function primarily as recommendations and guidance for Member States rather than as a joint action plan or a mandatory regulatory framework.

Therefore, when discussing AI policy, it is essential to view the EU not only as a single political entity but also as an alliance of 27 independent states. This requires analyzing each country’s AI policies individually and identifying both the common elements that unite them and the distinctive factors that set them apart.

The First wave of National AI Strategies in Europe (2017–2021)

The first European country to join the global wave of AI national strategies launched in 2017 was Finland, followed by Germany, Denmark, the Czech Republic, and Estonia in 2018–2019. Today, every EU Member State – except Croatia and Greece – has adopted a national AI strategy.

It can be said that between 2019 and 2021, Europe experienced the first major wave of national artificial intelligence strategies. During this period, 24 of the 27 EU Member States adopted official documents outlining their plans for the introduction and development of AI technologies. Given the rapid pace of technological advancement, however, it is nearly impossible to design a strategy that remains fully relevant even five to ten years, and in some cases even two to three years after its adoption. This makes it particularly important to examine how states that introduced multi-year AI policies have responded to ongoing technological changes and adapted their strategies accordingly.

Adapting Long-term strategies to rapid technological changes

Currently, out of 27 EU member states, 2 have not yet approved an AI strategy, 3 countries consider AI development under a general technological development strategy, and 3 countries have short-term 2-3-year strategies, while 19 states prefer to develop long-term policy documents, of which 12 countries did not specify a deadline in the original document.

Consequently, the initial wave of “AI strategization” in 2018–2020 was followed by a second, extensive wave of updates in 2024–2025. Of the 19 countries that had adopted long-term AI strategies, 18 either fully replaced their documents, issued updated versions, or introduced supplementary policy adjustments. This pattern highlights how quickly long-term strategies become outdated and demonstrates the need for continuous adaptation to rapid technological change.

At the EU level, both long-term and open-ended strategies share a common weakness: they struggle to keep pace with the rapid dynamics of technological development and, in practice, tend to have a functional lifespan of only 4–5 years. As a result, it appears that the AI policies of most EU Member States have effectively adopted a “rolling-strategy approach,” shaped by the speed and evolving nature of technological progress.

Recommendations

The findings may serve as a useful guide for countries that share European values and are still in the process of developing their first national AI strategy. These insights can be translated into the following recommendations:

• The strategy should be short-term, flexible, and non-static.
• It should align with the principles of the EU AI Act, OECD, and UNESCO.
• It should remain concise and priority-driven, with clearly defined KPIs.
• It should include continuous monitoring and evaluation mechanisms to enable timely updates.

About  the Author

Vladimer (Lado) Napetvaridze, one of the founders of the Georgian AI Association (GAIA), is a Senior Researcher at the Institute of Political Sciences at Ilia State University in Tbilisi, Georgia. He also serves as an Assistant Professor at Tbilisi State University and an Associate Professor at the Georgian Institute of Public Affairs (GIPA).

Lado specializes in the intersection of artificial intelligence, public administration, and governance. Holding a PhD in Political Science with a focus on e-governance implementation, his research interests include AI policy development, digital governance, and the transformative role of emerging technologies in public administration across Eastern Europe. He has served as a Visiting Scholar at Ghent University, where he delivered lectures on building critical AI capabilities in small European states. Lado has authored multiple peer-reviewed publications on AI integration in public administration and state-level AI policymaking, and he actively participates in international conferences addressing artificial intelligence, disinformation, and political processes.

His recent work includes a comparative analysis of AI policy implementation in Belgium, Estonia, and Georgia, supported by the National Science Foundation, as well as a comprehensive monograph on the topic of integrating artificial intelligence into public administration.