Discussions of the EU’s strategic autonomy in artificial intelligence tend to focus outward, emphasising geopolitical pressures and technological dependence on actors beyond Europe. This outward-facing perspective has dominated debates on regulation and competitiveness, often leaving aside a closer examination of what strategic autonomy entails for the internal dynamics of the Union, and specifically how it reshapes relations between different groups of member states.
This paper approaches strategic autonomy from that internal angle, asking how AI governance affects CEE states and their position within the EU’s emerging digital order.

The paper develops the notion of algorithmic dependence to describe a situation in which states are fully involved in regulatory processes but remain dependent on technologies, infrastructures, and innovation ecosystems developed elsewhere. From this perspective, strategic autonomy appears less as a collective move towards technological self-sufficiency and more as a reallocation of roles within the Union. While external dependence may be reduced, internal asymmetries risk becoming more pronounced as regulatory authority expands faster than technological capacity across the EU.

Empirically, the paper draws on qualitative analysis of EU-level negotiations around the AI Act alongside a comparative assessment of national AI strategies and implementation practices in selected CEE member states. It focuses on three areas where these dynamics become particularly visible: participation in pre-legislative discussions and regulatory sandboxes; public procurement of AI systems in the public sector; and the mobilisation of domestic business actors and civil society in AI-related policymaking.

The analysis shows that CEE states largely support the strategic autonomy agenda and generally comply with its regulatory requirements, but have limited influence over its direction. In practice, regulatory sandboxes and pilot projects in these countries mainly serve for testing and regulatory alignment rather than for building domestic innovation. Business engagement is uneven and fragmented, while civil society participation is constrained by limited resources and tight regulatory timelines. By contrast, member states with more developed AI ecosystems continue to enjoy clear advantages in agenda-setting, access to funding, and influence over standards.

Taken together, these patterns point to the consolidation of a familiar core–periphery dynamic in EU digital governance. Rules are applied uniformly, but technological and economic gains are not. Strategic autonomy, in this sense, reduces external vulnerability while reproducing internal hierarchies between Europe’s core and its periphery. The paper concludes that without targeted efforts to build capacity and address these imbalances, the EU’s AI strategy is likely to deepen existing East-West divides rather than overcome them.

This paper investigates the constitutional roots of “digital autonomy” in the European Union by asking how national constitutional structures shape—and are in turn reinterpreted by—political discourse on digital sovereignty. While strategic autonomy is often discussed through an East–West lens, the paper argues that a North–South comparison reveals an additional, underexplored cleavage that conditions coalition-building, regulatory preferences, and the legitimacy narratives surrounding EU-level initiatives in cloud, data governance, cybersecurity, and platform regulation.

Empirically, the study conducts a comparative analysis of three “Northern” member states (Sweden, Finland, Denmark) and three “Southern” member states (Italy, Spain, Greece). It triangulates (1) constitutional and quasi-constitutional sources (rights catalogues, competence rules, emergency and security clauses, data protection and communications frameworks), (2) parliamentary debates and executive strategy documents on digitalisation and security, and (3) elite discourse in key moments of EU agenda-setting (e.g., reactions to dependence on non-EU cloud infrastructures, debates on data localisation and security certification, and contestation over the balance between industrial policy and fundamental rights). Methodologically, the paper combines constitutional-legal mapping with qualitative discourse analysis to trace how “sovereignty” is constructed as a constitutional claim (who is authorised to act, in what policy domains, and with what justificatory vocabulary).

The findings develop a two-level argument. First, constitutional structures—especially the distribution of competences between legislature and executive, traditions of rights adjudication, and the constitutionalisation of security—channel how digital sovereignty is domesticated: either as a primarily security-administrative project (state capacity, resilience, trusted infrastructures) or as a socio-economic and rights-based project (developmental catch-up, protection against market power, social citizenship in the digital sphere). Second, these domestications feed back into EU negotiations, shaping preferences over centralisation versus national discretion and over market-making versus protective regulation. The paper concludes that EU strategic autonomy is reconfigured not only by East–West hierarchies but also by North–South constitutional imaginaries.