The Artificial Intelligence Act Between the EU and National Levels: The Slovenian Case Study
Sažetak
This article examines the complex legal and policy dimensions of the European Union’s Artificial Intelligence Act (AIA), focusing on its implementation in Slovenia. Although the AIA is a binding regulation directly applicable in all EU Member States, its hybrid nature—part legal mandate, part policy guideline—creates interpretative and practical challenges, especially for national administrative authorities. Using a three-stage methodology (normative analysis, expert survey, and focus group discussions), the study examined how Slovenian authorities perceive the AIA’s legal character, scope of application, and implementation obligations. Respondents included both institutional (ministries) and operational (tax and social services) authorities. While most recognised the formal status of the AIA as law, significant discrepancies were found in terms of awareness and readiness. Operational authorities such as the Centre for Social Work often viewed the AIA as a guideline rather than a binding regulation. Knowledge of related EU legislation (e.g., GDPR, Digital Services Act) varied from authority to authority, which impacted implementation strategies. The findings reveal three critical gaps: legal misinterpretation, institutional readiness, and actual application. Despite a shared understanding of the importance of the AIA, implementation measures, such as training, clear role definitions, and inter-agency collaboration, were insufficient or delayed. In addition, the authorities did not sufficiently recognise their role and the need for national coordination. The study concludes that effective implementation of the AIA requires not only legal clarity, but also structured action plans, sustained intergovernmental cooperation, and practical support mechanisms. These findings are especially relevant for smaller EU states, where legal formalism and limited resources hinder policy implementation. Slovenia’s experience serves as a cautionary tale and a learning opportunity for other Member States grappling with the multi-level governance of AI regulation.