From the AI Act to a European AI Agency: Completing the Union's Regulatory Architecture

A new academic paper argues that the European Union's landmark AI Act, while a crucial first step, requires a stronger enforcement body to be fully effective. Authored by Georgios Pavlidis and published in the Croatian Yearbook of European Law and Policy, the analysis contends that the European AI Office, established by the Act, should be evolved into a full-fledged European AI Agency. This move would create a centralized, supranational authority with greater capacity for risk assessment, regulatory oversight, and ensuring policy coherence across member states.

The proposed agency would address gaps in the current framework by providing more robust technical expertise and a unified enforcement mechanism. Pavlidis highlights that a strengthened EU-level body is essential not just for ethical governance and consumer protection, but also for the bloc's strategic goal of achieving 'digital and technological sovereignty.' This would allow Europe to better shape global AI standards and maintain competitiveness, rather than merely reacting to innovations developed elsewhere. The paper frames this as the necessary next phase in completing the EU's regulatory architecture for the AI era.