BEYOND COMPLIANCE: EMBEDDING ETHICS-BY-DESIGN IN ARTIFICIAL INTELLIGENCE SYSTEMS UNDER THE EU AI ACT

Abstract

This article explores the evolution of artificial intelligence governance in the European Union, focusing on the shift from regulatory compliance toward ethics-by-design. While the EU Artificial Intelligence Act establishes a comprehensive risk-based framework, the study argues that compliance alone is insufficient to address deeper ethical challenges related to human autonomy, dignity, fairness, and power asymmetries in algorithmic systems. The article conceptualizes ethics-by-design as a form of operational normativity that embeds ethical principles directly into the lifecycle of AI systems. It identifies key mechanisms for such integration, including ethical impact assessments, algorithmic auditing, transparency tools, and human oversight. At the same time, it highlights challenges such as translating abstract values into technical practices and avoiding superficial formalization. The study concludes that ethics-by-design is a necessary condition for the legitimacy of AI governance in the EU. By integrating ethical reasoning into system design, it becomes possible to bridge the gap between legal compliance and substantive justice, reinforcing a human-centric model of digital governance.