The role of codes of conduct and best practices in the EU Artificial Intelligence Regulation: between self-regulation, co-regulation and regulatory effectiveness

Abstract

This study examines the space that the European Union Artificial Intelligence Regulation reserves for self-regulation and co-regulation instruments, as governance mechanisms for controlling the risks arising from the use of artificial intelligence systems, with particular attention to codes of conduct and best practices. First, the doctrinal and regulatory debate surrounding the different alternatives for the regulation of AI is contextualized, contrasting traditional normative intervention models with strategies based on soft law instruments. Next, the concepts of self-regulation and co-regulation are defined, and their promotion within the framework of European technology governance policies is analyzed. Subsequently, the role that the AI Regulation assigns to codes of conduct and best practices is studied, identifying their structural and functional differences, as well as their suitability and effectiveness in complementing binding regulation. Finally, reference is made to the Code of Practice for General Purpose AI Models (2025) as a case study, and the main reflections and conclusions of the work are presented, aimed at assessing the effectiveness and limits of these instruments and formulating proposals in this regard.