ABSTRACT
The EU AI Act is a shift in European technology governance. It embeds co-regulation and soft-law mechanisms alongside binding legal obligations. 1 Co-regulation, as conceptualised in the AI Act, refers to a hybrid regulatory model in which public authorities and private actors, such as industry consortia and standardisation organisations, jointly develop and refine rules and standards for AI systems. 2 This approach is intended to harness the technical expertise of private actors and ensure public oversight and alignment with fundamental rights and societal values. 3 The co-regulatory turn in AI governance introduces legal and institutional challenges that question its efficiency and enforceability within the broader context of soft law. 4 The choice to embed co-regulation in the AI Act is driven by the recognition that AI technologies evolve rapidly, outpacing the capacity of traditional, command-and-control regulation to respond to new risks and opportunities. 5 The European Commission has explicitly acknowledged that fostering innovation and competitiveness in AI requires regulatory flexibility and stakeholder trust. 6 Co-regulation is seen as a way to balance these competing objectives by involving private actors in the regulatory process, leveraging their technical knowledge and capacity for rapid adaptation. 7 This model is also intended to address the “pacing problem,” which is the persistent lag between technological change and legislative response. It allows the standards and codes of practice to evolve more quickly than statutory law.
