Flexibly unpredictable? : The AI Act as a regulator of a moving target

This chapter has its origins in a Swedish version but has been updated through studies conducted as part of a Nordic research project on the 0governance of AI and automated decision-making in the public sector, and within the framework of a research cluster on the AI welfare state. The analysis aims to deepen our understanding of how European AI regulation is grappling with regulating a moving target as exceptional as AI development, and to show how these insights are relevant from a technology governance and supervisory perspective. In the chapter's conclusion, I highlight the need for structures for skills development, collaboration between public authorities, and active interpretation and communication of both existing legislation and future application of the AI Act. The emphasis here is on the flexible elements of the AI Act that defer certain normative issues by broadening the Commission's mandate, or by commissioning standards to be developed to assist regulatory compliance. Particular attention is paid to generative AI, as this is a fast-moving field of technology that has significantly influenced the legislative process.