Brussels effect or experimentalism? The EU AI Act and global standard-setting

Abstract

Supporters of the EU’s Artificial Intelligence Act have presented it as promising to define the global standard for AI regulation, thus emulating the widely heralded ‘Brussels effect’. This paper juxtaposes this expectation with the alternative position of experimentalist governance, which rather sees the EU’s AI Act as one approach to AI regulation among others and that conceives of its interaction with other regulatory approaches across the world in a more cooperative and open-ended way. The paper explores the differences between these two theories along two lines. First, comparing the nature of AI with established digital technologies, it finds that the fundamental uncertainty that accompanies AI calls for regulators across the world to adopt a more interventionist approach that balances between its promises and perils. Secondly, looking at the contents of the AI Act, the analysis establishes that it remains very procedural, reflecting a distinctively incremental and open-ended approach. On the whole, these analyses suggest that the external impact of the AI Act is more in line with experimentalist governance than with the theory of the Brussels effect.