# The Brussels Effect and Artificial Intelligence: How EU regulation will impact the global AI market

**Authors:** Charlotte Siegmann, Markus Anderljung  
**arXiv:** [2208.12645](https://arxiv.org/abs/2208.12645)  
**Date:** 2022-08-23  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2208.12645v1/

## Key points

- The EU is introducing one of the most stringent and comprehensive regulatory regimes for AI globally.
- The report discusses two possibilities for regulatory diffusion: de facto Brussels Effect and de jure Brussels Effect.
- The proposed EU AI Act is expected to have both de facto and de jure Brussels effects for certain aspects of the regulation.
- Large US tech companies with high-risk AI systems will likely be affected by the regulation.
- The upcoming EU regulation could provide an influential operationalization of trustworthy and human-centered AI development and deployment.
- Three case studies on previous instances of the Brussels Effect are included in the report: data privacy regulation, product liability regime, and product safety scheme.
- Tables summarize key points from the proposed AI Act and its potential to produce a Brussels Effect.
- The report highlights the potential of the EU's AI regulation to shape global standards in AI governance.

## Summary

The European Union (EU) is set to introduce one of the most stringent and comprehensive regulatory regimes for artificial intelligence (AI) among major jurisdictions worldwide. Drawing on Anu Bradford's work, the report outlines the mechanisms through which regulatory diffusion may occur. It considers two possibilities: a de facto Brussels Effect, where the EU regulation incentivizes changes in products offered outside the EU, and a de jure Brussels Effect, where other jurisdictions adopt similar regulations. Focusing on the proposed EU AI Act, the report tentatively concludes that both de facto and de jure Brussels effects are likely for certain aspects of the EU regulatory regime. Particularly, large US tech companies with AI systems classified as "high-risk" under the AI Act are expected to experience a de facto effect. The report argues that this upcoming regulation could be crucial in providing an influential operationalization of trustworthy and human-centered AI development and deployment. If significant diffusion of the EU regime is anticipated, ensuring its well-designed becomes a matter of global importance. To provide further context, three case studies on previous instances of the Brussels Effect are included: data privacy regulation, product liability regime, and product safety scheme (including CE marking). Additionally, tables summarize key points from the proposed AI Act and conclusions regarding its potential to produce a Brussels Effect. Overall, this report highlights how the EU's forthcoming AI regulation has significant potential to shape global standards in AI governance and emphasizes the need for careful design to ensure its effectiveness on a global scale.