Belgium proposes EU Agency for algorithm screening in response to AI Act

The Belgian government intends to propose the establishment of an EU agency responsible for evaluating algorithms during its term as Council presidency in the upcoming year. This initiative is a response to the forthcoming EU regulations on AI and content moderation, which demand effective enforcement.

With a focus on ensuring effective enforcement, Belgium intends to advocate for the establishment of an EU agency specifically tasked with evaluating algorithms. This proposal is driven by the recognition that a more unified approach to AI governance is needed, despite some data protection authorities within the EU already possessing algorithm-surveillance powers. To facilitate this vision, Belgium suggests upgrading the European Centre for Algorithmic Transparency (ECAT) in Seville, which currently provides valuable technical expertise in the realm of content moderation. Extensive discussions with multiple EU countries have taken place, and Belgium is committed to prioritising the inclusion of this proposal on the agenda throughout its Council presidency, commencing in January 2024.

The proposal comes amidst concerns that the EU lacks the necessary technical expertise to govern AI and enforce these regulations adequately. Belgian policymakers emphasised the importance of comprehending algorithm functioning to determine their acceptability. They argue that a specialised agency, similar to existing entities such as the European Medicines Agency or ENISA (the EU’s cybersecurity agency), is needed to ensure consistent and comprehensive algorithm evaluation across the EU.