Standardizing AI
Martin Ebers
Abstract
The EU Artificial Intelligence Act proposal is based on a risk-oriented approach. While AI systems that pose an unacceptable risk will be banned, high-risk AI systems will be subject to strict obligations before they can be put on the market. Most of the provisions deal with high-risk systems, setting out obligations on providers, users and other participants across the AI value chain. At the heart of the proposal is the notion of co-regulation through standardization based on the New Legislative Framework. Accordingly, this chapter provides a critical analysis of the proposal, with particular focus on how the co-regulation, standardization and certification system envisaged contributes to European governance of AI and addresses the manifold ethical and legal concerns of (high-risk) AI systems.