Rethinking the List-Based Approach to High-Risk Systems under the AI Act

Abstract

In this article, I critically analyse the expedited procedure for amending the list of high-risk systems under the AI Act. I conclude that the expedited procedure, along with the list-based approach in general, are suboptimal solutions as they fail to safeguard two key objectives: (i) protection of individuals’ fundamental rights; and (ii) legal certainty for businesses. The option of carrying out a revision of the legal instrument through the ordinary legislative procedure, while always a possibility, may be too slow for its purpose and its success is far from certain. As such, I argue: that a test-based approach would have been a better option to future-proof the AI Act; that its building blocks are already include in the AI Act; and that it would have been advantageous both for individuals and businesses.