Il diritto antidiscriminatorio al test di intelligenza artificiale

Abstract

Moving from the analysis of the case concerning the algorithm Frank, as adjudicated by the Tribunal of Bologna, the essay points out the complex challenges that the application of antidiscrimination law faces when the systems of Artificial Intelligence are based on machine-learning approaches. In particular, the problems arising as a consequence of algorithmic opacity, namely the so called «black box» effect, are highlighted. First, the topic of transparency is examined as it is addressed by multiple regulatory fields, with regard to Data, Computer Programs and Trade Secrets Protection. Then, the investigation focuses on the reconstruction of possible solutions aimed at straightening and ensuring the antidiscrimination protection. The reflection builds on the assessment of the measures contained in the proposal for a Regulation on Artificial Intelligence and of the techniques than could be borrowed from the proposal of Directive on pay transparency in the context of gender equality, both presented by the European Commission in 2021.