The explainability of automated decision-making: a historical perspective through EU Legislation

Abstract

There has been much discussion about the existence of a right to explanation of automated decision-making (ADM) in the General Data Protection Regulation (GDPR). However, little attention has been given to the evolution of the regulation of ADM, within the European Union, over the past thirty years. This paper aims to fill this gap in the literature, providing the reader with a look at this topic through the lens of a historical perspective, starting from the very first regulation of ADM in the Data Protection Directive, continuing with the GDPR and, finally, analysing how the right to explanation has ultimately been established in the Artificial Intelligence Act. We will also see how the EU has addressed the issue of transparency and explainability of ADM in other recent pieces of legislation (the 2019 reform of the EU consumer protection law, the Digital Services Act and the Platform Work Directive). Starting from this historical reconstruction of the EU regulation, a common thread will be identified: the tendency to impose increasingly stringent rules regarding the transparency and explainability of ADM. Lastly, three possible explanations for this regulatory development will be proposed.