A criterion-based approach to GDPR’s explanation requirements for automated individual decision-making

Abstract

Automation of decision-making processes represents an essential element of the digital transformation. However, automated data processing based on machine learning methods poses increased threats to the fundamental rights of data subjects. One main reason for this is the fact that tracing and explaining the solution path responsible for a certain machine output requires high technical effort. The new European data protection law provides a framework for explanation requirements that apply to users of the new – automated – technologies. This article outlines the current state of discussion on explanation requirements for automated decisions and advocates a restrictive interpretation of the corresponding provisions in the GDPR.