The politics of algorithmic profiling of unemployed: A framework for analysing legitimation,contestation,and compromises

Abstract

The proliferation of algorithmic profiling in public employment services has given rise to significant controversy, eliciting widespread concern and calls for more legitimate governance. So far, most studies addressing the issue of the legitimacy of algorithmic profiling of the unemployed have addressed it from a normative perspective asking whether such systems and tools are legitimate or not based on various principles such as effectiveness, fairness, data protection, and discrimination. Rather than evaluating the legitimacy of algorithmic profiling, the article explores how algorithmic profiling is legitimized by mapping the moral yardsticks used to evaluate algorithmic profiling. The main aim of the article is to develop a conceptual framework to identify and study the multiple yardsticks that are mobilized to justify and criticize algorithmic profiling. We term these yardsticks ‘worlds of profiling’. We develop and demonstrate the framework through a longitudinal document study of the justification and contestation over an algorithmic profiling tool targeting unemployed persons in Denmark from 2014 to 2022. The article shows how five worlds of profiling are mobilised to evaluate the tool. These are labelled the worlds of accuracy, efficiency, equity, privacy, and discretion.