Beyond bias: algorithmic machines, discrimination law and the analogy trap

Abstract

ABSTRACT This article shows how the ‘challenge of regulatory connection’ in an increasingly algorithmic society triggers legal responses guided by analogies between the 'human' and the 'digital' realms. Focusing on non-discrimination law, it argues that analogical reasoning, however, masks how the affordances of algorithmic systems challenge the epistemological foundations of the law. This article unpacks the ‘analogy’ black box to shed light on the normative implications of its abstraction operations. It articulates three central 'traps' that result from attempts to transpose discrimination laws to algorithmic machines and their biases. Thereby, it shows how algorithmic artefacts mediate, displace and erode the regulatory objects, forms of subjectivity and modes of reasoning traditionally constructed by the law. The aim is to expose how algorithmic rationality unsettles the patterns of power distribution and the allocation of burdens and benefits enacted by legal techniques and to give visibility to alternative normative options.