Contested delegation: Understanding critical public responses to algorithmic decision-making in the UK and Australia

Abstract

In public and private sectors alike, decision-making is increasingly carried out through the employment of ‘algorithmic actors’ and artificial intelligence. The apparent efficiency of these means in the eyes of politicians and the public has made recourse to them possible. Along with this belief in their efficiency, however, fears emerge that nonhuman actors have displaced judicious human decision-making. This article examines this belief and its contestation, drawing on overlapping notions of ‘delegation’ in the political sociologies of Bruno Latour and Pierre Bourdieu. We undertake two case studies of attempts to delegate decision-making to algorithms: the 2020 UK ‘A-level’ grade determination and the Australian ‘robodebt’ welfare funds recovery scheme. In both cases, the decision-making delegated to algorithms was publicly discredited as critics invoked a different form of fairness than the one used by those deploying the technology. In the ‘A-level’ case, complainants drew on a grammar of individual merit, while complainants in the ‘robodebt’ case made a technical critique of the algorithm’s efficiency. Using a theory of delegation, we contribute to understanding how publics articulate resistance to automated decision-making.