In Defense of Algorithmic Governance

Adam Lovett (Australian Catholic University)

Government bureaucracies routinely make decisions about individual cases: whether to grant an applicant benefits, whether to investigate a business for rule-breaking, or what costs to impose on a convict for a crime. In the twentieth century, this was all done by individual human decision-makers: caseworkers, inspectors, judges. Now, it is increasingly done by algorithms. Scholars have been largely critical of this change. In this paper I defend it. My main novel line of argument is that using algorithms to make decisions reduces domination. It replaces the tyranny of street-level bureaucrats with the non-dominating influence of algorithms. Additionally, such algorithms often reduce everyone’s risk of getting their rights violated in public administration. I also address the concerns that such algorithms violate people’s rights to explanation, that they inadmissibly rely on statistical evidence, and that they are discriminatory. The result is a general defense of algorithmic governance.