Predicting Proportionality: Algorithmic Decision-Making in Sentencing
Sentencing in many jurisdictions remains quite discretionary, with significant variability in how judges approach otherwise similar cases, raising concerns of both arbitrariness and bias. This paper proposes systematizing judgments of proportionality in sentencing by means of an algorithm. The aim of such an algorithm would be to predict what a typical judge in that jurisdiction would regard as a proportionate sentence in a particular case. Notably, unlike most discussions of algorithmic decision-making in the criminal law, the objective of the algorithm would be on predicting the behavior of judges rather than defendants. I show that endorsing such an algorithm does not come at the cost of case-specific justice, that it is consistent with a highly particularistic account of moral judgment, and that it is attractive even despite pervasive uncertainty as to the point of punishment.
Law & Criminology
University of Toronto
Tue, Mar 6, 2018
04:00 PM - 06:00 PM
Centre for Ethics, University of Toronto
Rm 200, Larkin Building