Algorithms and Recidivism: A Multi-disciplinary Systematic Review

Abstract

The adoption of algorithms across different jurisdictions have transformed the workings of the criminal justice system, particularly in predicting recidivism risk for bail, sentencing, and parole decisions. This shift from human decision-making to statistical or algorithmic tool-assisted decision-making has prompted discussions regarding the legitimacy of such adoption. Our paper presents the results of a systematic review of the literature on criminal recidivism, spanning both legal and empirical perspectives. By coalescing different approaches, we highlight the most prominent themes that have garnered the attention of researchers so far and some that warrant further investigation.