Predictive Algorithms in Sentencing and Bail Decisions: A Systematic Review

• In Event: Data and Risk Assessments in Courts

Abstract

The present study is a systematic review of the existing scholarly literature focusing on the use of predictive algorithms in sentencing and bail decisions in criminal courts. The study finds that scholarly literature on this topic is lacking, both in size as well as diversity of the methodologies used to address the topic, as there is an utter lack of experimental design or rigorous qualitative fieldwork. More specifically, literature tends to orient toward either how algorithms are currently used (if so, case law from key examples is usually analyzed) or how algorithms may come to be used. The most prominent themes or points of controversy in this body of literature are found to be data input and output, due process, productivity, and the importance of humans.

Author

• Joseph McDermott, Queen's University