Data-Driven Criminal Justice in the age of algorithms: epistemic challenges and practical implications

Abstract

The emergence of algorithmic tools and Artificial Intelligence and their use in criminal justice has raised a relevant theoretical and political debate. This article unpacks and synthesizes the debate on the role of causality for the scientific method to analyze predictive decision support systems, their practical value and epistemic problems. As a result of this discussion, it is argued that the measured usage of theory and causation-based algorithms is preferable over correlational (i.e., causally opaque) algorithms as support tools in the penal system. At the same time, the usage of the latter is supported when it is critically accompanied by abductive reasoning. Finally, the arguments put forth in this article suggest that the field of criminology needs a deeper epistemological understanding of the scientific value of data-driven tools in order to entertain a serious debate on their use.

The persistent attractions of low-tech: Challenging the efficiency paradigm of forensic technology

“…Scholars in Criminology and Science and Technology Studies have long been interested in the relationships between technological phenomena and crime and control (Aas et al, 2009; Brown, 2006; Cole, 2001; Powell et al, 2018). Precisely the rapid development within technological innovation mandates a continued critical engagement with the field (Castro-Toledo et al, 2023; Hopman and Bleumink, 2023; Kaufmann and Vestad, 2023). However, even though low-tech innovations are often interconnected with high-tech development (Hirsch-Kreinsen et al, 2008), their influence remains largely under-researched, especially in the fields of forensics and policing.…”