University of Twente Student Theses


Machine learning in criminal justice : a philosophical enquiry

Andreoli, Chiara (2019) Machine learning in criminal justice : a philosophical enquiry.

[img] PDF
Abstract:Forecasting in criminal justice can be dated back to at least the 1920s, while the machine-learning version of it is a fairly recent development. This thesis focuses on the specific case of the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) in State of Wisconsin v. Eric L. Loomis, and asks: how can philosophy of science, philosophy of technology and philosophy of criminal justice contribute to the understanding of machine learning in criminal justice? This thesis suggests that philosophy of science, through the discussion of value-ladenness, induction, the difference between explanation and prediction, and inductive and epistemic risk, can contribute to the understanding of the value of the knowledge associated to COMPAS, and the implications of the purposes that this knowledge is expected to serve. Philosophy of technology in turn can provide insight into the cost-benefit analysis that risk assessment entails, by taking a perspective that sees beyond an instrumental view of technology, and addresses panopticism and normalization in machine learning through the development of a Foucauldian argument. Last, philosophy of criminal justice, through an analysis that draws from Angela Davis’ insights into detention in the United States, can provide the analytical tools needed to address the political and social implications of machine learning in criminal justice, and therefore contribute to the understanding of the role machine learning plays in relation to the goals of criminal justice reform.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:08 philosophy, 86 law
Programme:Philosophy of Science, Technology and Society MSc (60024)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page