Algorithmic Injustice: Predictive Policing and Profiling in the Netherlands – the Top600 and Top400 approaches

Abstract

This thesis is located within the emerging intersections between feminist critique and algorithmic injustice, with a specific focus on predictive policing. By employing feminist theoretical perspectives on governmentality, the non-neutrality of science and statistics, phenomenology and structural injustice, it problematizes Top600 and Top400 (AKA Veilig Alternatief), two algorithmic tools that Amsterdam police and institutions use for the identification and risk assessment of potential perpetrators and repeat offenders of high impact crimes. These tools are marked by a known over-representation of Dutch-Moroccan men and minors, with a lower socio-economic status, mild-to-borderline intellectual disabilities, and from segregated neighbourhoods. Through a reading, inspired by critical law studies, of the official regulations of personal data usage by Top600 and Top400, and through a review of literature concerning them, the thesis problematizes the gap between their promised and factual outcomes, questioning the fairness of their targeting. It asks what practices and structures specific to algorithmic usage and policing are involved and potentially aggravating, how the relations between the tools and human actors are articulated and what biases and distortions they entail. The thesis concludes with considerations about the legal frameworks in which they function, especially in light of the new AI Act in the EU.