This article describes an ethnographic study of the use of statistical predictions for policing (i.e., predictive policing). In this research we ‘decipher’ both the input and the output of these predictions by examining how the use of this technology is achieved in practice. This starts with the data scientist, who influences the input and output with the design of the algorithm, and ends with officers on the street, who may or may not take the predictions seriously. Our research goes against the general assumption that a predictive policing algorithm can be used as an objective and independent instrument to increase efficiency and effectiveness. Instead, we argue that it sometimes requires months of work and depends on the actions and contextual knowledge of various actors (for example, police officers, intelligence specialists, police management, municipality) who interweave their efforts and judgments with the technology.
WAARDENBURG, L., SERGEEVA, A. et HUYSMAN, M. (2020). Predictive policing ontcijferd: Een etnografie van het ‘Criminaliteits Anticipatie Systeem’ in de praktijk. Cahiers Politiestudies, (54), pp. 69–88.