Predictive Policing as 'Chain of Translation'

• In Event: STS and Computational Knowledge Production in Policing and Criminal Justice - I

Abstract

Since several years now, analytical techniques mainly consisting of data mining and algorithmic decision making, commonly referred to as ‘predictive policing’, are increasingly used by police departments throughout the world to predict possible future offences in order to prevent them. Predictive policing is a topical example of the many ways of security authorities in general, but especially of law enforcement agencies in particular, which currently strive to “defuturize” (Luhmann, 1976: 141) the future by generating future-related knowledge via sociotechnical means. Indeed, as I will show, they do not merely generate anticipative insights about the future, but they also shape it by stimulating corresponding interventions already in the present. Grounded on empirical insights derived from qualitative interviews, participant observations and document analyses, I therefore aim to reconstruct the practices of predictive policing as sociotechnical processes of constructing and shaping crime-related futures (Kaufmann, 2018: 144).
By grasping the utilization of crime prediction software as genuinely sociotechnical or hybrid contexts of interaction (e. g., Bijker, 1993; Latour, 1994), I especially want to specify the concrete distribution of agency between the human operators and the prediction system: how is predictive policing technical and how is it social? To do so, I will reconstruct predictive policing, with special recourse on the crime prediction software PRECOBS, as “chain of translation” (Latour 1999: 70), tracing the journey of the predictions from data input to the execution on the street by police officers. A journey, which is hallmarked by manifold translations, thus epistemic transformations.

Author

• Simon Egbert, Technische Universität Berlin