Individual Submission Summary
Share...

Direct link:

Predictive Policing: On Algorithms and Community Policing

Fri, September 8, 9:30 to 10:45am, Palazzo Affari, Floor: second floor, Affari 2

Abstract

We present empirical data from our field research on predictive policing in Germany, focusing on an organisational unit using machine learning algorithms to analyse the connection between space and crime, trying to make these findings useful and actionable for the police units on the street. This organisational unit was founded out of a predictive policing project and meanwhile no longer only pursues the classic predictive policing approach of showing increased patrol presence in predicted risk areas in order to deter (in the short term) inclined offenders. Rather, it also pursues approaches which focus much more on long-term prevention strategies and in which the police work very closely with local actors and communities in the crime problem-solving process. This empirical observation leads us to critically question the thesis, familiar from the discussion on Abstract Police, that digital technologies in general and predictive policing in particular lead to more abstract police work. Hence, we aim to reconsider the debate on abstract policing, understood as a tendency for policing to become increasingly rigid, formalized, and distant from citizens. Partly because of the persistent ambiguity in the use of the concept of community policing, the debate in our view is vitiated by confusion between two different distinctions: the distinction anonymous/personal and the distinction acontextual/contextual (local). The notion of abstractness brings them both together, while the conditions and consequences of the two distinctions are significantly different. In the first case it is a matter of whether or nor police officers have direct contacts with citizens and know them in person, in the second case the presence or absence of a local reference that considers the specificity of each context. Thus, police activity can be impersonal and contextual, or personal and non-local, adding complexity to the notion of abstract police with regard to predictive policing practices.

Authors