Epistemologies of predictive policing: Mathematical social science, social physics and machine learning

Abstract

Predictive policing has become a new panacea for crime prevention. However, we still know too little about the performance of computational methods in the context of predictive policing. The paper provides a detailed analysis of existing approaches to algorithmic crime forecasting. First, it is explained how predictive policing makes use of predictive models to generate crime forecasts. Afterwards, three epistemologies of predictive policing are distinguished: mathematical social science, social physics and machine learning. Finally, it is shown that these epistemologies have significant implications for the constitution of predictive knowledge in terms of its genesis, scope, intelligibility and accessibility. It is the different ways future crimes are rendered knowledgeable in order to act upon them that reaffirm or reconfigure the status of criminological knowledge within the criminal justice system, direct the attention of law enforcement agencies to particular types of crimes and criminals and blank out others, satisfy the claim for the meaningfulness of predictions or break with it and allow professionals to understand the algorithmic systems they shall rely on or turn them into a black box. By distinguishing epistemologies and analysing their implications, this analysis provides insight into the techno-scientific foundations of predictive policing and enables us to critically engage with the socio-technical practices of algorithmic crime forecasting.

Ambidextrous Utilisation of Artificial Intelligence in Policing: A Conceptual Framework

“…We claim that a focus on exploration inside the task-dependent environment of today might become counterproductive if it does not allow rethinking of problems in current structures. Although this claim is compatible with recent studies on AI in policing (Berk 2021;Egbert & Leese 2021;Hälterlein 2021;Leese 2021), we propose that Holmqvist's model helps develop a nuanced approach to policing that could open itself up to exploration. Next, we will elaborate on this thought by examining the opportunities and challenges related to exploration with AI inside and outside today's task-dependent environment, before proposing a framework for ambidextrous use of AI in policing in our conclusions.…”

“…Understandably, technology has been a key component in many strategic innovations and reforms of police service, but how technologies and innovations relate to policing is yet to be thoroughly scrutinised (Huotari 2021). Policing becomes embedded in socio-technical structures, as technologies and their users both consciously and unconsciously contribute to pre-determined policing goals (Hälterlein 2021;Leese 2021). Important questions lie in what sense technology pushes policing on those tracks it already runs (Berk 2021), how much it contributes to standardisation of risks which shrinks 'the space for alternative perspectives and solutions' (Olsen et al 2020, 11), and in what sense it enables the construction of new tracks for the police to better serve the very purposes of policing in contemporary societies -i.e., maintaining the societal order and, when needed, adapting to changes to fulfil this purpose (Skogan 2006).…”

“…AI could well increase the value of administration for the citizen by enabling officials to improve service provision (Mehr 2017), by helping them do something that was not possible before (Anastasopoulos & Whitford 2019), or by helping them mitigate resource scarcity by making administration tasks more efficient (Wirtz et al 2020;Mehr 2017). However, utilising AI is far from unproblematic and requires critical scrutiny (e.g., Hälterlein 2021;Leese 2021;Autioniemi 2020;Kaufmann 2019).…”

“…Rather, we wish to inspire the conceptual analysis of AI precisely in the context of policing and thus pave the way for AI-supported policing capable of solving issues that are not even known of yet. This insight will contribute to developing the rationale and epistemic background of AI in policing (Hälterlein 2021). We know AI has implications for policing and ambidexterity therein (Bland 2020;Macnish et al 2020) -yet there is still inadequate understanding about how to build new amidst the existing (Huotari 2021).…”

“…(Mehr 2017, 3. ) In this article, we consider AI as a technology that does not require ex ante theoretical models, but rather, it uses data to detect possible patterns and associations within e.g., crime-related, individual-related or societal data (Hälterlein 2021). Despite this, the algorithms that AI runs on have been written, trained, and operated by people which means that AI is not completely free from human agency either (Kaufmann 2019).…”