Crime prediction: a police force's friend or foe? Examining predictive analytics within police work

Data-driven decision-making forms an essential part of police work. Across the last decade, the use of predictive analytics in policing has increased substantially. So-called ‘predictive policing’ (PP) attempts to identify places or people at heightened risk of crime or criminal involvement using historical data and statistical methods. Interest in PP, its accompanying ethical, social, and technical challenges, and the association of predictive analytics with artificial intelligence, has resulted in a rich literature base. Despite the significant attention, research on PP use in practice (and related organisational implications) is limited. The works of Brayne (2017), Waardenburg (2021), and Marciniak (2022) demonstrate vast differences in the application of PP in police work; including the use of intermediaries to make PP actionable, the non-use or replacement of predictive information, and increased tension between police personnel. Research thus far has been conducted in different countries and varied policing contexts. As the use of PP is only set to continue, a holistic exploration of the use of PP in police work within England and Wales is needed. This project uses a case-study approach to investigate PP in a police force in the south of England. Employing semi-structured interviews with police personnel (including police staff, officers, and leadership) this project contributes to the topic area by examining how police use, integrate, and interact with predictive analytics. This research aims to provide practical considerations for police forces in England and Wales regarding PP decision-making processes, training deficiencies and opportunities, and management best practices