Authors: Nick Lally*, University of Kentucky
Topics: Social Geography, Urban Geography
Keywords: policing, digital geographies
Session Type: Paper
Start / End Time: 9:55 AM / 11:35 AM
Room: Washington 3, Marriott, Exhibition Level
Presentation File: No File Uploaded
Vendors of predictive policing software promise scientific solutions to urban social problems, leveraging imaginaries of big data to sell a vision of a future that is governable. While critics have expressed numerous concerns about algorithmically-driven policing, how systems integrate with existing infrastructures and practices remains largely unknown. In this talk, I show how uncertainties that attend the implementation of predictive policing systems cultivate doubt amongst software developers. Based on interviews with predictive policing developers in the US and UK, I show how moments of doubt haunt the implementation of software at every turn. While developers can easily point to how well a particular computational model fits a given ‘crime’ dataset, uneven implementations of predictive systems, opaque practices of policing, everyday resistances, and critical issues with data all undermine confidence in the efficacy of such systems. In contrast to the often grandiose claims made in marketing materials, doubt produced through everyday practices of software development provide starting points for asking important questions about predictive policing systems and how they become integrated with already messy, fraught, and contested infrastructures of policing.