The The Making of Crime Predictions: Sociotechnical Assemblages and the Controversies of Governing Future Crime

We are witnessing an upsurge in crime forecasting software, which supposedly draws predictive knowledge from data on past crime. Although prevention and anticipation are already embedded in the apparatuses of government, going beyond a mere abstract aspiration, the latest innovations hold out the promise of replacing police officers’ “gut feelings” and discretionary risk assessments with algorithmic-powered, quantified analyses of risk scores. While police departments and private companies praise such innovations for their cost-effective rationale, critics raise concerns regarding their potential for discriminating against poor, black, and migrant communities. In this article, I address such controversies by telling the story of the making of CrimeRadar, an app developed by a Rio de Janeiro-based think tank in partnership with private associates and local police authorities. Drawing mostly on Latour’s contributions to the emerging literature on security assemblages, I argue that we gain explanatory and critical leverage by looking into the mundane practices of making and unmaking sociotechnical arrangements. That is, I address the chain of translations through which crime data are collected, organized, and transformed into risk scores. In every step, new ways of seeing and presenting crime are produced, with a significant impact on how we experience and act upon (in)security.