This presentation addresses predictive policing in the U.S. through the lenses of STS, critical geography and surveillance studies. These approaches highlight the normative assumptions about criminal risk encoded in algorithms while illuminating the ways that this technology perpetuates policing inequities.
Paper long abstract:
Predictive policing is the latest effort in a long line of technological and scientific (e.g. physiognomy, phrenology, genetics) attempts to predict crime. This presentation will discuss the contemporary proliferation of predictive policing products used in the United States (U.S.), illuminating their role as part of a dual strategy for managing populations through calculation and profiting from the criminalization of everyday life. Bringing STS, critical geography, and surveillance studies into conversation, this paper shows that the historical data used in predictive algorithms is marked by a legacy of inequity in which already marginalized and criminalized populations continue to be engaged in the cycle of recriminalization. Given this, predictive policing normalizes particular geographies, populations, and bodies as always already "at risk" and "criminal." In addition to illuminating the ways predictive policing constructs normativities and shapes the boundaries of criminality, the paper also imagines ways to push back against the touted "neutrality" of these technologies. This presentation concludes with an evaluation of the ethical and legal implications of predictive policing and offers a roadmap for future empirical work.