Algorithmic Police Reform: 'Reading' a Police Early Intervention Algorithm

This chapter 'reads' the semiotics of contemporary police reform practice and dogma through the algorithms that the police reform industry, police leaders, and court appointed police monitors create and deploy in response to calls to reduce police violence and abuse. Early Interventions Systems (EIS) are algorithmic processes that are key to police reform practice. EIS' are used to predict employee bad behavior and flag problem officers to supervisors. In the United States, as part of the country's federal consent decree police reform regime, court-appointed consent decree monitors often require the agencies that they oversee to acquire and implement EIS. But do the algorithms that underlie this software actually reduce state violence and discourage police illegality and abuse? Using the example of an EIS in widespread use in the United States by American police agencies undergoing consent decree reforms, this chapter reads this EIS’ police violence early warning algorithm, finding that, contrary to how the police reform industry and police leaders frame EIS as a key tool of reform, EIS, as it is currently designed, implemented, and governed may actually foreclose the possibility of meaningful organizational change; normalize state violence; promote accountability only for low-level officers; inhibit public knowledge and democratic deliberation about policies and practices; reduce public transparency; and divert attention from policing policies and police and city officials’ decision making. The chapter also explores whether EIS and other accountability processes, if made and governed differently, could be democratically responsive and promote organizational transparency and change.