I’ve been reading an interesting journal article from a few years ago that I thought might provide some background for a piece on predictive policing. While that’s true, I think there might be a more interesting story.
This paper has provided an overview of predictive policing – what it is, how it is supposed to work, some of the analytic tools used, and three key issues that must be considered for this model of policing. First of all, we draw attention to the assumptions being made at every stage of the predictive policing cycle. We then highlight the importance of independent evaluations of this model of policing and propose evaluation designs that are not simply focused on outcomes. Finally, we discuss some of [the] problems of accountability in predictive policing, and the need for some combination of transparency and evaluation in order to explain police deployment decisions.
Lyria Bennett Moses & Janet Chan (2018) Algorithmic prediction in policing: assumptions, evaluation, and accountability, Policing and Society, 28:7, 806-822, DOI: 10.1080/10439463.2016.1253695
I think the more interesting story might be about what’s changed in the five years since this was published. Have Bennett Moses and Chan published more on this subject since then? Where are we now? What’s next?
This might be more of a reporting-on-science story.
I’m not saying that I’ll tackle it, just “putting it down on paper.”