Police Technology Experiments

To be a modern local police force means embracing new surveillance technologies that promise to amass the ever-enlarging universe of data around us and to produce actionable inferences about it. Whether described as using algorithms, artificial intelligence, or automated decision-making, all of these surveillance technologies involve some degree of computational analysis of data that creates new forms of knowledge and permits new types of policing. The usual way we discuss the use of these technologies, however, is limiting. Not only is it limiting, but it also obscures the human costs that are an inevitable consequence.

We should reconsider how we approach these new policing tools. This article makes one straightforward claim: algorithmic surveillance tools piloted by the police function as technology experiments on communities. These police technologies are experiments in the sense that they pose potential harms on those subject to the technology in the service of a theoretical but typically unproven benefit: more effective policing. With this framework, two observations follow. First, the model of technological experimentalism provides a more nuanced and comprehensive framework for understanding the use of algorithmic surveillance tools in policing. Second, experimentalism foregrounds both ethical considerations and group harms that are ill-suited to traditional legal analysis.

Speaker: Elizabeth Joh, UC Davis