This conversation, sponsored by the Crimes of Prediction project at the Neubauer Collegium, brought together experts on crime and urban life using a wide variety of approaches, ranging from ethnography and survey analysis to machine learning, in order to investigate our improved capacity to predict crime, and the social, ethical and policy implications that arise in the context of enhanced prediction.

How predictable is urban crime? Are patterns from historical crime events general enough to allow to foresee novel crimes before they occur? What are the consequences of acting on this capacity? Could we unsuspectingly create a world of preventative policing characterized by Minority Report? And is it possible to evaluate and counter biases in these systems as, but not after they occur? Participants discussed results from a range of investigations into criminal prediction and remediation, with an eye to the consequences of their own findings. For example, we discussed a novel deep-learning architecture that suggests enhanced prediction is possible, but can it be used and not abused? Does crime prediction make law enforcement more efficient? Does it increase enforcement bias? The ability to predict events opens the door to a new generation of ethical questions regarding the nature of crime prediction and enforcement. Here we sought not only to understand and anticipate, but curb future crimes of prediction.

Speakers

Luís Bettencourt (University of Chicago)
Kathleen Cagney (University of Chicago)
Ishanu Chattopadhyay (University of Chicago)
John Eason (University of Wisconsin)
James Evans (University of Chicago)
Andy Papachristos (Northwestern University)
Andrea Tentner (University of Chicago)