An excerpt in Fast Company from a new book by Jon Fasman offers a concise overview of some of the issues raised by the use of algorithms for predictive policing, particularly when those algorithms are not transparent. Drawing on the use of predictive systems in New York City and Chicago, Fasman compares policing prior to the introduction of data systems to guide the distribution of effort with the more focused efforts possible when better data on the distribution of crime and criminals is available. He also explains how decisions to include or exclude data on certain crimes and criminal profiles can impact the degree of bias produced by such systems. With predictive systems likely to become more widely used over time, the benefits and the dangers are worth further examination, but such examinations are only possible if the algorithms are publicly available for review.
Be the first to comment