Campuses:

Fairness, Accountability, and Transparency: Adventures with predictive models in criminal justice

Wednesday, June 19, 2019 - 10:00am - 10:50am
Lind 305
Kristian Lum (Human Rights Data Analysis Group (HRDAG))
The related topics of fairness, accountability, and transparency in predictive modeling have seen increased attention over the last several years. Arguably, nowhere are these topics more important than for models pertaining to the criminal justice system, as these models tend to inform highly consequential decisions (e.g. who gets released before their trial, where police should target their efforts, et cetera.). In this talk, I will present several examples of how these issues have come up in my work with criminal justice data, and I will discuss what technical and non-technical solutions might exist.