Speaker: Jennifer Doleac, Associate Professor of Economics, Director of Justice Tech Lab, Texas A&M University
Algorithmic risk assessment is widely used throughout the criminal justice system. These tools have great promise to help courts use incarceration more efficiently and fairly, but many have raised concerns about whether they will increase racial disparities in the criminal justice system. There has been much focus on the role of the machine (the algorithm creating the risk score), but little discussion of how man and machine interact – that is, how those risk scores actually affect judges’ decisions in practice. We provide empirical evidence on the real-world effects of these tools. In this talk, I’ll discuss the effects of implementing risk assessment for criminal sentencing in Virginia, on incarceration rates, recidivism, and disparities in who is incarcerated. The results highlight the important role that human actors play in determining the results of criminal justice reforms.