In their new article “Uncertainty, risk and the use of algorithms in policy decisions: a case study on criminal justice in the USA”, Kathrin Hartmann and Georg Wenzelburger investigate (1) how risk assessment tools, based on machine learning (Correctional Offender Management Profiling for Alternative Sanctions, COMPAS) are implemented in daily public administration decision-making and (2) how legal officials in the local criminal justice system of Eau Claire County, Wisconsin experience the changes the use of risk assessment tools bring to their individual working routines.
Through qualitative expert interviews we conducted with legal officials in the local criminal justice system of Eau Claire County, Wisconsin as well as carefully analyzed insight documents provided by them, we could show that especially the individual notion of uncertainty in decision-making processes has changed through the introduction of COMPAS. While, before, a fundamental uncertainty concerning the outcome of proceedings seemed to be dominant, the introduction of the risk assessment tool generated a notion of statistical prediction to a situation. This, consequently, shifted the actors’ decision-making from a state of using incarceration when in doubt to a strong reliance on the risk score of COMPAS.
As a result, we see a need to carefully think about the consequences the use of algorithms might have in decision-making processes.
The article was published in Policy Sciences and can be accessed through: https://link.springer.com/article/10.1007%2Fs11077-020-09414-y