The paper "Algorithmic Decision Making and the Cost of Fairness" by Sam Corbett-Davies, Emma Pierson, Avi Feller, Sharad Goel, and Aziz Huq explores the tension between achieving algorithmic fairness and maximizing public safety in criminal justice decisions. The authors reformulate algorithmic fairness as a constrained optimization problem, aiming to maximize public safety while satisfying formal fairness constraints that reduce racial disparities. They find that for several definitions of fairness, the optimal algorithms require applying race-specific thresholds, which can lead to racial disparities in detention rates. In contrast, the optimal unconstrained algorithm applies a single, uniform threshold to all defendants, maximizing public safety while satisfying equality of treatment. The paper demonstrates that this trade-off is significant in practice, using data from Broward County, Florida. The authors conclude that there is an inherent tension between fairness and public safety, and policymakers must carefully consider the implications of their choices.The paper "Algorithmic Decision Making and the Cost of Fairness" by Sam Corbett-Davies, Emma Pierson, Avi Feller, Sharad Goel, and Aziz Huq explores the tension between achieving algorithmic fairness and maximizing public safety in criminal justice decisions. The authors reformulate algorithmic fairness as a constrained optimization problem, aiming to maximize public safety while satisfying formal fairness constraints that reduce racial disparities. They find that for several definitions of fairness, the optimal algorithms require applying race-specific thresholds, which can lead to racial disparities in detention rates. In contrast, the optimal unconstrained algorithm applies a single, uniform threshold to all defendants, maximizing public safety while satisfying equality of treatment. The paper demonstrates that this trade-off is significant in practice, using data from Broward County, Florida. The authors conclude that there is an inherent tension between fairness and public safety, and policymakers must carefully consider the implications of their choices.