Identifying biases in legal data: An algorithmic fairness perspective
This work addresses representation biases and sentencing disparities in legal data for stakeholders in criminal justice, but it is incremental as it applies existing fairness concepts to a specific dataset.
The study tackled the problem of identifying and measuring biases in legal case data by comparing a baseline model representing typical judge decisions with a fair judge model using fairness concepts, demonstrating quantification of biases across demographic groups in four criminal data case studies from Cook County, Illinois.
The need to address representation biases and sentencing disparities in legal case data has long been recognized. Here, we study the problem of identifying and measuring biases in large-scale legal case data from an algorithmic fairness perspective. Our approach utilizes two regression models: A baseline that represents the decisions of a "typical" judge as given by the data and a "fair" judge that applies one of three fairness concepts. Comparing the decisions of the "typical" judge and the "fair" judge allows for quantifying biases across demographic groups, as we demonstrate in four case studies on criminal data from Cook County (Illinois).