Identifying biases in legal data: An algorithmic fairness perspective

Abstract

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).

Mixed-domain Language Modeling for Processing Long Legal Documents

“…Recently, there has been growing interest in utilizing Machine Learning in the legal domain (Legal Artificial Intelligence), including for judgment prediction (Chalkidis et al, 2019;Medvedeva et al, 2020), the analysis of fairness in legal proceedings (Kleinberg et al, 2020;Ciocanel et al, 2020;Avery and Cooper, 2020;Sargent and Weber, 2021), as well as legal document analysis (Zhong et al, 2020;Grover et al, 2003;Sulea et al, 2017). The development of specialized legal language models can aid in the latter.…”