Identification and analysis of bias in judicial decision-making: a systematic review

• In Event: Courts and Sentencing and Penal Decision Making 1. Judicial Bias in Sentencing

Abstract

The project "IusMachina" aims to explore the regulatory framework for the use of predictive algorithms in the judicial and penitentiary fields, as well as investigate their real impact. The project also focuses on the ethical implications of implementing these technologies, as algorithmic solutions can manage and process data with great precision, which is especially important given that human errors and biases in judicial decisions are a reality. Discrimination generated by algorithms is a known problem, since a significant portion of algorithmic biases can come from non-random patterns of reality or those generated by already biased judicial practices. Stereotypes and biases of those who design algorithms are reflected in the tools, creating enormous ethical and normative challenges that surround the implementation and design of algorithms in the field of criminal justice.

A systematic review was conducted to empirically analyze the incidence of human biases in the judicial field and how they affect legal decision-making. The review followed the criteria of the PRISMA guide for the conduct and presentation of systematic reviews and meta-analyses. A search was conducted on three databases (Proquest, Web of Science, and Science Direct) using specific keywords ("Bias OR heuristics") AND ("justice OR judge").

Once completed, the systematic review will be a useful tool for detecting the most frequent biases in the judicial field and, subsequently, analyzing how they can be transferred to the development of algorithmic tools. The ultimate goal is to contribute to the development of regulatory frameworks and ethical guidelines for the implementation and use of algorithms in the field of criminal justice, which take into account the potential biases and limitations of these technologies, while also leveraging their potential benefits.

Author

• Sandra Pérez, Universidad Miguel Hernandez - Crímina