Algorithms Have Built Racial Bias in Legal System-Accept or Not?

DOI

10.2991/assehr.k.220105.224How to use a DOI?

Keywords

Algorithm; COMPAS; LSI-R; Racial Bias

Abstract

Algorithms have been applied in various fields and increased efficiency in most workplaces. Correctional offender Management Profiling for Alternative Sanctions (COMPAS) and Level of Service Inventory-Revised (LSI-R) are the two algorithms that are used to do risk assessment in the legal system. However, both algorithms (COMPAS and LSI-R) have built-in racial bias, and the risk scores (the result of algorithms) are affected by racial bias, which finally affects judges’ decisions because risk score is a factor that can influence defendants’ sentencing and bail. This paper describes the mechanism of COMPAS and LSI-R, analyses the reasons why algorithms have built-in racial bias: developer embed their inherent racial bias into the algorithm, and many static factors that are included into consideration in algorithms can lead to racial bias; why built-in racial bias that appears from algorithm in the legal system is not acceptable; and the four methods of eliminating the racial bias in algorithms and make sure the algorithms are impartial with high efficiency.

Copyright

© 2022 The Authors. Published by Atlantis Press SARL.

Open Access

This is an open access article under the CC BY-NC license.

Download article (PDF)

TY  - CONF
AU  - Junkai Zhang
AU  - Yuxuan Han
PY  - 2022
DA  - 2022/01/17
TI  - Algorithms Have Built Racial Bias in Legal System-Accept or Not?
BT  - Proceedings of the 2021 International Conference on Social Development and Media Communication (SDMC 2021)
PB  - Atlantis Press
SP  - 1217
EP  - 1221
SN  - 2352-5398
UR  - https://doi.org/10.2991/assehr.k.220105.224
DO  - 10.2991/assehr.k.220105.224
ID  - Zhang2022
ER  -