Chapter 6 Learning to Discriminate

It is often thought that traditional recidivism prediction tools used in criminal sentencing, though biased in many ways, can straightforwardly avoid one particularly pernicious type of bias: direct racial discrimination. They can avoid this by excluding race from the list of variables employed t...

Descrición completa

Gardado en:
Detalles Bibliográficos
Main Authors: Davies, Benjamin, Douglas, Thomas
Formato: Online
Idioma:inglés
Publicado: Oxford University Press 2024
Subjects:
Acceso en liña:https://library.oapen.org/handle/20.500.12657/90555
Tags: Engadir etiqueta
Sen Etiquetas, Sexa o primeiro en etiquetar este rexistro!
_version_ 1869527618914615296
author Davies, Benjamin
Douglas, Thomas
author_browse Davies, Benjamin
Douglas, Thomas
author_facet Davies, Benjamin
Douglas, Thomas
author_sort Davies, Benjamin
collection Directory of Open Access Books
description It is often thought that traditional recidivism prediction tools used in criminal sentencing, though biased in many ways, can straightforwardly avoid one particularly pernicious type of bias: direct racial discrimination. They can avoid this by excluding race from the list of variables employed to predict recidivism. A similar approach could be taken to the design of newer, machine learning-based (ML) tools for predicting recidivism: information about race could be withheld from the ML tool during its training phase, ensuring that the resulting predictive model does not use race as an explicit predictor. However, if race is correlated with measured recidivism in the training data, the ML tool may ‘learn’ a perfect proxy for race. If such a proxy is found, the exclusion of race would do nothing to weaken the correlation between risk (mis)classifications and race. Is this a problem? We argue that, on some explanations of the wrongness of discrimination, it is. On these explanations, the use of an ML tool that perfectly proxies race would (likely) be more wrong than the use of a traditional tool that imperfectly proxies race. Indeed, on some views, use of a perfect proxy for race is plausibly as wrong as explicit racial profiling. We end by drawing out four implications of our arguments.
format Online
id doab-20.500.12854ir-138244
institution Directory of Open Access Books
language eng
publishDate 2024
publishDateRange 2024
publishDateSort 2024
publisher Oxford University Press
publisherStr Oxford University Press
record_format ojs
spelling doab-20.500.12854ir-1382442024-07-11T04:11:24Z Chapter 6 Learning to Discriminate Davies, Benjamin Douglas, Thomas Discrimination; Profiling; Machine Learning; Algorithmic Fairness; Racial Bias; Redundant Encoding; Criminal Recidivism; Crime Prediction; Artificial Intelligence; AI thema EDItEUR::J Society and Social Sciences::JK Social services and welfare, criminology::JKV Crime and criminology::JKVF Criminal investigation and detection thema EDItEUR::J Society and Social Sciences::JK Social services and welfare, criminology::JKV Crime and criminology It is often thought that traditional recidivism prediction tools used in criminal sentencing, though biased in many ways, can straightforwardly avoid one particularly pernicious type of bias: direct racial discrimination. They can avoid this by excluding race from the list of variables employed to predict recidivism. A similar approach could be taken to the design of newer, machine learning-based (ML) tools for predicting recidivism: information about race could be withheld from the ML tool during its training phase, ensuring that the resulting predictive model does not use race as an explicit predictor. However, if race is correlated with measured recidivism in the training data, the ML tool may ‘learn’ a perfect proxy for race. If such a proxy is found, the exclusion of race would do nothing to weaken the correlation between risk (mis)classifications and race. Is this a problem? We argue that, on some explanations of the wrongness of discrimination, it is. On these explanations, the use of an ML tool that perfectly proxies race would (likely) be more wrong than the use of a traditional tool that imperfectly proxies race. Indeed, on some views, use of a perfect proxy for race is plausibly as wrong as explicit racial profiling. We end by drawing out four implications of our arguments. 2024-05-24T04:06:51Z 2024-05-24T04:06:51Z 2024-05-23T12:05:33Z 2022 chapter https://library.oapen.org/handle/20.500.12657/90555 https://directory.doabooks.org/handle/20.500.12854/138244 eng open access image/jpeg image/jpeg Attribution 4.0 International Attribution 4.0 International https://library.oapen.org/bitstream/20.500.12657/90555/1/DAVIES%20AND%20DOUGLAS%20Learning%20to%20Discriminate%20OUP%20preproof.pdf https://library.oapen.org/bitstream/20.500.12657/90555/1/DAVIES%20AND%20DOUGLAS%20Learning%20to%20Discriminate%20OUP%20preproof.pdf Oxford University Press db4e319f-ca9f-449a-bcf2-37d7c6f885b1 Sentencing and Artificial Intelligence H2020 European Research Council European Research Council (ERC) 26 open access
spellingShingle Discrimination; Profiling; Machine Learning; Algorithmic Fairness; Racial Bias; Redundant Encoding; Criminal Recidivism; Crime Prediction; Artificial Intelligence; AI
thema EDItEUR::J Society and Social Sciences::JK Social services and welfare, criminology::JKV Crime and criminology::JKVF Criminal investigation and detection
thema EDItEUR::J Society and Social Sciences::JK Social services and welfare, criminology::JKV Crime and criminology
Davies, Benjamin
Douglas, Thomas
Chapter 6 Learning to Discriminate
title Chapter 6 Learning to Discriminate
title_full Chapter 6 Learning to Discriminate
title_fullStr Chapter 6 Learning to Discriminate
title_full_unstemmed Chapter 6 Learning to Discriminate
title_short Chapter 6 Learning to Discriminate
title_sort chapter 6 learning to discriminate
topic Discrimination; Profiling; Machine Learning; Algorithmic Fairness; Racial Bias; Redundant Encoding; Criminal Recidivism; Crime Prediction; Artificial Intelligence; AI
thema EDItEUR::J Society and Social Sciences::JK Social services and welfare, criminology::JKV Crime and criminology::JKVF Criminal investigation and detection
thema EDItEUR::J Society and Social Sciences::JK Social services and welfare, criminology::JKV Crime and criminology
topic_facet Discrimination; Profiling; Machine Learning; Algorithmic Fairness; Racial Bias; Redundant Encoding; Criminal Recidivism; Crime Prediction; Artificial Intelligence; AI
thema EDItEUR::J Society and Social Sciences::JK Social services and welfare, criminology::JKV Crime and criminology::JKVF Criminal investigation and detection
thema EDItEUR::J Society and Social Sciences::JK Social services and welfare, criminology::JKV Crime and criminology
url https://library.oapen.org/handle/20.500.12657/90555
work_keys_str_mv AT daviesbenjamin chapter6learningtodiscriminate
AT douglasthomas chapter6learningtodiscriminate