TY - JOUR

T1 - Bias in context

T2 - What to do when complete bias removal is not an option

AU - Holm, Sune

AU - Petersen, Eike

AU - Ganz, Melanie

AU - Feragen, Aasa

PY - 2023

Y1 - 2023

N2 - It is widely recognized that machine learning algorithms may be biased in the sense that they perform worse on some demographic groups than others. This motivates algorithmic development to remove algorithmic bias, which in turn might lead to a hope—even an expectation—that algorithmic bias can be mitigated or removed (1). In this short comment, we make three points to qualify Wang et al.’s suggestion: 1) It may not be possible for algorithms to perform equally well across groups on all measures, 2) which inequalities count as morally unacceptable bias is an ethical question, and 3) the answer to the ethical question will vary across decision contexts.

AB - It is widely recognized that machine learning algorithms may be biased in the sense that they perform worse on some demographic groups than others. This motivates algorithmic development to remove algorithmic bias, which in turn might lead to a hope—even an expectation—that algorithmic bias can be mitigated or removed (1). In this short comment, we make three points to qualify Wang et al.’s suggestion: 1) It may not be possible for algorithms to perform equally well across groups on all measures, 2) which inequalities count as morally unacceptable bias is an ethical question, and 3) the answer to the ethical question will vary across decision contexts.

KW - Bias

U2 - 10.1073/pnas.2304710120

DO - 10.1073/pnas.2304710120

M3 - Letter

C2 - 37252997

VL - 120

JO - Proceedings of the National Academy of Sciences of the United States of America

JF - Proceedings of the National Academy of Sciences of the United States of America

SN - 0027-8424

IS - 23

M1 - e2304710120

ER -