Not So Fair: The Impact of Presumably Fair Machine Learning Models

MacKenzie Jorgensen; Hannah Richert; Elizabeth Black; Natalia Criado; Jose Such

Not So Fair: The Impact of Presumably Fair Machine Learning Models

MacKenzie Jorgensen, Hannah Richert, Elizabeth Black, Natalia Criado, Jose Such

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

Abstract

When mitigation methods are applied to make fairer machine learning models in fairness-related classification settings, there is an assumption that the disadvantaged group should be better off than if no fairness mitigation method was applied. However, this is a potentially dangerous assumption because a ``fair'' model outcome does not automatically imply a positive impact for a disadvantaged individual---they could still be negatively impacted. Modeling and accounting for those impacts is key to ensure that mitigated models are not unintentionally harming individuals; we investigate if mitigated models can still negatively impact disadvantaged individuals and what conditions affect those impacts in a loan repayment example. Our results show that most mitigated models negatively impact disadvantaged group members in comparison to the unmitigated models. The domain-dependent impacts of model outcomes should help drive future mitigation method development.

Original language	English
Title of host publication	Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society
Place of Publication	Montreal, Canada
Publisher	ACM
Publication status	Published - Aug 2023

Keywords

Fairness
Impact
Machine Learning
Synthetic data

Cite this

@inbook{b6a0d814a3a044668b5e7603944696ae,

title = "Not So Fair: The Impact of Presumably Fair Machine Learning Models",

abstract = "When mitigation methods are applied to make fairer machine learning models in fairness-related classification settings, there is an assumption that the disadvantaged group should be better off than if no fairness mitigation method was applied. However, this is a potentially dangerous assumption because a ``fair'' model outcome does not automatically imply a positive impact for a disadvantaged individual---they could still be negatively impacted. Modeling and accounting for those impacts is key to ensure that mitigated models are not unintentionally harming individuals; we investigate if mitigated models can still negatively impact disadvantaged individuals and what conditions affect those impacts in a loan repayment example. Our results show that most mitigated models negatively impact disadvantaged group members in comparison to the unmitigated models. The domain-dependent impacts of model outcomes should help drive future mitigation method development.",

keywords = "Fairness, Impact, Machine Learning, Synthetic data",

author = "MacKenzie Jorgensen and Hannah Richert and Elizabeth Black and Natalia Criado and Jose Such",

year = "2023",

month = aug,

language = "English",

booktitle = "Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society",

publisher = "ACM",

}

TY - CHAP

T1 - Not So Fair: The Impact of Presumably Fair Machine Learning Models

AU - Jorgensen, MacKenzie

AU - Richert, Hannah

AU - Black, Elizabeth

AU - Criado, Natalia

AU - Such, Jose

PY - 2023/8

Y1 - 2023/8

N2 - When mitigation methods are applied to make fairer machine learning models in fairness-related classification settings, there is an assumption that the disadvantaged group should be better off than if no fairness mitigation method was applied. However, this is a potentially dangerous assumption because a ``fair'' model outcome does not automatically imply a positive impact for a disadvantaged individual---they could still be negatively impacted. Modeling and accounting for those impacts is key to ensure that mitigated models are not unintentionally harming individuals; we investigate if mitigated models can still negatively impact disadvantaged individuals and what conditions affect those impacts in a loan repayment example. Our results show that most mitigated models negatively impact disadvantaged group members in comparison to the unmitigated models. The domain-dependent impacts of model outcomes should help drive future mitigation method development.

AB - When mitigation methods are applied to make fairer machine learning models in fairness-related classification settings, there is an assumption that the disadvantaged group should be better off than if no fairness mitigation method was applied. However, this is a potentially dangerous assumption because a ``fair'' model outcome does not automatically imply a positive impact for a disadvantaged individual---they could still be negatively impacted. Modeling and accounting for those impacts is key to ensure that mitigated models are not unintentionally harming individuals; we investigate if mitigated models can still negatively impact disadvantaged individuals and what conditions affect those impacts in a loan repayment example. Our results show that most mitigated models negatively impact disadvantaged group members in comparison to the unmitigated models. The domain-dependent impacts of model outcomes should help drive future mitigation method development.

KW - Fairness

KW - Impact

KW - Machine Learning

KW - Synthetic data

M3 - Conference paper

BT - Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society

PB - ACM

CY - Montreal, Canada

ER -

Not So Fair: The Impact of Presumably Fair Machine Learning Models

Abstract

Keywords

Fingerprint

Cite this