Position: Fairness Failure in Generative Models is an Evaluation Problem

Mariia Vladimirova, Jean-Yves Franceschi, Thibaut Issenhuth
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:172379-172407, 2026.

Abstract

Despite groundbreaking advancements in generative models during the last decade, concerns about their lack of fairness, reinforcing societal inequalities and harming marginalized groups, remain under-addressed and difficult to act upon. This position paper argues that fairness failures in generative models, albeit driven by multiple factors, are ultimately stemming from an evaluation problem: fairness findings are rarely comparable across papers or actionable for deployment decisions. This paper diagnoses recurring empirical and conceptual failure modes in current practice and motivates a shift from ad-hoc bias checks to standardized, generative-specific evaluation. We propose Fairness Cards as a minimal reporting artifact that makes evaluation choices explicit (prompt families, counterfactual protocols, metrics, and refusal handling) enabling reproducibility, comparability, and accountability. We conclude with additional recommendations towards a paradigm shift in evaluation standards. Our project page can be found at https://mariiavladimirova.github.io/fairness-cards.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-vladimirova26a, title = {Position: Fairness Failure in Generative Models is an Evaluation Problem}, author = {Vladimirova, Mariia and Franceschi, Jean-Yves and Issenhuth, Thibaut}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {172379--172407}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/vladimirova26a/vladimirova26a.pdf}, url = {https://proceedings.mlr.press/v306/vladimirova26a.html}, abstract = {Despite groundbreaking advancements in generative models during the last decade, concerns about their lack of fairness, reinforcing societal inequalities and harming marginalized groups, remain under-addressed and difficult to act upon. This position paper argues that fairness failures in generative models, albeit driven by multiple factors, are ultimately stemming from an evaluation problem: fairness findings are rarely comparable across papers or actionable for deployment decisions. This paper diagnoses recurring empirical and conceptual failure modes in current practice and motivates a shift from ad-hoc bias checks to standardized, generative-specific evaluation. We propose Fairness Cards as a minimal reporting artifact that makes evaluation choices explicit (prompt families, counterfactual protocols, metrics, and refusal handling) enabling reproducibility, comparability, and accountability. We conclude with additional recommendations towards a paradigm shift in evaluation standards. Our project page can be found at https://mariiavladimirova.github.io/fairness-cards.} }
Endnote
%0 Conference Paper %T Position: Fairness Failure in Generative Models is an Evaluation Problem %A Mariia Vladimirova %A Jean-Yves Franceschi %A Thibaut Issenhuth %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-vladimirova26a %I PMLR %P 172379--172407 %U https://proceedings.mlr.press/v306/vladimirova26a.html %V 306 %X Despite groundbreaking advancements in generative models during the last decade, concerns about their lack of fairness, reinforcing societal inequalities and harming marginalized groups, remain under-addressed and difficult to act upon. This position paper argues that fairness failures in generative models, albeit driven by multiple factors, are ultimately stemming from an evaluation problem: fairness findings are rarely comparable across papers or actionable for deployment decisions. This paper diagnoses recurring empirical and conceptual failure modes in current practice and motivates a shift from ad-hoc bias checks to standardized, generative-specific evaluation. We propose Fairness Cards as a minimal reporting artifact that makes evaluation choices explicit (prompt families, counterfactual protocols, metrics, and refusal handling) enabling reproducibility, comparability, and accountability. We conclude with additional recommendations towards a paradigm shift in evaluation standards. Our project page can be found at https://mariiavladimirova.github.io/fairness-cards.
APA
Vladimirova, M., Franceschi, J. & Issenhuth, T.. (2026). Position: Fairness Failure in Generative Models is an Evaluation Problem. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:172379-172407 Available from https://proceedings.mlr.press/v306/vladimirova26a.html.

Related Material