Position: Fairness Failure in Generative Models is an Evaluation Problem

• URL: http://arxiv.org/abs/2608.16974v1
• Date: Mon, 17 Aug 2026 14:41:28 GMT
• Title: Position: Fairness Failure in Generative Models is an Evaluation Problem
• Abstract summary: This paper argues that fairness failures in generative models are ultimately stemming from an evaluation problem.<n>We propose Fairness Cards as a minimal reporting artifact that makes evaluation choices explicit.
• Score: 5.939858158928473
• License: http://creativecommons.org/licenses/by/4.0/
• 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 .

Related papers

• Moral Safety in LLMs: Exposing Performative Compliance with Puzzled Cues [43.033440772619485]
  We show that current fairness evaluations substantially overestimate moral safety.<n>Models appear fair when demographic identity is stated as an explicit label, yet become measurably less fair when the same identity must be inferred.<n>We propose a model-agnostic robustness metric that can be added to any existing fairness benchmark to separate genuine from performative moral safety.
  arXiv  Detail & Related papers  (2026-06-30T13:25:29Z)
• To Compare, or Not to Compare: On Methodological Practices in Evaluating Social Bias [61.40435736418359]
  We introduce a unified framework that standardizes heterogeneous benchmarks to contrast isolated demographic assessments with forced-choice comparative settings.<n>We show that comparative settings act as aggressive catalysts for latent discrimination, a shift primarily driven by underspecified contexts.<n>Finally, we demonstrate that this comparative prejudice is a generalized phenomenon that scales positively with model size.
  arXiv  Detail & Related papers  (2026-06-23T13:53:50Z)
• Do Fair Models Reason Fairly? Counterfactual Explanation Consistency for Procedural Fairness in Credit Decisions [0.0]
  We show that existing outcome-fair models can still apply fundamentally different reasoning to individuals.<n>We propose Counterfactual Explanation Consistency ( CEC), a framework that detects and mitigates this bias.<n>Key contributions include a nearest-neighbor counterfactual generation method, a modified baseline for integrated gradient comparisons, and a corresponding training loss.
  arXiv  Detail & Related papers  (2026-05-12T19:54:25Z)
• Operationalizing Fairness in Text-to-Image Models: A Survey of Bias, Fairness Audits and Mitigation Strategies [5.336565378315399]
  Text-to-Image (T2I) generation models have been widely adopted across various industries, yet are criticized for frequently exhibiting societal stereotypes.<n>This paper provides a comprehensive systematic review of T2I fairness literature, organizing existing work into a taxonomy of bias types and fairness notions.<n>We propose a new framework for operationalizing fairness that moves beyond descriptive metrics towards rigorous, target-based testing.
  arXiv  Detail & Related papers  (2026-04-15T11:11:39Z)
• CDRRM: Contrast-Driven Rubric Generation for Reliable and Interpretable Reward Modeling [61.75914342638658]
  CDRRM (Contrast-Driven Reward Model) is a framework for high-quality rubric generation and guided preference judgment.<n>This work offers a scalable, interpretable, and data-efficient path for reward modeling.
  arXiv  Detail & Related papers  (2026-03-09T07:15:23Z)
• From Bias to Balance: Fairness-Aware Paper Recommendation for Equitable Peer Review [1.9029881798999504]
  We propose a framework for post-review paper selection that preserves, and in some settings can even enhance, measured scholarly quality.<n>By first analyzing the behavior of the fairness parameters under controlled conditions and then validating them on real submissions, Fair-PaperRec offers a practical, equity-focused framework for post-review paper selection that preserves, and in some settings can even enhance, measured scholarly quality.
  arXiv  Detail & Related papers  (2026-02-25T21:57:07Z)
• A Coin Flip for Safety: LLM Judges Fail to Reliably Measure Adversarial Robustness [57.510025257780306]
  We show that existing validation protocols fail to account for substantial distribution shifts inherent to red-teaming.<n>We propose ReliableBench, a benchmark of behaviors that remain more consistently judgeable, and JudgeStressTest, a dataset designed to expose judge failures.
  arXiv  Detail & Related papers  (2026-02-04T15:13:35Z)
• Where Fact Ends and Fairness Begins: Redefining AI Bias Evaluation through Cognitive Biases [77.3489598315447]
  We argue that identifying the boundary between fact and fair is essential for meaningful fairness evaluation.<n>We introduce Fact-or-Fair, a benchmark with (i) objective queries aligned with descriptive, fact-based judgments, and (ii) subjective queries aligned with normative, fairness-based judgments.
  arXiv  Detail & Related papers  (2025-02-09T10:54:11Z)
• On the Fairness, Diversity and Reliability of Text-to-Image Generative Models [68.62012304574012]
  multimodal generative models have sparked critical discussions on their reliability, fairness and potential for misuse.<n>We propose an evaluation framework to assess model reliability by analyzing responses to global and local perturbations in the embedding space.<n>Our method lays the groundwork for detecting unreliable, bias-injected models and tracing the provenance of embedded biases.
  arXiv  Detail & Related papers  (2024-11-21T09:46:55Z)
• Fair Enough: Standardizing Evaluation and Model Selection for Fairness Research in NLP [64.45845091719002]
  Modern NLP systems exhibit a range of biases, which a growing literature on model debiasing attempts to correct. This paper seeks to clarify the current situation and plot a course for meaningful progress in fair learning.
  arXiv  Detail & Related papers  (2023-02-11T14:54:00Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.