Measuring Social Bias in Vision-Language Models with Face-Only Counterfactuals from Real Photos

• URL: http://arxiv.org/abs/2601.06931v1
• Date: Sun, 11 Jan 2026 14:35:06 GMT
• Title: Measuring Social Bias in Vision-Language Models with Face-Only Counterfactuals from Real Photos
• Abstract summary: Real-world images entangle race and gender with correlated factors such as background and clothing, obscuring attribution.<n>We propose a textbfface-only counterfactual evaluation paradigm<n>We generate counterfactual variants by editing only facial attributes related to race and gender, keeping all other visual factors fixed.
• Score: 79.03150233804458
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Vision-Language Models (VLMs) are increasingly deployed in socially consequential settings, raising concerns about social bias driven by demographic cues. A central challenge in measuring such social bias is attribution under visual confounding: real-world images entangle race and gender with correlated factors such as background and clothing, obscuring attribution. We propose a \textbf{face-only counterfactual evaluation paradigm} that isolates demographic effects while preserving real-image realism. Starting from real photographs, we generate counterfactual variants by editing only facial attributes related to race and gender, keeping all other visual factors fixed. Based on this paradigm, we construct \textbf{FOCUS}, a dataset of 480 scene-matched counterfactual images across six occupations and ten demographic groups, and propose \textbf{REFLECT}, a benchmark comprising three decision-oriented tasks: two-alternative forced choice, multiple-choice socioeconomic inference, and numeric salary recommendation. Experiments on five state-of-the-art VLMs reveal that demographic disparities persist under strict visual control and vary substantially across task formulations. These findings underscore the necessity of controlled, counterfactual audits and highlight task design as a critical factor in evaluating social bias in multimodal models.

Related papers

• Bias in the Picture: Benchmarking VLMs with Social-Cue News Images and LLM-as-Judge Assessment [8.451522319478512]
  We introduce a news-image benchmark consisting of 1,343 image-question pairs drawn from diverse outlets.<n>We evaluate a range of state-of-the-art VLMs and employ a large language model (LLM) as judge, with human verification.<n>Our findings show that: (i) visual context systematically shifts model outputs in open-ended settings; (ii) bias prevalence varies across attributes and models, with particularly high risk for gender and occupation; and (iii) higher faithfulness does not necessarily correspond to lower bias.
  arXiv  Detail & Related papers  (2025-09-24T00:33:58Z)
• Prompting Away Stereotypes? Evaluating Bias in Text-to-Image Models for Occupations [9.58968557546246]
  We frame representational societal bias assessment as an image curation and evaluation task.<n>Using five state-of-the-art models, we compare neutral baseline prompts against fairness-aware controlled prompts.<n>Results show that prompting can substantially shift demographic representations, but with highly model-specific effects.
  arXiv  Detail & Related papers  (2025-08-31T13:46:16Z)
• Vision-Language Models under Cultural and Inclusive Considerations [53.614528867159706]
  Large vision-language models (VLMs) can assist visually impaired people by describing images from their daily lives. Current evaluation datasets may not reflect diverse cultural user backgrounds or the situational context of this use case. We create a survey to determine caption preferences and propose a culture-centric evaluation benchmark by filtering VizWiz, an existing dataset with images taken by people who are blind. We then evaluate several VLMs, investigating their reliability as visual assistants in a culturally diverse setting.
  arXiv  Detail & Related papers  (2024-07-08T17:50:00Z)
• GenderBias-\emph{VL}: Benchmarking Gender Bias in Vision Language Models via Counterfactual Probing [72.0343083866144]
  This paper introduces the GenderBias-emphVL benchmark to evaluate occupation-related gender bias in Large Vision-Language Models. Using our benchmark, we extensively evaluate 15 commonly used open-source LVLMs and state-of-the-art commercial APIs. Our findings reveal widespread gender biases in existing LVLMs.
  arXiv  Detail & Related papers  (2024-06-30T05:55:15Z)
• SocialCounterfactuals: Probing and Mitigating Intersectional Social Biases in Vision-Language Models with Counterfactual Examples [6.084482865688909]
  We employ text-to-image diffusion models to produce counterfactual examples for probing intersectional social biases at scale. Our approach utilizes Stable Diffusion with cross attention control to produce sets of counterfactual image-text pairs. We produce SocialCounterfactuals, a high-quality dataset containing 171k image-text pairs for probing intersectional biases related to gender, race, and physical characteristics.
  arXiv  Detail & Related papers  (2023-11-30T18:32:14Z)
• Probing Intersectional Biases in Vision-Language Models with Counterfactual Examples [5.870913541790421]
  We employ text-to-image diffusion models to produce counterfactual examples for probing intserctional social biases at scale. Our approach utilizes Stable Diffusion with cross attention control to produce sets of counterfactual image-text pairs. We conduct extensive experiments using our generated dataset which reveal the intersectional social biases present in state-of-the-art VLMs.
  arXiv  Detail & Related papers  (2023-10-04T17:25:10Z)
• VisoGender: A dataset for benchmarking gender bias in image-text pronoun resolution [80.57383975987676]
  VisoGender is a novel dataset for benchmarking gender bias in vision-language models. We focus on occupation-related biases within a hegemonic system of binary gender, inspired by Winograd and Winogender schemas. We benchmark several state-of-the-art vision-language models and find that they demonstrate bias in resolving binary gender in complex scenes.
  arXiv  Detail & Related papers  (2023-06-21T17:59:51Z)
• Stable Bias: Analyzing Societal Representations in Diffusion Models [72.27121528451528]
  We propose a new method for exploring the social biases in Text-to-Image (TTI) systems. Our approach relies on characterizing the variation in generated images triggered by enumerating gender and ethnicity markers in the prompts. We leverage this method to analyze images generated by 3 popular TTI systems and find that while all of their outputs show correlations with US labor demographics, they also consistently under-represent marginalized identities to different extents.
  arXiv  Detail & Related papers  (2023-03-20T19:32:49Z)
• Vision-Language Models Performing Zero-Shot Tasks Exhibit Gender-based Disparities [19.03751960721954]
  We explore the extent to which zero-shot vision-language models exhibit gender bias for different vision tasks. We evaluate different vision-language models with multiple datasets across a set of concepts.
  arXiv  Detail & Related papers  (2023-01-26T13:44:31Z)
• DALL-Eval: Probing the Reasoning Skills and Social Biases of Text-to-Image Generation Models [73.12069620086311]
  We investigate the visual reasoning capabilities and social biases of text-to-image models. First, we measure three visual reasoning skills: object recognition, object counting, and spatial relation understanding. Second, we assess the gender and skin tone biases by measuring the gender/skin tone distribution of generated images.
  arXiv  Detail & Related papers  (2022-02-08T18:36:52Z)

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.