Cultural Counterfactuals: Evaluating Cultural Biases in Large Vision-Language Models with Counterfactual Examples

2026-03-02

AI Summary

This work addresses the challenge of evaluating cultural biases—such as those related to religion, nationality, and socioeconomic status—in current large vision-language models, which is hindered by the absence of datasets annotated with cultural context and the impossibility of inferring cultural identity from visual appearance alone. To overcome these limitations, the authors introduce the first systematically constructed synthetic dataset comprising nearly 60,000 counterfactual image pairs, generated via image editing techniques that embed the same individual into diverse authentic cultural settings. This approach enables controlled manipulation of otherwise invisible cultural dimensions, thereby transcending conventional bias assessment paradigms that rely solely on visual representations. The study successfully quantifies cultural biases across multiple attributes in mainstream models and demonstrates the effectiveness and novelty of the proposed dataset for detecting culturally grounded biases.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Machine Learning: Ethics, Bias, and Fairness

Application Category

• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Large Vision-Language Models (LVLMs) have grown increasingly powerful in recent years, but can also exhibit harmful biases. Prior studies investigating such biases have primarily focused on demographic traits related to the visual characteristics of a person depicted in an image, such as their race or gender. This has left biases related to cultural differences (e.g., religion, socioeconomic status), which cannot be readily discerned from an individual's appearance alone, relatively understudied. A key challenge in measuring cultural biases is that determining which group an individual belongs to often depends upon cultural context cues in images, and datasets annotated with cultural context cues are lacking. To address this gap, we introduce Cultural Counterfactuals: a high-quality synthetic dataset containing nearly 60k counterfactual images for measuring cultural biases related to religion, nationality, and socioeconomic status. To ensure that cultural contexts are accurately depicted, we generate our dataset using an image-editing model to place people of different demographics into real cultural context images. This enables the construction of counterfactual image sets which depict the same person in multiple different contexts, allowing for precise measurement of the impact that cultural context differences have on LVLM outputs. We demonstrate the utility of Cultural Counterfactuals for quantifying cultural biases in popular LVLMs.

Problem

• cultural bias
• large vision-language models
• counterfactual examples
• cultural context
• bias evaluation

Innovation

• Cultural Counterfactuals
• Large Vision-Language Models
• cultural bias
• counterfactual image generation
• image editing