OASIS Uncovers: High-Quality T2I Models, Same Old Stereotypes

2025-01-01

AI Summary

Text-to-image (T2I) models exhibit pervasive sociologically grounded visual stereotypes—particularly toward cultural groups, occupational categories, and populations from low-internet-footprint countries—while existing fairness metrics fail to distinguish harmful bias from neutral stereotypic cognition. To address this, we propose OASIS, the first sociology-informed quantitative evaluation framework, formalizing stereotypes as measurable phenomena along two orthogonal dimensions: *distributional violation* and *spectrum variance*. We introduce StOP (Stereotype Association Localization) to identify spurious attribute correlations and SPI (Latent-Space Progression Interrogation) to track representational drift across diffusion steps. Integrating statistical distribution analysis, spectral analysis of latent representations, and concept disentanglement, our empirical study reveals persistent stereotyping in state-of-the-art models—including FLUX.1 and SDv3—with markedly amplified biases against marginalized nationalities.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Philosophy and Ethics of AI: Bias, Fairness & Equity
• 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 evaluation methodologies and metrics

Abstract

Images generated by text-to-image (T2I) models often exhibit visual biases and stereotypes of concepts such as culture and profession. Existing quantitative measures of stereotypes are based on statistical parity that does not align with the sociological definition of stereotypes and, therefore, incorrectly categorizes biases as stereotypes. Instead of oversimplifying stereotypes as biases, we propose a quantitative measure of stereotypes that aligns with its sociological definition. We then propose OASIS to measure the stereotypes in a generated dataset and understand their origins within the T2I model. OASIS includes two scores to measure stereotypes from a generated image dataset: (M1) Stereotype Score to measure the distributional violation of stereotypical attributes, and (M2) WALS to measure spectral variance in the images along a stereotypical attribute. OASIS also includes two methods to understand the origins of stereotypes in T2I models: (U1) StOP to discover attributes that the T2I model internally associates with a given concept, and (U2) SPI to quantify the emergence of stereotypical attributes in the latent space of the T2I model during image generation. Despite the considerable progress in image fidelity, using OASIS, we conclude that newer T2I models such as FLUX.1 and SDv3 contain strong stereotypical predispositions about concepts and still generate images with widespread stereotypical attributes. Additionally, the quantity of stereotypes worsens for nationalities with lower Internet footprints.

Problem

• Bias
• Stereotypes
• Text-to-Image Models

Innovation

• OASIS
• stereotype measurement
• T2I bias analysis