AutoDebias: Automated Framework for Debiasing Text-to-Image Models

2025-08-01

AI Summary

Text-to-image (T2I) models frequently reproduce societal biases—such as gender and racial stereotypes—especially in cases of intersectional and latent bias; existing debiasing methods rely heavily on predefined bias categories and exhibit poor generalization. This paper introduces the first framework that automatically discovers and mitigates complex, intersecting biases without requiring prior knowledge of bias types. Our approach leverages vision-language models (VLMs) for end-to-end bias detection, integrates CLIP-guided fairness-aware fine-tuning, and employs inclusive prompt generation—all while preserving image quality. Evaluated across a benchmark encompassing 25+ bias scenarios, our method achieves 91.6% bias detection accuracy and reduces biased outputs from 90% to under 1%, with no degradation in visual fidelity. The core contribution is the first fully automated, controllable identification and mitigation of unknown, overlapping, and latent social biases in T2I generation.

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

Text-to-Image (T2I) models generate high-quality images from text prompts but often exhibit unintended social biases, such as gender or racial stereotypes, even when these attributes are not mentioned. Existing debiasing methods work well for simple or well-known cases but struggle with subtle or overlapping biases. We propose AutoDebias, a framework that automatically identifies and mitigates harmful biases in T2I models without prior knowledge of specific bias types. Specifically, AutoDebias leverages vision-language models to detect biased visual patterns and constructs fairness guides by generating inclusive alternative prompts that reflect balanced representations. These guides drive a CLIP-guided training process that promotes fairer outputs while preserving the original model's image quality and diversity. Unlike existing methods, AutoDebias effectively addresses both subtle stereotypes and multiple interacting biases. We evaluate the framework on a benchmark covering over 25 bias scenarios, including challenging cases where multiple biases occur simultaneously. AutoDebias detects harmful patterns with 91.6% accuracy and reduces biased outputs from 90% to negligible levels, while preserving the visual fidelity of the original model.

Problem

• Detects and mitigates social biases in text-to-image models automatically
• Addresses subtle and overlapping biases without prior knowledge
• Preserves image quality while reducing biased outputs significantly

Innovation

• Automatically identifies and mitigates biases
• Uses vision-language models for detection
• CLIP-guided training for fair outputs