Interactive Discovery and Exploration of Visual Bias in Generative Text-to-Image Models

2025-04-28

AI Summary

Generative text-to-image (T2I) models exhibit latent visual biases, yet their systematic detection and interpretation remain challenging. To address this, we propose ViBEx—the first interactive visual bias exploration tool enabling unified exploratory and confirmatory analysis. Methodologically, ViBEx introduces a flexible prompt-tree interface, integrates zero-shot CLIP probing, and supports three bias query types: forward, cross, and inverse. Its model-agnostic architecture enables plug-and-play compatibility across diverse T2I models. In four case studies involving AI and ethics experts, ViBEx successfully uncovered novel, previously undocumented visual biases in state-of-the-art models. The open-source implementation establishes a reproducible, scalable analytical paradigm for fairness assessment of T2I systems, facilitating rigorous, human-in-the-loop bias auditing.

Technology Category

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

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Bias in generative Text-to-Image (T2I) models is a known issue, yet systematically analyzing such models' outputs to uncover it remains challenging. We introduce the Visual Bias Explorer (ViBEx) to interactively explore the output space of T2I models to support the discovery of visual bias. ViBEx introduces a novel flexible prompting tree interface in combination with zero-shot bias probing using CLIP for quick and approximate bias exploration. It additionally supports in-depth confirmatory bias analysis through visual inspection of forward, intersectional, and inverse bias queries. ViBEx is model-agnostic and publicly available. In four case study interviews, experts in AI and ethics were able to discover visual biases that have so far not been described in literature.

Problem

• Interactive exploration of bias in text-to-image models
• Novel prompting tree for zero-shot bias discovery
• Model-agnostic tool for confirmatory visual bias analysis

Innovation

• Interactive prompting tree for bias exploration
• Zero-shot bias probing using CLIP
• Model-agnostic forward and inverse bias queries