Investigating Disability Representations in Text-to-Image Models

2026-02-04

AI Summary

This study addresses the significant yet underexplored bias in the representation of people with disabilities in current text-to-image generation models. It presents the first systematic evaluation of disability representation in Stable Diffusion XL and DALL-E 3, employing a multimodal methodology that integrates structured prompt engineering, image similarity analysis, sentiment polarity detection, and human-AI hybrid assessment. The investigation comprehensively examines representational diversity, fairness, and affective framing across generated outputs. Findings reveal pervasive imbalances in how disability is depicted, with certain groups consistently underrepresented or stereotyped. Furthermore, the study demonstrates that targeted mitigation strategies can effectively improve the emotional valence and inclusivity of generated imagery. These results provide empirical grounding and actionable directions for developing more equitable and representative generative AI systems.

Technology Category

• Humans and AI: AI for Accessibility
• Computer Vision: Bias, Fairness & Privacy
• Philosophy and Ethics of AI: Bias, Fairness & Equity

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: Web query analysis, representation and understanding

Abstract

Text-to-image generative models have made remarkable progress in producing high-quality visual content from textual descriptions, yet concerns remain about how they represent social groups. While characteristics like gender and race have received increasing attention, disability representations remain underexplored. This study investigates how people with disabilities are represented in AI-generated images by analyzing outputs from Stable Diffusion XL and DALL-E 3 using a structured prompt design. We analyze disability representations by comparing image similarities between generic disability prompts and prompts referring to specific disability categories. Moreover, we evaluate how mitigation strategies influence disability portrayals, with a focus on assessing affective framing through sentiment polarity analysis, combining both automatic and human evaluation. Our findings reveal persistent representational imbalances and highlight the need for continuous evaluation and refinement of generative models to foster more diverse and inclusive portrayals of disability.

Problem

• disability representation
• text-to-image models
• generative AI
• social bias
• inclusive AI

Innovation

• disability representation
• text-to-image generation
• affective framing
• prompt design
• inclusive AI