Disability Representations: Finding Biases in Automatic Image Generation

TL;DR

This paper investigates biases in AI image generation models, revealing a tendency to stereotypically depict disabled individuals as old, sad, and wheelchair-bound, emphasizing the need for more inclusive AI development.

Contribution

It provides a comprehensive analysis of representation biases in popular text-to-image models regarding people with disabilities, highlighting significant stereotypical portrayals.

Findings

Most images depict disabled individuals as old and sad

Disabled individuals are mainly shown using manual wheelchairs

Biases highlight the need for inclusive AI development

Abstract

Recent advancements in image generation technology have enabled widespread access to AI-generated imagery, prominently used in advertising, entertainment, and progressively in every form of visual content. However, these technologies often perpetuate societal biases. This study investigates the representation biases in popular image generation models towards people with disabilities (PWD). Through a comprehensive experiment involving several popular text-to-image models, we analyzed the depiction of disability. The results indicate a significant bias, with most generated images portraying disabled individuals as old, sad, and predominantly using manual wheelchairs. These findings highlight the urgent need for more inclusive AI development, ensuring diverse and accurate representation of PWD in generated images. This research underscores the importance of addressing and mitigating biases in AI models to foster equitable and realistic representations.