Beyond Content: How Grammatical Gender Shapes Visual Representation in Text-to-Image Models

TL;DR

This paper demonstrates that grammatical gender in language significantly influences visual representations in text-to-image models, revealing biases linked to language structure across multiple languages.

Contribution

It introduces a cross-linguistic benchmark to study grammatical gender effects on image generation, highlighting language structure's role in bias beyond stereotypical content.

Findings

Grammatical gender markedly affects gender representation in generated images.

High-resource languages exhibit stronger grammatical gender effects.

Language structure influences AI bias and fairness in multilingual multimodal systems.

Abstract

Research on bias in Text-to-Image (T2I) models has primarily focused on demographic representation and stereotypical attributes, overlooking a fundamental question: how does grammatical gender influence visual representation across languages? We introduce a cross-linguistic benchmark examining words where grammatical gender contradicts stereotypical gender associations (e.g., ``une sentinelle'' - grammatically feminine in French but referring to the stereotypically masculine concept ``guard''). Our dataset spans five gendered languages (French, Spanish, German, Italian, Russian) and two gender-neutral control languages (English, Chinese), comprising 800 unique prompts that generated 28,800 images across three state-of-the-art T2I models. Our analysis reveals that grammatical gender dramatically influences image generation: masculine grammatical markers increase male representation to 73% on average (compared to 22% with gender-neutral English), while feminine grammatical markers increase female representation to 38% (compared to 28% in English). These effects vary systematically by language resource availability and model architecture, with high-resource languages showing stronger effects. Our findings establish that language structure itself, not just content, shapes AI-generated visual outputs, introducing a new dimension for understanding bias and fairness in multilingual, multimodal systems.