Gender Bias in Text-to-Video Generation Models: A case study of Sora

TL;DR

This paper investigates gender bias in OpenAI's Sora text-to-video model, revealing that it perpetuates societal stereotypes by associating genders with stereotypical behaviors and professions.

Contribution

It provides a detailed analysis of gender bias in a state-of-the-art text-to-video model, highlighting societal prejudices embedded in training data.

Findings

Sora shows significant gender bias in generated videos.

The model associates genders with stereotypical roles.

Bias reflects societal prejudices in training data.

Abstract

The advent of text-to-video generation models has revolutionized content creation as it produces high-quality videos from textual prompts. However, concerns regarding inherent biases in such models have prompted scrutiny, particularly regarding gender representation. Our study investigates the presence of gender bias in OpenAI's Sora, a state-of-the-art text-to-video generation model. We uncover significant evidence of bias by analyzing the generated videos from a diverse set of gender-neutral and stereotypical prompts. The results indicate that Sora disproportionately associates specific genders with stereotypical behaviors and professions, which reflects societal prejudices embedded in its training data.