# VEAT Quantifies Implicit Associations in Text-to-Video Generator Sora and Reveals Challenges in Bias Mitigation

arXiv: https://arxiv.org/abs/2601.00996 · HTML: https://mlsanity.com/paper/2601.00996/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 35 references that are in the map.
- First posted on arXiv 2 Jan 2026.

## Builds on (most cited first, 23 of 23)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Semantics derived automatically from language corpora contain human-like biases](https://mlsanity.com/paper/1608.07187/) (2016)
- [WorldSimBench: Towards Video Generation Models as World Simulators](https://mlsanity.com/paper/2410.18072/) (2024)
- [Bias in Bios: A Case Study of Semantic Representation Bias in a High-Stakes Setting](https://mlsanity.com/paper/1901.09451/) (2019)
- [Towards Understanding and Mitigating Social Biases in Language Models](https://mlsanity.com/paper/2106.13219/) (2021)
- [Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scale](https://mlsanity.com/paper/2211.03759/) (2022)
- [The Capacity for Moral Self-Correction in Large Language Models](https://mlsanity.com/paper/2302.07459/) (2023)
- [Image Representations Learned With Unsupervised Pre-Training Contain Human-like Biases](https://mlsanity.com/paper/2010.15052/) (2020)
- [Are Gender-Neutral Queries Really Gender-Neutral? Mitigating Gender Bias in Image Search](https://mlsanity.com/paper/2109.05433/) (2021)
- [OpenBias: Open-set Bias Detection in Text-to-Image Generative Models](https://mlsanity.com/paper/2404.07990/) (2024)
- [ROBBIE: Robust Bias Evaluation of Large Generative Language Models](https://mlsanity.com/paper/2311.18140/) (2023)
- [VisoGender: A dataset for benchmarking gender bias in image-text pronoun resolution](https://mlsanity.com/paper/2306.12424/) (2023)
- [Bias Against 93 Stigmatized Groups in Masked Language Models and Downstream Sentiment Classification Tasks](https://mlsanity.com/paper/2306.05550/) (2023)
- [BERTScore is Unfair: On Social Bias in Language Model-Based Metrics for Text Generation](https://mlsanity.com/paper/2210.07626/) (2022)
- [Causally Testing Gender Bias in LLMs: A Case Study on Occupational Bias](https://mlsanity.com/paper/2212.10678/) (2022)
- [Breaking Bias, Building Bridges: Evaluation and Mitigation of Social Biases in LLMs via Contact Hypothesis](https://mlsanity.com/paper/2407.02030/) (2024)
- [ValNorm Quantifies Semantics to Reveal Consistent Valence Biases Across Languages and Over Centuries](https://mlsanity.com/paper/2006.03950/) (2020)
- [VAST: The Valence-Assessing Semantics Test for Contextualizing Language Models](https://mlsanity.com/paper/2203.07504/) (2022)
- ["I don't see myself represented here at all": User Experiences of Stable Diffusion Outputs Containing Representational Harms across Gender Identities and Nationalities](https://mlsanity.com/paper/2408.01594/) (2024)
- [Social Norm Bias: Residual Harms of Fairness-Aware Algorithms](https://mlsanity.com/paper/2108.11056/) (2021)
- [Intrinsic Bias is Predicted by Pretraining Data and Correlates with Downstream Performance in Vision-Language Encoders](https://mlsanity.com/paper/2502.07957/) (2025)
- [Bias Amplification in Stable Diffusion's Representation of Stigma Through Skin Tones and Their Homogeneity](https://mlsanity.com/paper/2508.17465/) (2025)
- [Gender Bias in Nobel Prizes](https://mlsanity.com/paper/1810.07280/) (2018)

## Built on by (most cited first, 2 of 2)

- [Calibrating WEAT Against Anisotropy: ZCA Whitening as a Geometric Pre-Processing Step for Embedding Association Tests](https://mlsanity.com/paper/2608.06908/) (2026)
- [DiVid: Diagnosing Dimension-Specific Diversity Collapse in Video Generation Models](https://mlsanity.com/paper/2610.01661/) (2026)

## Abstract (arXiv)

Text-to-Video (T2V) generators such as Sora raise concerns about whether generated content reflects societal bias. We extend embedding-association tests from words and images to video by introducing the Video Embedding Association Test (VEAT) and Single-Category VEAT (SC-VEAT). We validate these methods by reproducing the direction and magnitude of associations from widely used baselines, including Implicit Association Test (IAT) scenarios and OASIS image categories. We then quantify race (African American vs. European American) and gender (women vs. men) associations with valence (pleasant vs. unpleasant) across 17 occupations and 7 awards. Sora videos associate European Americans and women more with pleasantness (both d>0.8). Effect sizes correlate with real-world demographic distributions: percent men and White in occupations (r=0.93, r=0.83) and percent male and non-Black among award recipients (r=0.88, r=0.99). Applying explicit debiasing prompts generally reduces effect-size magnitudes, but can backfire: two Black-associated occupations (janitor, postal service) become more Black-associated after debiasing. Together, these results reveal that easily accessible T2V generators can actually amplify representational harms if not rigorously evaluated and responsibly deployed.