mlsanityData to 7 Oct 2026

Stable Bias: Analyzing Societal Representations in Diffusion Models

arXiv 2303.11408 ↗First posted 20 Mar 2023cs.CYcited by 225

Not on the map. Why?

Cited by 78 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

As machine learning-enabled Text-to-Image (TTI) systems are becoming increasingly prevalent and seeing growing adoption as commercial services, characterizing the social biases they exhibit is a necessary first step to lowering their risk of discriminatory outcomes. This evaluation, however, is made more difficult by the synthetic nature of these systems' outputs: common definitions of diversity are grounded in social categories of people living in the world, whereas the artificial depictions of fictive humans created by these systems have no inherent gender or ethnicity. To address this need, we propose a new method for exploring the social biases in TTI systems. Our approach relies on characterizing the variation in generated images triggered by enumerating gender and ethnicity markers in the prompts, and comparing it to the variation engendered by spanning different professions. This allows us to (1) identify specific bias trends, (2) provide targeted scores to directly compare models in terms of diversity and representation, and (3) jointly model interdependent social variables to support a multidimensional analysis. We leverage this method to analyze images generated by 3 popular TTI systems (Dall-E 2, Stable Diffusion v 1.4 and 2) and find that while all of their outputs show correlations with US labor demographics, they also consistently under-represent marginalized identities to different extents. We also release the datasets and low-code interactive bias exploration platforms developed for this work, as well as the necessary tools to similarly evaluate additional TTI systems.

Key facts. Cited by 225 papers in the map (data through 7 Oct 2026). First posted on arXiv 20 Mar 2023.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 45; 2024: 87; 2025: 52; 2026: 41.

225 papers.

Sort:

By area:

  1. 2023Unified Concept Editing in Diffusion Modelscited by 377
  2. 2023OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documentscited by 316
  3. 2023LLM-grounded Diffusion: Enhancing Prompt Understanding of Text-to-Image Diffusion Models with Large Language Modelscited by 242
  4. 2023Sociotechnical Safety Evaluation of Generative AI Systemscited by 154
  5. 2023Art and the science of generative AI: A deeper divecited by 93
  6. 2023Finetuning Text-to-Image Diffusion Models for Fairnesscited by 90
  7. 2024Red-Teaming for Generative AI: Silver Bullet or Security Theater?cited by 81
  8. 2024Large-scale Reinforcement Learning for Diffusion Modelscited by 78
  9. 2024SafetyPrompts: a Systematic Review of Open Datasets for Evaluating and Improving Large Language Model Safetycited by 75
  10. 2023The Bias Amplification Paradox in Text-to-Image Generationcited by 69
  11. 2024Survey of Bias In Text-to-Image Generation: Definition, Evaluation, and Mitigationcited by 66
  12. 2023ITI-GEN: Inclusive Text-to-Image Generationcited by 64
38 more
  1. 2024Red Teaming Visual Language Modelscited by 62
  2. 2023On the Design Fundamentals of Diffusion Models: A Surveycited by 59
  3. 2024Design Principles for Generative AI Applicationscited by 57
  4. 2024PAL: Pluralistic Alignment Framework for Learning from Heterogeneous Preferencescited by 53
  5. 2023VisoGender: A dataset for benchmarking gender bias in image-text pronoun resolutioncited by 52
  6. 2023TIBET: Identifying and Evaluating Biases in Text-to-Image Generative Modelscited by 51
  7. 2024ConceptPrune: Concept Editing in Diffusion Models via Skilled Neuron Pruningcited by 51
  8. 2024Balancing Act: Distribution-Guided Debiasing in Diffusion Modelscited by 50
  9. 2023Photoswap: Personalized Subject Swapping in Imagescited by 47
  10. 2023Analyzing Bias in Diffusion-based Face Generation Modelscited by 45
  11. 2023Inspecting the Geographical Representativeness of Images from Text-to-Image Modelscited by 44
  12. 2023The Journey, Not the Destination: How Data Guides Diffusion Modelscited by 42
  13. 2023Towards Safe Self-Distillation of Internet-Scale Text-to-Image Diffusion Modelscited by 41
  14. 2023Fast Diffusion-Based Counterfactuals for Shortcut Removal and Generationcited by 41
  15. 2024Critical windows: non-asymptotic theory for feature emergence in diffusion modelscited by 41
  16. 2023COCO-Counterfactuals: Automatically Constructed Counterfactual Examples for Image-Text Pairscited by 39
  17. 2023The Hidden Language of Diffusion Modelscited by 34
  18. 2024FairRAG: Fair Human Generation via Fair Retrieval Augmentationcited by 34
  19. 2023Feedback-guided Data Synthesis for Imbalanced Classificationcited by 33
  20. 2023'Person' == Light-skinned, Western Man, and Sexualization of Women of Color: Stereotypes in Stable Diffusioncited by 33
  21. 2024Learning Action and Reasoning-Centric Image Editing from Videos and Simulationscited by 32
  22. 2024GenderBias-VL: Benchmarking Gender Bias in Vision Language Models via Counterfactual Probingcited by 31
  23. 2024Beyond Aesthetics: Cultural Competence in Text-to-Image Modelscited by 30
  24. 2023Balancing the Picture: Debiasing Vision-Language Datasets with Synthetic Contrast Setscited by 27
  25. 2024ViSAGe: A Global-Scale Analysis of Visual Stereotypes in Text-to-Image Generationcited by 27
  26. 2024Safety Alignment for Vision Language Modelscited by 27
  27. 2024Identifying Implicit Social Biases in Vision-Language Modelscited by 27
  28. 2023Stable Diffusion Exposed: Gender Bias from Prompt to Imagecited by 25
  29. 2024From Lived Experience to Insight: Unpacking the Psychological Risks of Using AI Conversational Agentscited by 25
  30. 2024Would Deep Generative Models Amplify Bias in Future Models?cited by 24
  31. 2024Improving Geo-diversity of Generated Images with Contextualized Vendi Score Guidancecited by 24
  32. 2024Advancing Social Intelligence in AI Agents: Technical Challenges and Open Questionscited by 21
  33. 2024FAIntbench: A Holistic and Precise Benchmark for Bias Evaluation in Text-to-Image Modelscited by 21
  34. 2023FABRIC: Personalizing Diffusion Models with Iterative Feedbackcited by 19
  35. 2023Mitigating stereotypical biases in text to image generative systemscited by 19
  36. 2024SCoFT: Self-Contrastive Fine-Tuning for Equitable Image Generationcited by 19
  37. 2024Constrained Diffusion Models via Dual Trainingcited by 19
  38. 2024Scores as Actions: a framework of fine-tuning diffusion models by continuous-time reinforcement learningcited by 19

50 most cited of 225 All 225 (JSON)