mlsanityData to 7 Oct 2026

Identifying Implicit Social Biases in Vision-Language Models

arXiv 2411.00997 ↗First posted 1 Nov 2024cs.CV cs.AI cs.CLcited by 27

On the map: Multimodal LLMs / vision-language models → Social bias and fairness in VLMs

Abstract (arXiv)

Vision-language models, like CLIP (Contrastive Language Image Pretraining), are becoming increasingly popular for a wide range of multimodal retrieval tasks. However, prior work has shown that large language and deep vision models can learn historical biases contained in their training sets, leading to perpetuation of stereotypes and potential downstream harm. In this work, we conduct a systematic analysis of the social biases that are present in CLIP, with a focus on the interaction between image and text modalities. We first propose a taxonomy of social biases called So-B-IT, which contains 374 words categorized across ten types of bias. Each type can lead to societal harm if associated with a particular demographic group. Using this taxonomy, we examine images retrieved by CLIP from a facial image dataset using each word as part of a prompt. We find that CLIP frequently displays undesirable associations between harmful words and specific demographic groups, such as retrieving mostly pictures of Middle Eastern men when asked to retrieve images of a "terrorist". Finally, we conduct an analysis of the source of such biases, by showing that the same harmful stereotypes are also present in a large image-text dataset used to train CLIP models for examples of biases that we find. Our findings highlight the importance of evaluating and addressing bias in vision-language models, and suggest the need for transparency and fairness-aware curation of large pre-training datasets.

Key facts. Cited by 27 papers in the map (data through 7 Oct 2026). Placed in Social bias and fairness in VLMs, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 59 of its 107 references that are in the map. First posted on arXiv 1 Nov 2024.

Builds on

59 of 107 references.

Closest ancestors · 1

Same direction, newest first

  1. 2022A Prompt Array Keeps the Bias Away: Debiasing Vision-Language Models with Adversarial Learning22 Mar

Other references · 58

  1. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  2. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  3. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,539
  4. 2022BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generationcited by 4,310
  5. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
  6. 2021Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervisioncited by 3,848
  7. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  8. 2021CLIPScore: A Reference-free Evaluation Metric for Image Captioningcited by 2,388
  9. 2021LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairscited by 1,544
  10. 2016Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddingscited by 1,304
  11. 2020RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Modelscited by 1,235
  12. 2019Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalizationcited by 1,232
38 more
  1. 2021CLIP-Adapter: Better Vision-Language Models with Feature Adapterscited by 1,021
  2. 2021Ethical and social risks of harm from Language Modelscited by 957
  3. 2020StereoSet: Measuring stereotypical bias in pretrained language modelscited by 911
  4. 2016Semantics derived automatically from language corpora contain human-like biasescited by 894
  5. 2024DeepSeek-VL: Towards Real-World Vision-Language Understandingcited by 718
  6. 2020Contrastive Learning of Medical Visual Representations from Paired Images and Textcited by 636
  7. 2023DataComp: In search of the next generation of multimodal datasetscited by 635
  8. 2017Age Progression/Regression by Conditional Adversarial Autoencodercited by 426
  9. 2019On Measuring Social Biases in Sentence Encoderscited by 397
  10. 2020Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditingcited by 380
  11. 2019Social Bias Frames: Reasoning about Social and Power Implications of Languagecited by 377
  12. 2019Bias in Bios: A Case Study of Semantic Representation Bias in a High-Stakes Settingcited by 346
  13. 2021Multimodal datasets: misogyny, pornography, and malignant stereotypescited by 326
  14. 2021How Much Can CLIP Benefit Vision-and-Language Tasks?cited by 321
  15. 2022Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scalecited by 262
  16. 2019Evaluating Gender Bias in Machine Translationcited by 247
  17. 2021Quality at a Glance: An Audit of Web-Crawled Multilingual Datasetscited by 241
  18. 2023Stable Bias: Analyzing Societal Representations in Diffusion Modelscited by 225
  19. 2019Identifying and Reducing Gender Bias in Word-Level Language Modelscited by 214
  20. 2020Fairness in Machine Learningcited by 194
  21. 2021General Facial Representation Learning in a Visual-Linguistic Mannercited by 193
  22. 2020Detecting Emergent Intersectional Biases: Contextualized Word Embeddings Contain a Distribution of Human-like Biasescited by 167
  23. 2019Queens are Powerful too: Mitigating Gender Bias in Dialogue Generationcited by 152
  24. 2019Reducing Sentiment Bias in Language Models via Counterfactual Evaluationcited by 148
  25. 2020Towards Controllable Biases in Language Generationcited by 112
  26. 2020Image Representations Learned With Unsupervised Pre-Training Contain Human-like Biasescited by 107
  27. 2022Quality Not Quantity: On the Interaction between Dataset Design and Robustness of CLIPcited by 97
  28. 2021Are Gender-Neutral Queries Really Gender-Neutral? Mitigating Gender Bias in Image Searchcited by 90
  29. 2019FairFace: Face Attribute Dataset for Balanced Race, Gender, and Agecited by 87
  30. 2022CLIP-Event: Connecting Text and Images with Event Structurescited by 86
  31. 2022A Whac-A-Mole Dilemma: Shortcuts Come in Multiples Where Mitigating One Amplifies Otherscited by 85
  32. 2020Hurtful Words: Quantifying Biases in Clinical Contextual Word Embeddingscited by 79
  33. 2018Achieving Fairness through Adversarial Learning: an Application to Recidivism Predictioncited by 77
  34. 2019On Measuring Gender Bias in Translation of Gender-neutral Pronounscited by 59
  35. 2023Into the LAIONs Den: Investigating Hate in Multimodal Datasetscited by 54
  36. 2021Worst of Both Worlds: Biases Compound in Pre-trained Vision-and-Language Modelscited by 52
  37. 2020SensitiveLoss: Improving Accuracy and Fairness of Face Representations with Discrimination-Aware Deep Learningcited by 50
  38. 2020Mitigating Gender Bias in Captioning Systemscited by 50

50 most cited of 58 All 58 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 2; 2025: 14; 2026: 11.

27 papers.

Sort:

By area:

  1. 2024BendVLM: Test-Time Debiasing of Vision-Language Embeddingscited by 17
  2. 2025Hyperbolic Safety-Aware Vision-Language Modelscited by 17
  3. 2024Gender Bias in Text-to-Video Generation Models: A case study of Soracited by 8
  4. 2025Bias Is a Subspace, Not a Coordinate: A Geometric Rethinking of Post-hoc Debiasing in Vision-Language Modelscited by 7
  5. 2026Interpretable Debiasing of Vision-Language Models for Social Fairnesscited by 6
  6. 2025EmoNet-Face: An Expert-Annotated Benchmark for Synthetic Emotion Recognitioncited by 5
  7. 2025Vision-Language Models display a strong gender biascited by 5
  8. 2025Robust Fairness Vision-Language Learning for Medical Image Analysiscited by 4
  9. 2025Person-Centric Annotations of LAION-400M: Auditing Bias and Its Transfer to Modelscited by 4
  10. 2025TPC: Cross-Temporal Prediction Connection for Vision-Language Model Hallucination Reductioncited by 2
  11. 2025BiasConnect: Investigating Bias Interactions in Text-to-Image Modelscited by 2
  12. 2025Breaking Language Barriers or Reinforcing Bias? A Study of Gender and Racial Disparities in Multilingual Contrastive Vision Language Modelscited by 2
15 more
  1. 2025Data Matters Most: Auditing Social Bias in Contrastive Vision Language Modelscited by 1
  2. 2025Discovering Divergent Representations between Text-to-Image Modelscited by 1
  3. 2025Surgeons Are Indian Males and Speech Therapists Are White Females: Auditing Biases in Vision-Language Models for Healthcare Professionalscited by 1
  4. 2026FAIRT2V: Training-Free Debiasing for Text-to-Video Diffusion Modelscited by 1
  5. 2026SEM: Sparse Embedding Modulation for Post-Hoc Debiasing of Vision-Language Modelscited by 1
  6. 2026Bias at the End of the Scorecited by 1
  7. 2025Enhancing Performance of Explainable AI Models with Constrained Concept Refinementcited by 0
  8. 2025Exposing Hidden Biases in Text-to-Image Models via Automated Prompt Searchcited by 0
  9. 2026IndicFairFace: Balanced Indian Face Dataset for Auditing and Mitigating Geographical Bias in Vision-Language Modelscited by 0
  10. 2026GLEaN: A Text-to-image Bias Detection Approach for Public Comprehensioncited by 0
  11. 2026Neutral-Reference Prompting for Vision-Language Modelscited by 0
  12. 2026Mod-Guide: An LLM-based Content Moderation Feedback System to Address Insensitive Speech toward Indigenous Ethnic and Religious Minority Communitiescited by 0
  13. 2026Fairness-Aware Test-Time Prompt Tuningcited by 0
  14. 2026Guardrail-Agnostic Societal Bias Evaluation in Large Vision-Language Modelscited by 0
  15. 2026FairLens: Benchmarking Fairness in Vision-Language Models for High-Stakes Decision-Makingcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Uncovering Bias in Large Vision-Language Models at Scale with Counterfactuals
  2. 2024Uncovering Bias in Large Vision-Language Models with Counterfactuals
  3. 2025VisBias: Measuring Explicit and Implicit Social Biases in Vision Language Models
  4. 2023MultiModal Bias: Introducing a Framework for Stereotypical Bias Assessment beyond Gender and Race in Vision Language Models
  5. 2024Dataset Scale and Societal Consistency Mediate Facial Impression Bias in Vision-Language AI
  6. 2025Hidden Bias in the Machine: Stereotypes in Text-to-Image Models
  7. 2024A Unified Framework and Dataset for Assessing Societal Bias in Vision-Language Models
  8. 2025Bias in the Picture: Benchmarking VLMs with Social-Cue News Images and LLM-as-Judge Assessment