mlsanityData to 14 Oct 2026

Going PLACES: Participatory Localized Red Teaming for Text-to-Image Safety in the Global South

arXiv 2605.19190 ↗First posted 18 May 2026cs.CY cs.AI cs.HCcited by 1

On the map: LLM safety and security → Safety alignment, refusal and guardrails

Abstract (arXiv)

Despite the global deployment of text-to-image (T2I) models, their safety frameworks are largely calibrated to a Western-centric default, creating significant vulnerabilities for the rest of the world. To embrace cultural pluralism and bring historically under-represented perspectives in T2I safety, we conduct localised community-centered red teaming studies in the Global South. Our two-fold approach prioritizes localization and participation, by focusing on secondary urban centers in these regions, and conducting community engagement and training workshops to contextualize local norms. As a result, we present PLACES, a dataset comprising over 26,000 examples of T2I model failures collected in partnership with universities in Ghana, Nigeria, and two regions of India (Karnataka and Punjab). Analysis of prompts collected reveals a wide-ranging diversity in socio-cultural and linguistic attributes, when compared to existing geography-agnostic crowdsourced red-teaming data. We observe unique adversarial patterns enabled by local cultural and linguistic nuances, and distinct clusters within region around specific themes, such as religion in India. Moreover, we uncover structural contextual gaps in existing safety frameworks by identifying novel harms showing normative dissonance (e.g., violating religious norms, ignoring local customs, and ominous symbolism). This work argues that expanding T2I safety requires moving beyond mere scale to incorporate deeply localised, participatory methodologies for data collection and contextualization. Content warning: This paper includes examples containing potentially harmful or offensive content.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security. Builds on 58 of its 101 references that are in the map. First posted on arXiv 18 May 2026.

Builds on

58 of 101 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025Toxicity Red-Teaming: Benchmarking LLM Safety in Singapore's Low-Resource Languages18 Sep
  2. 2025Attributional Safety Failures in Large Language Models under Code-Mixed Perturbations20 May
  3. 2024LlavaGuard: An Open VLM-based Framework for Safeguarding Vision Datasets and Models7 Jun

Other references · 55

  1. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  2. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,355
  3. 2018UMAP: Uniform Manifold Approximation and Projection for Dimension Reductioncited by 2,898
  4. 2022Red Teaming Language Models with Language Modelscited by 1,026
  5. 2023Muse: Text-To-Image Generation via Masked Generative Transformerscited by 648
  6. 2023Bias and Fairness in Large Language Models: A Surveycited by 639
  7. 2021Dynabench: Rethinking Benchmarking in NLPcited by 459
  8. 2020Artificial Intelligence, Values and Alignmentcited by 448
  9. 2024ShieldGemma: Generative AI Content Moderation Based on Gemmacited by 243
  10. 2023Stable Bias: Analyzing Societal Representations in Diffusion Modelscited by 226
  11. 2023Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignmentcited by 223
  12. 2024Rainbow Teaming: Open-Ended Generation of Diverse Adversarial Promptscited by 205
38 more
  1. 2021The Values Encoded in Machine Learning Researchcited by 172
  2. 2024Towards Measuring and Modeling "Culture" in LLMs: A Surveycited by 167
  3. 2024The PRISM Alignment Dataset: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language Modelscited by 144
  4. 2022Sociotechnical Harms of Algorithmic Systems: Scoping a Taxonomy for Harm Reductioncited by 138
  5. 2021Re-imagining Algorithmic Fairness in India and Beyondcited by 104
  6. 2020Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligencecited by 101
  7. 2024Benchmarking Vision Language Models for Cultural Understandingcited by 101
  8. 2024Red-Teaming for Generative AI: Silver Bullet or Security Theater?cited by 83
  9. 2024Clio: Privacy-Preserving Insights into Real-World AI Usecited by 72
  10. 2023AfriSenti: A Twitter Sentiment Analysis Benchmark for African Languagescited by 68
  11. 2023Goal-Driven Explainable Clustering via Language Descriptionscited by 58
  12. 2023AI's Regimes of Representation: A Community-centered Study of Text-to-Image Models in South Asiacited by 54
  13. 2023AART: AI-Assisted Red-Teaming with Diverse Data Generation for New LLM-powered Applicationscited by 54
  14. 2022Towards Afrocentric NLP for African Languages: Where We Are and Where We Can Gocited by 47
  15. 2023Supporting Human-AI Collaboration in Auditing LLMs with LLMscited by 46
  16. 2023Inspecting the Geographical Representativeness of Images from Text-to-Image Modelscited by 44
  17. 2022Handling and Presenting Harmful Text in NLP Researchcited by 43
  18. 2024Code-Switching Red-Teaming: LLM Evaluation for Safety and Multilingual Understandingcited by 34
  19. 2024Participation in the age of foundation modelscited by 32
  20. 2024CulturalBench: A Robust, Diverse, and Challenging Cultural Benchmark by Human-AI CulturalTeamingcited by 30
  21. 2025ShieldGemma 2: Robust and Tractable Image Content Moderationcited by 28
  22. 2024ViSAGe: A Global-Scale Analysis of Visual Stereotypes in Text-to-Image Generationcited by 27
  23. 2024Do Generative AI Models Output Harm while Representing Non-Western Cultures: Evidence from A Community-Centered Approachcited by 26
  24. 2023Prompting Large Language Models for Topic Modelingcited by 25
  25. 2023Making Harmful Behaviors Unlearnable for Large Language Modelscited by 20
  26. 2024Adversarial Nibbler: An Open Red-Teaming Method for Identifying Diverse Harms in Text-to-Image Generationcited by 20
  27. 2021Dissonance Between Human and Machine Understandingcited by 18
  28. 2024Akal Badi ya Bias: An Exploratory Study of Gender Bias in Hindi Language Technologycited by 17
  29. 2025The Case for "Thick Evaluations" of Cultural Representation in AIcited by 17
  30. 2024Towards Geographic Inclusion in the Evaluation of Text-to-Image Modelscited by 16
  31. 2025Charting the Landscape of African NLP: Mapping Progress and Shaping the Road Aheadcited by 16
  32. 2018A Fast, Compact, Accurate Model for Language Identification of Codemixed Textcited by 11
  33. 2025AfriHate: A Multilingual Collection of Hate Speech and Abusive Language Datasets for African Languagescited by 11
  34. 2025Whose View of Safety? A Deep DIVE Dataset for Pluralistic Alignment of Text-to-Image Modelscited by 11
  35. 2025The Multilingual Divide and Its Impact on Global AI Safetycited by 10
  36. 2025Culture in Action: Evaluating Text-to-Image Models through Social Activitiescited by 10
  37. 2024The World Wide recipe: A community-centred framework for fine-grained data collection and regional bias operationalisationcited by 7
  38. 2025Taxonomizing Representational Harms using Speech Act Theorycited by 7

50 most cited of 55 All 55 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Quantifying the Salience of Geo-Cultural Values for Pluralistic Safety Alignmentcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Bridging the Multilingual Safety Divide: Efficient, Culturally-Aware Alignment for Global South Languages
  2. 2026Culturally-Adapted Red-Teaming Across East and Southeast Asian Contexts: A Methodological and Comparative Analysis
  3. 2026CAGE: A Framework for Culturally Adaptive Red-Teaming Benchmark Generation
  4. 2025Multimodal Cultural Safety: Evaluation Framework and Alignment Strategies
  5. 2024DOSA: A Dataset of Social Artifacts from Different Indian Geographical Subcultures
  6. 2024Exploring the Boundaries of Content Moderation in Text-to-Image Generation
  7. 2026Red Teaming LLMs as Socio-Technical Practice: From Exploration and Data Creation to Evaluation
  8. 2026Pluralis v0.1: Towards a Multicultural, Multimodal, Multilingual Benchmark for AI Risk and Reliability