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

2026-05-18

AI Summary

This study addresses the Western-centric bias in safety mechanisms of current text-to-image models, which overlooks cultural diversity in the Global South and creates structural blind spots regarding religious taboos, local customs, and symbolic meanings. To bridge this gap, the work introduces a novel approach combining deep localization with community engagement, conducting red-teaming exercises and localized workshops in secondary cities across Ghana, Nigeria, and the Indian states of Karnataka and Punjab. Through these efforts, the project collected multilingual adversarial prompts and developed a culturally sensitive evaluation framework. The resulting PLACES dataset comprises over 26,000 failure cases, revealing distinct risk patterns rooted in regional cultural differences and identifying clusters of region-specific adversarial themes.

Abstract

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.

Problem

• text-to-image safety
• cultural pluralism
• Global South
• localized red teaming
• normative dissonance

Innovation

• localized red teaming
• text-to-image safety
• cultural pluralism
• participatory methodology
• Global South