FairI Tales: Evaluation of Fairness in Indian Contexts with a Focus on Bias and Stereotypes

2025-06-29

AI Summary

Existing fairness research predominantly operates within Western sociocultural contexts, limiting its applicability to India’s pluralistic society. To address this gap, we introduce INDIC-BIAS—the first systematic, India-specific benchmark for evaluating large language model (LLM) fairness. It covers 85 socially salient identity groups—including caste, religion, region, and tribal affiliation—and comprises 20,000 culturally grounded, real-world scenario templates. We define three evaluation tasks—plausibility judgment, value assessment, and generative output—to quantify both allocative and representational harms. Our methodology integrates domain expert input with human-validated template generation, enabling multidimensional fairness measurement. Experiments across 14 state-of-the-art LLMs reveal pervasive negative biases against marginalized communities, persisting even under explicit reasoning prompting. The INDIC-BIAS benchmark is publicly released to advance localized bias detection, auditing, and mitigation research in multilingual, multicultural settings.

Technology Category

• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Machine Learning: Ethics, Bias, and Fairness
• Philosophy and Ethics of AI: Bias, Fairness & Equity

Application Category

• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Existing studies on fairness are largely Western-focused, making them inadequate for culturally diverse countries such as India. To address this gap, we introduce INDIC-BIAS, a comprehensive India-centric benchmark designed to evaluate fairness of LLMs across 85 identity groups encompassing diverse castes, religions, regions, and tribes. We first consult domain experts to curate over 1,800 socio-cultural topics spanning behaviors and situations, where biases and stereotypes are likely to emerge. Grounded in these topics, we generate and manually validate 20,000 real-world scenario templates to probe LLMs for fairness. We structure these templates into three evaluation tasks: plausibility, judgment, and generation. Our evaluation of 14 popular LLMs on these tasks reveals strong negative biases against marginalized identities, with models frequently reinforcing common stereotypes. Additionally, we find that models struggle to mitigate bias even when explicitly asked to rationalize their decision. Our evaluation provides evidence of both allocative and representational harms that current LLMs could cause towards Indian identities, calling for a more cautious usage in practical applications. We release INDIC-BIAS as an open-source benchmark to advance research on benchmarking and mitigating biases and stereotypes in the Indian context.

Problem

• Evaluating fairness in LLMs for Indian cultural contexts
• Addressing biases and stereotypes across 85 Indian identity groups
• Assessing allocative and representational harms in LLM outputs

Innovation

• Introduces INDIC-BIAS for Indian fairness evaluation
• Generates 20,000 validated real-world scenario templates
• Structures evaluation into three distinct tasks