Re-contextualizing Fairness in NLP: The Case of India

By Shaily Bhatt, Sunipa Dev, Partha P. Talukdar, Shachi Dave, Vinodkumar Prabhakaran. 2022

84 citations recorded by the source corpus.

Recent research has revealed undesirable biases in NLP data and models. However, these efforts focus of social disparities in West, and are not directly portable to other geo-cultural contexts. In this paper, we focus on NLP fairness in the context of India. We start with a brief account of the prominent axes of social disparities in India. We build resources for fairness evaluation in the Indian context and use them to demonstrate prediction biases along some of the axes. We then delve deeper into social stereotypes for Region and Religion, demonstrating its prevalence in corpora and models. Finally, we outline a holistic research agenda to re-contextualize NLP fairness research for the Indian context, accounting for Indian societal context, bridging technological gaps in NLP capabilities and resources, and adapting to Indian cultural values. While we focus on India, this framework can be generalized to other geo-cultural contexts.

Subfields: Artificial Intelligence, Natural Language Processing, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

In Safety, Robustness & Interpretability

Citations in the corpus

Llama 2: Open Foundation and Fine-Tuned Chat Models (2023, background): “For Nationality, Race and Ethnicity, and Religion, we observe a Western skew (Bhatt et al., 2022).”
Scaling Instruction-Finetuned Language Models (2022, background): “Additionally, we highlight that improvements in this work related to toxic language and identity groups remain limited to a biased subset of identity terms and in English only (Smith et al., 2022; Bhatt et al., 2022; Dev et al., 2021b).”
PaLM 2 Technical Report (2023, background): “For the wide category of race, religion, ethnicity, and nationality we see a Western skew (Bhatt et al., 2022) in the representations of "americans" (42% of references).”
This Prompt is Measuring : Evaluating Bias Evaluation in Language Models (2023, background): “Future work in other subfields of NLP may, in using our taxonomy as scaffolding, be able to see reflected back the assumptions that limit the scope and the predictive power of their research, and will have a roadmap for correcting them.”
Sociotechnical Harms of Algorithmic Systems: Scoping a Taxonomy for Harm Reduction (2022, background): “There are examples of such work already (e.g., [31, 127]).”
Bias and Fairness in Large Language Models: A Survey (2023, background): “Bhatt et al. (2022) and Malik et al. (2022) provide examples of such work for Indian society.”
The Tail Wagging the Dog: Dataset Construction Biases of Social Bias Benchmarks (2022, background): “Also, the bias benchmarks themselves imbibe the notion of fairness with the Western value system (Bhatt et al., 2022), and future explorations of benchmarks should diversify culturally as well.”