AI Diffusion in Low Resource Language Countries

2025-11-04

AI Summary

This study identifies linguistic resource scarcity as an independent constraint on the equitable global diffusion of artificial intelligence (AI), particularly in low-resource language countries (LRLCs), where AI adoption rates remain disproportionately low. Method: We empirically disentangle the effect of language from socioeconomic confounders—such as population size and GDP—using a weighted regression model calibrated on cross-national AI usage data. Contribution/Results: Controlling for these factors, we find that LRLCs exhibit a statistically significant ~20% lower AI user penetration rate than predicted by socioeconomic baselines, confirming linguistic accessibility as a substantive barrier to AI adoption. Furthermore, state-of-the-art large language models suffer degraded performance on low-resource languages due to insufficient training data, undermining AI utility and user uptake. This work introduces a novel mechanism—language-driven technological exclusion—to explain global AI inequality and underscores the critical importance of linguistic inclusivity for equitable global AI governance.

Technology Category

• Humans and AI: AI for Accessibility
• Natural Language Processing: (Large) Language Models
• Philosophy and Ethics of AI: Bias, Fairness & Equity

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Artificial intelligence (AI) is diffusing globally at unprecedented speed, but adoption remains uneven. Frontier Large Language Models (LLMs) are known to perform poorly on low-resource languages due to data scarcity. We hypothesize that this performance deficit reduces the utility of AI, thereby slowing adoption in Low-Resource Language Countries (LRLCs). To test this, we use a weighted regression model to isolate the language effect from socioeconomic and demographic factors, finding that LRLCs have a share of AI users that is approximately 20% lower relative to their baseline. These results indicate that linguistic accessibility is a significant, independent barrier to equitable AI diffusion.

Problem

• AI adoption lags in low-resource language countries due to linguistic barriers
• Large language models underperform in low-resource languages from data scarcity
• Language accessibility independently reduces AI user share by approximately 20%

Innovation

• Weighted regression isolates language effect
• Measures AI adoption gap in low-resource countries
• Quantifies linguistic accessibility as adoption barrier