The Multilingual Divide and Its Impact on Global AI Safety

TL;DR

This paper examines the persistent language gap in AI safety, analyzing its causes and impacts, and offers policy recommendations to promote multilingual AI safety and reduce disparities across languages.

Contribution

It provides a comprehensive analysis of the causes and consequences of the language gap in AI safety and offers actionable policy guidance to address these challenges.

Findings

Language gap persists in AI capabilities and safety across languages.

Barriers include data scarcity and lack of transparency.

Policy actions can mitigate safety disparities.

Abstract

Despite advances in large language model capabilities in recent years, a large gap remains in their capabilities and safety performance for many languages beyond a relatively small handful of globally dominant languages. This paper provides researchers, policymakers and governance experts with an overview of key challenges to bridging the "language gap" in AI and minimizing safety risks across languages. We provide an analysis of why the language gap in AI exists and grows, and how it creates disparities in global AI safety. We identify barriers to address these challenges, and recommend how those working in policy and governance can help address safety concerns associated with the language gap by supporting multilingual dataset creation, transparency, and research.