AI Only in English? Bill Gates Launches Global AI Language Data Alliance [Global AI Briefing]
- Input
- 2026-09-22 08:16:59
- Updated
- 2026-09-22 08:16:59
The Bill & Melinda Gates Foundation, a nonprofit foundation established by Bill Gates, officially announced on the 21st (local time) that it would launch the AI Language Data Alliance, with 60 organizations from around the world taking part. Within the next five years, the alliance aims to enable approximately 3.4 billion people who speak languages that are not currently adequately supported by AI to use AI in their own languages and with their own voices.
The AI Language Data Alliance includes AI companies and research institutions such as Anthropic, Google, and the OpenAI Foundation, as well as government agencies, charitable organizations, and local community groups from various countries.

The alliance will jointly build datasets for underdeveloped and low-resource languages, create benchmarks for evaluating AI performance, and connect them to real-world AI models and services. It also presented privacy protection, user consent, and data sovereignty as shared principles.
The AI Language Data Alliance is an extension of the Gates Foundation's recently announced $1 billion AI investment plan, worth approximately KRW 1.3 trillion. On the 15th, the foundation announced that it would invest at least $1 billion in AI-related initiatives over the next two years to expand the use of AI in healthcare, education, and agriculture while building language datasets for underserved regions.
Using this investment, the AI Language Data Alliance plans to build the data infrastructure needed to make AI work for people in impoverished countries. Its strategy is to enable the use of AI in healthcare, education, and agriculture through local languages and dialects in every country around the world.
Industry experts expect the AI Language Data Alliance to mark a turning point, expanding the foundation of AI competition—which has so far centered on GPUs, data centers, and model performance—to local-language data from around the world. In other words, the race to build high-performing AI is evolving into a race to create AI that can be used by more diverse people and across a wider range of language environments.
[email protected] Lee Gu-soon Reporter