IDRC joins global coalition to close AI's language gap

Canada's International Development Research Centre (IDRC) has joined a coalition, led by the Gates Foundation, committing to five years of work to close what organisers describe as one of artificial intelligence's most significant access gaps – the failure of most AI models to serve the majority of the world's languages.

The coalition's founding statement argues that while AI can now advise, diagnose, tutor and translate at levels unimaginable a few years ago, billions of people are not benefiting, because the information farmers, educators, health workers and students need already exists but often cannot reach them in languages their tools support.

Leading AI models currently serve only a small share of the roughly 7,000 languages spoken worldwide, leaving speakers of under-represented languages served by systems that may be less accurate or less able to understand how they communicate – a gap the coalition says risks mistranslation, exclusion and poor decision-making.

IDRC said it works with AI researchers across low- and middle-income countries as a long-standing supporter of responsible AI research for development, and framed its participation as part of that existing work.

The coalition's stated goal is for an estimated 3.4 billion people who speak currently under-represented languages to be able to use AI tools in their own language and voice within five years, through gathering language data responsibly and with consent, building open, non-proprietary datasets and benchmarks, and investing in local research ecosystems so that value flows back to the communities involved.

Organisers said no single company, country or foundation could close the gap alone, and called on AI labs, researchers, community organisations, governments and funders to contribute – whether by building open datasets, developing benchmarks, integrating data into models, funding the work or deploying applications.

Priorities for the coming year include building shared open-language data infrastructure, tracking progress against the coalition's goals, turning language data into tools usable beyond the largest AI developers, and ensuring the work protects privacy, consent and data sovereignty throughout.