Global AI initiative targets 3.4 billion people through local languages

Sixty organisations from across the global artificial intelligence ecosystem have committed to a five-year initiative aimed at making AI services more accessible to people who speak languages currently underrepresented in AI models.

The initiative targets an estimated 3.4 billion people, who the organisations say should be able to use AI tools in their own languages and voices.

The coalition brings together technology companies, AI research organisations, government ministries and departments, philanthropies, developers and other implementing organisations.

The initiative is particularly significant for countries such as Kenya, where people communicate in English and Kiswahili alongside numerous local languages and dialects.

The organisations say only a small proportion of the world’s approximately 7,000 languages are currently well-resourced enough to support advanced AI capabilities.

Many other languages have limited representation in the datasets, tools and benchmarks used to develop and test AI systems. This can result in AI tools being less accurate or less effective for speakers of those languages.

The initiative goes beyond translating text, with participating organisations also seeking to improve AI’s ability to understand and respond through voice.

This could be particularly important in communities where text-based interfaces are less practical, while better recognition of dialects, slang, idioms and cultural context could improve the usefulness of AI systems.

For Kenya and other African countries, improved AI support for local languages could expand access to AI-powered services in sectors such as healthcare, education, agriculture, financial services and public services.

The organisations will collaborate across four key areas: building open language data infrastructure, developing benchmarks to track progress, turning language data into AI models and applications, and ensuring the technology is developed responsibly.

The coalition says privacy, consent and data sovereignty will guide the initiative as organisations work to expand access to AI.

The 60 initial signatories have also invited other organisations, governments, researchers and technology developers to join the five-year effort and contribute resources and expertise.

The initiative builds on existing efforts to improve AI language and voice capabilities and seeks to bring together work on datasets, models, benchmarks and applications under a common global goal.

The post Global AI initiative targets 3.4 billion people through local languages appeared first on KBC Digital.