The Bill & Melinda Gates Foundation has formed a new coalition with major technology companies, including Anthropic, Google, and OpenAI, along with dozens of other organizations, to make artificial intelligence (AI) more accessible in underrepresented languages. This initiative brings together 60 partners, including AI research labs, corporations, and philanthropies, all focused on expanding the number of languages supported by AI tools. The goal is to reach over 3 billion people in five years by better coordinating current efforts and ensuring more inclusive language coverage. Gates Foundation CEO Mark Suzman emphasized the need to continue this work "full speed ahead," even as some of the largest AI companies have called for a slowdown in the development of advanced models. He argued that governments should regulate AI's impact on areas like cybersecurity and children's development, while also expanding its use to benefit underserved communities. Suzman added that even if AI development were to pause, the foundation would continue working on language sets and making them compatible with existing tools. This initiative follows the Gates Foundation's recent Goalkeepers report, which pledged $1 billion for AI-focused projects aimed at improving health, education, and farming practices in developing regions. The report warned that using unrepresentative language data could lead to serious errors, such as a pregnant woman in Malawi saying her "water has broken" being mistranslated into "thrown away water" in English. This highlights the importance of accurate and culturally relevant language data in AI systems. Mozilla Data Collective CEO E.M. Lewis-Jong pointed out that many AI tools are trained on language data collected from the internet, which often lacks diversity and representation. Her organization, now part of Mozilla.org, is working on a data-sharing platform that allows communities to upload their own cultural and linguistic data, ensuring that data collection respects local needs and preferences. This approach aims to give communities more control over how their language and cultural information are used. Details about the coalition’s governance are still being finalized, according to Suzman. A secretariat will be responsible for tracking each member's commitments, and the Gates Foundation may push partners to address larger gaps in language coverage as needed. Google, one of the coalition members, has been supporting Project Vaani, an initiative to collect over 150,000 hours of audio recordings across all districts in India. This effort highlights the need to capture speech data from various dialects within a single language. Google is working with local partners to record this data in the field. Anthropic, another coalition member, has been collaborating with the Gates Foundation to accelerate vaccine development and improve its chatbot’s knowledge of local crops. Elizabeth Kelly, head of beneficial deployments at Anthropic, noted that the company's AI tools still lag behind in many African languages. She stressed the importance of ensuring accurate language support to achieve benefits like improved healthcare and education outcomes worldwide.