Google has partnered with top African research institutions to unveil WAXAL, a massive open-access speech dataset aimed at giving over 100 million people across Sub-Saharan Africa a stronger voice in the AI future.
The initiative, announced Monday, targets a key digital gap by supplying high-quality data for 21 local languages, such as Hausa, Yoruba, Luganda, and Acholi. It includes 1,250 hours of transcribed natural speech and more than 20 hours of studio-grade recordings for advanced synthetic voice tech.
Voice AI has exploded globally, yet Africa’s 2,000-plus languages suffer from severe data shortages, sidelining hundreds of millions from native-tongue tech access. WAXAL, three years in the making with Google funding, changes that equation.
“This dataset lays the groundwork for students, researchers, and entrepreneurs to craft tech in their own languages, impacting over 100 million lives,” said Aisha Walcott-Bryantt, Head of Google Research Africa.
She highlighted its potential for African innovators to develop educational tools and voice services that spark economic growth continent-wide.
Community-led from the start, the project involved Makerere University (Uganda), University of Ghana, Digital Umuganda (Rwanda), and others. These partners own the data outright, pioneering fair AI collaboration models.
Covered languages: Acholi, Akan, Dagaare, Dagbani, Dholuo, Ewe, Fante, Fulani (Fula), Hausa, Igbo, Ikposo (Kposo), Kikuyu, Lingala, Luganda, Malagasy, Masaaba, Nyankole, Rukiga, Shona, Soga (Lusoga), Swahili, Yoruba.
WAXAL goes live today. Details at goo.gle/IntroducingWaxal.
