In a landmark move for digital sovereignty, Google and a consortium of leading African research institutions announced the launch of WAXAL, a large-scale, open-access speech dataset.
This initiative is designed to bridge the digital divide for over 100 million people across Sub-Saharan Africa by providing the foundational data needed to build inclusive AI technologies.
While voice-enabled AI is common globally, a lack of high-quality data has historically sidelined Africa’s 2,000+ languages. WAXAL directly addresses this scarcity by offering resources for 21 languages, including Hausa, Yoruba, Luganda, and Swahili.
Developed over three years with Google funding, the project features 1,250 hours of transcribed natural speech and over 20 hours of high-quality studio recordings for high-fidelity synthetic voices.
Aisha Walcott-Bryantt, Head of Google Research Africa, stated that the ultimate impact of WAXAL is the empowerment of people in Africa, noting that the dataset provides a critical foundation for innovators to build technology on their own terms and expressed excitement for creators to develop tools ranging from education to voice-enabled economic services.
A central pillar of WAXAL is its community-led framework, with leading institutions like Makerere University, the University of Ghana, and Digital Umuganda spearheading the data collection.
Crucially, these partner institutions retain full ownership of the data, setting a new standard for equitable AI development. Joyce Nakatumba-Nabende of Makerere University noted that for AI to have a real impact, it must speak local languages and understand unique contexts.
Also, Prof. Isaac Wiafe from the University of Ghana added that over 7,000 volunteers joined the effort to ensure their languages belong in the digital future, already sparking innovation in health, education, and agriculture.
The dataset covers a diverse linguistic range, including Swahili, Luganda, Acholi, Lingala, Kikuyu, Dholuo, Masaaba, Nyankole, Rukiga, and Soga in East and Central Africa. In West Africa, it covers Yoruba, Hausa, Igbo, Akan, Dagaare, Dagbani, Ewe, Fante, Fulani, and Ikposo, while also including Shona and Malagasy from Southern Africa.