Breaking Language Barriers: Google’s WAXAL Project Brings African Voices to AI

For billions around the world, asking Siri for directions or dictating a text message is second nature. But for hundreds of millions of Africans, this convenience has remained frustratingly out of reach. The reason? Most of Africa’s 2,000+ languages simply don’t exist in the datasets that train artificial intelligence systems.

That’s changing with WAXAL, a groundbreaking project recently launched by Google in partnership with leading African research institutions. Taking its name from the Wolof word for “speak,” this initiative represents more than just another AI dataset—it’s a fundamental shift in how technology development happens on the African continent.

The Data Divide

The digital exclusion of African languages isn’t just an inconvenience; it’s a barrier to economic opportunity, education, and healthcare access. While voice-enabled technologies have become common in much of the world, a profound scarcity of high-quality speech data has prevented their development for most of Africa’s languages. This has left over 100 million speakers unable to access technology in their native tongues.

The problem runs deep. According to reports, fewer than 5% of Sub-Saharan African languages have the resources needed for Natural Language Processing, the technology that allows computers to understand and generate human language.

What Makes WAXAL Different

WAXAL isn’t just impressive in scale—though it certainly is, with more than 11,000 hours of voice recordings drawn from nearly two million audio files. What truly sets this project apart is its approach to ownership and partnership.

Partner institutions at Makerere University in Uganda and the University of Ghana led data collection for a combined 13 languages, while Digital Umuganda in Rwanda headed the effort for five major languages. But here’s the crucial difference: these African institutions retain full ownership of the data they collected.

This represents a significant departure from traditional models where African data has often been extracted and controlled by outside entities. As one expert noted, this rare move toward digital sovereignty means the data set is owned by African partners who worked on the project, not Google.

The Languages and the Data

The initial release covers 21 Sub-Saharan African languages, including Hausa, Yoruba, Igbo, Luganda, Swahili, Acholi, Akan, Fulani, Kikuyu, Lingala, and Shona, among others. The dataset includes 1,250 hours of transcribed, natural speech, and over 20 hours of high-quality, studio recordings designed for building high-fidelity synthetic voices.

But Google isn’t stopping there. While WAXAL provides a critical foundation for 21 Sub-Saharan languages, the project is already expanding its reach and is currently at 27 languages, with several more in the pipeline.

Real-World Impact Already Emerging

The practical applications are already materializing. At the University of Ghana, over 7,000 volunteers contributed their voices to the project, and researchers are using the data for maternal health-care research. Other applications span education, agriculture, and public services—particularly important for reaching populations with limited literacy.

Joyce Nakatumba-Nabende, Senior Lecturer at Makerere University, captured the significance: “For AI to have a real impact in Africa, it must speak our languages and understand our contexts. The WAXAL dataset gives our researchers the high-quality data they need to build speech technologies that reflect our unique communities”.

Building on a Growing Movement

WAXAL joins a broader push to localize AI for African languages. In 2025, Nigeria launched N-ATLAS, an open-source model for Yoruba, Hausa, Igbo, and Nigerian English. Private sector startups are also entering the space, with companies like South Africa’s Lelapa AI building speech recognition and translation tools.

The collaborative approach is expanding too. Google is now partnering directly with the Masakhane African Languages Hub, a grassroots initiative that ensures African languages are developed by and for Africans.

The Path Forward

While WAXAL provides essential infrastructure, datasets alone don’t guarantee outcomes. Building reliable voice systems requires sustained investment, local deployment, and commercial pathways that keep value within African countries. Questions also remain about how the data will be used by global companies in the future.

Still, the project represents a significant step forward. By making the dataset openly accessible on platforms like Hugging Face, WAXAL lowers barriers for local startups and researchers who previously lacked resources to collect data at scale. It also reduces dependence on foreign APIs that rarely support African languages effectively.

As Aisha Walcott-Bryant, Head of Google Research Africa, stated: “This dataset provides the critical foundation for students, researchers, and entrepreneurs to build technology on their own terms, in their own languages, finally reaching over 100 million people”.

For a continent where language diversity is a defining feature, WAXAL offers something that’s been missing from the AI revolution: the foundation for African voices to be heard, understood, and empowered in the digital age—on their own terms.

---

The WAXAL dataset is now publicly available on Hugging Face for researchers, developers, and entrepreneurs worldwide.