A woman dressed in traditional African clothing interacts with voice AI technology, highlighting advancements in language support.
Google has teamed up with African universities and research centers to launch WAXAL, an open-source speech database built to power voice-based AI in African languages. Covering 21 languages and featuring over 11,000 hours of recordings, the dataset aims to improve tools like speech recognition, voice assistants, and text-to-speech systems across education, healthcare, and public services. The project tackles a major gap, as fewer than 5% of Africa’s 2,000-plus languages currently have enough data for effective AI development. By keeping the dataset open while letting institutions retain ownership, WAXAL is fueling homegrown innovation. Together with initiatives like Nigeria’s N-ATLAS and startups such as Lelapa AI, it marks a growing push to bring African languages fully into the digital era.
TechCabal