Artificial Intelligence is often described as the great equaliser of the digital age, yet for millions of Africans, it risks becoming just another tool of exclusion.
Despite the continent’s rich linguistic diversity, with over 2,000 estimated languages, more than a quarter of Africa’s languages are absent from AI development.
The problem lies in both investment gaps and the scarcity of usable data. Most mainstream AI tools, including the popular ChatGPT, are built on vast datasets primarily in English, European, and Chinese languages. These languages dominate the internet, providing AI systems with endless text to learn from. By contrast, many African languages remain oral traditions rather than written ones, leaving too little text for machines to train on. The result? Africa risks being left behind in the so-called next level of AI progression.
Prof. Vukosi Marivate of the University of Pretoria captures the dilemma succinctly: “We think in our own languages, dream in them and interpret the world through them. If technology doesn’t reflect that, a whole group risks being left behind.”
Recognising this gap, researchers across the continent have launched an ambitious response. The Africa Next Voices project, a collaboration between linguists and computer scientists, has created what is believed to be the largest dataset of African languages to date. In just two years, they recorded 9,000 hours of speech across Kenya, Nigeria, and South Africa. Everyday conversations—from farming instructions to health discussions and classroom interactions were captured in 18 languages including Kikuyu, Dholuo, Hausa, Yoruba, isiZulu, and Tshivenda.
Though 18 may seem small compared to Africa’s thousands of tongues, the effort is a critical first step. Supported by a $2.2 million Gates Foundation grant, the dataset is designed as open access, enabling developers to build translation tools, transcription services, and interactive AI assistants that actually “speak” African languages.
Already, start-ups are moving quickly.
South Africa’s Lelapa AI, for instance, is building indigenous-language tools for banks and telecoms companies, aiming to transform customer service and accessibility. These applications show how language-aware AI could help solve real-world problems, from financial inclusion to healthcare communication.
But experts warn that data alone will not close the gap. Without sustained investment, Africa risks falling further behind in AI innovation, with its populations depending entirely on technologies that fail to reflect their identities.
Marivate puts it bluntly: “We’re going through this AI revolution, imagining all that can be done with it. Now imagine there’s a part of the population that just doesn’t have that access because all the information is in English.”
The challenge, then, is not just technical but cultural. If AI continues to overlook African languages, it risks erasing voices, histories, and ways of knowing that have shaped the continent for centuries. But with projects like Africa Next Voices, there is growing hope that the future of AI in Africa will be one where technology doesn’t just speak, but speaks our languages too.
