Across Africa, a growing number of initiatives are trying to address one of the continent’s most important digital gaps: the underrepresentation of African languages in artificial intelligence.
The 2026 LINGUA Africa Open Call, backed by Microsoft AI for Good Lab, the Gates Foundation, Masakhane African Languages Hub and Google.org, invited African-led projects to develop datasets, models, tools and sector applications for African languages. The selected projects span areas such as health, agriculture, justice, journalism and cultural preservation.
This is encouraging. But it also raises a more fundamental question.
What happens when a language is spoken, but not properly documented, digitised or archived?
AI systems depend on data. If African languages are poorly represented in text, audio, dictionaries, terminologies, oral histories and other digital resources, then technology will continue to understand some communities better than others. The consequences go beyond convenience. They affect access to education, public services, healthcare information, finance and participation in the digital economy. The issue is therefore not simply whether African languages appear in AI. It is whether we have invested seriously enough in the foundations that make that possible.
Africa is home to more than 2,000 living languages, yet many remain underrepresented in digital systems and training data. Recent initiatives such as GSMA’s African language AI work and the launch of CommonLingua, which supports 61 African languages, show that the technology is beginning to move.
But technology cannot archive what has never been collected.
Governments, universities, cultural institutions and communities need to treat language documentation as cultural infrastructure. That means recording oral traditions, building open corpora, developing terminology for contemporary fields, supporting translation, preserving community knowledge and ensuring that young speakers are involved in creating new digital content.
The urgency is not only about preservation. Languages must also remain useful.
They need to be present in education, media, science, public services, creative industries and technology. If young people can speak a language at home but cannot use it to search, create, learn, publish or interact with emerging technologies, its place in the future becomes increasingly fragile.
AI presents Africa with an opportunity. But it also exposes an old weakness.
Before we ask whether machines can speak our languages, we should ask whether we have done enough to document, develop and carry those languages into the future.