By Jonathan Mbiriyamveka
VICTORIA FALLS — If artificial intelligence cannot understand Kiswahili, isiZulu, Shona, Yoruba, Hausa, Amharic, or Arabic, the Director General of Zimbabwe’s postal and telecommunications regulator warned on Tuesday, then Africa risks becoming invisible in the future knowledge economy.
Dr. Gift Machengete, addressing the ITU Regional Development Forum for Africa 2026 and the African Preparatory Meeting for PP-26 in Victoria Falls, issued an urgent call for African nations to ensure that AI systems are built to understand and speak the continent’s indigenous languages — or face the cultural erasure of millions of people.
“A continent whose languages and values are absent from Artificial Intelligence systems risks becoming invisible in the future knowledge economy,” Machengete said.
The POTRAZ chief laid out a clear and uncomfortable reality for delegates: today’s leading large language models and generative AI systems are overwhelmingly trained on English, Mandarin, and a handful of other global languages.
African languages, which are spoken by hundreds of millions of people across the continent, remain largely absent from the datasets that power the AI revolution.
That absence is not merely a technical inconvenience, Dr Machengete argued. It is a form of cultural exclusion with profound consequences.
“We must ensure that Artificial Intelligence systems understand African realities, African cultures, and African languages,” he said. “AI systems must speak and understand Kiswahili, isiZulu, Shona, Yoruba, Hausa, Amharic, Arabic, and the many indigenous languages that carry the identity, wisdom, and cultural heritage of our people.”
The stakes, according to the regulator, could hardly be higher. As AI becomes embedded in education, healthcare, government services, commerce, and communication, speakers of African languages who do not also speak dominant global languages risk being left behind. Children who are taught in their mother tongue may find that the AI tools they depend on cannot understand them.
Farmers seeking agricultural advice, patients describing symptoms, and citizens accessing government services may all be forced to navigate systems that do not recognize their primary language.
Beyond practical access, Dr Machengete warned of a deeper loss: the gradual disappearance of linguistic and cultural heritage.
“Language carries identity, wisdom, and cultural heritage,” he said. “If African languages are absent from AI, then African ways of knowing, African storytelling traditions, African proverbs, and African wisdom risk being erased not by conquest but by omission.”
Linguists and technologists who work on African natural language processing have long warned that without deliberate investment, the AI revolution will bypass hundreds of African languages.
A 2024 study by the African Languages Technology Initiative found that fewer than 20 of Africa’s more than 2,000 languages have any meaningful representation in major AI training datasets.
But Machengete was not merely delivering a warning. He was also issuing a challenge to the continent’s governments, regulators, universities, and technology companies.
“Africa must not arrive late to that conversation,” he said. “We must converge as a continent to ensure that our voice is heard clearly and collectively within the global digital ecosystem.”
The POTRAZ chief’s remarks positioned language inclusion as a core component of what he called digital sovereignty — the ability of African nations to determine their own technological futures rather than having futures imposed upon them.
Some progress is already underway. The African Union has endorsed the Decade of African Languages (2024–2034), and several countries, including Kenya, Nigeria, and South Africa, have launched initiatives to develop local language datasets for AI training.
But Machengete made clear that these fragmented efforts must become a coordinated continental priority.
As the African Preparatory Meeting for PP-26 continues in Victoria Falls, delegates are expected to discuss specific proposals for including language requirements in continental AI governance frameworks.
Among the ideas being floated: mandatory funding for African language dataset development, requirements that AI systems deployed on the continent support a minimum number of indigenous languages, and the creation of a pan-African AI language council.
For now, Dr Machengete left the audience with a simple but powerful image: a child in a rural village, speaking her mother tongue, asking an AI a question — and being understood.
“That is the Africa we must build,” he said. “Not an Africa that consumes. An Africa that creates. An Africa that innovates. An Africa that influences. And an Africa that leads.”
About Author
