Recently, a Kenyan teacher recounted at a conference I was attending that she had initially been optimistic about using an artificial intelligence (AI) powered learning application brought to her school by an international non-governmental organization. She had been told the technology would respond to the needs of every child, whatever their situation, at their own pace and in their own way. But when she asked the application to help explain a cultural concept to her students, using a local proverb about trees that her grandmother had taught her, the application had no answer.
Beyond the anecdote, this scene reflects a growing disconnect experienced by many African teachers today. Over the past decade, the continent has imported an increasing volume of AI-powered educational technologies, which are often built, shaped, and trained around Western realities, culture, and languages. The African education sector must now reflect on how these technologies could profoundly reshape and even disrupt indigenous ways of learning.
This is most apparent in the realm of language, as the majority of educational AI systems operate in English. Studies show that in Nigeria, ChatGPT is among the most prevalent AI tools used for learning. Yet the application only recognizes 20 per cent of written Hausa sentences, even though more than 80 million people in the country speak the language.
ChatGPT only recognizes 20 per cent of sentences written in Hausa, a language spoken by over 80 million people in Nigeria
This forces millions of Hausa-speaking students to forgo learning in their indigenous language and switch to English when using this tool. However, language is more than mere words; it encodes the culture and worldview of its speakers. When AI-powered technologies marginalize local languages, they weaken the cultural frameworks through which students interpret knowledge.
The data gap
Across the continent, African educators have not stood idle in the face of this cultural erasure. In countries like Ghana, developers have built GhanaGPT, a customized AI created to support localized curricula, allowing learners to receive guidance both in English and local languages such as Twi. But, despite the progress made, this technology is still prone to hallucinations, fabricated translations, and failure to adequately interpret some concepts, because indigenous languages have not been sufficiently incorporated into the training of AI systems.
This challenge is not unique to Ghana; even well-intentioned AI initiatives, such as “Abstracts into Indigenous Voices” in South Africa, struggle with limited indigenous language data. This initiative aimed to use AI to translate academic texts into local languages, but performed very poorly in translating educational material text into Zulu. This is not because Zulu is overly complicated, but because there is simply not enough text on the internet to train these systems; traditionally, much of Zulu education is carried out orally. In such cases, the very tools meant to democratize knowledge end up reinforcing foreign languages like English. This limits the students’ ability to think and express themselves fully in their own tongue.
In many parts of Africa, education is not merely a transfer of academic information; it is a profoundly relational process through which communities pass down moral values, rituals, and worldviews to learners. It has been documented that in some schools in Kenya, Namibia, and South Africa, elders and community members are co-teaching in classrooms with teachers to share their traditional insights, wisdom, proverbs, oral tradition, and stories. Education thus becomes a communal activity. But this pedagogical approach is now challenged by AI, which promotes individual screen-based learning.
Misfitting content
In Zaria, northern Nigeria, religious educators have voiced concerns that AI systems could erode cultural identities and threaten interpersonal relationships between mentors and mentees. They argue that this technology lacks kirki (moral agency in Hausa ethical thought) and therefore cannot serve as a genuine moral or cultural instructor to students. Similarly, a study of a high school in Lesotho found that teachers feared that AI would increasingly isolate them from guiding students in their spiritual growth.
Students in Africa who use AI may get content that does not suit their context and needs
Furthermore, imported AI systems in Africa tend to produce homogenized learning content that reflects Western realities instead of their own. Here’s a very simple example to illustrate this: when researchers asked ChatGPT and Google Gemini how many seasons there are, these systems replied four. In West Africa, however, there are primarily two seasons: wet and dry. This demonstrates how students in Africa who use AI may get content that does not suit their context and needs. When students internalize these decontextualized facts, it undermines the development of culturally rooted critical thinking.
African institutions are increasingly realizing that if Africa does not shape AI, then AI will shape Africa. To address this, language specialists from the “African Next Voices” project have digitized 9,000 hours of spoken languages from Kenya, South Africa, and Nigeria, creating a dataset for anyone to train their own AI models on local languages. Educational technology companies in these countries can take advantage of this data.
Similarly, Rori – an AI-powered maths tutor that operates through WhatsApp in Sierra Leone and Ghana – was designed specifically to help local students. A study involving over 1,000 students showed that after six months of using the application, their academic results tended to improve. The benefits were especially pronounced for students who had been struggling to meet grade-level expectations. Such success stories underscore the pressing need for the African education sector to move beyond imported models and develop homegrown AI systems rooted in local languages, culture, and context.
James Maisiri
Researcher and Ph.D. candidate at the University of Johannesburg (South Africa), specializing in the study of digital technologies and their impact on the African labour market.
This story was originally published by UNESCO.org