By Business Insider Reporter
As artificial intelligence (AI) continues to reshape the global education landscape, East Africa faces a critical crossroads: embrace imported AI tools as-is, or redefine them to reflect local values, languages and pedagogies.
A recent thought-provoking essay by James Maisiri and Solomon Musonza – published by UNESCO – warns that while AI promises to enhance education through personalised learning and teacher support, its unchecked adoption risks a cultural and epistemic displacement in African education systems.
“Technology is never neutral. It reflects the values and ideologies of the societies that create it,” argue Maisiri and Musonza, urging African education policymakers to look beyond the hype and critically assess the cultural origins of imported AI tools.
From Nairobi to Kampala, AI is being piloted in classrooms to help improve literacy, automate marking, and bridge teacher shortages.
However, as the authors note, most tools – such as ChatGPT or Gemini – are built on datasets and design philosophies from the Global North.
This cultural origin, they caution, often excludes Indigenous knowledge systems, communal learning approaches, and African languages.
A key example cited is the One Laptop Per Child (OLPC) project in Nigeria, which undermined communal, face-to-face learning models by introducing self-directed digital learning.
Similarly, Maasai and Kipsigis communities in Kenya value interpersonal and intergenerational education, yet AI platforms promote solitary, screen-based learning.
“When students turn to AI tools, they often receive a cultural transfer of external methods while being deprived of their own,” Maisiri and Musonza write.
These concerns echo across East Africa. In Uganda, where education reforms have focused on digitisation, local teachers express concern over foreign edtech tools eroding the role of oral traditions and community engagement.
“We risk losing not just language but our worldview,” one teacher noted during a recent Makerere University seminar.
Language remains a key battleground. Many AI tools offer multilingual functionality but still frame knowledge through a Western lens.
When Google’s Gemini was asked to translate “God is good” into Gurune, a Ghanaian language, it failed.
“Languages are not just for communication. They embody knowledge systems. If these systems are not embedded in AI, they will be erased,” warn the authors.
In Tanzania, where Swahili is the dominant medium of instruction, experts argue that educational AI tools must be locally trained to reflect ecological, spiritual and cultural realities.
For instance, AI-generated agricultural lessons often ignore traditional farming wisdom passed down orally.
“Western rationality sees plants as objects; East African cultures see them as beings,” the essay notes.
Yet, hope lies in homegrown efforts. The Masakhane initiative, a pan-African volunteer project, is building AI language models using African languages and participatory research.
Its “Decolonise Science” programme translates scientific abstracts into multiple local languages, empowering grassroots ownership of knowledge.
“The presence of grassroots initiatives alongside national strategies can assist African education systems to create equitable and culturally relevant AI solutions,” write Maisiri and Musonza.
As East Africa deepens its investment in edtech, experts argue that it must prioritise building its own AI tools, informed by local values.
Governments must pair STEM education with cultural literacy and ensure national AI policies include guidelines on educational tools’ cultural appropriateness.
“We cannot just be consumers of AI. We must be co-creators,” the authors urge. With general elections approaching in Tanzania and growing regional interest in tech-driven development, the future of education in East Africa must be guided not just by innovation – but by identity.
