By Oluwaseyi Akintola
In June 2024, some 800 students were enrolled in a World Bank-funded pilot after-school programme that uses generative artificial intelligence to aid learning in Benin City, the capital of southern Nigeria’s Edo state.
The students asked questions, received immediate responses, and progressed at their own speed. Within six weeks, students in the pilot programme showed academic improvements that would typically take nearly two years to achieve. These outcomes exceeded the vast majority of educational programmes evaluated in developing nations.
Yet while students in Nigeria are experimenting with AI as a learning tool, only a few schools are teaching students what they actually need to know about artificial intelligence itself. In fact, some schools are not equipped with the human and technical resources needed to handle such tasks. This gap matters more than we realise, especially in Africa and the global south, where the stakes are highest.
The irony is stark. AI is being used in some classrooms, but students don’t understand how it works, when to trust it, or how to use it critically. In Sub-Saharan Africa, where average class sizes exceed 40 students, and only 24 percent of secondary teachers are trained to use digital tools, AI could be transformative. But without proper education about AI, we risk creating a generation of passive consumers rather than informed creators.
Consider the numbers. Globally, the vast majority of students now use AI tools, yet formal guidelines for AI use exist in only a small fraction of schools. This disconnect is even more pronounced in Nigeria and elsewhere in the global south, where the curricula are yet to embrace the real essence of AI.
The situation is even more dire in rural areas where access to digital tools is grossly limited. In Nigeria, only 23% of rural communities have internet access. More than 70% of rural areas across the African continent do not have access to the internet. The digital divide that we thought could be solved with connectivity and basic computer skills now requires something deeper: AI literacy.
The consequences of this education gap are already visible. AI tools are predominantly trained on English language data and Western contexts. When Microsoft Pilot, which was used in the Benin pilot programme or similar systems, is asked basic questions, they often fail students in the global south. Ask these systems how many seasons exist, and they confidently answer “four.” In West Africa, there are two main seasons. Ask about farming practices, and you get Western industrial methods, not the contextual ecological knowledge that African communities have developed over generations.
This cultural blindness in AI systems matters because it shapes what students learn and how they see themselves. AI educational tools trained on Western curricula frequently overlook indigenous knowledge and values. For the Maasai and Kipsigis communities in Kenya, education emphasises interpersonal skills and communal learning. Yet AI tools push individualised, screen-based approaches that clash with these traditions.
The language barrier compounds the problem. Hindi, spoken by over 600 million people, is considered a “low-resource language” in AI development. So are Igbo, Quechua, and dozens of other languages spoken across the global south. Students learning in these languages receive lower-quality AI responses and have fewer resources available. This creates an unfair advantage for English speakers and widens existing inequalities.
Some African nations are pushing back with locally driven solutions. Ghana’s Ministry of Education has developed AI tools specifically designed around Ghanaian educational content and cultural values. Teachers using these homegrown applications reported dramatic improvements in preparation time and content accuracy. Rwanda has established the Rwanda Coding Academy and community-based AI clubs to ensure that even rural students gain exposure to AI principles.
But tool development is only part of the solution. What’s missing in most curricula is critical AI literacy. Students need to understand how algorithms make decisions, recognise bias in training data, and question AI-generated content. They need to know that AI systems can perpetuate stereotypes, make errors, and reflect the values of their creators.
This knowledge gap has real consequences for the future workforce. While developed nations benefit from designing and deploying AI algorithms, the global south is increasingly relegated to low-skilled data labelling and correction work within the AI value chain. Without comprehensive AI education, African students will remain on the consumption side of the AI economy rather than the creation side.
The homework divide illustrates what’s at stake. During the COVID pandemic, 10 to 20 percent of students in wealthy countries lacked home internet access, but in the global south, this gap reached as high as 90 percent in some areas. Now we face an AI literacy divide that could be even more consequential.
What should schools be teaching? First, the basics of how AI works, what it can and cannot do, and when to be sceptical of its outputs. Second, hands-on experience with AI tools, but paired with critical analysis of their limitations and biases. Third, ethical frameworks for thinking about AI’s impact on society, privacy, and fairness. Lastly, practical skills in prompting, evaluating, and using AI effectively while maintaining academic integrity.
The encouraging news is that effective AI education doesn’t require massive infrastructure investments. The Nigerian programme worked with basic tablets and internet access. Ghana leverages existing teacher training sessions held weekly in all 712 senior high schools. Rwanda uses community centres and mobile platforms.
What it does require is intentionality. Schools must move beyond treating AI as just another tech tool and recognise it as a fundamental shift in how knowledge is created and accessed. Curriculum developers must work with local experts to ensure AI education reflects local contexts and values. Teachers need training not just in using AI, but in teaching students to think critically about it.
The students in that Benin City classroom demonstrated that with proper support, AI can democratise access to quality education. But only if we also democratise understanding of what AI is, how it shapes our world, and how to use it wisely. The question is not whether AI will transform education in the global south. It already is. The question is whether our schools will prepare students to be informed participants in that transformation or passive recipients of someone else’s technology.
Oluwaseyi Akintola is a Talent Management professional with over ten years of experience in organisation development, performance management, learning, and change management. At the International Monetary Fund in Washington, D.C., she manages enterprise-wide learning initiatives, coaching, and leadership programs, driving strategic talent policies and advancing workforce development through innovation and continuous improvement.
In this article