The concept of “disruptive innovation,” introduced by Clayton Christensen in The Innovator’s Dilemma, explains how newcomers using cutting-edge technology can displace entrenched corporations. Disruption in higher education can be rare, due to high barriers to entry and the rooted influence of prestigious institutions. However, in the Global South, the dynamics of disruption are different, and artificial intelligence (AI) could be the key to disrupting and even leapfrogging traditional educational systems.
The Global South, which includes Africa, Latin America, Asia, and the Middle East, houses 7.2 billion people, about 85% of the world’s population. Unlike the U.S. and Europe, over 55% of the Global South’s population is under the age of 25. The increasing share of population from Africa is especially notable, as the continent is expected to contribute 62% of global population growth between now and 2050, with countries like Nigeria and Tanzania expecting 50-90% growth by 2050. This demographic surge has created an unprecedented demand for education and have a unique opportunity to introduce disruptive models, with AI at the forefront.
The potential for AI to transform education is not significant due to the expansion of internet and smartphones across the Global South. In countries like India, Brazil and Mexico, smartphone ownership in 2023 reaches over 60% of the population. In Africa, smartphone access has been growing rapidly, but with vast regional differences, creating a “digital skills gap” among populations, with the majority of the population in countries such as Egypt, Ghana, Kenya and Nigeria owning smartphones, but less than 40% in some countries and in many rural areas. Despite these gaps, widespread adoption of smartphones with fast internet in the Global South has “leapfrogged” landline infrastructure, and provides a transformative opportunity to leapfrog outdated education systems.
AI as a Disruptor and Leapfrog Agent in Brazil
In Brazil, for example, the “Teacher-Present High School with Technological Mediation” system was built with videoconferencing technology to deliver education to remote villages near the Amazon. This system has democratized access to education for over 300,000 students in underserved areas. Harvard scholar Seiji Isotani’s “AIED Unplugged” program has reached over 250,000 students using AI and smartphones. AIED unplugged provides advanced AI tools to teachers, empowering them to efficiently evaluate student writing and improve outcomes. This work is being extended to Mexico, Peru and the Philippines, where Isotani is helping to develop policies to expand access to AI in remote regions. In a recent article Isotani notes that AIED Unplugged “empowers the teacher with knowledge so they can act in class in a better way,” according to Isotani, who adds that “when we are thinking of adapting technology, it’s not about the technology, it’s about the people.”
Case Study: AI in African Entrepreneurship
A recent study conducted by Harvard Business School professor Rem Koning in Kenya highlights both the potential and limitations of AI in developing economies. The five-month experiment involved 640 entrepreneurs divided into two groups, one which received standard business guidance in printed form, while the other accessed AI-powered advice through WhatsApp. The results revealed a nuanced picture: high-performing entrepreneurs saw a 20% improvement in business outcomes, while low-performing entrepreneurs experienced a 10% decline.
This suggests that AI has the potential to widen the gap between individuals, particularly in environments where infrastructure and digital literacy are lacking. Koning observed in an interview that AI may adversely affected some business owners who “lacked judgement” to discern between options as AI will “flood you with options.” Koning also suspects that AI can cause lower performing businesses to consider the AI to be “a new friend” and rely on it for information in lieu of “talking to other people.” Koning notes that combining the use of AI with a human mentor, provides “massive improvements” by “keeping you connected to other people.”
The notion of AI producing detrimental effects has reinforced by a controlled study of students in Turkey, that showed 17% lower performance compared to a control group when an AI system is implemented without proper guidance by teachers.
Challenges of AI Deployment
Scholar Dan Björkegren notes that AI systems developed in the West often fail to meet the unique needs of poorer countries due limited datasets in the local language and gaps in infrastructure. Björkegren advocates building localized datasets, cross-border collaboration among entrepreneurs, and engaging international organizations to bridge resource gaps. Björkegren’s research on digital credit systems illustrates how users in developing nations can game AI algorithms to improve their credit scores, leading to unreliable outcomes. Studies of AI systems used in Kenya, Togo and Sierra Leone show that these issues can be overcome by adapting AI to local conditions, suggesting that AI systems need to consider the local context to be effective.
Building Local Capacity for AI in Africa
Progress is being made across many sectors in Africa. African universities and organizations are taking the lead in responsibly developing AI applications by incorporating local languages and regional datasets. African universities have produced over 1000 publications on AI, with the number of papers and books produced doubling roughly every two years. Honoris United Universities is partnering with EdTech companies to develop AI-based learning environments that reach thousands of African students. The African Union is developing its own Continental Strategy on AI that prioritizes education and preservation of African languages. The African Observatory on Responsible AI is working to “promote African voices, experiences and value systems in global debate around responsible AI,” and private foundations like the Mastercard Foundation have funded numerous EdTech initiatives to build AI education applications in Africa.
Equitable AI Requires Local Context
Governments and education leaders in the Global South must be cautious about scaling AI-driven solutions too quickly. The cautionary tale of AI’s potential negative impacts in education and business mirrors the environmental damage caused by the rapid and unchecked adoption of plastics. Cheap, efficient, and scalable solutions are not always the best, especially when they fail to consider long-term sustainability. Studies of business in the developing world show that engaging the population as co-producers and building community-based systems enable commercial and social value to be developed together, “like two blades of a pair of scissors,” according to a study by Harvard scholar V. Kasturi Rangan of emerging markets. The AI systems in the Global South must similarly be rooted in the local cultures and communities to be sure they are effective and equitable.
The Global South has a unique opportunity to use AI to leapfrog traditional educational systems and spark rapid economic growth. However, this leap must be taken with caution, ensuring that AI is used to empower communities rather than widen inequalities.