If MIT Can’t Keep Up With AI, What Chance Does Africa Have?

A new MIT report reveals something unsettling: even the world’s leading universities are struggling to keep up with AI. What does that mean for Africa, where hundreds of millions of young people will need to be educated in the coming years?

On 25 August, MIT published a report that caught my attention. The university’s president, Sally Kornbluth, had asked an ad hoc committee to look at how AI was actually being used across the institution: in classrooms, laboratories, dorm rooms and, importantly, in the work students were submitting.

The committee spent several months looking at what was happening. What it found was striking. AI can now produce credible solutions to almost any written assignment in MIT’s undergraduate curriculum. Essays, proofs, coding assignments, problem sets. A finished piece of student work can no longer, by itself, tell you what the student who submitted it actually understands.

Kornbluth called it a watershed moment for MIT.

What caught my attention was not simply that AI can do these things. It was what MIT discovered about its own ability to respond.

The problem was never the technology

MIT is not short of money, talent or access to technology. Its problem is that institutions move slowly.

The committee found that changing courses and curricula can take a year or more, moving through multiple committees and layers of approval. That may have worked when universities could assume that the world would remain reasonably stable between one reform cycle and the next.

AI has broken that assumption. The committee is now calling for departments to have more freedom to make AI-related changes to courses on a rolling basis.

Read that carefully. One of the institutions that helped create the AI revolution is discovering that its own institutional machinery cannot move at the speed of what it helped create.

And the report is worried about something deeper than cheating. It points to declining attendance at office hours, less participation in online discussions and reports of fewer in-person study groups. It worries about the quieter parts of education that are difficult to measure: learning how to ask a good question, receiving a mentor’s correction, struggling through a difficult problem and slowly developing judgment.

AI can give you an answer. But it cannot give you the experience of becoming the person who knows what question to ask. That distinction may become one of the most important questions in education.

What this means beyond Cambridge

I have been speaking with education leaders across Africa about what AI is doing to education, and I keep hearing some version of the same thing:

Things are moving too fast. We cannot adapt quickly enough.

They see the technology. They see students using it. They understand that teaching, assessment and the skills young people need are changing.

What they struggle with is the speed.

By the time a policy has been written, consulted on, approved and implemented, the technology has moved again.

This is not really a story about plagiarism or whether students should use AI for their assignments. It is a story about institutional time.

For generations, education has operated on the assumption that the institution could move more slowly than the world around it. A curriculum could take years to develop because the knowledge inside it would still be relevant when the students graduated.

That assumption is becoming harder to defend.

And nowhere is that more important than Africa.

Why this lands differently in Africa

MIT is trying to adapt an education system that already exists.

Africa is still expanding and building its education systems at enormous scale.

According to UNICEF, Africa’s primary school-age population is projected to grow by 33%, from 189 million in 2015 to 251 million by 2030.

UNESCO’s International Institute for Educational Planning estimates that the population aged 6 to 15 in sub-Saharan Africa will grow from just over 280 million in 2020 to approximately 450 million by 2050. To provide quality basic education for all children, the region would need around 9 million additional classrooms and 9.5 million additional teachers.

So think about the position we are in. We are going to build millions of classrooms and train millions of teachers at exactly the moment when AI is forcing the world to reconsider what a classroom and a teacher are actually for.

That is not simply a challenge of catching up. It is a design problem.

I wrote earlier in this series that many African universities are still working from an imported blueprint, a model designed for another time and another set of conditions.

But MIT’s report adds another layer to that argument. Even one of the world’s leading universities is now having to rethink how education should work in an AI era. There is no finished, working model left to import.

For decades, when we asked what a good university should look like, we could look to universities elsewhere and copy what appeared to work.

That becomes much harder when those universities are themselves asking what comes next.

We are still building

Africa is not less constrained.

We still need teachers, classrooms, connectivity, infrastructure and funding. We have no shortage of problems to solve.

But we are still building. And there is something important about building after the rules have changed.

If we are building a classroom today, why should we assume the teacher’s main job is to deliver information that a machine can increasingly provide?

If we are designing a curriculum today, why should memorising information remain the centre of how we decide whether someone has learned?

And if we know that AI will continue changing faster than our institutions can predict, why should our systems be designed around five-year assumptions about what students will need?

Perhaps the opportunity is not to build an African version of the old education system.

Perhaps it is to build something that belongs to the new one.

What should we protect?

The answer is not to reject AI. It is to become much clearer about what we do not want to lose.

Assessment should move towards judgment. If a machine can produce the answer, we need to become better at seeing how a student arrived at it, why they believe it and what they do when the answer is wrong.

Governance should be built for adaptation. We cannot design institutions that require years to respond to technologies that change in months.

And we should protect apprenticeships. Thankfully, in Africa, apprenticeship is not an imported idea. It is part of how knowledge, craft and enterprise have long been transmitted across the continent.

The mentor who watches you work. The master who corrects you. The person who lets you make a mistake and then asks you to try again. These things look inefficient beside a machine that can produce an answer in seconds. But perhaps that is the point.

Education was never only about producing answers.

It was about producing people who could judge which answers mattered.

AI makes that distinction harder to ignore.

In Closing

MIT has spent several months asking what AI means for education and has come to a conclusion that should concern every education system: the technology is moving faster than some of the institutions built to respond to it.

Africa is facing the same problem, but from a different starting point.

We have hundreds of millions of young people to educate. We need millions more classrooms and teachers. We have enormous constraints. But we are also still building.

And that means we have a choice.

We can take the old education model, add AI tools to it and call that transformation. Or we can ask the more difficult question:

If we were building education today, knowing what we now know about AI, would we build it the same way?

I suspect the answer is no.

And if that is true, then perhaps Africa’s greatest opportunity is not to catch up with the education systems of the past.

It is to help build the education system of the future.