AI Won’t Replace Thinkers — But It Will Redefine Knowledge Work

My name is Jeffrey Mdala, and I am an AI Engineer, founder of Zambian Online Education Company (ZOEC), and the builder of eskulu, an AI-powered learning platform designed around the Zambian ECZ curriculum. Over the years, I have worked across software engineering, education technology, and applied AI in Zambia, and one thing is becoming impossible to ignore: the nature of knowledge work is changing very fast.

As advanced models continue to improve, we are entering a period where a very large share of research, reporting, document analysis, and even programming can be handled by AI systems. That does not mean human beings become irrelevant. It means our role is shifting. More and more, the real job is becoming oversight, judgment, direction, and monitoring.

From where I stand as someone building AI products in Zambia, this shift is not theoretical. It is already happening.

The New Reality of AI and Knowledge Work

Recent advances in frontier models have made one thing very clear to me: a huge percentage of what we call knowledge work can now be done by machines. Tasks that once required hours of manual effort, such as reading documents, summarizing findings, comparing reports, drafting responses, organizing research, and generating code, can now be completed in minutes.

When you see how powerful the latest models have become, it becomes easier to understand why many people are starting to say that perhaps 90% of knowledge work is becoming automatable. Whether that number is exact or not, the broader point stands: AI is no longer just a tool for simple automation. It is becoming an active production system for intellectual work.

That should get the attention of every student, founder, teacher, developer, and policymaker in Africa.

Programming Is Changing Faster Than Many Expected

I started coding in Grade 12 and have been building continuously since 2016. I built Zedpastpapers, which now serves more than 200,000 users every month, and I later built eskulu during the COVID-19 period to help Zambian learners access notes, past papers, marking schemes, quizzes, and AI support. So when I say programming is changing, I say it as someone who genuinely loves building software.

For a long time, programming was seen as the act of manually writing every line of logic yourself. That picture is becoming outdated. Today, AI can generate boilerplate, explain bugs, suggest architecture, write functions, refactor code, and even help design systems. In many cases, the bottleneck is no longer typing code. The bottleneck is knowing what should be built, how it should behave, and how to verify that it is correct.

That is why I increasingly see the programmer’s role moving from pure implementation to supervision. The work is becoming less about writing every line from scratch and more about:

• Defining the problem clearly
• Giving the AI the right instructions
• Testing outputs carefully
• Checking for logic errors and hallucinations
• Ensuring security, ethics, and reliability
• Making final product decisions

In other words, the job is not disappearing. But the job description is being rewritten.

Why Human Oversight Still Matters

As capable as these systems are, they do not remove the need for human judgment. In fact, stronger models make human judgment even more important. An AI system can produce something that looks polished, confident, and complete while still being wrong in subtle but dangerous ways.

That is especially important in African contexts, where local realities are often underrepresented in global datasets. If you are building for Zambia, you cannot blindly accept outputs that do not reflect our curriculum, our legal environment, our business realities, or the way our institutions actually work.

This is something I take seriously in my own work. At eskulu, the goal is not just to use AI because it is trendy. The goal is to make AI genuinely useful for Zambian learners. That means grounding systems in the ECZ curriculum, making content accessible, and ensuring that technology solves a real educational problem rather than creating confusion.

AI can accelerate work, but humans still provide context, responsibility, and accountability.

What This Means for Zambia and Africa

In Zambia and across Africa, this transition creates both risk and opportunity. The risk is obvious: many people are preparing for jobs as if the old model of work will remain unchanged. The opportunity is even bigger: we have a chance to leap forward if we adapt quickly.

Africa does not need to wait for permission to participate in the AI era. We can build local tools, local datasets, local products, and local businesses that solve African problems. That is part of what has driven me in my own journey. I did not start from a massive lab or a giant company. I started by learning, building, experimenting, and staying consistent.

That consistency has opened important doors for me. I reached the Top 5 of the ZICTA Innovation Programme with eskulu, won Business With a Purpose at the X Pitchathon by Accessbank and MTN in 2023, and placed 3rd at the Yango and Zindi Data Science Hackathon in 2024. Those milestones matter to me not just as personal wins, but as proof that serious innovation can come from Zambia.

If AI is going to reshape work, then African builders must be part of shaping what comes next.

From Doing Everything Manually to Directing Intelligent Systems

I believe one of the most important skills going forward will be the ability to direct intelligent systems effectively. That includes prompt design, workflow thinking, model evaluation, deployment, and knowing when not to trust an output.

This is one reason I have continued sharpening my technical foundation through both formal study and practical work, including certifications such as AWS Lambda Foundations and Amazon Bedrock. Modern AI is not just about chatting with a model. It is about building systems that are reliable, scalable, and useful in the real world.

For students and young professionals, this means the winning mindset is no longer simply “learn to do the task manually.” It is “learn the task deeply enough that you can supervise, validate, and improve what AI produces.” That difference is huge.

The future belongs to people who can combine domain expertise with AI leverage.

What I Am Seeing as a Builder

As someone building products in education and consulting around AI, I am seeing a pattern: the people who benefit most from AI are not always the people with the most technical titles. They are often the people who understand workflows, ask better questions, and know how to evaluate results.

That matters for founders, schools, businesses, and institutions in Zambia. If you are running a company, AI can help with reporting, customer support, internal knowledge systems, and software development. If you are in education, AI can support tutoring, revision, content delivery, and personalization. If you are a developer, AI can dramatically increase speed, but only if you know how to guide it well.

The real competitive advantage is moving away from raw effort alone and toward intelligent orchestration.

My View of the Future

I do not think the future is one where humans stop working. I think it is a future where routine intellectual tasks become increasingly automated, while human beings focus more on strategy, verification, creativity, ethics, and systems thinking.

That future is highly relevant to what I am building through ZOEC. My long-term vision is not just to create another app. It is to help build an ecosystem where every learner in Zambia can access intelligent educational support, and where African technology is built with local purpose. Through eskulu, we have already reached 500,000+ students across Zambia, and I believe this is only the beginning.

If AI can now handle much of the heavy lifting in knowledge work, then our responsibility is to make sure it is applied in ways that expand opportunity rather than concentrate it.

Conclusion

We are moving into a world where AI can perform a large share of research, reporting, analysis, and coding. That shift is real. But the deeper story is not that human work is ending. It is that human work is evolving.

The people who thrive in this new era will be those who learn how to guide AI, monitor AI, and build with AI responsibly. For Zambia and Africa, this is a moment to adapt boldly, build locally, and think beyond consumption toward ownership and innovation.

I am committed to that mission through eskulu, ZOEC, and my broader AI and software work.

If you want to explore eskulu, collaborate on an AI project, or need consulting in AI systems, edtech, or software development, feel free to reach me at jeffmdala@gmail.com.