By Dr Etuna Simon
A student sits in front of a laptop with an assignment due the next morning.
The question is not easy. It requires reading, interpretation and a carefully constructed argument.
A few years ago, that might have meant hours of searching through articles, making notes, deleting weak paragraphs and starting again.
Today, the first instinct may be different.
The student opens an AI tool, types in the assignment question and waits.
Within seconds, an answer appears.
It is structured. It sounds confident. The language is polished. There is an introduction, an argument and even a conclusion.
The assignment suddenly looks much easier.
But imagine what happens next.
The student reads the response, changes a few sentences, adds a reference or two and submits the work. The final product may be stronger than anything the student could have produced alone.
Yet one question remains. What exactly did the student learn? That question is becoming increasingly important for education in Namibia.
Artificial Intelligence (AI) is giving learners access to tools that can explain difficult concepts, generate ideas, improve writing and support problem-solving.
Used well, these tools can expand learning in ways that would have been difficult to imagine only a few years ago. But the same technology can also remove some of the intellectual struggle through which learning normally takes place. The challenge is therefore no longer simply whether students should use AI. They already are.
The more difficult question is whether our education system can distinguish between better output and deeper learning.
Because an assignment can improve while the learner does not. And that may become one of the most important educational questions of the AI era.
A recent analysis by IJG Securities, the stockbroking and research arm of Namibia’s IJG Group, drawing on anonymised data from Anthropic’s Economic Index, found that coursework represents the largest share of Claude usage in Namibia and accounts for more than twice the global proportion.
The same analysis suggests that Namibian users are more inclined than the global average to give AI direct instructions and automate tasks rather than use it collaboratively.
The report is careful to note that Claude users are not representative of the entire Namibian population. Still, the pattern deserves our attention.
If education is becoming one of the main spaces in which Namibians encounter generative AI, then our classrooms and universities may be where the country learns its most important habits about how humans should work with intelligent machines.
Completing a task is not the same as learning
This is where the conversation needs to move beyond whether AI use is “good” or “bad”. The evidence is more complicated.
A 2026 randomised study involving 275 university programming students found that AI support improved task performance and reduced frustration, yet the improvement in completing the task did not translate into significantly better conceptual learning.
Another randomised study published in 2025 found that students using unrestricted ChatGPT as a study aid performed worse on a surprise knowledge-retention test 45 days later than students who had studied using traditional approaches.
But other research has found benefits. A July 2026 randomised experiment found that students given access to generative AI achieved better immediate knowledge-test results, with gains persisting a week later.
Importantly, the strongest benefits appeared among students who used AI for augmentation asking it to explain concepts and support their thinking rather than simply asking it to generate the work for them.
That distinction may be the most important lesson for education. AI itself is not necessarily the problem. What matters is what intellectual work remains with the learner.
Namibia’s usage pattern should make us curious
The recent Namibia data becomes particularly interesting when viewed through that lens. If coursework dominates AI use, and if users are leaning more towards direct instruction and automation, what exactly are we automating?
Are learners asking AI to explain something they do not understand? Are they using it to interrogate an argument? Are they comparing its response with their textbooks and lecturers?
Or are we increasingly outsourcing the difficult parts of learning reading, organising ideas, constructing arguments and solving problems?
We do not yet have enough evidence to answer those questions conclusively. That is precisely why universities and schools should start asking them. The wrong response would be panic. The equally wrong response would be to assume that because an assignment looks better, learning must also have improved. AI has made that assumption unsafe.
Our assessment systems were built for another era
For a long time, education could reasonably assume that the person who submitted an essay had done most of the intellectual work behind it. That relationship is weakening.
A polished essay no longer automatically demonstrates that the student can construct an argument. Working computer code does not necessarily demonstrate that the learner understands why it works.
A beautifully structured research proposal may conceal very limited understanding of the methodology it describes.
This does not mean written assignments should disappear. It means assessment must become more interested in the process of thinking, not only the final product. We need more oral defence of submitted work.
More supervised problem-solving. More drafts that demonstrate how an argument developed. More assessment based on authentic Namibian problems.
And where AI has been used, students should increasingly be expected to explain what they used it for, what they accepted, what they rejected and why. The critical assessment question may no longer be simply, “What did you submit?” It may become: “Can you explain and defend what you submitted without the machine?”
We need AI literacy, not merely AI access
The National Student Research Symposium earlier this month placed artificial intelligence at the centre of discussions about the next generation of Namibian researchers. Education Minister, Hon. Sanet Steenkamp offered a principle that should extend well beyond research: “AI can assist a researcher, but it cannot relieve that researcher of the responsibility to think critically, verify evidence and defend conclusions.” That is exactly the kind of AI literacy education needs. Teaching students to write prompts is not enough.
They should understand how to verify AI-generated information, recognise confident but inaccurate responses, protect sensitive information, interrogate sources and know when a human expert is required.
Most importantly, learners must recognise when AI is helping them think and when it is quietly thinking instead of them. The difference may not always be obvious.
Perhaps we should measure learning differently
Namibia does not need to choose between AI and education. We need to become better at designing education for a world in which AI is present.
The new data showing substantial coursework use should therefore not provoke a campaign against students using AI. It should provoke better questions.
Are our assignments still testing what we think they are testing? Are students learning how to collaborate with AI rather than merely delegate to it? Are lecturers redesigning assessment? Are teachers being prepared to recognise the difference between polished output and genuine understanding?
And perhaps most importantly, are we teaching learners that productive struggle still has educational value?
Sometimes the difficult paragraph, the problem that takes an hour to solve, or the argument that must be rewritten three times is not wasted effort. That struggle may be where the learning actually happens.
AI can make education faster. It can make it more personalised, accessible and productive. But speed should never become our measure of learning.
As Namibia embraces AI, the question should therefore not be how much coursework the technology can help us complete. The more important question is whether, after the assignment is submitted and the screen is closed, the knowledge still belongs to the learner.
*Dr Etuna Simon is a researcher, strategist and writer with a keen interest in education, innovation, artificial intelligence and national development. The views expressed in this article are solely those of the author. She can be reached at etunasimon@icloud.com
