The illusion of learning in the age of AI

Walk into almost any university in Pakistan and you will see students producing assignments, presentations and exam scripts that look impressively neat and complete. Yet beneath this appearance lies a quieter question: how much real learning has taken place? Over time, our academic incentives have gradually shifted toward the polished output, the essay submitted, and the test score achieved, rather than the cognitive effort required to produce them.

When outcomes become the dominant measure of success, the intellectual struggle that makes learning meaningful often fades from view. Long before artificial intelligence made such shortcuts effortless, this emphasis on completion over comprehension had already begun to blur the distinction between producing the right answer and truly understanding it.

Seen this way, AI is not only a tool. It is a new cognitive environment. Artificial intelligence has made access cheap and immediate. Information, explanations, summaries and even polished arguments now arrive in seconds — a remarkable achievement. But it has also made it easier to confuse exposure with understanding. When answers appear fluently, we mistake recognition for knowledge. We assume that because something makes sense on the screen, it has taken root in the mind.

The journey from screen to mind is what we call learning. It is an internal process. Cognitive science defines it as durable changes in long-term memory.

Turning information into understanding takes time. The mind must stay with an idea and work through it until it becomes usable knowledge. That process depends on sustained attention. Yet attention is now the scarce resource. Herbert Simon, a Nobel laureate and pioneer of artificial intelligence, warned that wealth of information creates poverty of attention. We are now living inside that paradox. Information is abundant, but the attention required to digest it is becoming scarce. And without full attention, access to that information does not translate into learning.

Learning is not a smooth transfer of information. It involves the breaking and rebuilding of mental structures. When an idea challenges an existing mental model, the mind resists. We feel confusion, difficulty and discomfort. That friction is not a defect in the learning process. It is the signal that old structures are being questioned and new ones are being built. Consider a child learning to walk. No amount of watching demonstrations produces balance. The child must wobble and fall, repeatedly, until instability becomes coordination. The same applies to intellectual growth. When a student wrestles with a proof or struggles to articulate an argument, they are reorganising cognitive structures.

Still, friction alone is not enough. Learning also demands emotional investment. The brain does not consolidate information into long-term memory unless it is tagged as meaningful. Curiosity, interest, relevance, or even mild frustration: these emotional signals tell the mind that the effort is worth sustaining. Without intellectual and emotional attention, difficulty leads not to growth but to withdrawal. Durable learning is cognitive strain sustained by commitment.

The real danger here is not artificial intelligence itself. It is a human tendency that AI happens to satisfy exceptionally well. We are psychologically drawn to cognitive ease. When a result looks polished and arrives quickly, we treat it as our own proof of competence. That bias makes it easy to confuse performance with learning, and fluency with understanding. The question becomes how to overcome this bias while keeping AI in the loop without taking the learner out of the work.

The answer is not to ban the tool. It is to redesign the learning experience so meaningful friction is preserved. Students still need to retrieve ideas from memory rather than recognise them on a screen. They still need to attempt solutions before seeing polished ones. They still need to explain, apply and revise.

This is where AI becomes less a shortcut and more a design partner. A teacher's task is not merely to explain clearly. It is to design experiences in which productive friction occurs, where students must think, attempt, revise and persist. AI can help create those experiences at scale by producing varied practice, generating counterexamples and giving immediate feedback that keeps effort from turning into despair. The goal is to calibrate difficulty, to find the tension between comfort and confusion where growth occurs.

Classroom design matters, but it sits inside an institutional incentive system. This responsibility cannot rest with teachers and students alone. Students respond rationally to the environments we place them in. When grades reward completion, they will optimise completion. When assessment rewards polish, they will seek polish. In the age of AI, the gap between what is rewarded and what is learned becomes impossible to ignore.

Universities have a choice. They can treat AI as a disciplinary problem and chase detection. Or they can treat it as a design signal and rebuild the learning environment around what cognitive transformation requires. That means shifting the academic focus from the outcome to the learning progression itself. Equally important, it means consciously placing value on meaning-making and its authentic, subjective expression over the seamless, objective generation of content.

For Pakistan, the stakes are particularly significant. With more than 60 per cent of our population under the age of 30, our universities are shaping the intellectual capacity of one of the world's youngest societies. The question, then, is not whether machines will produce better answers. It is whether our universities will continue to create the conditions in which young minds learn how to think, question and understand. In a country as young as Pakistan, that question may well determine whether artificial intelligence becomes a shortcut to superficial knowledge, or a catalyst for deeper learning and stronger nation-building.