Some of the most foolish people I know are highly educated. This may sound provocative but it points towards a distinction we increasingly refuse to make.
Education and intelligence are not the same thing. Knowledge and thought are not the same thing. The ability to reproduce an answer is not the same as the ability to generate, examine or defend it.
A person may possess degrees, qualifications and an impressive command of specialised language, yet struggle to think beyond the boundaries within which that knowledge was acquired.
They can cite authorities, repeat recognised theories and operate competently inside familiar institutional procedures.
But give them a problem that requires synthesis, independent judgement or connecting several domains, and their apparent intellectual confidence often exceeds their actual capacity.
Their education has supplied them with knowledge. It has not necessarily formed them as thinkers.
KNOWLEDGE AND KIZOMBA
I encounter this whenever I speak about kizomba. People are often surprised that a dance can reveal profound philosophical questions about difference, responsiveness, leadership, receptivity, tension, freedom and coordination.
The framework I use to examine kizomba is the same one I use to examine institutional corruption, including the Fishrot scandal.
The visible subjects are radically different but both can be investigated through questions of structure, formation, power, role, misalignment and consequence.
That is what thinking should make possible. It should allow us to recognise a structural pattern across apparently unrelated manifestations.
The test of thought is not merely how much it can recall but whether it can travel. Knowing several subjects is not the same as synthesising them.
A person may know a great deal about governance and a great deal about human relationships while remaining incapable of recognising the same formative dynamics operating within both.
Artificial intelligence (AI) now threatens to widen the gap between possessing knowledge and knowing how to think.
It can retrieve information, organise arguments, summarise complex material and produce polished language within seconds.
It can give users the visible product of reasoning without requiring them to undergo the reasoning through which that product would ordinarily be formed.
This creates a dangerous illusion: that possessing the answer is equivalent to possessing the thought. It is not.
MANIFESTATION
Within Structural Polarity, a framework I established in ‘The Dance of Difference: Structural Polarity and the Architecture of Human Coordination’, unfolds through formation into manifestation: S = F M
We usually encounter reality at the manifestation level. We see the political scandal after corruption becomes visible, the relationship after it has deteriorated, or the institution after its internal weaknesses have produced public dysfunction.
But the manifestation does not explain itself.
To understand it, we must move backwards towards the formation that produced it and the structure within which that formation became possible.
The same applies to thought. An answer is a manifestation. The observations, distinctions, assumptions, contradictions, comparisons and inferences through which that answer emerged constitute its formation.
When people retrieve a conclusion without reconstructing the reasoning behind it, they acquire the manifestation of thought without acquiring the thought itself.
They may know the answer but not why it follows, which assumptions support it, its boundaries, what evidence could overturn it, or whether it applies when the context changes.
Their knowledge is brittle. It functions when the question arrives in a familiar form. Change the language, introduce a contradiction or move the problem into another domain, and they become intellectually disoriented.
They possess the answer but cannot regenerate it.
OLD FAILURES ARE EXPOSED
Long before generative AI, algorithmic culture trained us to receive conclusions rather than form them.
Search engines determine what appears first. Recommendation systems repeat what retains our attention. Social platforms reward immediate reaction before sufficient understanding has had time to develop.
AI did not create the appetite for ready-made thought. It entered a culture already organised around satisfying it.
AI has exposed a weakness that already existed within modern education.
Universities should be institutions of intellectual formation. At best, they introduce students to accumulated knowledge, rigorous methods and arguments that no individual could develop independently.
Students learn the accepted vocabulary, approved authorities and expected conclusions of a discipline. They learn how to present the answer in the form the institution recognises.
They do not always learn how the answer was formed, what pressures it has survived, which assumptions remain contestable or how to recognise when an established framework no longer explains reality adequately.
A student can cite an argument without ever truly entering it.
This educational model functioned imperfectly even before generative AI. Students memorised, reproduced and forgot. AI has simply made the reproduction easier and more sophisticated.
THE PROBLEM
A student can now generate an essay without wrestling with the subject, produce an argument without discovering its distinctions, and submit polished language whose intellectual depth exceeds their ability to defend it.
The problem is not merely cheating. It is the separation of the manifestation of authorship from the formation of an author.
Writing has never been only the transfer of completed thoughts onto a page. It is one of the processes through which thought becomes clear.
The resistance of the sentence reveals where the idea remains vague. Revision exposes contradictions.
The search for the correct word forces the writer to distinguish what they mean from what they merely intended to mean.
When that process is entirely outsourced, the student does not merely avoid labour. They lose a site of intellectual formation.
They may possess the finished paper while remaining unchanged by its production.
Universities cannot respond to this crisis merely by trying to detect whether AI produced a piece of writing.
That approach remains trapped at the level of manifestation. It asks who generated the visible text without sufficiently examining whether the student possesses the reasoning it expresses.
PROCESS AND PRESENCE
The more important question is whether students can explain, defend and extend the argument.
Can they identify its assumptions? Can they respond to a serious objection? Can they apply the same reasoning to an unfamiliar case? Can they recognise where an apparently coherent answer becomes disproportionate or false?
Assessment must move beyond the finished product. The task of assessment is now to make intellectual formation visible.
Students should be required to defend important arguments orally, reconstruct the reasoning beneath their conclusions and apply their ideas in contexts not rehearsed in advance.
They should be asked to compare competing explanations, identify what would change their position and explain where an argument ceases to apply.
AI should also be brought into the educational process openly, not merely treated as a forbidden presence.
Students can be required to interrogate AI-generated answers, identify conceptual weaknesses, test claims and improve weak reasoning.
The capacity we need to develop is not the ability to produce language more quickly than a machine. It is the ability to judge what the machine produces.
WHAT TO OFFLOAD
None of this requires the rejection of AI. Human beings have always used external cognitive tools.
Writing stores memory. Maps externalise spatial knowledge. Calculators perform operations. Books allow us to think with people who lived centuries before us.
The real question is not whether we offload cognition. It is what we surrender.
There is a fundamental difference between offloading cognitive labour and offloading cognitive judgement.
AI can help organise material, expose contradictions, generate alternative formulations and apply pressure to an emerging argument. Used in this way, it can strengthen thought.
But when a person accepts a generated conclusion without testing its formation, assistance becomes dependency.
The tool no longer carries intellectual weight under human direction. It carries judgement in the human being’s place.
My own epistemic method is simple: Apply pressure. Observe. Apply more pressure. Watch what survives.
A claim should not be accepted because it arrives fluently, confidently or with institutional authority.
It must survive contradiction, evidence, consequence and alternative explanation. That discipline matters even more in an age when intelligent language has become cheap.
THE REAL CRISIS
We are approaching a future in which people may possess more answers than any previous generation while becoming progressively less capable of explaining how those answers were formed.
That is not collective intelligence. It is collective foolishness supported by unprecedented access to information.
The real AI crisis is not that machines can produce answers. It is that our institutions may continue rewarding the appearance of thought while neglecting the formation of thinkers.
The decisive question is not whether we can obtain the answer. It is whether, after the answer arrives, we still know how to think.
– Ismael Ashipembe is a Namibian lawyer, author and founder of the Structural Polarity Advisory.
