The next twenty years will change the world, though not in the manner currently advertised by those who mistake a product demonstration for destiny.

AI will not “take over”. Large language models will improve. Reasoning systems will become more capable. Tools will become faster, more precise and more deeply embedded in ordinary work.

But tools do not abolish knowledge. They increase its value.

A person trained in art history, visual composition and the proper language of style can produce a far better image with an AI system than someone who merely types “make it beautiful”. The machine supplies execution. The human supplies vocabulary, judgement, context and purpose.

The same is true in law, science, engineering, literature and economics. An untrained person may obtain an answer. A trained person can recognise whether it is coherent, incomplete, derivative or wrong.

The divide will not be between humans and machines.

It will be between people who understand their field and can direct the tools, and people who use the tools without understanding what they are asking for.

Education must therefore change.

We do not need to preserve every manual trick merely because earlier generations lacked calculators, computers or models. We do need people who understand the underlying ideas well enough to use those tools intelligently, test their outputs and improve upon them.

The purpose of education is not to make the student imitate a machine.

It is to ensure that, when the machine produces something, the student knows what it means.