When artificial intelligence eliminates the junior work we need in order to create the senior professionals of the future, a paradox emerges — not with a bang, but in silence, slipping through the doors of offices everywhere. It can be stated in a single sentence: the senior no longer needs the junior to do the work that once taught the junior how to become a senior.
Artificial intelligence is multiplying the productivity of experienced professionals. But it may simultaneously be dismantling the quiet mechanism through which our industries have always produced the next generation of talent.
Consider a real episode. A senior law professor was discussing with fellow lawyers the use of AI in drafting legal opinions. A young trainee — a former student of his and an outstanding one — was listening. At a certain point, she asked:
'Professor, you know how to write a legal opinion. You use AI to do it faster, and then you know how to check whether it is right or wrong. I have never written one. How am I supposed to learn?'
'You will have to find a way.'
The answer is simple. The question is not.
For decades, young professionals began their careers by doing the work that senior colleagues knew how to do but no longer had the time or need to execute themselves. They researched legislation, organised information, ran calculations, prepared first drafts, made mistakes, corrected them, and made mistakes again. It was inefficient. But that inefficiency was not a flaw in the system — it was the system. It was how people learned.
Artificial intelligence is now automating precisely those tasks.
For the senior professional, this is a liberation. For the firm, it is a productivity gain. For the junior, it may be the disappearance of the ladder.
Because nobody starts as a senior.
There is an important distinction here from previous technological revolutions. Not long ago, a newcomer could rapidly overtake a veteran in specific tasks by mastering Excel, pivot tables, internet research, or new databases. The senior had experience; the junior brought the tool. They needed each other, and knowledge flowed in both directions.
AI breaks that equilibrium. Its primary interface is language. An experienced professional does not need to learn to code or master complex formulas in order to use it. What is required, above all, is the ability to articulate the problem. And here the senior has an inherent advantage: they know what to ask, what context to provide, which criteria matter — and, critically, when an answer does not make sense.
AI functions as an experience multiplier. Those without experience have little to multiply.
There is therefore a professional generation in a singular position: it learned without AI and can now produce with AI. Those entering the labour market today face the inverse problem. They can produce a great deal before they have learned what would allow them to evaluate what they produce.
And that cannot be resolved with a prompt-writing course.
A young lawyer can obtain a legal opinion in seconds. But if they have never constructed one, how do they distinguish a sound analysis from an argument that is merely persuasive? A designer can generate fifty proposals in an afternoon. But volume does not create repertoire, context, or judgement.
The question therefore shifts. It is no longer what AI can do. It becomes what we stop learning when AI does it for us.
This is not about preserving pointless tasks out of pedagogical nostalgia. Nobody needs to spend an afternoon searching for a regulation that a machine can locate in seconds. But eliminating the task cannot mean eliminating the learning that used to happen within it.
This is where firms and universities face a genuinely new problem. For a long time, organisations taught young professionals because they needed their work. They may now no longer need much of that work — but they will continue to need the professionals that work helped to create. When the senior leaves, someone will have to be capable of taking their place. And that someone cannot be improvised.
The prescription is straightforward: do not train young professionals to compete with AI on execution. Train them to think before it, to work alongside it, and to interrogate it afterwards.
Let them use it — but require them to explain the problem, justify the choices, identify the errors, and defend the conclusions. Deliberately create the learning opportunities that previously arose, effortlessly, from the work itself.
Because today's productivity cannot be purchased at the cost of tomorrow's competence.
That is the continuity paradox.
It has never been so easy to work like a senior. It has never been so uncertain how to become one.
Source: Diário Económico
Original article: https://www.diarioeconomico.co.mz/?p=534472
