We talk a great deal about learning AI. We rarely talk about unlearning, and yet it is the skill that seems to me the most decisive today, and by far the hardest.
To unlearn is not to forget. It is accepting that certainties which worked yesterday no longer work, and letting go before they become blind spots. A trade you thought was safe. A way of working you judged unbeatable. A line between what was human and what could not be. AI shifts these lines faster than our habits update, and our habits die hard.
The danger is not ignorance, it is the expired certainty
For a leader, the risk is not failing to know. Ignorance you spot, you fill it. The real risk is knowing things that are no longer true, and deciding on them without realising it. Yesterday’s confidence becomes today’s trap.
I see it often. A leader who built their success on a right intuition for twenty years applies that intuition to a world that has changed under their feet, and does not understand why it no longer works. It is not a lack of intelligence, it is the opposite: an intelligence that learned too well, and did not unlearn in time. Solid knowledge is the most dangerous when the ground moves, because it is the knowledge we question the least.
Why it is so hard, and so rare
Unlearning takes a particular kind of courage, because it forces you to admit you had settled into an idea, sometimes comfortably, sometimes for a long time. It is uncomfortable, it touches the ego, it gives the feeling of going backwards.
That is exactly what separates those who get through a change from those who endure it. The first revise their maps when the territory changes. The second defend the old maps until they get lost, blaming the territory for being wrong. AI, right now, is redrawing the territory at full speed. Never has the ability to revise your maps been worth so much.
And it is rare, because everything in a successful career pushes towards confidence rather than doubt. You are rewarded for being right, not for changing your mind. Unlearning runs against that mechanism.
How to keep it sharp, in practice
The good news is that it can be trained like a muscle.
Ask yourself regularly, honestly, what you take for granted and what might have moved. Not once a year at a seminar, but as a mental hygiene. Go and see what is being done elsewhere, in other sectors, before it reaches yours, because ruptures travel from one trade to another. Prefer, in a meeting, the awkward question to the reassuring answer. And surround yourself with people who dare to tell you that you are wrong, rather than people who confirm what you already think.
There is a direct link with AI itself. A leader who knows how to unlearn uses AI well, because they accept that it unsettles their certainties about what was possible. A leader braced against their own knowledge rejects it or endures it, depending on their mood, without ever really using it.
AI does not reward those who know the most. It rewards those who accept the fastest to revise what they know. Learning, we all do it, it has even become a reflex praised everywhere. Unlearning, almost no one dares. It is precisely there, in that space most people avoid, that the lead of the coming years is won.