I regularly meet leaders who almost apologise. You know, I am not technical. As if that admission disqualified them from deciding anything at all about AI. It is a widespread belief, and it is not only false, it is dangerous, because it pushes them to delegate decisions that are theirs to make.
Steering and coding are two trades
Steering AI and building AI are two different things. We do not ask a leader to know how to make the engine. We ask them to know where they want to go, what they accept handing to the machine, and where they keep control. That is not technical. It is judgement, and it is precisely their job.
What a leader must understand about AI comes down to a few things, and none of them requires an engineering degree. That the tool produces the plausible and not necessarily the true, so you check what commits you. That its data is an asset to protect, so you watch where it goes. That a tool serves a decision and never the other way round, so you start from the need, not the fashion. With these three ideas, you steer very well. Better, often, than the person who masters the technology but loses sight of the stake.
The opposite mistake, more dangerous still
There is a symmetrical mistake, and I find it worse. Handing the whole decision to someone technical, on the grounds that they understand the tool.
Understanding how it works says nothing about what should be done with it. The technician knows what is possible. The leader must know what is desirable. The two do not replace each other, they complete each other. A leader who abdicates their judgement because they feel illegitimate lets deeply strategic choices, about their data, their customers, their organisation, be settled by a purely technical logic. And what is feasible is not always what is good for the company.
I have seen projects drive into the wall exactly like that. Not through technical incompetence, on the contrary, through excess confidence in the technology and abdication of judgement. No one, in the room, was asking the simple question: do we really want to do this, and what for?
Becoming legitimate, not technical
That is also why, in my work with leaders and managers, I never try to turn them into technicians. I try to make them legitimate. Able to ask the right questions of a provider or a team, to spot an empty promise, to tell real value from showmanship, to decide knowingly without being impressed by jargon.
This legitimacy is built fast, faster than those concerned believe. A few hours well spent understanding what this technology really is, what it can and cannot do, are enough to turn an intimidated leader into one who leads the discussion. Jargon is often just a barrier, and once you have grasped the few principles that count, it falls.
The technology, you will always find people to handle it, and that is fine. The judgement, no one will carry it for you. Feeling illegitimate about AI because you do not code is like feeling illegitimate to run a company because you cannot fix the photocopier. Your value was never there. It is in your ability to decide what is desirable. AI does not change that rule. It makes it more important than ever, because it multiplies the number of feasible things, and therefore the number of decisions where someone has to say what is worth doing.