Everyone debates whether AI will replace managers.
Almost no one debates what happens when it keeps new managers from ever being developed.
And that's the real problem.
The figure that should keep any CEO awake
The KPMG CEO Outlook 2026, published in March 2026 with 100 CEOs of large US companies, revealed something that isn't in the headlines about artificial intelligence.
Asked about their biggest concern regarding AI's impact on leadership development:
- 31% cited fewer opportunities for early-career professionals to develop judgment through experience
- 30% pointed to over-reliance on AI for decisions, limiting critical thinking
- 22% mentioned less exposure to ambiguity, failure and trial-and-error learning
Together, 83% of CEOs are worried about something that isn't technology. It's human development.
And here's the paradox: 67% of those same CEOs admitted they still haven't redefined roles or career paths for the AI context. In other words: they know the problem exists. They're doing nothing about it.
The silent pact that broke
There was a tacit agreement in every organization.
The junior professional did the operational work, repetitive, draining, but formative. In return, they got mentoring, exposure to real decisions and a clear path to leadership.
AI broke that pact.
The operational work that used to form people is now done by the machine. And no one redesigned the path.
The result? The junior no longer makes mistakes, because AI answers for them. Doesn't face ambiguity, because the model suggests the way out. Doesn't develop judgment, because they never had to decide without a safety net.
As a study published in March 2026 in the Journal of Experimental Orthopaedics put it, the risk isn't only deskilling, the loss of skills already acquired. It's never-skilling: when AI enters training so early that the foundational skills are never developed at all.
Today's productivity, tomorrow's fragility
The short-term numbers are seductive.
A Forbes survey showed that 77% of employees say AI lets them do more work in less time. KPMG's CEO, Tim Walsh, described the new metric dominating boardrooms: the labor cost margin, how much labor costs per unit of work delivered. The pressure is clear: cut cost, raise volume.
But while global AI spending heads toward US$500 billion in 2026, training budgets are shrinking. And employee motivation hit a six-month low, according to Fortune.
The math doesn't add up.
You gain efficiency now. But you're emptying the leadership pipeline that will sustain your operation five years from now.
In industry, the risk is even more concrete
On the shop floor, judgment isn't abstract. It's the coordinator who looks at the line and senses something's wrong before any sensor flags it. It's the manager who decides to resequence production when the plan won't hold. It's the maintenance lead who prioritizes an intervention based on accumulated experience, not just an algorithm.
That kind of competence isn't transferred through a dashboard. It's built over years of exposure to real problems, imperfect decisions and consequences lived through.
If the next industrial manager never had to make a hard decision because AI always suggested the path, what happens the day the system goes down, the scenario shifts or the model gets it wrong?
How Magellan handles this in practice
At Magellan Consulting Group, we use AI as a project tool, to speed up diagnosis, model scenarios and anticipate bottlenecks precisely.
But at no point does AI replace the manager's judgment. It feeds the decision. The person decides.
In our transformation programs, the logic is deliberate:
- AI does the heavy lifting of data and simulation
- The manager interprets, challenges and decides
- The junior team takes part in the build, not just the execution, so they learn to think, not just to operate
Because we know technology changes fast. But a leader who never learned to decide under pressure isn't formed by a software upgrade.
The uncomfortable point
The question few are asking isn't how many jobs will AI eliminate?
It's: how many leaders will AI keep from ever existing?
Because if you automate the whole path, from analysis to recommendation to decision, what's left for the human isn't leadership. It's passive supervision.
And passive supervision forms no one.
57% of US workers already say the erosion of human skills will be the biggest problem in the labor market in 2026, above even the fear of losing their jobs.
The risk isn't the machine taking the manager's place.
It's the manager never having learned to be one.
Final question: In your company, are the professionals being developed still genuinely deciding, or are they already just validating what AI suggests?