Students who resist AI are raising the right concern. Uncritical AI use can weaken judgment. It can make people faster at producing answers but less capable of deciding whether those answers are worth having. It also trains students to outsource attention, interpretation, and discernment. Those aren’t minor losses. They are at the center of what education is supposed to develop.
But the concern and the response aren’t the same thing. The concern is real, but the response is too weak. Resistance alone doesn’t protect judgment.
It begins with the wrong question: What will AI do to learners? That question matters, but it gives the tool too much authority. It treats the learner as a passive object and leaves only two positions: adoption or refusal.
That isn’t how judgment develops.
Judgment develops when people decide what deserves attention, what can be delegated, what must be verified, what must be resisted, and what consequences they’re willing to own. AI exposes that work.
Every general-purpose technology changes the conditions under which human capacity is tested. Calculators, television, the internet, and smartphones all produced warnings about attention, memory, reasoning, social life, and authority. Some warnings were crude. Others were right. Critics often saw that something human was at stake, even when they were wrong about the remedy.
Refusal didn’t preserve the old capacity. It usually left resisters with less influence over the new environment in which that capacity had to operate. The calculator didn’t make mathematical judgment irrelevant. It changed which forms of mathematical work deserved time. The internet didn’t make knowledge unnecessary. It made source judgment, synthesis, and disciplined attention more important.
AI’s likely to follow the same pattern, with higher stakes.
The question isn’t whether AI will enter education and organizational life. It already has. Will institutions teach people to use it with judgment? Will students and executives learn by convenience, imitation, and fatigue?
The student who resists AI sees one failure clearly. Many institutions haven’t done the hard work. They haven’t explained what AI changes about learning. And they haven’t distinguished between work in which AI should be prohibited, work in which AI should be required, and work that should force students to defend the line between the two.
A writing assignment, for example, shouldn’t simply say “AI allowed” or “AI prohibited.” It should ask students to show their work.
What did they ask the system to do?
Which sentence did they keep?
Which one did they rewrite?
Which claim sounded plausible but turned out to be unsupported? Where did the machine flatten the argument, miss the tension in a source, or make a confident claim that direct reading didn’t justify?
That kind of assignment treats AI as material for judgment.
Most institutions still aren’t there. They haven’t taught verification as a discipline or delegation as an intellectual act, nor have they required students to explain what they handed to the machine, what they kept for themselves, what they checked, what they rejected, and where accountability remained.
That’s institutional abdication.
And abdication is answered by a more serious pedagogy.
Carlota Perez’s account of technological installation clarifies the mistake. Early in a general-purpose technology’s life, institutions tend to swing between enthusiasm and panic. One side treats the tool as liberation; the other treats it as corruption. Both miss the harder fact: technology is becoming part of the infrastructure of work, coordination, production, and thought.
We must ask who is shaping AI infrastructure.
That distinction matters for students, and even more so for leaders. A student who refuses AI may protect some habits for a time. The student may write more slowly, face the blank page more directly, and preserve a stronger sense of ownership over their first thoughts.
Those gains matter.
But refusal alone doesn’t teach the student how to judge machine-generated work nor does it teach students to detect false fluency. We must ensure that students learn how to use machine output without absorbing its assumptions, and see when speed is damaging the work.
Refusal can preserve effort, but it doesn’t necessarily build judgment.
The same problem now appears inside enterprises. Senior leaders are being asked to approve AI pilots, redesign workflows, evaluate vendors, manage model risk, protect customer trust, and decide where automation changes the moral architecture of the organization. Anti-AI purity isn’t a serious executive position, neither is uncritical enthusiasm.
The serious position is disciplined delegation.
A customer service team can use AI to draft responses, but someone still has to decide which complaints deserve escalation, which promises the company shouldn’t make, and which customers need a human being rather than a faster answer. A legal team can use AI to summarize documents, but someone has to know when a summary hides uncertainty. A hospital system can use AI to identify patterns, but someone has to decide where efficiency ends and care begins.
Those are critical leadership questions.
What can be delegated to AI?
Under what conditions?
With what visibility?
What must be verified by a human being?
What shouldn’t be automated, even if it can be? Where must accountability remain explicit because the decision carries reputational, legal, civic, or human consequences?
These questions can’t be answered by technologists and ethicists alone.
They belong to leaders.
This is where students who resist AI are closest to the truth. They understand that something valuable can atrophy when the tool is used without discipline. They’re right to worry about the substitution of fluent output for thought. They’re right that institutions have too often answered with policy rather than teaching.
But they aim that insight in the wrong direction.
The goal is to make AI part of the terrain on which judgment is deliberately developed and tested. A serious AI assignment wouldn’t ask only for a final answer. It would require students to show their work: what was delegated, what was kept, what was verified, what was rejected, and what changed after human review.
A serious executive program would do the same. It wouldn’t teach leaders to use AI as a productivity trick. It would teach them to decide where intelligence should reside inside the organization, how authority should move, and what consequences can’t be handed to a machine. That’s the harder educational task, and it’s also the only one equal to the moment.
The danger isn’t that students will use AI. Many will. And even more will use it badly. The concern is that institutions will allow a false choice between unconscious adoption and principled refusal. Both leave judgment underdeveloped.
The future belongs to people who can think clearly in the presence of powerful tools. That requires knowing when to use AI and when not to, recognizing the difference between assistance and dependency, seeing when speed has become a substitute for thought, and accepting that human accountability doesn’t disappear because a machine helped produce the answer.
The student resisters are right to defend judgment.
They’re just wrong to think resistance is enough.
Explore AI EDGE to learn how RGI helps leaders build the judgment, governance, and operating discipline to use AI responsibly.