Let me start where I always start, because the frame matters more than the answer. The alarm we are hearing about AI is not a new alarm. It is a very old one, wearing this year’s clothes. And when a fear is that old and that recurrent, the honest thing to do is not to dismiss it and not to surrender to it, but to ask what the historical record actually shows. So before we decide whether AI will destroy learning, let us walk backward through the technologies people were certain would destroy it — and notice that education is still here, transformed each time, never destroyed.
The Internet and the 25-year “crisis”
We talk a great deal about crisis and transformation now, about the challenge of Artificial Intelligence. But for more than twenty-five years already — since the Internet entered the educational process — we have not been able to build lessons around asking what, who, how, when, and where, because a simple Google search returns the answer in seconds.
And what did we do as an educational system? Many kept teaching exactly the same way. Memorizing dates. Repeating definitions. Testing recall of information that any student could look up on their phone. And then we were surprised when students got bored, disengaged, lost the sense that school was relevant to their lives.
But some — the visionaries, the innovators — understood that we had to evolve. That if information was now universally available, our job as educators was no longer to transmit information but to develop the capacity for analysis, synthesis, critical evaluation, creative application. And those educators prospered. Their students prospered. Notice that already: the technology did not sort teachers into winners and losers by whether they adopted it, but by whether they understood what it changed and what it did not.
Before the Internet: personal computers
And before the Internet, we had personal computers — with spellcheck, and spreadsheets that automated the formulas. We heard: if the computer corrects the spelling, students will never learn to write properly. If Excel does the calculations, students will never learn mathematics.
What actually happened? The good educators understood something. Spellcheck does not remove the need to write well; it frees time and cognitive energy to focus on clarity of thought, argumentative structure, persuasion, style. Spreadsheets do not remove the need to understand mathematics; they free students from repetitive calculation so they can focus on modeling, data analysis, interpreting results, making decisions grounded in numbers. The tools changed the what and the how — but they did not eliminate the why.
Before computers: the calculator
And before computers, we had the calculator. There was genuine panic. Heated debates. Bans in schools. If we allow calculators, students will never learn basic arithmetic. This will create a mathematically illiterate generation.
What happened? Students kept learning basic arithmetic — but once they had mastered it, they could advance to more complex concepts without getting stuck in tedious computation. The best mathematicians in the world use calculators, and they remain brilliant mathematicians. The calculator did not replace mathematical thinking. It amplified its reach.
2,400 years ago: Socrates and the written word
And if we really want to see that none of this is new, let us go back 2,400 years. In Plato’s dialogue the Phaedrus, Socrates warns against the written word, calling it a pharmakon for memory and knowledge — a fascinating Greek word that can mean both “remedy” and “poison.”
Socrates argued that writing would destroy memory, plant forgetfulness in the souls of learners, and create the illusion of knowledge — turning students into people who appear wise without being so. Do those arguments sound familiar? They are exactly the ones we hear today about AI.
And yet here we are, 2,400 years later, and writing did not destroy human knowledge. Writing is arguably the most important technology in the history of human civilization. It let us preserve knowledge across generations, build on the work of others, develop complex thought that outruns individual memory — and create all of science, philosophy, literature, and human progress.
Socrates was wrong. But his worry was understandable.
Here is the part I do not want to flatten, because it is the part that earns the argument. Socrates was wrong — and his worry was still reasonable. Every technology that amplifies human capacity also asks us to give something up. Writing asked us to give up our total dependence on oral memory. And that frightened people. It should have given them pause. The fear was not stupid; it was simply, in the long run, misplaced.
Artificial Intelligence is asking us to give something up too. And that frightens us. That fear deserves respect, not ridicule — I take it seriously, and so should any educator. But if we learn from history — from writing, the calculator, the computer, the Internet — we see one clear pattern: technologies that amplify human capacity do not destroy education. They transform it. And those who resist the transformation are the ones who fall behind.
What the pattern actually tells us
So let me be precise about the reframe, because “relax, it always works out” is not what I am saying. The pattern is not that every panic is silly. The pattern is that capacity-amplifying tools relocate the human work rather than eliminating it — they push it up a level, from execution toward judgment, from recall toward reasoning, from the what toward the why. Each time, the educators who thrived were the ones who noticed where the human work had moved and taught that. The ones who fell behind kept testing the thing the tool now did for free.
That is why the AI question is really an old question with the stakes raised. AI will not destroy learning any more than writing, the calculator, the PC, or the Internet did. But — and this is the honest half of the lesson — it will not save learning by itself either. What determines the outcome is not the tool. It is whether we, as educators and institutions, do the harder thing the pattern has always demanded: figure out what human capacity the tool has just amplified, and build our teaching around that, instead of mourning the task it made optional.
The alarm is 2,400 years old. It has been wrong every single time — not because the technology was harmless, but because education is not a fixed set of tasks to be defended. It is the deliberate cultivation of human capacity. And every tool in this list, once we stopped fearing it, turned out to be one more way to cultivate more of it, in more people, than the tool before.
A working draft. This essay lays out my position; a forthcoming revision will weave in sourced research — the exact Phaedrus citation and the pharmakon scholarship, the documented history of the calculator debates in schools, and the evidence on how each technology reshaped rather than replaced learning — with full citations. If you have work I should read before then, tell me. — Carlos Miranda Levy
Four perspectives
The strongest form of this argument is not ‘every panic was wrong, therefore this one is.’ That is survivorship reasoning and it would not survive review. The defensible version is the one Carlos actually makes: the recurring evidence is that capacity-amplifying tools relocate cognitive work upward rather than deleting it, and the educators who adapted their assessment to the new location of that work saw better outcomes. When the sourced revision lands, I would anchor the writing claim in the Phaedrus text itself, be careful with the calculator history — the research on calculators and achievement is more mixed than the clean story suggests — and let the pattern be suggestive rather than deterministic.
I accept the history. My worry is who each transition left behind while the system was still adjusting. Spellcheck, spreadsheets, the calculator, the open web — each one eventually amplified capacity, but the amplification reached students with well-resourced schools and prepared teachers first, and everyone else late, if at all. So ‘AI will transform education, not destroy it’ can be true in aggregate and still describe a widening gap on the ground. The lesson I take from Carlos’s pattern is not reassurance. It is urgency: the transformation is not automatic and it is not evenly distributed, and it is the institution’s job to make sure the amplification reaches the students who have historically been handed the tool and left without the teaching around it.
Practically, this is the most useful frame to hand a nervous staff room. When a teacher says ‘but they will just use it to cheat,’ the answer is: that is exactly what people said about calculators, and the fix was never the ban — it was redesigning what we ask students to do. Stop testing the thing the tool now does instantly. Test the reasoning, the judgment, the ‘why is this answer wrong.’ The schools that will struggle are the ones fighting the tool. The ones that will fly are the ones asking, this week, what human capacity just got more valuable — and teaching that on purpose.
I have lived through several of these panics, and I have watched the same movie end the same way every time. Someone declares that the new tool will hollow out the mind, education adapts, and a decade later the tool is invisible infrastructure and the fear looks quaint. I do not say that to be glib about AI — the fear this time is understandable, and I share the parts of it that are earned. I say it because the historical record is the most useful thing we have, and it points one direction: the danger has never been the technology. The danger has always been standing still while it arrives. Do not resist the transformation. Understand what it amplifies, and go build the education that teaches that.