What’s the point of learning if AI is going to do everything?

by Dustin Steeve

“What’s the point of learning if AI is going to do everything?”

My friends who teach are hearing this question more and more, and the students asking it aren’t being cynical. They’re demoralized. I sit on the board of trustees for a well-established and growing private Christian school in Southern California, and I can tell you the administrators, the teachers, and the board are wrestling with the same question. Maybe your son or daughter is asking you some version of it right now. It’s a good question, and it deserves a serious answer, especially if your child is nearing college age and wondering whether a degree is worth the time or the cost.

Here’s how I answer it. I start by separating two things we tend to lump together: learning and memorizing.

Think back to your own schooling. For most of us, the curriculum was one long, building exercise in memorization: historical dates, scientific classifications, math formulas, the parts of a cell. Teachers “taught to the test,” and the test usually asked us to reach back into memory and either fill in the blank or pick the right fact off a list. If we remembered, we wrote it down or filled in the right bubble. If we remembered correctly, we scored well, earned good grades, and eventually were handed a diploma.

That diploma was useful. It was proof that we were good at remembering things and acting on what we remembered, which made us employable, and employment made income, and income let us buy a house, raise a family, and enjoy the benefits of a stable society. Memorize, recall, act, get rewarded. Call it the MRAR cycle. That’s what we’ve equated to “learning” based on our life experience.

Technology has long enabled part of this cycle. Books helped us memorize. Computers added recall. But a human was always necessary to complete it, to act on what was recalled and to be rewarded for it. AI is profoundly different from its predecessors in that it can run the whole cycle on its own. It memorizes, recalls, acts, and gets rewarded.

Today you can text a few sentences to a computer, tell it to build you a pretty website, and it will not only “understand” what a website is, but it will have criteria for whether the thing is “pretty.” Then it takes all the data it has absorbed about “websites” and “pretty” and builds a “pretty website” in front of you while you watch. And it is rewarded much the way we were: your edits reinforce the model, and your payments for tokens reward its maker.

So if learning is only the MRAR cycle, then there seems to be little reason to learn in the age of AI.

This is seemingly the assumption Dario Amodei was operating under when he warned last year that AI could cut U.S. entry-level white-collar jobs in half by 2030. Amodei runs Anthropic, the company that builds Claude. And the bottom line is hard to argue with. The vast majority of us will never out-memorize an AI, leaving aside the edge cases, the pockets of knowledge so rare that the machine has almost no data to train on. And before long we won’t out-enact it either, not on the strength of memorized data. Step for step, the MRAR cycle now belongs to the machine.

Laid out like that, the future feels heavy and bleak.

But we only arrived here because we started with the wrong question. The right one is shorter.

Not “what’s the point of learning if AI can do everything.” Just: what is the point of learning?

It wasn’t until the 1900s that “learning,” at least the kind that happens in a school, arrived at this functional but shallow formulation, the MRAR cycle. Prior to that, in the Western tradition, learning aimed at something else entirely: the cultivation of virtue. The idea runs deep. In Book II of the Laws, Plato defines education as “that training which is given by suitable habits to the first instincts of virtue in children.” Centuries later, in On Christian Teaching, Augustine sharpens it:

“Now he is a man of just and holy life who forms an unprejudiced estimate of things, and keeps his affections also under strict control, so that he neither loves what he ought not to love, nor fails to love what he ought to love, nor loves that more which ought to be loved less, nor loves that equally which ought to be loved either less or more, nor loves that less or more which ought to be loved equally.”

Tie the ideas of these and other formative philosophers of Western Civilization and the principle is clear: the point of learning is to love what ought to be loved, and to act in accordance with that love. That sounds abstract and difficult to apply to the real world, so let’s get concrete about how going back to this principle inspires us to learn in the age of AI.

Humans interact with AI using large language models. These models depend on natural language processing. The common denominator, the thing the whole edifice sits on, is language. Plain language, the kind you are reading right now. Language is how people will unleash the power of AI to build humanity’s future. It is how they are going to execute.

And notice what that assumes. It assumes a human being who is harnessing language. So the real questions move up a level. What is that human going to say? How will they decide what is worth saying, and what is better left unsaid? A head full of memorized facts doesn’t answer that. Data on its own has never once produced the creative spark it takes to build something that matters in the real world.

When a human sits down and prompts an AI, look at what they are actually bringing to the table. Genius. Instinct. The knowledge that comes only from having done the thing. Tacit knowledge, the kind you carry but can’t quite put into words. Wisdom, insight, judgment, intuition. The read you have on a person because you have sat across the table from them. An AI can mimic these, but only to a point, and only by mapping statistically probable relationships between words. It does not understand anything. Not really.

That fact matters. We call it “generative AI,” but that’s a lie. AI is iterative. God has granted one species on earth the power to be truly generative: humans. There are things about being human that we will simply never compress into a bit-based format for a model to run through its trillions of weights, no matter how vast and inscrutable the calculation becomes.

So, what is the point of learning?

The point of learning is to help a human become more fully who God made them to be: an image bearer, a creature gifted by God with the capacity to create and to steward.

Take a science classroom. When we teach our kids biology, of course we want them to understand how and why humans have classified the creatures of the earth the way we have. But we should want more than that. We should want them to love those creatures, and to feel the weight of the role God gave us in caring for them. Learning science, done well, motivates us to grow in knowledge, discernment, gratitude, and an ordered sense of duty. Learning is an act of love. It opens the mind. It makes room to wonder. It lights the imagination in ways we can’t fully predict, and out of that lit imagination come new ways of loving God, loving our neighbor, and stewarding creation.

The real world applications of this are manifest in countless ways. It’s manifested when we capture energy to heat our homes and power our devices more sustainably. It’s manifested when we innovate ways to reduce food waste. It’s manifested when we use water more wisely while still beautifying our cities and refreshing those who are thirsty. It’s manifested when we invent new fertilizers and farming techniques that coax higher yields out of the same acre of farmland. To get there, we will reach for tools like AI. In fact AI will become something like a superpower for us, not the thing that generates or holds the vision, but the tool that supercharges us as we move from vision to execution.

Which brings me back to language. In a world where useful work increasingly flows out of natural language, the people who will drive humanity forward will be the ones who have become masters of it: people who can capture a vision of the future and form it into words, words that summon goods and services shaped by a rightly ordered love.

That is the point of learning: to help us love what we ought, then to order that love, and finally to harness the generative spark born of love into language that will metaphorically and literally shape the world of tomorrow. Any future in which AI genuinely advances humankind, and serves a Godly vision of stewardship over creation, depends on learning.

Dustin Steeve is a serial entrepreneur, the founder of Inaltum.agency, and the creator of Magnetic Brand. He’s a graduate of Biola University and Torrey Honors College, serves on the governing board for Hildegard College, resides in north San Diego County with his wife and two children, and attends New Life Presbyterian Church.