Why I Think Artificial Intelligence Could - But Probably Won't - Revolutionize Schools

Why I think Artificial Intelligence Could - But Probably Won't - Revolutionize Schools

4/19/2026

Artificial Intelligence is here.

It has changed everything, so I am told.

Except, of course, for the things it hasn't yet changed; but those things, I have been assured, will also be changed in short order.

Schools are one of those apparently few remaining institutions that - aside from improving kids' cheating capacity and efficiency - haven't yet been reshaped the way that the rest of the world allegedly has been.

I think, however, that they could be soon. There are real reasons to be extremely optimistic that A.I. could truly revolutionize school and learning.

But.

I also think that it probably won't.

I'm open to being wrong, and I hope that I am. I've thought we were due for a revolution for quite some time, and I want to get to be part of one before I retire! I really do! However, I've seen "the next big thing" rise and fall in a predictable pattern more times than I can count, and I'm guessing that this one will do the same.

From where I sit now, I can see at least three reasons why A.I. could revolutionize schools and learning, but that it probably won't.

1. ARTIFICIAL Intelligence can Fill the Biggest Need In Schools

During the pandemic, I wrote a piece called Students Who Fall Behind Rarely Catch Up... Why Not? If you want to know the biggest reason A.I. could be the educational prince that was promised, please take a few minutes to read it.

The biggest reason, I posit in that piece, that students who fall behind rarely catch up is this:

When it comes to discussing struggling students, we have a tendency to make two assumptions. The first assumption is that their struggles are the result of their behaviors and habits in the classroom. Do they study? Do their work? Pay attention? Complete homework? Take notes? Etc. The second assumption is that if - at any point in time - the student gets his or her habits and behaviors in order, they’ll be successful in the classroom.

Both of these assumptions are usually wrong.

Regarding the first assumption, it turns out that if a student is struggling to learn something (in a good learning environment), the reason is usually a skill gap - he or she is missing preceding skills that new learning is built upon. A student who gets a C in reading in 2nd grade, say, is passed up to 3rd grade despite missing 25% of the prior year’s reading skills, upon which new learning will build. Learning gaps accumulate, grow, and cause major issues down the road.

Regarding the second assumption, getting a student’s work habits in order won’t do anything to fill in skill gaps. It might help to limit the gaps’ growth in the future, but becoming a strong and organized student in 7th grade, say, won’t do anything to fill in the skill gaps that came before. It might prevent them from growing any larger, but they're still there, and they still matter.

Finding a way to fill in skill gaps is the Holy Grail of improving schools. Nationally, why are two-thirds of students below proficient in math and reading? Primarily because they didn't learn a lot of needed skills in prior grades that the content they're being taught in their classes today depends on.

Most schools do not have a good way of dealing with this. They have quite a few bad ways of dealing with this, but obviously nothing that it is actually working.

"How exactly do we go about filling this gap?"

A.I. has the capacity to change this. If it can determine what isolated skills each of my 6th graders didn't learn in elementary school and help fill those gaps in a differentiated way for each student - a task A.I. is well suited to handle - that would change everything.

Truly.

Fill those gaps, and the kids won't have any problem learning 6th grade math.

Fill those gaps, and the kids will catch up.

You may be thinking "sure, but that's only helpful to the students who need to catch up. What about everybody else?" Some things, however, to keep in mind that are not often advertised:

Most students are in the "need to catch up" camp. About two-thirds of them, data would indicate.
Good grades and high graduation rates are not indications that students are on track. We tend to assume that things are largely going well educationally because so many students earn good grades and the graduation rate is very high, but grades and actual educational attainment are largely decoupled at this point. Famously, one of the nation's most selective public universities disclosed that a huge percentage of its students aren't able to do middle school math despite having the top-tier grades in their high school math classes needed to get admitted.
Even the students who don't need catching up attend schools where many - even most - of the students do, and the struggling masses are driving most of the decision-making in these schools, not the successful ones. The on-track students would be learning more and would have more opportunities if there were more of them and if schools didn't have to focus so sharply on the students that the system is failing.

A.I., again, is uniquely suited to fix this. It can diagnose the kids who are behind, tutor them individually to close their skill gaps, catch them up, and make them ready to thrive in on-grade-level classes, opening the floodgates for everyone to flourish in classes that aren't restrained by missing foundational learning.

The first reason I feel like A.I. could revolutionize schools and learning is that it can serve the one function in schools that is needed most and that nothing else currently can.

But

But... the fact that A.I. can serve the most glaring need in the educational system doesn't mean that it will.

Why not?

Motivation, frankly.

One of the things that I think the general public underestimates is the extent to which teachers have to educate kids against their will.

I find that, when people imagine kids in schools, they imagine kids doing everything they can to be as successful in school as possible. They imagine schools full of kids working, trying, paying attention, pleading for extra tutoring sessions, and doing everything in their power to earn a fabulous education, but the school and the system are failing them.

The reality is, however, that at every school, teachers are dragging a lot of their students kicking and screaming toward educational attainments that they're actively working against.

The fact that A.I. could provide individualized, gap-filling tutoring to students who badly need it only matters if we can get students to engage with A.I. in that way, which I suspect we largely cannot. I surmise that the opportunity to go back and learn 3rd grade math skills that my 6th graders never learned initially would be met with the same response that most other learning opportunities are:

"I don't want to do that."

My reaction when my teacher suggests that I take the time to learn skills I didn't learn in past years.

In a classroom, I have certain techniques I can employ, levers I can pull, and a force of will I can impose to make my students learn despite their best efforts to avoid it.

A.I., on the other hand, can only tutor the willing.

But my son has taught himself so many things on YouTube! Isn't this proof that kids can learn math and science and reading independently given the right tools, too? It is proof that they can, yes. But not proof that they will. The fact that my friend's son has taught himself leather crafting on YouTube (yes, really!) doesn't mean that he's also teaching himself math using available tools, even though he could be. He is motivated to teach himself leather crafting. He is not so motivated to teach himself math in this way.

A.I. could genuinely solve education's biggest problem.

But I just don't think that we'll be able to get enough kids who need it to learn from it.

and

And... educational technology already seems to be doing more harm than good in the first place. To say, in this particular day and age, that a new computer-based tool is going to rescue us when existing computer-based tools are contributing heavily to the crisis to begin with seems like a stretch. It is reasonable to assume that doubling down on a technological solution is doubling down on the causes of the problem, too. It feels a little like saying that eating too much cake is causing me to gain weight, so I'm going to fix it by eating pie.

Kids' mindset about what digital devices are for isn't compatible with educating them right now. Not only is A.I. ill-equipped to change that, but it is likely to make it worse.

2. Artificial Intelligence Drastically Improves Human Generative Capability IN Every Field

The second reason I think A.I. courld revolutionize schools and learning is that... it is revolutionizing every other organization though its capacity to expand human generative capability. In the hands of a capable mind, A.I. is able to do countless, high-level, generative tasks much faster than, and in some cases better than, that mind could do them on its own.

A.I. can make everything from slide decks to computer programs to architectural plans to medical diagnoses for us. A good journalist can use A.I. to generate his or her primary output faster and better than ever. So can a good sitcom writer. So can a good accountant.

It can unleash and expand human intelligence in previously unthinkable ways, and it is doing so in nearly every profession and sector and organization.

It stands to reason that it will unleash a world of possibility in school, too, right? It seems to be only a matter of time before the experts in every field - ours included - transform what they do, how they do it, and their creative and generative capacity.

The second reason I think that A.I. could transform education is that... of course it will. It expands human capacity everywhere else, so it will surely expand ours as well.

But

But... extending human intelligence isn't the same as building it.

If I'm a computer programmer, I'm adding artificial intelligence to my own intelligence to increase my productive capacity. Same if I'm an ad-writer. Or a data analyst. In these uses, A.I. isn't making me an inherently better programmer or ad-writer or data analyst, it is a supplement to make me a more productive one. It isn't developing my abilities, just my productive output.

As educators, our output is the creation of human intelligence itself, in a sense.

It makes sense that a computer program that could add A.I's generative capacity to a human's would be better.

But what if your "output" is the human ability side of that equation?

We aren't trying to extend human intelligence, we are trying to create it. I'm not sure the formula works the same way in that case.

Extending human intelligence is one thing. Building it may be another thing completely.

And

And... A.I. is also uniquely suited to "get in its own way" in the creation of human intelligence because of its capacity to do for us the very thing that we need to do ourselves in that process.

When students use A.I. to do assignments for them, we call it "cheating." We are used to thinking of that in a moral sense - they are cheating the system. They are also, however, cheating our attempts to build their intelligence.

Reading a novel or writing a paper or doing math homework is all done in the name of building students' intelligence. It doesn't serve that purpose if A.I. does it for them.

A.I. could become a powerful tool for student learning.

But it is already a powerful tool for avoiding learning. It is already in its own way, in that respect.

In this section, I've been trying to say that A.I. expands human intelligence, but it doesn't create it. Now I want to add on the fact that it actually reduces it the way it is most often used by learners.

It just may not be the same tool in the education sector that it is in others based on the nature of the output we're trying to generate.

3. Artificial Intelligence Can Make Teachers More productive

The third reason I think that artificial intelligence could revolutionize schools and learning is that, even if I'm skeptical that it can unleash the potential of kids, it could certainly unleash the potential of teachers.

I think most people are aware that teachers have to do a ton of work beyond teaching, namely lesson planning, communicating with parents, and grading/feedback of student work. I think most people are not aware that they get next to no time to do any of that work during their contract hours. As a 5th grade teacher, on a great week, I might get three 45-minute blocks of time to do all of that during the school day. One or two blocks a week is more common. Zero isn't unusual.

There are three ways that teachers compensate for this lack of time to do core professional tasks:

They do that work outside of school hours
They do that work instead of teaching the kids at times (silent reading time everybody!)
They don't do that work at all

It is a genuinely brutal choice to make. None of the options are good ones. They all have their trade-offs.

Enter A.I.

Planning, communicating, and grading, since I've mentioned those, are well within A.I.'s capacity to drastically speed up. Just think of the possibilities with those three tasks accelerated, automated, and removed from the time sink! The teachers choosing option #1 above could get their personal time back, the students whose teachers are choosing option #2 could get more instructional time, and the students whose teachers are choosing option #3 could have an educational experience where those things are happening.

Everybody wins!

The third reason I think that artificial intelligence could revolutionize schools and learning is that it could unleash teachers' in the way that it has unleashed other professionals. It could reduce the time needed for key professional tasks, freeing them up to spend more time on their most important work.

But

But... speed, in all things, usually comes at the expense of quality.

A.I. can make teachers faster.

It can't make them better.

And there is a fair chance it will make them worse.

One of the primary fears we are hearing about artificial intelligence is that, while it is a powerful tool in the hands of someone who already has expertise in a certain generative field, it also makes novices in that field far less likely to ever generate expertise.

An expert writer, given A.I., is an even better writer. He or she knows when to use the tool to generate quick, low-quality text, and he or she has the expertise needed to make that low-quality text into high-level text.

A novice writer who uses A.I. to write for them has no such judgment, and he or she is far less likely to ever develop it since they are offloading expertise-building opportunities onto A.I.

It's like a baseball player hitting off of a tee. Professional players use tees to fine-tune the final, minute details of their swings. In the hands of a professional, a tee makes a great hitter even better. Novices, however, given a tee and told that it will make them a better hitter are not going to use it in a way that makes them better. They'll just become hitters that are good at hitting off of a tee.

This is what I fear for teachers as well. I've used A.I. to write lesson plans. They stink. But I've planned thousands of lessons on my own, so I have the judgment to a) tell that they stink in the first place, and b) decide if it is worth starting with them and turning them into something better or throwing them out and planning something myself.

I only developed that expertise, however, by planning thousands of lesson from scratch myself. Not only could A.I. not have helped me develop that expertise, it likely would have prevented me from ever doing what it took to develop it in the first place.

Speed is wonderful, but we have to put that speed in the right hands and apply it to the right tasks.

"This lesson stinks. I'll bet he farmed it out to A.I. so he could go home earlier."

And

And... any efficiency gains made by teachers are unlikely to be used to revolutionize school and learning, anyways. Quickly generating bad lesson plans is no revolution. Having a computer grade students' work so that I never get a feeling for their understanding or misconceptions isn't either. Having A.I. manage my parent emails isn't either.

It'll get me home sooner.

It might revolutionize my personal life.

But it won't revolutionize the kids' learning experience.

Conclusion

There's so much to be excited about when it comes to A.I. and learning. I genuinely think it has the potential to revolutionize what happens in schools.

I also think that it probably won't.

Thirty years ago, email had the potential to streamline communication, reduce meetings, and make the office experience so much better.

It didn't.

Fifteen years ago, social media had the potential to make us more connected, better informed, and happier.

It didn't.

A tool's potential doesn't always foretell its ultimate use.

There's a near-future out there where A.I. unlocks learning for a better academic future. There's another one out there where it has the complete opposite effect.

I'm hoping for the first.

I'm preparing for the second.