Finding the Sensible Middle: AI Literacy, Cognitive Offloading, and Student Voice in the Classroom
In his junior year English class, Matt Miller sat watching a lesson drag and quietly redesigned it in his head. This would be so much better if she did this. This would be better if she did that. He never said anything to his teacher, but he did go home and tell his mom about it.
Years later, after he'd switched out of a journalism major and into teaching, she brought that afternoon back up, reminding him, “You've been thinking like a teacher since you were a student all the way back in high school.”
The lecture that emptied the room
First, though, he had to become the teacher.
Matt was wrapping up a lecture on reflexive verbs, and by his own account it was an absolutely riveting one, described in detail, fantastically, he's sure. His Spanish students had gone glassy. With five minutes left he did what he thought was the generous thing and told them to start their homework. When nobody moved and no notebooks opened, he walked to the front of the room. "You're not working on your homework, I see. That must mean that you got it all done and that I didn't give you enough. So let me give you some more homework." He owns it now without flinching: "Yes, I went there. Unfortunately, I went there."
The bell rang and every student in that room sprinted for the door. Matt stood alone in a space he had just emptied, holding one thought: babies learn how to speak a language without verb conjugation drills and long vocabulary lists. There has to be a better way to do this.
What Ditch That Textbook actually means
The name gets misread constantly, so Matt corrects it himself. "Contrary to popular belief, Ditch That Textbook wasn't created because I hate textbooks or I think that they're evil or terrible or anything," he told us. "I was just like, if we can do better than them, then we should."
The blog came before the mission. Going from journalism major to classroom educator, he already liked to write, and once he had his sea legs in the classroom he started publishing what he was trying, including the parts that flopped. Alongside that, he kept hunting for better things to do than the questions at the end of the chapter. One at a time, little by little. As he added more of them he reached for the textbook less, until he looked at the shelf one day and wondered why the things were still there.
That test, if we can do better than this, then we should, is also the test he applies to AI.
Why he's still in the room
Matt taught Spanish full time for 11 years, then left in 2015 to do workshops and keynotes, which he did for eight years. Seven books later, he's back in a classroom teaching one period of Spanish a day at the high school his own kids attend, then turning back into Ditch That Textbook Matt for the rest of the day.
"I'm going through the creative struggle of trying to engage them every single day," he said. He's in the hallway. He's with colleagues. Problems confront him and he has to solve them on the spot, which means he never runs out of material, because the classroom hands it to him.
The roller coaster
That's his description of AI in the classroom. "You have the highs and the lows, you know, all of a sudden things are going great and then you go down a terrifying hill and you're screaming."
The promise, for him, is arithmetic. He has 31 kids in one Spanish 1 class and one of himself. He's good at whole-group direct instruction and he'll tell you so. He's equally clear that 31 students in that arrangement are not all getting meaningful, personalized feedback, and that they can't. Immediate feedback has been one of the real barriers to the results teachers want for decades. If safe, reliable tools start closing that gap, that isn't a small thing. And then there’s the tedium. Plenty of what fills a teacher's day adds nothing a human uniquely brings, and it's part of why good people keep leaving a profession that can't afford to lose them.
The concerns are also specific. Cognitive offloading, and students who ask before they think. Hallucinations, bias, and data privacy. The energy, water, and resources that data centers consume. And the version of this that families are actually voicing: Matt presented recently to parents in a district that had surveyed them about their beliefs on AI, and in a lot of different words, the answers came back the same. We're afraid it's going to make our kids stupid. They won't be able to solve problems. They won't be able to think critically.
"There's no easy answer," he said, "and it's complicated for people who understand it, let alone for teachers who are just getting to know it."
Why this one feels different
Two things make AI unlike the technology waves teachers have absorbed before.
The first is order of arrival. Jody described the pattern every veteran teacher knows: you get the tool first, you get comfortable with it, then you introduce it to students. That’s by and large not what happened with generative AI. A lot of students had it before their teachers did, or at the exact same moment, so the usual runway for getting comfortable never existed. Ryan Lufkin heard from educators recently that some students say they trust AI more than they trust their teacher. Others won't touch the tools their school provides, because they assume they're being watched or recorded.
The second is that the thing itself doesn't behave like software. "Traditional software has a repeatable process where you can expect the same thing out of it over and over again," Matt said. "And if you ask an AI model the same question 10 times, you might get 10 different answers." So it’s no wonder plenty of teachers are struggling with navigating the “idea” use and impact of AI.
The sensible middle
The loudest voices sit at the edges: Technology is the savior and you have to use all of it, or technology is the devil and you should never touch it.
"I feel like the ones of us that are in schools on a regular basis, we live in the sensible middle," he said, "where we go, if it'll help me, I'll use it. But if it doesn't, I won't." The edges are exhausting precisely because they contradict each other and both point at teachers. As he put it, it's easy to wag a finger and ask why you aren't keeping up with the times, then turn around and ask why in the world you're using something so harmful and dangerous.
He has no patience for the nostalgia either. When people propose turning the screens off and going back to paper, pencil, and textbooks, his answer is that we're looking at it through rose-colored glasses.A textbook isn't responsive. A textbook gets outdated. A textbook is one shot to get it right at one moment in time, and if it doesn't engage a kid, that kid can simply switch off.
What actually holds teachers back
Matt has worked with hundreds of schools and districts, so we asked what stops teachers from making the shifts he made. The first answer is the obvious one: time.
We take professional development time and pack it full of presentations about things, instead of giving teachers time to plan. Add the obligations that eat prep periods and the hours before and after school, and the room to actually build something new never appears.
His second answer is the one nobody talks about: comfort, noting he knows plenty of teachers who say, "I've done it this way for a while and it's predictable and I know how it's gonna be." And plenty of the teachers using settled methods get genuinely good results. His point is about ceilings, not competence. "We're sort of fooling ourselves if we think, okay, I've arrived. This is it. This is the curriculum that I will teach until I retire." Then he turned the harder question back on us: do we have teachers who really want to take it to the next level? He left it open.
Ask your students
When asked for one concrete thing a teacher could try tomorrow without needing anyone's permission, Matt answered, “Ask your students” with the conviction of the kid in that English class who was never asked.
He lays out the kind of conversation (think: not a survey) in his recent book, AI Literacy in Any Class.
- Start wide with general concerns about AI.
- Narrow to specific fields, medicine, law, and creative work, and press on what's unfair, what's acceptable, and how they can tell the difference.
- Narrow again to schools in general. Should it be allowed, in what ways, and why?
Only after all of that do you get to how this class is going to handle it. Open with "how should we use this in Spanish class" and students answer to protect their own access. They're calculating which answer keeps the door open, which is human nature, but also useless data. Build down from the wide questions instead, and by the time you reach your own room, they've put a lot on the record that they now have to reconcile. They see the implications of the whole thing, and the norms belong to them rather than to you.
Teachers have been taught for decades to build classroom norms with students rather than for them. The reason it matters more here is that the alternative isn't a rule students resent. It's a rule they route around.
Everything that was going to replace us
Matt has a list. YouTube and Khan Academy were going to make teachers obsolete. So were Wikipedia, encyclopedias, and printed books. He can take it back to the printing press and mentioned ancient Greek philosophers worrying that written text would ruin memory. Printed books are now the hallmark of a good education.
So maybe AI won’t replace teachers, like all tech before it didn’t, but it’s important to think about the evolution of modern teaching and learning. Start at the one-room schoolhouse, where the curriculum lives in the teacher's brain plus whatever handful of books they have, and the teacher is the source of all information. Then printed books, then the internet. Each step decentralizes knowledge away from the teacher and toward the student. AI is another couple of big leaps in that same direction.
So the question he thinks every educator should be asking about the next decade is what remains uniquely ours. Part of his answer is about relationships and experience, the adult in the room with real adult life behind them. Part of it is stranger and better than that. "Our brains are still wired to engage with other humans because, you know, as humans, we are we're unpredictable," he said. "We're unique and complex and messy and unhinged sometimes. And it is fascinating and entertaining."
The rest of his answer is a list of places rather than features. The hallway. The lunchroom. The principal's office, when you didn't make great decisions. Indiana is a year into a statewide phone ban, and what they're noticing at Matt's school is that kids are talking to each other, and that there's suddenly serious demand for decks of cards so they can play Euchre at lunch.
"Let's center that and preserve it," Matt said. "It's sacred. We've got to protect it at all costs." He ditched textbooks because they weren't engaging, and AI can very much be the same way, because it isn't real human to human connection either.
Every episode of the Educast 3000 show opens by asking the guest for a favorite learning moment. Not one guest has ever named a piece of technology. Every answer has been about a person.
Listen to the full conversation with Matt Miller on our latest episode of Educast 3000, “Planet of the Apps: Every Technology That Was Going to Replace Teachers, and Why None Did”.
