AI: What Kind of Thinkers Are We Building?

Artificial intelligence has quickly become the newest source of anxiety in education. At the same time, books like Jonathan Haidt's The Anxious Generation have sparked a national conversation about technology, social media, and children's mental health. States are adopting phone bans. Districts are limiting device use. Parents are rethinking screen time at home.

These are important steps. I've seen firsthand how removing phones from classrooms changes the culture almost overnight. Students are more present. Teachers aren't competing with devices. Everyone is far less distracted.

But I worry that we're beginning to mistake removing distractions for rebuilding attention. Those aren't the same thing and nowhere is that more true than in the earliest grades.

Older students who grew up with fewer devices, at least, had some years to build a baseline of sustained attention before screens entered the picture. Many of today's youngest students don't have that baseline yet. They're forming their capacity to focus for the first time in an environment full of tablets, videos, and instant response. A phone ban in a high school hallway removes a distraction. In a kindergarten or third-grade classroom, the question isn't just removing a distraction, it's whether we're giving young children the chance to build sustained attention at all, before habits solidify around constant switching.

Phone bans don't rebuild the habits we've already lost or build the habits that haven't formed yet. Those habits require intentional teaching. 

We're Debating the Wrong Question

A lot of our conversations about AI in schools center on whether students should use it, what our policies should say, or how to catch AI-generated work. Those questions matter, but they skip over a more fundamental question, one that applies from kindergarten on up: what kind of thinkers are our students becoming, and are we giving them the chance to build the attention that thinking depends on?

In a recent conversation with literacy educators Mary Ehrenworth and Philip Seyfried for The Invitation, Philip put another question on the table: rather than continuing to debate whether AI belongs in learning, perhaps we should be thinking more carefully about when and how AI enters the learning process.

That distinction matters. If AI enters before students have had a chance to notice, question, generate ideas, or struggle with something themselves, it can remove the very thinking we're trying to develop. If it enters after students have knowledge, questions, or ideas of their own, it can do something quite different: extend the thinking rather than replace it.

That question didn't start with ChatGPT. Researchers including Maryanne Wolf, Gloria Mark, and Nicholas Carr had already documented a shift in sustained attention and deep reading well before AI arrived — a shift that shows up earliest in the youngest readers, who are still building the neural pathways for focus in the first place. AI didn't create that pattern, it exacerbated it, because it's remarkably good at removing the discomfort of not knowing.

None of this happened by design. Computers weren't built to weaken attention, and AI wasn't built to replace learning. But every tool changes the habits of the people who use it. When the easiest path is always available — when a device will simply show you the answer, sound out the word, or finish the thought — it's tempting to take it.

Attention Is the Skill Underneath All the Other Skills

This is why attention, not AI use itself, is the dangerous skill to lose. Every literacy skill rests on it: from a kindergartner noticing that a word starts with the same sound as their name, to a fifth grader catching that a character's actions don't match what they said earlier in the story.

Attention isn't just staying in your seat or staying on task. It's the ability to notice what deserves your thinking: a pattern in the letters, a surprising choice a character makes, an idea that doesn't quite fit. If young students rarely get to practice sustained attention, we shouldn't be surprised when sustained thinking is hard for them later. This isn't a problem that begins in high school. The habits that support sustained thinking are built much earlier. 

Here's the irony: the students who will benefit most from AI later on are the ones who build strong attention and reading habits now, years before they ever touch the tool themselves. A student who has practiced noticing, questioning, and sitting with uncertainty will know what a good answer looks like when AI eventually enters their world. A student who hasn't will simply accept whatever they're given, and that gap is being set now, not in tenth grade.

AI Rewards Precision — and Precision Is a Literacy Skill

AI rewards precision, but using it well also requires students to bring something of their own to the interaction. Students who can recognize ambiguity, define exactly what they mean, and ask a good follow-up question will get more out of it than students who accept the first vague answer they receive. 

Those aren't tech skills. They're the same skills we've always taught — and they're built years before a student ever opens an AI tool, through the everyday work of noticing, questioning, and sitting with a hard sentence instead of rushing past it.

There's an equity issue here, too. Students don't arrive at AI with equal access to background knowledge, vocabulary, strong reading and writing instruction, or adults who can help them learn to use these tools well. If AI increasingly rewards the students who already know how to ask precise questions, evaluate an answer, and push back when something doesn't make sense, it has the potential to widen gaps rather than close them.

Mary Ehrenworth captured part of this idea simply in our conversation: “You always want to come to AI with something.” That “something” matters. Students need knowledge, ideas, questions, and language of their own if they're going to do more than simply accept what AI gives them.

What This Means for the Classroom

If we want students to develop that kind of attention, we have to build room for it deliberately, and starting early. Not every lesson needs to be fast, gamified, or run through a screen. Young students especially need time to sound through a hard word instead of having it solved for them, retell a story in their own words, sit with a picture book long enough to notice something on a second look, and wrestle with a question that doesn't resolve immediately. That's not time taken away from learning. It is the learning and it's also the foundation that will determine whether, years from now, AI is something a student can use well or something that quietly does the thinking for them.

Emerging research gives us another reason to pay attention to that sequence. In our conversation, Philip pointed to an MIT study that compared people writing with an AI chatbot, a search engine, or no digital assistance. Researchers found differences in cognitive engagement, recall, and ownership of the writing. But one part of the study was especially interesting: when participants who had initially worked without AI were later given access to it, they appeared to engage with the tool differently than those who had begun there.

Mary and Philip use a simple phrase for the principle behind this: brain first.

The distinction is important. Brain first doesn't mean students should do everything without assistance. It means they need opportunities to build knowledge, form ideas, try, struggle, and create something of their own before we hand over parts of that work to a tool. As Philip said during our conversation, “AI can really steal the learning from us if we just start with it.”

The Choice in Front of Us

AI isn't going away. Neither are computers, digital tools, or online learning. What we can still shape is how we prepare students to live inside that reality.

We can keep reacting to whatever tool shows up next — banning it, chasing it, writing policy around it. Or we can recommit to the habits that have always mattered, starting in the earliest grades: paying attention, reading closely, communicating clearly, asking good questions, revising our thinking. Those habits mattered before AI. And they may matter even more in a world where students have to decide when to trust an answer, when to question it, and when to keep thinking for themselves.   

Continue the Conversation

When should AI enter the learning? is the focus of an upcoming conversation on The Invitation with literacy educators and authors Mary Ehrenworth and Philip Seyfried. We talk about brain-first learning, cognitive load and ownership, smarter screen time, and what it might mean to prepare students to use AI without allowing it to do the learning for them. 

Suggested Reading

• Naomi S. Baron. How We Read Now: Strategic Choices for Print, Screen, and Audio.

• Naomi S. Baron. Who Wrote This? How AI and the Lure of Efficiency Threaten Human Writing.

• Nicholas Carr. The Shallows: What the Internet Is Doing to Our Brains.

• Mary Ehrenworth and Philip Seyfried. AI and Reading Instruction: Digital Partners for Deepening Comprehension and Discussion in K–12 Classrooms.

• Johann Hari. Stolen Focus: Why You Can’t Pay Attention—and How to Think Deeply Again.

• Ethan Mollick. Co-Intelligence: Living and Working with AI.

• Ethan Mollick. Co-Existence: The Next Phase of AI. Coming October 2026. 

• Maryanne Wolf. Reader, Come Home: The Reading Brain in a Digital World.