Four Ways to Think About AI in Schools

The idea of AI in schools is inspiring either alarm bells or excitement, depending on who you're reading or following on social. But since it's already here, I invite you to consider four frames for considering it: Helping Teachers Reduce Friction, AI with Training Wheels, Naming What Students Are Already Doing, and Conversations, Not Lessons.

I'm no lackey for AI, and I've been deeply skeptical of its rapid advance on schools. But I also recognize that it's here and educators need to be prepared to think and discuss with students and, in some cases, engage with the technology itself.

Transcript

SPEAKER_00 0:11

Welcome to What's the Big Idea? I'm your host, Dan Carney. If you've been paying attention to the conversation around AI in education, you've probably noticed that a lot of it sounds like it does in other parts of life, either a fire alarm or a press release. On one side, there's genuine anxiety about kids outsourcing their thinking, about the erosion of authentic work, about a technology moving faster than schools can respond. On the other side, there's breathless enthusiasm about personalized learning and unlocking potential and the future of everything. And somewhere in the middle, teachers and administrators are just trying to figure out what to do on Monday morning. But this isn't hypothetical. AI is already a major force in the lives of young people. They're using it at home, on their phones, sometimes right there in the classroom. So the question isn't whether AI is coming to school, it's already here. The question is whether schools are going to engage it thoughtfully or keep treating it like a problem to be managed from a distance. What I want to do today is get practical. Based on things I've been reading, watching, exploring, and experiencing firsthand, I'd like to offer four concrete frames for how AI is showing up in schools right now, and how teachers and administrators might think about each one as we head into next school year. And honestly, this is probably what AI was made for. According to Edweek Research Center, the percentage of teachers using AI tools in K-12 classrooms nearly doubled in just two years, from 34% in 2023 to 61% in 2025. So it's happening and it's accelerating. The ways teachers are using it are pretty varied. Lesson planning and curriculum design are the obvious ones. There's also adapting materials for different reading levels, translating messages into home languages, and even using AI as a kind of reflective thought partner in putting a lesson plan and asking it to anticipate where students might struggle. One math teacher I read about uploads activities he's used before and asks AI to suggest new ways to make them more engaging. He gets five or six options back, picks what makes sense, and builds from there. And that's a really healthy model. AI is a brainstorming partner, not a replacement. And I'd encourage teachers to keep exploring this. The one area I'd pump the brakes on, though, is using AI to write feedback for students or drafting family communications. In short, be careful of using AI to create outward-facing items. Like in the case of feedback, kids are savvy. They can tell. I actually came across a teacher who wrote about exactly this. His students recognized that the feedback they were receiving was AI generated and it felt inauthentic. And honestly, I think they're right to feel that way. Feedback is one of the most human things a teacher creates and provides for students. It's one worth protecting. This is when schools use subscription platforms, tools like Magic School, or maybe you're a title school. And these tools give teachers a structured boundaried way to bring AI into the classroom. Students aren't really using AI here. They're using a tightly prescribed version of it inside of a framework that the teacher controls. And this is pretty widespread. Magic School alone has contracts in Atlanta, Denver, New York City, Seattle, among others. And I've heard some pretty solid examples of this. Small bursts of AI to reinforce a concept or practice a skill. I mean, I do think some of the features of these programs are a bit hokey and potentially problematic. You've probably heard of these chatbots impersonating historical figures so the students can have a conversation with them. I'm not a huge fan of that. But I think that there's real value when these programs are used in moderation with intention inside guardrails. A tightly boundaried version of AI deployed to reinforce skills or spark curiosity. I think there's some real value. And connecting back to my first frame, programs like Magic School also provide lots of support for teachers for tasks like differentiating assignments. Frame three, naming what students are already doing. Here's the reality: kids are using AI. A recent survey from NPR found that 40% of K-12 teachers say their students are using AI in the classroom at least once a week, and that's just what's visible to teachers. Outside of school, they're using chat, Claude, Gemini, all of it. So the third frame is the approach that acknowledges this directly and draws a clear ethical line. Cheating is cheating. This approach says, yes, you can use AI, but you know what it means to submit work that isn't yours. And I actually believe students do know. I've had enough conversations with kids to be convinced that they understand the ethical complexities here. They get it. So this frame leans into that. Let's talk about it openly, lay out what responsible use looks like, and trust students to navigate it. The catch is that this approach requires a teacher who really understands how AI works and can recognize AI-generated work. It's not for everyone. But for savvy teachers who are ready to engage with AI on its own terms, this might be the most honest and real-world approach out there. And if if you're someone, if you're listening and thinking, yeah, but kids will just cheat with AI, students have always cheated. They always cheat for the same reasons. Perceived lack of skill, lack of time, and I think in a lot of cases, just lack of interest. And AI is just the latest tool that allows students to cheat, but it didn't invent cheating. So approaching AI, just like we'd approach copy and paste from Google or copying off a friend's paper, i.e., you can't do it, is consistent with how schools have handled cheating. Frame four. Conversations, not lessons. This dovetails off my last point, and it's one that I care about most. And it's actually the simplest. Have conversations. I'm a little skeptical of the idea that we need to design formal AI literacy lessons, partly because, frankly, kids already know how to use the stuff. Building a lesson around that can feel a bit condescending, but there's a version of this I find really compelling: conversations with students. In my own school's weekly Town Hall assembly this past year, we had such conversations. One was prompted by student questions about how much their teachers were using AI. And it led to an awesome, honest, and insightful conversation that really gave me a lot of hope. And I came across a story about a high school that created an elective called Driver's Education for AI. So it wasn't a how-to or an AI literacy, but the whole premise was getting students to think of themselves as drivers of the technology, not passengers. Are you actively directing what AI does for you, or are you just letting it happen to you? One student had this great moment of reckoning when she realized that her Spotify AI DJ was basically making all her musical choices for her. It's a small example, but it cracked something open in her thinking. And the students in that class were naive. They were thinking about creativity, authorship, what it means to rely on something. One student noticed his dad had started texting him in long formal paragraphs, clearly ChatGPT, and told him, I want to hear what you have to say, not what AI says. So these kind of sophistic conversations can happen. They can happen organically, and they can happen in settings where we provide the structure and space. This connects to work I've I've drawn on from an organization called Civics of Technology, which does exactly this. It helps people examine and question their relationship to technology. Their work is really excellent. A couple years back, I had a great conversation with Dan Krutka from the organization right here on the podcast. Recommend going back and checking that out because the way he frames how we interact with technology has really stuck with me. What I don't think kids have had enough space for is the deeper question. What does it mean to be the driver of your own learning when AI can do so much of it for you? What does it mean to be an authentic writer, authentic artist in this moment? Not here's how AI works or here's what AI-generated writing looks like. Kids know all that. What they need is space to think philosophically about what it means to live and learn alongside these tools. I am very skeptical of how this technology impacts the way we think, the way we create.