How should we bring AI into the classroom?

By: Carljoe Javier - 55 seconds ago

AI kind of just showed up in the classroom and decided it wouldn’t leave.

No, that’s not entirely true. Kids showed up with it first, and then teachers had to figure out how to adapt.

As a teacher who has been in the classroom for almost twenty years, I have to admit that it is one of the most disruptive things I’ve ever had to face. I can’t talk to another teacher these days without it coming up, and in fact so many of my conversations with fellow teachers revolve around student use, how we are meant to cope, what we can do, and so on. My own position continues to shift as it is challenged, with new ideas, new studies about its impact, new orders and directives are released, and new crises that continue to emerge.

This will be the first in a series of pieces about AI and teaching, and I’ll start off with laying some groundwork on the narratives, some of the policies, and how we can begin to think about education in a world filled with AI.

I think one of the prevailing narratives and one of the larger context pieces that we need to consider is that we as a country are dealing with an education and literacy crisis. The Edcom 2 reports point to challenges in foundational skills. Further, the PSA’s 2024 FLEMMS survey found that approximately 18 million Filipinos who completed basic education remain functionally illiterate.

When you look at this context then you know that with or without AI, we need to be rethinking education.

Starting from my own education

When I was in school, I felt like this just didn’t move at the speed that I needed. In some classes I was left behind, in others, the pace was so slow that I was bored, and being bored I would be distracted which might be cause for getting in trouble. In any case, this sort of mismatch and my eventually doing well when I was able to set my own pace is what makes me open to AI as a tool for improving education outcomes.

One of the big opportunities AI offers is customization. Not everyone learns at the same pace. But what if teachers, helped by AI assessments and tracking, could craft better lessons? They could provide those falling behind with more guided support, and then give more challenging work to those who are ready.

Even better, this empowers students who want to do more learning to choose to take on more challenging things, while at the same time allows students who are struggling to ask for more guidance or help. AI tutors show promising results; for example Khan Academy’s Khanmigo recorded a 6.1% improvement in whether students correctly answered the next problem after an AI tutoring interaction. That’s small, but it’s not nothing. Considering that not just regular students, but students who might have different kinds of learning needs might be helped by more tools and customization, there is a compelling argument to be made.

From Opportunity to Imposition back to Opportunity

One of the challenges I keep hearing from other teachers is that they feel like AI is “being forced” on them. I think this isn’t unique to teachers. In fact there is an entire inevitability narrative. But teachers are at the front lines of AI adoption, even though they never chose to be.

If you look at the DepEd’s DO3, which focuses on AI, it’s a strong document that lays a solid foundation for beginning AI adoption in schools. It has a focus on using AI to increase human agency, human-centered deployment of AI, and clear red-lines and protections. The document also outlines responsibilities at every level. But from my read of it, major challenging decisions are going to have to be made by teachers in the classroom. These are what teachers need to hold (with my annotations in parentheses):

Design and deliver AI-enhanced lessons (but do teachers have the training to use AI ethically and effectively? And what if lessons don’t need AI?)
Establish boundaries for AI use (important to do in every classroom, and a huge challenge)
Empower cognitive development (which is what teachers want to do, but given the technology it’s going to be a real balancing act)
Personalize instruction (doable with the right training)
Promote transparency (essential to any AI use)
Collaborate and train (this is with colleagues to establish best practice)
Monitor classroom usage (can you imagine how challenging this could be given large class sizes and how teachers usually teach more than one section?)
Exercise professional judgment

This provides freedom, in the sense that each classroom’s AI journey will be different and teachers are very empowered to craft their approach. But at the same time, the question becomes does each and every classroom have the resources it needs, both in terms of devices, connectivity, and other tech requirements, as well the AI training that teachers and students need to make full use of the technology–and to do so ethically? And again, this is a lot for teachers to add onto their already challenging workloads.

It’s no wonder then that AI can feel like an imposition. Teachers have to learn it, and the level at which they need to be competent with it would have to be high for them to fulfill these targets. On top of the need for more training, there is the larger challenge we face when looking at the situation teachers face. The ideal teacher to student ratio is 1:30, while we are at 1:40. And there’s reporting that we have a shortage of over 50,000 teachers. That’s alongside the need for 165,000 more classrooms.

But I think one reframe we need is to move from looking at these tactical and responsive approaches to AI and the classroom and to start thinking more broadly about education.

AI doesn’t create the problem, but it reveals

I’ve found in looking at AI issues that in many cases AI doesn’t create the problem. It reveals or even exacerbates an already existing issue, making it so that it cannot be ignored. In our case, it’s that there are so many problems with education that the powers that be are willing to look at anything. When kids can’t pass tests, we have literacy crises, we don’t have enough classrooms, all these problems, then a technology promising to address some of them is instantly appealing.

So I don’t take it against leaders that they would try and push AI forward. They are trying to solve a problem and this looks like a solution. However, the way that we adopt and implement AI is one of the real challenges. If we just plug AI into whatever is already there, either it won’t have any real impact, or it won’t actually help students perform better. It’ll just help them hit certain metrics.

For me the invitation is to rethink education overall. I know that becomes more philosophical and esoteric than what school administrators and teachers want to be thinking about. But given the opportunities the technology affords, I think progressive educators should start thinking not only about AI integration but entirely different models for education. I’ll also write an entire piece about an imagined AI-powered future, alongside a more practical and near-term piece.

To end this first piece, I think the invitation is rather than to rush into AI adoption, we need to think about what kind of education we want people to receive and then consider how we might best find parts of that where AI fits. Also taking into consideration that perhaps we need to be careful with how soon we put technology into kids’ hands. And how well-equipped teachers need to be before they are using it and rolling it out. There’s a lot more to discuss, and it’ll show up in the following pieces.