At the start of this year, OECD and Brookings each published reports exploring the role of AI in education. Both raised some concerns, but roughly pointed to the same conclusion: AI can improve outcomes with students completing tasks more efficiently, teachers saving time, and systems becoming more responsive.
But I’d argue there’s a deeper question underneath those findings: Are we improving performance—or are we improving learning? And more specifically, are students and teachers becoming more engaged in the work, or further removed from it?
The student impact
It goes without saying that AI makes it easier than ever to get to answers. That’s the promise and the appeal. But when it comes to classrooms, learning has never really been about just getting the answers right. It’s about the process of how you get there.
Take writing, for example. We often say that the strongest work comes from editing and rewriting—a process that involves iterating, refining, and rethinking. Yes, it can be slow and uncomfortable. But it’s where deep thinking actually develops.
AI, however, changes that dynamic. It can generate a first draft instantly. It can tell you where to make improvements or it can just make them itself.
This is why the challenge isn’t whether students can use AI—it’s how they use it. Do they treat it as a shortcut? Or do they use it as a starting point?
That difference comes down to something less technical and more human: maturity. The ability to stay engaged in the work, to question, revise, and improve what AI produces, rather than simply accepting it.
Some of the most important parts of learning are inherently inefficient.
Executive Vice President, HMH Labs
The teacher impact
There’s a similar shift happening on the teacher's side, too. AI is already proving its value in reducing workload, especially in areas like giving feedback and grading. Tasks that once meant bringing home stacks of papers, and hours reading at the dinner table, can now be done in minutes.
That efficiency matters, but it’s worth asking what skills and insight were embedded in that work before AI offered an automated route.
For many teachers, grading isn’t just evaluation. It’s one of the primary ways they understand their students—how they’re thinking, where they’re struggling, how they’re progressing, even how they are expressing themselves. If AI fully takes over that process, there’s a risk that teachers lose that visibility, even as systems become more efficient.
The goal, then, isn’t to remove the work entirely, but to use AI to support the role of the educator. In practice, this looks like AI acting more like a teacher’s aid, while keeping teachers in the driver’s seat, reviewing, interpreting, and staying close to student thinking and the learning process itself.
Designing AI for engagement, not just efficiency
It’s entirely possible to build systems that give students answers immediately, automate feedback end-to-end, and minimize the time required to complete tasks. But some of the most important parts of learning are inherently inefficient.
The same is true in classrooms. The goal isn’t perfect individual optimization, where each student moves independently through content. It’s a shared environment where students learn with and from each other, guided by a teacher who understands how to orchestrate that experience.
AI can support that orchestration by helping teachers connect instruction more precisely to student needs and surfacing insights that would otherwise be missed. But it shouldn’t replace the human dynamics that make learning meaningful.
What emerges from both the research and the early classroom experience is that AI will only improve learning if it keeps both students and teachers engaged in the work, not removed from it. For students, that means continuing to develop habits that lead to real thinking like persistence, revision, and the willingness to go beyond the first answer. For teachers, it means maintaining visibility into student thinking, even as tools become more powerful.
And for those designing systems, it means making intentional choices—sometimes even introducing friction with a purpose—to make sure that efficiency doesn’t come at the cost of meaningful engagement.
***
For more articles on trends in education from experts, subscribe to HMH's LinkedIn newsletter, Policy in Motion.
Discover best practices for integrating AI in the classroom.
