The New Student-Teacher Gap on AI
The Digital Education Council released survey data this month worth examining closely. Eighty-eight percent of students are already using AI as part of their learning—this is a baseline now, not theory or experimentation. The same survey found that 77 percent of faculty use AI in their teaching, so we’re not in a world where teachers have rejected the technology and students are rebelling. But here’s what this reveals: 57 percent of students say their classes don’t give them clear guidance on how to use AI appropriately, and only 29 percent believe their instructors are equipped to help them navigate it.
The real pressure point in education right now isn’t about access to AI or whether students will use it. It’s about the mismatch between what students are doing with these tools and what we’re equipped to teach them to do.
I think about the assignments I give. I give essays and analyses and debates—most of them can be approached with or without AI. When a student uses Claude or ChatGPT to help them think, that’s not automatically cheating, but it’s also not automatically learning. The quality of that learning depends on how the student wields the tool, whether they’re using it to sidestep thinking or to extend it. And right now, many of my students are figuring that out without clear guardrails. I’m not hostile to the technology. The real issue is that I haven’t given them a clear framework for when it’s appropriate, when it’s not, and what actually constitutes a shortcut versus an extension of their thinking.
One of the things the survey data tracks is what students are using AI for. Most commonly: homework help, study guides, brainstorming, and writing assistance. If I frame homework as “produce an essay by Friday,” and a student uses AI to draft a first pass, the conversation should focus on what the student does with it afterward. Did they interrogate the output? Did they fact-check it? Did they understand the argument well enough to defend or revise it? Those are the moves that separate using AI as a thinking tool from using it to outsource thinking.
The problem is I haven’t always been explicit about that with students. And so the students who would benefit most from clarity are often the ones leaving it to chance.
Some schools are moving faster on this than others. MSU Denver launched an elective this semester open to any major—it teaches practical AI literacy through applied case studies and ethical reasoning. UCCS opened ChatGPT Edu to its students on August 14, but with a requirement: you have to complete an AI literacy course before your account is provisioned. Those are structure-based solutions. They’re saying: if you’re going to use this on campus, we’re going to make sure you understand what you’re doing first.
There’s also integration happening at the tool level. Google added Gemini to Classroom this summer, and it now generates personalized study aids for students. Microsoft’s Copilot is positioning itself as a learning coach. When these tools land inside the systems students use every day, the question of whether AI is in the classroom becomes moot. The real question is what we do about it now that it’s there.
Most of these solutions, though, are happening outside the direct relationship between a teacher and their students. UCCS has a literacy course. MSU Denver has an elective. Google and Microsoft have tools. But in my classroom, on Tuesday, when I’m grading essays or asking students to defend an argument, I’m the one who sets the terms. I’m the one who decides whether an AI-assisted draft is acceptable starting material or whether it’s a shortcut that hasn’t earned its weight. And if I’m not clear about that decision, I’m leaving students in the position the survey data describes: using AI in their learning but unsure whether they’re doing it right.
That 29 percent statistic reveals something crucial: the gap isn’t about teacher competence, it’s about teacher communication. Some teachers are extraordinary. They’re thoughtful, rigorous, and they care about what students are learning. But if they haven’t articulated a stance on AI use, if they haven’t built assessment structures that make the role of AI explicit, then students won’t know whether their use is aligned with what the teacher expects.
So what does that look like in practice? It might mean being explicit in assignment language. Instead of “write an essay,” it might be “use AI to generate three possible thesis statements, choose one, and explain why you chose it and how you’d revise it.” Or it might mean front-loading a conversation: “In this class, if you use AI to help you draft something, I want you to annotate what you changed and why. That’s not cheating. That’s showing me how you’re using the tool.” Or it might mean building checkpoints into longer assignments—a rough draft review before the final submission, where we talk about what’s working and what needs to be rethought.
The barrier isn’t expertise—it’s expectations. Most teachers know more about their subject than ChatGPT does, but we haven’t updated how we assign, evaluate, and give feedback now that students can co-author with an AI. That’s a different teaching move. It requires us to think differently about what we’re evaluating and what counts as evidence of learning.
One more thing the survey mentioned: students in schools with explicit AI policies were more confident in their use. Rather than avoiding AI, they used it more thoughtfully, because they knew what the boundaries were. That’s the signal for what the next move is. Not to ban the tool. Not to pretend students aren’t using it. But to make our thinking transparent and to give students the framework they need to use it well.
The students are already using AI in their learning. The question for us is whether we’re going to guide that use or leave it to chance.
Takeaway for Teachers: Pick one assignment this month. Decide explicitly whether AI use is encouraged, forbidden, or allowed-with-conditions. Write that decision into the assignment prompt. If you allow it, specify what you want students to show you about their use—drafts, annotations, process notes. If you forbid it, explain why. If you’re unsure, that’s the conversation to have with your students before they start working, not a month in. Clarity beats assumption every time.
If you’re thinking through questions like this one—what it means to teach writing and thinking when AI tools are reshaping what students expect from the learning process—my book goes deeper. The AI Doesn’t Know Your Students is available on Amazon and at shouldiuse.ai/book. It walks through how to think about AI in terms of your specific students, your specific classroom, and the specific things you’re trying to teach.
David Jacobson is a high school history teacher. He writes about AI, education, and the messy intersection of the two at shouldiuse.ai.
