The Biggest Challenge in Bringing AI to the Classroom isn’t Tech

Turnitin’s Annie Chechitelli offers this commentary on how the biggest challenge in bringing AI to the classroom isn’t tech. This article originally appeared in Insight Jam, an enterprise IT community that enables human conversation on AI.

Artificial intelligence is finding its way into classrooms faster than most educators can keep up with, and how students use it varies more than many might expect. One student turns to AI to pressure-test their argument and fill gaps in their research. Another uses it to generate a finished product they barely read before submitting. Same tool, same technology, but a fundamentally different learning experience depending on what’s happening in the classroom around it.

As schools and universities adopt AI, much of the conversation has focused on logistics. Which tools should we use? How fast should we roll them out? But the biggest challenge isn’t access, it’s whether students and teachers know how to use these tools well.

The usage numbers make this clear: 94% of undergraduates report using AI tools in some capacity. But access hasn’t translated into competence. Many students feel they don’t know how to maximize the benefits of AI, with limited educational institutions offering formal support for developing AI skills.

AI adoption gets confused with AI skills development. A student who can get ChatGPT to produce an essay isn’t necessarily a student who can tell whether that essay is accurate, well-reasoned, or missing the point entirely. Students need to develop judgment about when AI is helpful and when it’s leading them astray. Educators face the same challenge – they’re being asked to incorporate tools that didn’t exist five years ago while still doing everything else their job requires.

For years, the fear was that students would use AI to cheat. That’s still a concern, but a more urgent one is surfacing, that students will graduate knowing how to use AI without knowing how to think with it. And students themselves sense this with concerns rising around AI reducing their critical thinking skills and whether they are becoming over-reliant on these tools.

Educators aren’t becoming less important here. If anything, they’re more central than before. They’re the ones helping students determine when to use AI in the first place. If using AI in an assignment, they are helping students evaluate AI-generated information, spot inaccuracies, ask better questions, and understand where these tools fall short. They’re teaching judgment, not just content.

But students are getting mixed signals. One instructor encourages experimentation, another bans AI entirely, and most land somewhere in between. Students end up guessing the rules instead of learning a framework. The data backs this up: fewer than half of institutions currently have an AI policy in place.

Institutions can fix this, but it takes more than making tools available and hoping for the best.

• Set clear expectations. Students and faculty need to know when AI use is appropriate, and those rules need to be consistent enough that students aren’t relearning them every semester.

• Train people, not just deploy tools. Knowing how to prompt an LLM is a start. Knowing how to evaluate what comes back, cite it properly, and recognize its limitations is the actual skill. Most faculty haven’t received any formal guidance on this yet.

• Guide how students engage with AI, not just what they turn in. Teachers may not always control the assignment, but they can shape the conversation around it. Asking students to explain their reasoning, walk through how they used AI, or reflect on what they’d do differently builds the kind of critical thinking that outlasts any single task.

What is clear: students are using AI in their assignments, and in many cases, to do the work for them. The issue facing educators and institutions is in understanding how students are using AI and to guide the if, when and how AI is used such that it enhances learning, not replaces it.