Generative AI has become part of daily life in classrooms, and for schools, the question has moved past whether students are using it to write assignments. That’s already happening. The harder question is what gets lost along the way, and what schools should be doing about it.
Writing Is Not Just a Way to Record Thoughts, It is a Way to Form Them
Researchers who study how people write have long understood something that is easy to overlook: the struggle of putting an idea into words is not just a hurdle to get past, it is part of how the idea gets developed in the first place. Composition is difficult, not despite its value, but because of it. When a student works through how to structure an argument or find the right words, they’re doing real cognitive work that shapes how well they actually understand the material, not just how well they can describe it.
What the Research Is Starting to Show
A growing body of research backs this up with real data. A widely discussed MIT study found that students who relied on AI to help write showed lower brain activity while writing and had a harder time recalling what they had supposedly written themselves, along with less sense of ownership over the finished product. Other studies have linked heavier AI reliance among students to measurable declines in critical thinking and, in some cases, weaker academic performance overall. None of this means occasional AI use is harmful. It does suggest that leaning on it to skip the hard part of thinking through a problem comes with a real cost.
It Is Not All One-Sided
The picture is not entirely negative. Research also shows AI can be genuinely useful for tasks like organizing research, analyzing data, or getting feedback on a draft, particularly when students treat it as a collaborator rather than a shortcut. Timing seems to matter quite a bit: students who use AI after they’ve already worked through their own ideas tend to fare very differently than those who let AI generate the ideas for them from the start.
Worth Acknowledging: Writing Is Genuinely Hard for a Lot of Students
It is also fair to acknowledge why AI is so tempting to reach for. Writing can be a source of real frustration and even embarrassment for students who struggle with it, and a significant share of students nationally read and write below grade level. For those students in particular, an AI tool that can instantly produce a polished result is an understandably appealing way to avoid a painful and high-stakes task. That reality is worth keeping in mind rather than dismissing, even as schools think about where to draw the line.
What This Means for Schools
None of this means schools need to ban AI outright, but it does suggest that policies drawing a hard line only at “cheating” may be missing the more important question. Several educators are moving toward approaches that ask students to generate their own ideas and a first draft before AI enters the picture at all, preserving the part of the process where the real thinking happens while still allowing AI to help with revision, feedback, or polish afterward. Clear, well-communicated expectations, for students, parents, and staff alike, go a long way toward avoiding confusion and disputes down the line.
Takeaway
AI is not going away, and treating it purely as an integrity problem misses what may be the bigger issue: how and when it enters a student’s process matters as much as whether it is used at all. Schools that think through that sequencing thoughtfully, rather than reactively, are likely to end up with policies that hold up better over time. We are glad to help schools think through how these considerations intersect with existing academic policies and expectations.