When the Essay Isn’t Theirs: Spotting AI in the Classroom (and What to Do About It)

By Carter F. Smith

It started with a few suspiciously polished paragraphs. Now, it’s practically a new academic season: AI Essay Spring, where bots bloom, syntax flows, and citations are mostly—if not entirely—made up.

Professor Mark Massaro, quoted recently in Gizmodo, offered a candid glimpse into the strange new world of grading papers that students didn’t exactly write. He’s not wrong: AI-generated essays are the new classroom contraband. From overused em dashes and suspiciously clean grammar to invented citations and oddly robotic meditations on “the importance of friendship,” the signs are everywhere. One student even left their ChatGPT prompt in the final draft. (Bold move.)

AI, of course, isn’t new. What’s new is the ease with which students can now outsource their voices, trade uncertainty for polish, and sidestep the very struggle that teaches them how to think.

And here’s the kicker: they often don’t know what they’re losing.

The Voice That Got Replaced

Massaro put it bluntly: this is the time students should be “finding their voice.” Instead, they’re letting algorithms speak on their behalf—turning what should be a messy, reflective, formative process into a clean, impersonal Google Doc.

Eric Detweiler, in his recent interviews and book Responsible Pedagogy, expands this idea further. He argues that good writing instruction is inherently human: it’s about conversation, revision, uncertainty, and meaningful engagement. Students don’t just learn to write by writing—they learn to write by getting things wrong, talking it through, and revising not just their words but their ideas.

AI can’t do that. It can simulate understanding, but it can’t think. It can polish, but it can’t care.

And this isn’t a Luddite complaint. Those who scorned the typewriter in favor of the quill pen could still write well—just not as fast. And that slower pace may have afforded them a chance to think more deeply. (Or maybe just more time to stare wistfully into the distance. Same difference.)

What’s a Teacher to Do? Pedagogical Judo

Let’s be honest. We can’t out-tech the tech. AI detectors are flaky. Privacy policies are real. And unless we plan to duct-tape ourselves to our students’ keyboards, we need a different approach—something like pedagogical judo: use the momentum of the tool, redirect its energy, and land softly in a teachable moment.

Here’s how:

1. Make the Process Visible

Require drafts, revision notes, peer feedback exchanges, and reflections. Ask students to write about their writing. Not just what they think, but how they got there. If the essay popped into existence like Athena from Zeus’s forehead—fully formed and citation-ready—you’ll know.

2. Assign the Un-Googleable

Personal reflection? Make it weird. “Describe a conversation you had at a gas station,” or “Connect a recent personal failure to a historical event we’ve studied.” Bonus: you’ll actually want to read these. AI is great at summarizing Plato. It’s terrible at recalling what your mom said to you during a thunderstorm.

3. Teach the AI

Yes, really. Let students use the tools—but teach them how. Have them critique an AI’s draft. Ask what it got wrong. Show them the fake citations. Make them revise an AI-generated paragraph and explain their changes. Make it clear: AI isn’t a cheat code—it’s a conversation starter.

4. Push Them Off Script

Ask for a spontaneous writing sample in class. Give short in-person assignments where they can’t hide behind the screen. You don’t need to catch them—you just need a baseline. If their homework sounds like Emerson and their in-class writing sounds like a group text… well.

5. Create Intellectual Ownership

If students feel seen, respected, and curious, they’re less likely to outsource their voice. Assignments that connect to their lives, beliefs, and experiences can’t be delegated. When they care, they write.

Final Thought: Don’t Fear the Quill

Let’s not mourn the essay just yet. Students still want to be heard. They’re just being seduced by the illusion of perfection—and that’s our opening. If we can show them that uncertainty is part of thinking, that voice matters more than polish, and that ideas are worth struggling for, we can steer this moment away from crisis and into growth.

AI might be fast, but thinking takes time. And that’s the point.

So yes, the bots are here. But so are we. And if we lean in—not just to catch them, but to teach them—we might find something better than a perfect essay. We might find students who think again.