Teachers Are Using AI to Catch AI — The Double Standard Nobody Wants to Talk About

There's an elephant in every classroom right now. Teachers are telling students not to use AI — while quietly running essays through Turnitin's AI detector, GPTZero, and Copyleaks. Nobody says it out loud. But the irony is hard to miss.

We sat down with Dr. Alex Morgan, an academic integrity researcher who has spent the last two years studying how schools implement AI detection policies. What they shared might change how you see the whole situation.

What Is the Irony, Exactly?

The irony is direct: teachers who prohibit student AI use are themselves relying on AI systems to enforce that prohibition. AI detectors like Turnitin's AI Writing Indicator are not human judgment — they are machine learning models making probabilistic guesses about your text.

Q: Let's start with the big picture. What's the irony you're seeing play out in schools right now?

A: "It's honestly fascinating from a research standpoint. You have educators telling students 'don't use AI, it's cheating' — and then the moment a student submits, the teacher runs it through an AI model. Turnitin, GPTZero, Originality.AI — these are all AI systems. So we've created a situation where AI is banned for students but institutionally endorsed for teachers. Nobody seems to want to name that directly."

Q: Why isn't this talked about more?

A: "Power dynamics. The teacher controls the grade. The student controls almost nothing. When a student points out 'hey, you're using AI too,' it sounds defensive — even when it's completely valid. There's also a cultural assumption that school tools are neutral and objective. They're not. Once you understand how AI detectors work, you see just how imperfect these systems really are."

Are AI Detectors Actually Reliable?

No — and this is where the irony becomes genuinely harmful. AI detectors carry documented false positive rates. Real human writing gets flagged regularly. ESL students, concise writers, and students who write in formal register are all disproportionately affected.

Q: How accurate are the tools teachers are actually using?

A: "Studies show error rates anywhere from 10% to over 30% depending on the tool and text type. GPTZero has been openly criticized for flagging human writing as AI. Turnitin's own documentation acknowledges a 1% false positive rate — which sounds small until you realize that's roughly 1 in 100 students being wrongly accused. At scale, that's thousands of students every semester. It's not fringe. It's systemic."

Q: So innocent students are getting accused?

A: "Constantly. A student writes in their second language — clear sentences, simple structure — and suddenly they're sitting in an academic integrity office. The problem of AI detection false positives is one of the most underreported stories in education right now."

What Can Students Actually Do?

If your work gets flagged, you have more options than you think. A detection score is not proof. It is a probabilistic output from a machine — not a verdict.

Q: What's your advice for a student who just got flagged?

A: "Document everything. Keep your drafts, your revision history, your notes app, your browser tabs. If you wrote it yourself, you can show your process. Know that most institutions have appeal processes — you have rights. There's genuinely useful guidance on what to do if accused of using AI that most students don't even know exists."

Q: What about students who do use AI and want to avoid detection?

A: "That gets philosophically complicated fast. But practically? Tools like WriteMask exist specifically to humanize AI-generated text — restructuring phrasing, varying sentence rhythm, making output read more naturally. WriteMask reports a 93% pass rate against major detectors. I'm not here to give a moral verdict. Students are making practical decisions inside a system that isn't particularly fair to begin with."

Is AI Detection Even the Right Answer for Schools?

Q: Do you think AI detection is actually solving anything?

A: "No. It's an arms race nobody wins. AI writing tools improve constantly and detectors struggle to keep pace. The schools handling this well are shifting focus to process — oral defenses, in-class writing, drafts with tracked revision history. Using a detector as your primary enforcement tool is like trying to fight a flood with a sponge."

Q: Last thought for students navigating this right now?

A: "Know the system you're operating in. Know its flaws. Before you submit anything, run your own work through a free AI detector to see what your teacher might see first. If you're producing AI-assisted work, revise it genuinely — not just to dodge a score, but to make it actually yours."

The irony of teachers using AI to catch AI is not going away. But students who understand how these tools work — their biases, their limitations, their inconsistencies — are far better positioned to protect themselves when it matters.