The New Sorting Hat

AI detection tools are sold as academic integrity measures. But what they actually do is decide which students get believed. That's not a plagiarism check. That's classification infrastructure with real consequences, and most institutions are deploying it with very little scrutiny of the error rate.

AI detection tools don't catch cheating. They classify students as credible or not. Those are very different things, and the difference matters most to the students most likely to be flagged.

The framing around AI detection in schools is almost entirely about academic integrity. The more accurate framing: these tools decide which students get believed. That's a classification problem with real stakes, not a plagiarism problem with a technical fix.

AI detection tools are marketed as a way to protect academic integrity. What they actually do is sort students into believed and not believed. The integrity framing makes the tools sound neutral and procedural. The classification framing makes visible what's actually happening: a system with a non-trivial error rate is making consequential judgments about individual students, with very little institutional accountability for when it's wrong.

AI detection tools don't catch cheating. They classify students as credible or not. Schools are treating those as the same thing.

AI detection isn't really about catching cheating. It's about deciding which students get believed. The integrity framing hides what the tool is actually doing.