Lesson 6 of 9

Is it cheating? Do detectors work?

5 min read

If a student uses AI, surely a detector can catch it? People bet grades and jobs on these tools — so let's see how well they actually work.

Detectors guess, and miss

AI-text detectors output a probability, not proof. They routinely do both bad things: flag genuine human writing as AI (a false positive) and wave real AI text through as human (a false negative). Plain, clear human writing gets flagged most — which punishes exactly the wrong people.

A detector's score is a guess — wrong often enough to ruin trust in it.

So is using AI cheating?

That depends on the rules you're under, not on a detector. The tool can't settle it, and leaning on one risks false accusations. The honest approach is about disclosure and intent: know the policy, be transparent about what AI did, and use it to learn rather than to skip the learning.

'Cheating' is a rules-and-honesty question — not something a detector decides.

Don't trust an 'AI detector' verdict as proof — false positives hit real human writing, and confident students get wrongly accused.