AI-detection software has a well-documented false-positive problem, and confidently accusing an honest student is worse than missing a dishonest one. A better toolkit combines lighter-touch detection with structural changes.
Treat detectors as a hint, not a verdict
If a detection tool flags something, use it as a reason to have a conversation — ask the student to explain their process — not as standalone proof.
Redesign assignments around what AI can't easily fake
In-class writing, oral defenses of submitted work, and process documentation (drafts, notes, revision history) are much harder to fully outsource than a single final essay.
Have the conversation directly
Asking a student to walk you through their reasoning, in person, reveals genuine understanding (or its absence) faster and more fairly than any software.
Make the policy clear before the assignment, not after
Ambiguity is where most disputes happen. State explicitly what AI use is and isn't allowed for a given assignment, ideally with examples, before students start working.