AI Detectors for Teachers: Using Them Fairly

Start with a clear policy

• State what AI use is allowed, if any: brainstorming, outlining, grammar help, or none.
• Explain how students should disclose AI use.
• Say whether and how detectors will be used, and that a score alone will never decide an outcome.
• Put the policy in the syllabus and repeat it on major assignments.

Use detectors as a screen, not a verdict

• Look at which passages were flagged, not just the overall number.
• Ignore flags on quotations, citations, and boilerplate.
• Compare with the student’s earlier work in class.
• Remember that non-native English writers and students taught strict structures are flagged more often.
• Never assign a penalty based on a score alone.

Talk before you decide

A short, private, non-accusatory conversation resolves most concerns. Ask the student to walk you through their argument, sources, and process. Students who wrote their work can usually explain it in detail.

Design assignments that make misuse harder

• Require process checkpoints: topic proposal, outline, annotated sources, rough draft, and final draft.
• Ask for reflection on how the student’s thinking changed.
• Connect prompts to class discussions, local issues, or personal experience.
• Include some in-class writing to establish each student’s voice.
• Ask students to submit version history or a writing record with major papers.

Ask for proof up front

Instead of detecting after the fact, you can ask students to write major assignments in a tool that records their process and gives them a certificate to submit. Honest students are protected from false accusations, and misuse becomes much harder. This is what Prove Human is built for.

Know your institution’s position

Some universities have disabled built-in AI detectors because of concerns about false positives and transparency. Check your institution’s current guidance before relying on any tool.