How to Write an AI Use Policy for Your School
If your school doesn’t have an AI policy yet, you need one. And even if it does, it might need updating. Here’s a framework for developing a practical, enforceable AI use policy based on what I’ve seen work at other schools.
Start by acknowledging reality. Any policy that simply says “AI is banned” will be ignored. Students have smartphones and home computers. You cannot prevent access. A good policy acknowledges that AI exists and focuses on defining appropriate use.
Define a spectrum of acceptable use. I recommend something like this:
Always allowed: Using AI for brainstorming, research assistance (with verification), spell checking, accessibility support, and learning about new concepts.
Allowed with disclosure: Using AI to generate outlines, get feedback on drafts, translate between languages, and create study materials. Students must disclose AI use and describe how they used it.
Not allowed: Submitting AI-generated text as your own work, using AI during in-class assessments (unless specifically permitted), using AI humanizers to disguise AI-generated content, and using AI to impersonate others.
Include consequences that are graduated. First offense should typically involve a conversation and a redo of the assignment. Repeated offenses follow your existing academic integrity progression.
Train your teachers. A policy is useless if teachers don’t know how to implement it. Professional development on AI detection, AI literacy, and assessment design should accompany any policy rollout.
Involve students. Seriously. They understand the tools better than most adults, and involving them in policy development increases buy-in.
Review annually. This technology changes fast. Your policy needs to evolve with it.
Does your school have an AI policy? What works well about it and what doesn’t?
7 Replies
Join the discussion.
Log In to Reply
THANK YOU for saying what I've been thinking. The ban approach is not working. We tried it for a full semester and all it did was make students sneakier. The transparency model is so much better. When I told my class they could use AI as long as they documented it, the quality of conversations about their work went way up.
Academic integrity in the AI era requires us to be more precise about what we're actually assessing. If the goal is to evaluate a student's understanding of a topic, then the assessment method should make AI use irrelevant, not just prohibited. Process portfolios, oral examinations, and in-class writing accomplish this far more effectively than detection tools.
this is so helpful. im a new teacher and had literally no idea what to do when a student's essay came back flagged. ended up just ignoring it because i wasn't confident enough to confront them. not great i know
Twenty years ago, the integrity conversation was about internet plagiarism. Before that, it was about copying from encyclopedias. The medium changes but the underlying questions remain the same: what does it mean for work to be 'your own,' and how do we assess genuine understanding? We've always muddled through these questions. We'll muddle through this one too.
i brought up the false positive issue at a staff meeting and got told to 'trust the technology.' this is the same admin that can't figure out how to unmute on zoom