Resources · For principals & coordinators
School AI Policy: A Practical Guide for Principals
A workable school AI policy is six named decisions, written down, published where parents can read them, with an owner and a review date. It is not a ban (bans move use out of sight) and not a values statement (values are not checkable at marking time). Here are the six decisions, with the free resources to start from.
Editorial guidance, not legal advice — regulatory requirements differ by jurisdiction. Last reviewed 31 August 2026.
What six decisions does an AI policy need?
1. Which tools are approved, by age?
Name the tools, not the category. The differences are large and vendor pages obscure them: Claude is 18+, Character.AI removed open-ended chat for under-18s, Snapchat’s Family Center cannot show a parent My AI conversations. Our AI tool age-limit table is checked against vendor terms and kept current — use it as the starting worksheet.
2. Where is the line between help and submission?
The rule that survives is behavioural, not tool-based: attempt first, use AI to explain or check, never submit text you did not write, disclose use the way the school specifies. A tool ban moves use out of sight; a submission rule is checkable at marking time.
3. What may students never enter into an AI tool?
Personal identifying details, health information, other students’ information, and anything covered by your safeguarding policy. This clause matters more than the tool list — it applies to whatever tool arrives next year.
4. What are the rules for staff?
Teachers use AI too — for marking support, lesson prep, report drafting. Decide what is acceptable, what needs disclosure, and what may never leave the building (student data in prompts is the usual failure). A policy that only binds students will be ignored by Easter.
5. How do you tell parents?
Publish the policy where parents can read it, and say what happens when it is breached. A short public page beats a PDF in a portal. Commitments written down and published are commitments a school can be held to — which is exactly why they build trust.
6. Who reviews it, and how often?
Name an owner and a review date each term. Vendor age policies moved at least three times in the last year; a policy referencing last year’s tool landscape reads as abandoned.
Which free resources help you draft it?
- •ORISE publishes a free, no-registration lesson plan — “Developing a Classroom AI Policy” (Grades 6-12) — useful for drafting the student-facing rules with students in the room. orise.orau.gov
- •Our AI tool age-limit table — every major tool’s real minimum age and what its parental and admin controls actually do, checked against vendor terms.
- •A worked example of publishing commitments: our own Child Safety Protocol — published so we can be held to it, which is the posture a school policy should take too.
- •The parent conversation, ready-made: our parents’ guide to teenagers and AI and the homework rule that works.
Where does teaching fit?
A policy controls behaviour; it does not create understanding. Students follow AI rules far better when they know how these systems actually work — why a model is confidently wrong, what training data does, where bias comes from. That is the curriculum layer, and it is what LittleAIMaster teaches for Grades 6-8 and Grades 9-12 — first chapters free, no card, per-seat pricing published.
Bring this curriculum to your school
LittleAIMaster runs structured AI lessons for Grades 6-12 — 540+ chapters, teacher dashboards, and per-student progress visibility. The app is freemium, so your teachers can test the first chapters free before any conversation about seat licensing.
Talk to us about your schoolResponse within 1 business day