Write your AI acceptable use policy around one principle: protecting authentic student learning. Set clear, tiered rules for when AI may and may not be used, how students disclose it, and how staff verify authenticity. Co-design it with staff, students and whānau, align it to NZQA authenticity rules, and review it each year.
Why does your school need an AI use policy now?
Your school needs an AI use policy now because generative AI is already in students’ hands, and without shared rules every teacher improvises a different line between help and cheating. A clear policy protects the one thing assessment depends on — authentic evidence of what each student can actually do — and gives staff, students and whānau a common understanding rather than a patchwork of classroom-by-classroom decisions.
The goal is not to ban a technology students will use throughout their working lives. It is to be deliberate about where AI supports learning and where it undermines the evidence a grade is meant to represent. A policy that only says ‘no AI’ quickly becomes unenforceable and ignored; one that treats AI as a tool to be used well, in the open, is far more likely to hold.
Getting ahead of this also lowers risk. NZQA holds schools responsible for the authenticity of work submitted for NCEA, so a documented policy and consistent staff practice matter if a question is ever raised. Writing your policy now, before an incident forces the conversation, lets you design something considered and educative rather than reactive and punitive.
What should an AI acceptable use policy actually cover?
A strong AI acceptable use policy covers purpose, permitted and prohibited uses, disclosure, authenticity checks, data privacy, and consequences — all written in plain language staff and students can apply. Start from your school’s values and the authenticity requirements you already work under, then translate them into concrete guidance for everyday tasks like essays, coding, research and revision.
Keep the document short enough to be read and specific enough to be used. Vague principles (‘use AI responsibly’) leave every hard case unresolved, so pair each principle with an example. It also helps to distinguish clearly between formative work, where AI can be a useful tutor, and summative NCEA assessment, where authenticity rules are strict and the bar for AI involvement is much higher.
- Purpose and values: why the school allows or limits AI, in one short paragraph.
- Permitted, conditional and prohibited uses, each paired with a concrete example.
- Disclosure: how and when students record the AI help they used.
- Authenticity: the checks staff use, aligned to NZQA requirements.
- Data and privacy: what student information must never be entered into AI tools.
- Consequences and support: what happens, educatively, when the policy is breached.
How do you set clear rules for when students can use AI?
Set clear rules by sorting AI use into a few named tiers rather than trying to list every possible task. A simple three-tier model — not permitted, permitted with disclosure, and actively encouraged — lets students and teachers quickly place any new situation. The tier that applies should depend on the purpose of the task and whether AI would replace the very thinking being assessed.
The same task can sit in different tiers depending on context. Using AI to brainstorm essay angles during a drafting lesson might be encouraged; using it to write sentences in a final NCEA English submission is not permitted, because the standard assesses the student’s own writing. Make the reason explicit — ‘AI is off-limits here because this standard assesses your analysis’ — so the rule teaches judgement rather than just compliance.
Publishing the tiers as a shared reference — on assessment cover sheets, in the LMS, on classroom walls — means teachers apply the same line across subjects. Consistency is what makes a policy credible to students: if one class treats a tool as normal and another treats it as cheating, the policy loses authority fast.
How does an AI policy protect NCEA authenticity?
An AI policy protects NCEA authenticity by making explicit that work submitted against achievement standards must be genuinely the student’s own, and by defining the checks that confirm it. NZQA requires schools to manage authenticity for internal assessment, and undisclosed AI-generated content directly threatens that evidence. Your policy turns that requirement into everyday classroom practice.
Practical authenticity measures work better than surprise detection. Break assessments into checkpoints, see drafts and process, and use short verification conversations where a student explains their thinking. These methods reveal whether the work reflects the student’s own understanding far more reliably than any AI-detection tool, which agencies caution can be inaccurate. Design tasks so that authentic thinking is visible, not just the final product.
This is also where internal and external assessment differ. External examinations are supervised, so AI risk is low; internal assessment, completed over time, is where clear rules and process evidence matter most. Aligning your policy to the moderation mindset NZQA already uses — consistent standards and documented evidence — keeps it defensible and familiar to staff.
How should students disclose the AI help they used?
Students should disclose AI use through a simple, routine declaration that names the tool they used, for what part of the task, and how. Normalising disclosure turns AI from something hidden into something discussed, which is exactly the culture an authenticity policy needs. A one-line statement — ‘I used an AI tool to check grammar in paragraph two’ — is often enough for formative work.
Make disclosure low-stakes and habitual. If admitting AI use is treated as a confession, students will hide it; if it is a normal part of submitting work, honesty becomes the default. Teach the difference between acceptable support (checking spelling, explaining a concept) and substitution (generating the analysis being assessed), so students can describe their use accurately rather than guessing what will get them in trouble.
How do you write and roll out the policy with your community?
Write and roll out the policy by co-designing it with staff, students and whānau rather than issuing it from the top. A policy people helped shape is one they understand and follow. Convene a small working group, draft against a template, test the rules on real assessment examples, then consult widely before you publish and build it into staff and student induction.
Treat the policy as a living document. AI tools change quickly, so schedule an annual review and a clear owner — often a deputy principal or a digital-technologies lead — responsible for updates. Pair the written policy with professional learning so staff can apply it confidently, and with student lessons so the rules are taught, not just posted. A policy no one is trained on is a document, not a practice.
- Form a small working group including teachers, students and a whānau voice.
- Draft from a template, then pressure-test it against real assessment tasks.
- Consult staff, students and the board before publishing.
- Build the policy into staff and student induction, not just a one-off email.
- Name an owner and set an annual review date to keep it current.
How can a tool like JeddAI support your AI policy in practice?
A tool like JeddAI supports your policy by keeping teachers in control of assessment while making feedback faster and more consistent. Teachers connect or upload student work, and JeddAI drafts feedback and marking aligned to the school’s own rubric, success criteria and comment banks; the teacher then reviews and edits every judgement. The professional decision — and the authenticity call — stays with the teacher.
Used within a clear policy, that model reinforces rather than undermines authenticity: AI drafts the repetitive first pass of marking, and staff apply their standards consistently across a class while reclaiming time for teaching. If you are building an AI approach that keeps educators firmly in charge, you can Get started with JeddAI.
| Tier | What it means | Example task | Disclosure |
|---|---|---|---|
| Not permitted | AI would replace the thinking being assessed | Writing a final NCEA essay or analysis | N/A — AI use here is a breach |
| Permitted with disclosure | AI supports the work but the student's own thinking stays central | Brainstorming angles or checking grammar on a draft | Required — note the tool and how it was used |
| Actively encouraged | AI builds skills without touching assessed evidence | Explaining a concept or generating practice questions | Good practice to note; low stakes |
Frequently asked questions
Should an AI policy ban ChatGPT outright?
Rarely. A blanket ban is hard to enforce and denies students skills they will need; a tiered policy that permits disclosed, purposeful use while protecting assessment is usually more workable and honest.
Who should own the AI policy in a school?
A named senior leader — often a deputy principal or digital-technologies lead — should own updates, supported by a working group, so the policy has clear accountability and a scheduled review.
Can we rely on AI detectors to enforce the policy?
No. AI-detection tools can be unreliable and produce false accusations. Process evidence — drafts, checkpoints and verification conversations — protects authenticity far more dependably.
Does the policy need to mention student privacy?
Yes. Under the Privacy Act 2020, students should never enter personal or identifying information into public AI tools, and your policy should state that clearly.
How often should we review the policy?
At least once a year, given how fast AI tools change, plus an unscheduled review if a major new tool or an assessment incident makes the current rules outdated.
Get started with Jeddle
Jeddle gives teachers and students instant, syllabus-aligned feedback powered by JeddAI.
Looking for study material? Browse Jeddle's Australian-English subject resources, or explore more articles on AI in Education.
