A clear AI policy for your school

1. Purpose and scope

This policy sets out how [School name] uses artificial intelligence tools, including generative AI. It covers staff use for professional purposes and student use for learning and assessment. It helps our school use AI safely, honestly, and in ways that strengthen teaching and learning rather than replace it.

Provided as general guidance, not legal advice. Have your Board review it against your school’s circumstances.

2. Guiding principles

1. People decide, AI assists. AI never makes final decisions about a student's learning, assessment, wellbeing, or discipline.
2. Learning comes first. Students may use AI where it supports learning, and may not use it where it does the learning for them.
3. Privacy is protected. Personal information is only entered into tools approved for that purpose.
4. We are honest about AI use. Staff and students are open about when and how AI was used.
5. Access is fair. Where AI is part of learning, the school plans for equitable access and reasonable adjustments.

3. Staff use of AI

Encouraged uses

Staff may use school-approved AI tools to support planning, resource creation, differentiation, routine communications, assessment preparation, and administration.

Conditions on all staff use

• Human review, always. Staff verify and take professional responsibility for AI-assisted output before it reaches students or families.
• No personal information in unapproved tools. Do not enter names, photos, assessment data, health, pastoral, or family information unless the tool is approved for that data.
• Professional judgement is not delegated. AI may inform, but must not determine, grades, support, behaviour, safeguarding, or reports about an individual.

Prohibited staff uses

• Using AI to make disciplinary, safeguarding, or other high-impact decisions
• Entering identifiable personal information into unapproved tools
• Presenting AI-generated evaluation as a staff member's own professional judgement without review
• Using AI detection as the sole evidence of academic misconduct

4. Student use of AI

Teachers state clearly for each task which traffic-light level applies. If no level is stated, students ask before using AI.

Younger students ([School to specify year or grade levels]) use AI only under direct staff supervision.

5. Assessment integrity and NCEA authenticity

• For NCEA and other formal assessment, [School name] follows NZQA’s current authenticity requirements. Where AI use is not explicitly permitted, submitting AI-generated work as the student’s own is a breach of authenticity.
• AI detection software is not used as sole evidence. The school considers drafts, version history, comparison with other work, and conversation with the student.
• Teachers make authenticity visible where practical through supervised components, oral checkpoints, drafts over time, and classroom-specific material.
• Suspected breaches follow the school's existing assessment and academic-honesty procedures, with the student given an opportunity to respond.

6. Privacy and data

• [School name] follows the Privacy Act 2020 and the school’s privacy policies in all use of AI tools.
• Personal information is entered only into approved tools and only for approved purposes.
• The school prefers tools that do not train on school data and that clearly explain storage, retention, deletion, access, and subprocessors.
• Any suspected privacy breach involving AI is reported immediately to [privacy or data-protection role] and handled under the school's breach procedure.

7. Approving new AI tools

Before an AI tool is used with student data or by students, [approver role] completes the AI Tool Approval Checklist (Appendix B). The review covers educational purpose, data collection and location, model training, retention and deletion, age requirements, accessibility, safeguarding, security, cost, contract terms, and exit. Approved tools are recorded on the school's register.

8. Responding to concerns and misuse

• Concerns about staff use are raised with [role] and handled through normal professional processes.
• Suspected student misuse in assessment is handled under Section 5.
• The school treats most first-time misuse as a learning opportunity while applying existing behaviour and academic-honesty procedures where appropriate.

9. Communication with parents and whānau

[School name] communicates openly with parents and whānau about how AI is used, how personal information is protected, the traffic-light framework, and which tools students use. The letter in Appendix A is shared when the policy is adopted and with new families.

10. Adoption and review

This policy is approved by the Board and reviewed at least annually, or sooner if the Ministry of Education, ERO, NZQA, or the Office of the Privacy Commissioner issues new requirements. [Role] keeps the approved-tools register current and proposes updates.

Appendix A: Letter to parents and whānau

Appendix B: AI Tool Approval Checklist

Before approving an AI tool, confirm:

1. Educational purpose: what problem does this solve?
2. Data: what personal information does it collect, infer, or generate?
3. Storage and access: where is data stored, retained, and processed—and by whom?
4. Model training: is school data used to train models? Prefer no.
5. Age and safeguarding: do terms, moderation, and supervision suit intended users?
6. Equity and accessibility: can all relevant learners use it fairly and accessibly?
7. Accuracy and bias: what testing, monitoring, and human review are required?
8. Contract and exit: what are the costs, renewal terms, deletion rights, and exit plan?
9. Approval: approved by [role], date, and added to the register.