AI, assessment and governance resources for real organisations.

Overview

This template helps a school explain how AI may be used, what is not allowed, who is responsible, and how safeguarding, data protection and assessment integrity will be protected.

Purpose

To support useful AI practice while reducing risks linked to personal data, over-reliance, unsafe content, unfair advantage, inaccurate outputs and learner misuse.

Who it is for

• Headteachers and senior leaders
• Teachers, support staff and pastoral teams
• Learners, parents and carers
• Governors, trustees and safeguarding leads

Why it matters

AI tools can support planning, accessibility and feedback, but they can also expose personal data, create inaccurate content, bypass learning and affect fairness in assessment.

Roles and responsibilities

• Governors: approve the policy and review risk reports.
• Senior leaders: decide approved use, training and monitoring.
• DSL: reviews safeguarding risks and reporting routes.
• DPO or data lead: advises on GDPR and personal data.
• Staff: use AI professionally and check outputs.
• Learners: follow acceptable use rules.

Key principles

• AI supports education; it does not replace professional judgement.
• Personal, sensitive or safeguarding information must not be entered into unapproved tools.
• AI output must be checked before use.
• Learners must not submit AI-generated work as their own.
• Parents and carers should receive clear guidance.

Staff acceptable use statement

Staff may use approved AI tools to support planning, resource drafting, accessibility and administration where this improves education and does not expose personal data. Staff remain responsible for checking accuracy, suitability, bias and age appropriateness.

Learner acceptable use statement

Learners may use AI only when a teacher allows it. Learners must explain how AI helped them and must not present AI-generated work as their own learning, thinking or assessment evidence.

Safeguarding and GDPR

• Do not enter names, addresses, health data, behaviour records or safeguarding details into public AI tools.
• Report unsafe, harmful or inappropriate AI output through the normal safeguarding route.
• Use privacy checks before approving AI tools.

Assessment integrity

Teachers should state when AI can be used, require learner declarations where needed, and use discussion, drafts, in-class work or questioning to confirm understanding.

Staff checklist

Learner checklist

Overview

This template helps centres and providers set expectations for AI use in delivery, assessment, evidence production, assessor judgement and IQA monitoring.

Purpose

To protect qualification integrity, learner fairness, awarding organisation requirements, data protection and the reliability of assessment decisions.

Who it is for

• Training providers and approved centres
• Assessors, tutors and IQAs
• Quality managers and curriculum leads
• Learners and apprentices

Key principles

• AI must not replace learner competence.
• Assessment evidence must remain authentic and traceable.
• Assessors must verify understanding before making decisions.
• IQA must sample AI-related risk.
• Awarding organisation rules must be followed.

Assessor responsibilities

• Explain permitted AI use before assessment.
• Review evidence for authenticity and consistency.
• Use professional discussion or questioning where needed.
• Record concerns clearly and follow malpractice procedures.

IQA responsibilities

• Check assessment plans include AI controls.
• Sample high-risk evidence types.
• Review assessor feedback and declarations.
• Record standardisation actions linked to AI risk.

Remote assessment considerations

Remote assessment should include identity checks, clear instructions, time boundaries, evidence logs, questioning and documented judgement. Where risk is high, add live discussion or practical demonstration.

Malpractice and plagiarism

Submitting AI-generated evidence as independent learner work may be treated as malpractice. Decisions should follow the provider and awarding organisation malpractice policy.

Overview

Use this checklist to identify strengths, gaps and immediate actions before rolling out AI tools or AI-supported working practices.

Scoring system

Score each area from 0 to 3. 0 means not started, 1 means early discussion, 2 means partly in place, and 3 means embedded and reviewed.

Checklist areas

• Leadership readiness: clear ownership and priorities.
• Governance readiness: policy, approval and oversight.
• Staff readiness: training, confidence and boundaries.
• Technology readiness: approved tools and access controls.
• Assessment readiness: authenticity and evidence checks.
• Risk readiness: data, safeguarding, bias and cyber controls.

Maturity model

Overview

This template helps organisations identify, score, control and monitor AI risks before tools are approved or used with staff, learners or customers.

Likelihood and impact scoring

Score likelihood from 1 to 4 and impact from 1 to 4. Multiply the scores to identify low, medium, high or critical risk.

Risk areas to review

• Data protection and privacy
• Cyber security and account access
• Safeguarding and harmful content
• Bias, fairness and discrimination
• Assessment integrity and malpractice
• Legal, copyright and compliance
• Reputation and public trust

Control measures

• Approved tool list
• Data entry rules
• Human review before decisions
• Staff training
• Learner declarations
• Incident reporting
• Regular review dates

Overview

AI governance means having clear responsibility, rules and review processes so AI tools are used safely, lawfully and in line with organisational values.

AI Governance Structure

Board or governor responsibilities

• Approve the strategic approach to AI.
• Ask for risk and compliance updates.
• Ensure safeguarding, data and equality duties are considered.

Senior leadership responsibilities

• Assign owners for AI policy and risk.
• Approve tools and training priorities.
• Ensure incidents are reviewed and acted on.

Risk reporting framework

Report high-risk AI use, data concerns, safeguarding concerns, assessment concerns, staff training gaps and incidents to the named AI governance lead and the relevant existing committee.

Governance maturity model

Why assessment integrity matters

Assessment must show what the learner knows, understands and can do. If AI produces the work and the learner cannot explain or demonstrate it, the evidence may not be valid.

Key principles

• Be clear about permitted AI use before assessment.
• Design tasks that require personal, practical or contextual evidence.
• Use more than one evidence source.
• Check understanding through discussion or questioning.
• Record judgement and concerns clearly.

Evidence verification methods

• Professional discussion
• Oral questioning
• Observation
• Practical demonstration
• Draft review
• Workplace records
• Learner declaration

Assessment redesign examples

• Replace a generic essay with a scenario linked to the learner's own setting.
• Add a short viva-style discussion after written evidence.
• Require annotated drafts showing decision-making.
• Use practical demonstration where competence is required.

Assessor checklist

IQA checklist