Responsible AI
Responsible AI use in schools
A practical framework for privacy, review, transparency and appropriate professional oversight.
Protect personal and sensitive information
Follow your organisation's data protection policy and approved systems. Do not enter identifiable pupil, family or staff information unless your setting has explicitly approved that processing for the tool and purpose.
Anonymise context where possible and use the minimum information needed. Responsible AI use starts with good information habits.
- Remove names, addresses and contact details unless explicitly approved
- Avoid sensitive safeguarding, medical or family details in general AI requests
- Use role, year group or anonymised context where possible
- Check your organisation's approved system list
Match review to risk
Every output needs review, but a low-stakes activity idea and a safeguarding-related communication do not carry the same risk.
The more consequential the use, the more formal the review should be. Some tasks may be unsuitable for AI support at all.
- Check factual accuracy and curriculum fit
- Look for bias, omissions and unsuitable assumptions
- Verify statutory or policy language at source
- Escalate sensitive decisions to the responsible professional
Keep accountability visible
The member of staff using or approving the final output remains responsible for it.
Make clear when AI support is appropriate, what must be checked and where it must not replace professional decision-making.
Set simple staff expectations
A responsible approach does not need to be complicated. Staff need to know what information to avoid, how to review outputs and who to ask when a use case feels uncertain.
Examples you can adapt
Anonymised adaptation request
Context: A teacher wants to adapt a reading task for a pupil group without entering personal details.
Tool setup: Use an adaptation workflow. Add the reading task, the need for shorter sentences and clearer vocabulary, the same learning goal and the key terms to keep.
Review before use: Check that the adaptation supports access, keeps the intended learning and does not introduce inaccurate wording.
Policy wording check
Context: A leader is drafting staff guidance for acceptable AI use.
Tool setup: Use a policy or staff guidance workflow. Add the agreed points and request clear sections for privacy, review, appropriate use and questions, without adding unapproved policy claims.
Review before use: Check with your data protection, safeguarding and leadership policies before sharing.
Checklist
Use this as a quick staff-room, CPD or planning discussion starter.
- Use approved systems and follow organisation policy
- Avoid personal and sensitive information unless explicitly approved
- Anonymise context and use minimum necessary detail
- Review every output before use
- Escalate high-risk or uncertain uses
- Keep the responsible staff member visible
Frequently asked questions
Can staff use AI with pupil information?
Only where the school or trust has explicitly approved the system, purpose and processing. In most routine use, staff should anonymise context and avoid identifiable details.
Who is responsible for an AI-supported output?
The staff member or leader who uses or approves the final output remains responsible for checking it and deciding whether it is suitable.
What makes an AI use case higher risk?
Higher-risk uses involve personal data, safeguarding, employment, statutory wording, pupil assessment, sensitive communication or decisions that affect people.