Trust Centre · Responsible AI
AI that assists, people who decide.
Every AI feature has a clear purpose, uses only the data it needs, is labelled when it contributes, and leaves the final decision with a named person.
Where AI is allowed
Green, amber and red, decided in advance
Every AI use falls into one lane, agreed with your school before anything is switched on.
AI can help
- Draft lessons and quizzes from your own material
- Translate and simplify messages
- Summarise long documents
- Suggest timetable solutions
A person must approve
- Suggested marks and feedback
- Messages to families
- Report-card comments
- Attendance follow-up drafts
Never allowed
- Final grades on its own
- Admissions or disciplinary decisions
- Safeguarding or health judgements
- Profiling children for advertising
Six commitments
Assistive by design, accountable by default
Assistive purpose
AI prepares, organises, translates and spots patterns; people stay in control.
Human authority
A named person reviews output wherever judgement or a learner is affected.
Minimum context
Each task uses only the information its purpose requires.
Age and role sensitivity
What AI may do depends on who it serves—children get more care, not less.
Explainability that fits the use
You can see when AI contributed and check the sources behind it.
Evaluation over assumption
Quality, safety and fairness are measured before and after release.
Prohibited and high-risk uses
Some decisions are never handed to a model
Always
- Draft material for a person to review
- Organise repetitive work
- Highlight patterns for trained people to interpret
Never
- Make a high-stakes decision about a child, employee or school on its own
- Profile children or use their data for marketing
- Make clinical, safeguarding or disciplinary determinations
- Work with unrestricted model or data access
Correction and monitoring
Oversight that continues after launch
Every AI use has an owner, a review cadence and a way to pause it.
Approve the use
Purpose, owner, data and risk are recorded before it's switched on.
Governance ownerA person decidesReview output
Consequential output is approved by a named person.
Teacher or leaderA person decidesCorrect
People correct and flag output; corrections are kept.
Anyone using itEvaluate
Quality, fairness and drift are checked; a use can be paused or withdrawn.
Governance ownerA person decides
Independent assurance
AI assurance, stated honestly
We publish only what is independently verified. Ask us anything your review needs.
No autonomous high-stakes decisions
No AI feature makes a high-stakes decision about a child, employee or school on its own.
Drafts stay drafts
AI output remains a draft wherever professional judgement or a learner is affected.
Your policy first
Prohibited uses and escalation routes are agreed with your school before a feature is switched on.
Evaluating School Niva?
Ask the questions your review needs answered.
Security, privacy, AI or accessibility—we'll answer in writing.