AI and children
Assistance with a boundary around it
Aurelia uses artificial intelligence carefully and narrowly, to help a young person think, plan and improve their own work. It is never used to generate finished work in a child's name, to replace human judgement, or to profile children for engagement.
Our position
AI should support authorship, never replace it
Every use of AI inside Aurelia is designed around a single principle: the work must remain genuinely the child's own, and everyone reviewing it must be able to tell what help, if any, was used.
Bounded, age-banded
What AI can do for an eight-year-old is deliberately narrower than what it can do for a fifteen-year-old, and both are narrower than general-purpose AI tools.
Authorship labelling
Where AI has contributed to a piece of work, that contribution is labelled clearly, so verification never mistakes assisted work for unassisted work.
Human verification stands
A teacher's judgement, not an AI score, remains the standard by which achievement is verified in the passport.
Bounded, age-banded assistance
Different help at different ages
AI capability inside the Creator Studio is tied to a child's age band and to settings a guardian or school has chosen, rather than offering the same open-ended assistant to every user.
- Younger children see the most limited forms of assistance, focused on prompts, structure and encouragement rather than generated content.
- Older children within the under-16 range may have access to more capable assistance, always within limits set for their age band.
- Assistance is scoped to specific tasks inside a project — it does not operate as an open, general-purpose chat companion.
- AI never initiates contact with a child; it responds only within a project the child has opened.
Authorship labelling
Verification depends on honesty about what helped
Because the Achievement Passport depends on teachers being able to trust what they are verifying, any AI contribution to a project is recorded and shown alongside the work.
A teacher reviewing a submission can see where a child used assistance to structure an idea, check spelling, or explore an alternative approach, distinct from the parts of the work the child produced unaided. This is not a penalty against using assistance — it is what allows verification to remain meaningful.
Labelled, not hidden
Assisted sections of a project remain visible as such throughout drafting, review and, where relevant, verification — they are never quietly folded into the finished piece as if a child produced them alone.
Guardian and school controls
Families and schools decide how much AI is available
AI assistance in Aurelia is opt-in at the family and school level, not a default a parent has to work out how to switch off.
Restrict by setting
Guardians can limit AI assistance to specific tasks or disciplines within the Creator Studio.
Disable entirely
AI assistance can be switched off completely for a child's account, at the guardian's discretion, or by a school across its own accounts.
Data
Children's data does not train third-party models
Aurelia does not send a child's work, conversations or personal data to third-party AI providers for the purpose of training their models.
Where third-party AI capability is used to power in-product assistance, it is configured to process the specific request needed and is not used to build or improve general models on the back of children's content.
Human verification
A person always stands behind a verified achievement
No amount of AI assistance changes the fact that only a named, accountable adult can verify an achievement for the passport. There is no automated approval path.
This is a deliberate limit on what AI is allowed to do inside Aurelia: it can support the making of work, but it never replaces the human judgement that turns that work into a recognised achievement.
See where AI sits inside the wider platform
AI assistance is one part of a much larger safety architecture. Read how the rest of Aurelia is built to protect children.