VEDI Global School

Safety & AI

What the AI does — and what it’s never allowed to do

VEDI teaches children with AI. That comes with real responsibilities — and with lines we’ve decided never to cross. Both are set out here, plainly.

The lines we never cross

These aren’t aspirations. They’re built into the software.

No reading emotions from faces or voices

VEDI never analyses a child’s camera or microphone to guess how they’re feeling — not for mood, not for engagement, not for anything. This is prohibited for schools under European law, and we agree with the law.

No labels on children

Children are never tagged as anxious, withdrawn or difficult. Our notes record what happened — “shared openly today”, “was quiet today” — never a verdict on who a child is.

A human approves everything

No grade reaches a parent or a record without a human Director having reviewed and agreed with it.

Always told it’s AI

The AI teacher is labelled as AI everywhere she appears. No child is ever led to believe they’re talking to a person.

Your child’s data

We collect only what teaching your child actually needs, we never sell it, and we never share it for advertising. It’s kept in secure, access-controlled systems, and you can see, correct or delete it at any time.

The full detail — exactly what we hold, why, and for how long — is in our Privacy Policy.

Keeping children safe

Every group session is led by a person — there is always an adult present, never software alone. Children work in small, settled squads that stay together for weeks, not open rooms that strangers can wander into.

A human mentor meets your child one to one every week, and is the person who notices when something is off. And anyone — a child or a parent — can raise a problem or a worry from any screen, at any time. If you ever have a concern about your child’s safety, a real person answers it.

Where we stand on AI rules

VEDI is built to the EU AI Act — among the strictest rules in the world on using AI with children — and we hold to it everywhere we teach, not only where the law requires it.

In practice that means: we never use AI to infer a child’s emotions from their face or voice; the AI is always clearly labelled as AI; and a qualified human stays responsible for the teaching, the grades and the decisions. Where we’re ever unsure whether something is allowed, we simply don’t do it — we would rather under-claim than over-claim when it comes to children.