For well-designed AI tutoring services in 2026, the answer is yes — but only if four questions check out. This guide is for UK parents weighing whether an AI tutor is safe for a Year 7–9 child, and how to verify the claim before paying.

What "safe" actually means

There are three dimensions, and a service has to pass all three:

  1. Data safety — what happens to your child's conversations and account data
  2. Content safety — what the AI says back, especially under pressure
  3. Safeguarding — how the AI responds when a child shares distress or harm signals

A service strong on one dimension and weak on the others is not safe overall. Score all three before signing up.

The four questions every UK parent should ask

1. Where is conversation data stored, and for how long?

UK GDPR requires a published lawful basis and a retention period. Ask the provider:

  • Where are the servers (UK / EU / US)?
  • What's the retention period? (Anything beyond 30 days needs justification.)
  • Can you delete data on demand?

aitutors.me stores conversations in the EU region and deletes them on a 30-day rolling window by default.

2. Will my child's conversations train AI models?

This is the single biggest privacy question. The privacy policy should contain an explicit "no training on user data" clause. If it's hedged ("we may use anonymised data to improve our service"), assume yes.

3. What happens on a safeguarding indicator?

The right answer: the tutor stops, surfaces Childline (0800 1111), ends the session, and (the safest services) notifies the parent. The wrong answer: "the AI tries to help" — AI is not a counsellor and shouldn't pretend to be.

4. Does the service require parental supervision?

For under-13s, the parent-account-only model is the cleanest GDPR posture. The parent holds the account; the child uses the parent's session. No child is the data subject directly.

Verification checklist

Before paying:

  • ☐ Privacy policy names a data controller and a lawful basis
  • ☐ Explicit "no training" clause
  • ☐ Safeguarding policy linked from the footer
  • ☐ UK ICO registration (search the ICO data protection register)
  • ☐ Cookieless analytics (no cookie banner needed = smaller data surface)

Why "safe" ≠ "harmless"

A safe AI tutor refuses harmful things — including doing your child's homework. Anti-cheating is part of safety, because educational outcomes are why you're paying. Professor Pi at aitutors.me uses a 4-level hint ladder and never gives the final answer.

FAQ

Is AI tutoring safe for under-13s in the UK?

It depends on the service. UK GDPR makes lawful processing of under-13 data complex. Parent-account-only services like aitutors.me sidestep this — the parent owns the account and supervises sessions.

Will my child be exposed to inappropriate content?

AI tutors built on frontier models (Claude, GPT-4-class) have strict safety policies built in. Risk is dramatically lower than open chatbots. Verify the safeguarding policy is published before signing up.

Are conversations used to train AI models?

It varies. The right answer is a clear "no" in the privacy policy. aitutors.me does not use conversations for training.

Written by Jason — founder of aitutors.me, built for his Year 8 daughter, now shared with other UK families. Updated 20 May 2026.