How We Use AI

1. Why we are publishing this

U Teach Mi teaches computing to primary-aged children, and artificial intelligence is now part of how many educational products are built. Parents, carers and schools are entitled to know exactly where it sits in ours: what it touches, what it does not, and who is accountable for what a child ends up seeing.

This page is written to be read by a parent, not a lawyer. It sits alongside our Safeguarding Children Policy, Privacy Policy and Terms of Service, and does not replace them.

2. Our curriculum is human-authored

This is the most important thing on this page, so we will be direct about it.

The U Teach Mi curriculum was designed by our founder, a teacher with years of experience teaching computing to primary-aged children. The sequence of courses, the progression ladders that carry a child from Year 1 through to Year 6, the concepts introduced at each stage, the order they are introduced in, and the judgement about what a seven-year-old is ready for and what should wait until they are ten — all of that is professional teaching expertise, built from real classroom practice.

It is deliberately aligned to the UK National Curriculum for Computing, so that what a child does with us reinforces and extends what their school is required to teach, rather than running alongside it unconnected.

AI did not devise our curriculum, and we would not ask it to. Deciding how a child learns to think computationally is a teaching judgement. It is the part of this business that cannot be automated, and it is the part we are least willing to hand over.

3. Where AI is actually used

With that established, here is where AI genuinely does the work. We use it in four areas, all of them downstream of the teaching decisions above.

Building the activities

Our lessons are delivered as interactive games, simulations and 3D worlds. AI assists our team in building those activities — writing the software, generating artwork and characters, and producing the audio and video that present a lesson. The activity is the vehicle; the teaching inside it is ours.

Supporting children as they learn

Some features respond to a child directly, such as our in-platform tutor and feedback on the projects children build. These are designed to prompt, hint and encourage — the same role a teaching assistant plays when circulating a classroom. They are scoped to the lesson the child is working on.

We do not use AI to grade a child, to form judgements about their ability or potential, or to make any decision that affects their access to the platform. Progress is tracked against the concrete activities a child has completed.

Marketing and social media

Video clips promoting the platform on social media are recorded from real gameplay and narrated using an AI-generated voice. No child appears in this content, and no child's work, name, image or data is used in it.

Running the business

We use AI for internal operational tasks such as summarising our own records and drafting routine communications. Where such a tool can read information relating to children or families, that access is restricted and read-only, and a member of staff reviews anything before it is acted on or sent.

4. Where we deliberately do not use AI

To state the boundaries plainly, we do not use AI to:

• Devise or sequence our curriculum, or set learning objectives.
• Make safeguarding decisions. Concerns are always handled by our Designated Safeguarding Lead, as set out in our Safeguarding Policy.
• Grade, rank, profile or assess a child's ability.
• Make decisions about admissions, membership, billing or a child's access to the platform.
• Replace a qualified adult in any of our in-person clubs or camps.
• Communicate with a child in an open-ended way outside the boundaries of a lesson.
• Generate content depicting weapons or combat within children's customisable characters and rewards.

5. How we check AI output before a child sees it

We treat AI as capable but fallible, and we design on the assumption that it can get things wrong. Our checks do not rely on the AI policing itself.

• Filtering runs independently of the AI.Where a feature draws on outside material, that material is checked against our own safety rules both before and after any AI processing — so an AI mistake is caught by a check the AI does not control.
• Children's creations are moderated by an adult. Anything a child makes and chooses to share is reviewed by our team before another child can see it, and shared work is stripped of identifying details.
• Creative features are constrained by design. Where children build games or characters, they configure options we have defined rather than producing unrestricted content, and free text between children is not part of these features.
• Lesson content is reviewed by a teacher. AI-assisted material that forms part of a lesson is checked against the learning objective it is supposed to serve before it is published.

6. Children's data and AI

Our full commitments on personal data are in our Privacy Policy. Three points are specific to AI:

• We do not sell or share children's personal data with AI companies for their own purposes.
• Children's work is not used to train external AI models. We use established AI providers under business terms, and we do not submit children's personal information for the training of third-party models.
• We send the minimum necessary.Where a feature needs to send something to an AI provider to work, it sends what that feature requires — not a child's wider profile or history.

7. Teaching children about AI

Children growing up now will use these tools throughout their lives. Our position is that they are better served by understanding them than by being kept away from them.

Where AI appears in our teaching, the aim is for children to understand that it is a tool built by people, that it can be confidently wrong, and that their own thinking is the thing being developed. We are explicit with children that a computer helping them is not the same as a computer knowing better than them.

8. Accountability

U Teach Mi Ltd is responsible for everything published on this platform, including material produced with AI assistance. “The AI generated it” is not an explanation we would offer for something that should not have reached a child, and it is not one we would accept.

Our Designated Safeguarding Lead holds overall responsibility for the safety of children using the platform, including in relation to AI-assisted features.

9. Questions, concerns and corrections

If you have a question about how AI is used in something your child has seen, or you believe an AI-assisted feature has produced something inappropriate, inaccurate or unfair, please tell us. We would rather hear about it.

10. Review

This page is reviewed at least annually, and sooner where our use of AI changes materially, new features are introduced, guidance or legislation changes, or feedback highlights something we should make clearer.