Grow with AI

Minors and AI tools

A practical position, not legal advice, checked before entering the classroom.

Choose the look:
See the styles full size →
Safeguarding

Minors and AI tools

A practical position, not legal advice, checked before entering the classroom:

  • No AI account is created in a minor’s name. Generation happens on the account and device of the adult who leads the program, projected for the whole classroom.
  • No personal data goes into any prompt: no full names, no addresses, no school name, no faces, no recognisable voices.
  • Pieces stay where they were made. Nothing is published outside the classroom without separate written consent, piece by piece.
  • Anyone can ask for their piece not to be played in front of the group, without giving any explanation.
  • The piece belongs to the team that made it. The program takes ownership of nothing.
  • Tool neutrality: tools are named as categories —text, image, audio, video, code—, never as required brands, and nothing has to be bought.
  • There is a named responsible adult, known to the group and to the families, to whom anything uncomfortable can be reported, including something a tool produced.
Can this go into a prompt? Something for the card And the piece, what happens to it? Does it name someone, or say where they live or study? full name, address, the school yes It stays out no Is it a recognisable face or voice? of anyone at all yes It stays out Does someone ask for their piece not to be played? in front of the group yes It is not played no explanation no Will the piece leave the classroom? be published or shared yes Only with written consent piece by piece no It goes on the card with nobody’s name and no address Generated on the Learning Facilitator’s account, projected for the room no It stays in the classroom the piece belongs to the team that made it This is not put to a vote. The group decides the how, never the whether. Something uncomfortable? Tell the responsible adult.
Safeguarding, as a path of decisions. First, what can go into a prompt; then, what happens to the piece. No AI account is created in a minor’s name: everything is generated on the adult’s account, projected for the room. The room agreement decides how AI is used; safeguarding is not up for discussion.

Tools and age: what was verified on 16 and 23 September 2026

Lane C. The program runs on the Learning Facilitator’s account, operated by the adult and projected for the room: no AI account is created in a minor’s name, no age is declared to any provider and no consent instrument is owed to any provider. It is a design decision that predates the review, and the review confirmed it.

What was verified on 16 September 2026 in each provider’s terms: Claude (Anthropic) requires users to be 18 or older for its consumer service, with no route for a minor, and its own offering for educators is built “for educators only”, exactly in the shape of lane C; ElevenLabs requires 18 or older; Character.AI withdrew open-ended chat for under-18s (company announcement, 29 October 2025). The only two routes that provide for someone aged 13 to 17 are the institution-administered ones, Microsoft’s K-12 education tenant and Google school accounts, not a consumer account with a permission form. The rest of the tools reviewed set a floor of 13, with a condition attached for everyone aged 13 to 17.

What was verified on 23 September 2026, in a fresh reading of the terms: D-ID and HeyGen, avatar video generators, require users to be 18 or older and are excluded, just like Claude and ElevenLabs. Grammarly offers no route through a legal representative: only someone old enough to consent by themselves may use it, 13 under most United States law and 16 under European Union law, so below that age no one gets in, not even with the family’s permission. And in Google school accounts, Gemini Notebook (formerly NotebookLM) is available to all ages under the control of the institution’s administration. None of these readings changes lane C.

These conditions change. The dates consulted are 16 and 23 September 2026, each fact with its own, and they are checked again at every tool audit. Tools continue to be named as categories, text, image, audio, video, code, never as required or recommended brands.

Comprehensive AI learning designed for educators, by educators. From awareness to mastery.