Why this programme needs a parent present
AI tools are powerful, interesting, and genuinely useful — and they were not designed for children. They can produce incorrect information confidently. They can drift into topics that are not appropriate. They respond to whatever is asked of them. None of this makes AI dangerous, but it does make adult supervision essential at this age.
The good news: when a parent is present, AI sessions become some of the best learning conversations you can have with a child. You notice things together. You question things together. You laugh at the AI's mistakes together. That is the whole point.
The 5 rules — and why each one matters
Rule 1 — Never share personal information
No real name, school name, age, city, address, phone number, or photos — ever. This applies even when the AI asks. AI systems log conversations and are sometimes used to train models. Personal information shared in a chat is not private in the way a conversation with a person is.
Rule 2 — AI makes mistakes. Always check.
AI sounds very confident when it is completely wrong. This is the most important thing a child can learn about AI at any age. Every session includes a deliberate "find the mistake" activity for exactly this reason. Model the behaviour yourself: when AI says something, say out loud "I wonder if that's actually right — let's check."
Rule 3 — AI is a tool, not a friend
Children can develop parasocial relationships with AI faster than adults. The AI is always friendly, always available, always agreeable — which is exactly the opposite of a real relationship. Be clear and consistent: "It is a very clever tool. Like a calculator, but for words and ideas. It does not know you. It does not care about you. That is what we are here for."
Rule 4 — Don't use AI to do your homework
This one is about the child's development, not safety. Using AI to get answers bypasses the learning process — the struggle, the thinking, the understanding. The sessions here always use AI to understand something, not to get an answer. Reinforce the distinction: "We ask the AI to explain it, then you tell me what you understood — in your own words."
Rule 5 — If it feels wrong, close the tab. No questions.
This is the most important operational rule. If at any point — now or in the future — your child is using any AI tool and something feels uncomfortable, wrong, or upsetting: close the tab immediately. No explanation needed. And when they tell you, do not interrogate them about what they saw or why they felt that way. Just thank them for telling you and move on. The goal is to make telling you feel safe.
Account setup — what to use
✅ Recommended
- Use your own parent account
- ChatGPT free tier (no sign-in needed for basic use) or Claude.ai
- Sit beside your child — same screen
- Parent types the prompts for younger children (under 8)
- Child types with parent watching for 8–10
- Child types, parent reviews output for 10+
❌ Avoid
- Creating an account in the child's name
- Letting the child use AI alone on their room
- Using AI as a reward or unsupervised entertainment
- Voice AI or AI companions (designed to build emotional attachment)
- Any tool that asks for the child's age to verify it
Conversation starters for before the session
"Before we start — what do you think AI is?"
"Can you think of anything that uses AI that we see every day?"
"If I asked AI what 2+2 is and it said 5 — would you believe it?"
"What's one thing you would NOT tell a stranger on the street? Same rule applies to AI."
Printable safety cards
Print these and stick them near the computer your child uses. One card per rule. Each session reinforces one.
Ready to start? Head to Session 1 of the AI Starter programme.
Start AI Starter →