Are AI chatbots safe for children?

Mainstream tools used with the right age account and clear rules can be useful. AI "companion" apps marketed as friends or partners are higher risk and are best avoided.

Will my child cheat on homework with AI?

It is a real risk. Ask to see how AI helped them, and check your school's policy. Most UK schools now teach acceptable AI use openly.

What if my child tells a chatbot about self-harm?

Take it seriously, speak to them calmly, and contact your GP or call NHS 111. Childline 0800 1111 is also there. Call 999 if there is immediate danger.

Ages 10-17

Your Child's First AI Chatbot Use

A UK pathway for introducing AI chatbots like ChatGPT and Gemini with realistic limits and safety conversations.

AI chatbots are now everywhere, from homework apps to social platforms. They can be useful, but they can also produce confident-sounding misinformation, age-inappropriate content and unhealthy emotional bonds.

The goal is informed use, not a ban. Children who understand the limits of AI use it more safely.

Readiness signs

Look for these before saying yes

✓Understands that AI can be wrong and make things up.
✓Will not paste personal details into online tools.
✓Can describe AI as a tool, not a real friend.
✓Talks to you about new apps they want to try.
✓Understands their school's rules on AI for homework.

Parent checklist

Step 1

Choose a mainstream tool (ChatGPT, Gemini, Copilot) over unknown apps.

Step 2

Confirm the tool's age limit (often 13 or 18) and create the account together.

Step 3

Turn off chat history or training where the option exists.

Step 4

Talk through your child's school policy on AI in homework.

Step 5

Block or restrict AI "companion" apps designed to be a friend or partner.

Step 6

Set device-level age ratings so unknown AI apps cannot be installed.

Step 7

Agree which subjects AI can help with and which it cannot.

Family agreement points

•I will not share my name, school, address or photo with a chatbot.
•I will check important answers in another source.
•I will not use AI as a replacement for talking to a real person about feelings.
•I will tell you if a chatbot says anything that scares or upsets me.
•I will follow my school's rules on AI in coursework.

What to say

Phrases that help

"AI can be helpful, but it does not know you and it is not your friend."
"If a chatbot tells you something disturbing, screenshot it and show me."
"Sharing your secrets with a chatbot is not private, the company can read them."
"Using AI to write your whole homework is cheating yourself and your teachers."
"If anything online ever encourages you to hurt yourself, we stop and talk straight away."

Settings to review

•Account age set correctly on the AI tool.
•Chat history and training data sharing switched off.
•Image generation restricted or off for younger users.
•Voice mode disabled or supervised on younger accounts.
•App Store / Play Store filters blocking 17+ AI apps.
•School-managed device policies for AI tools respected.

Red flags

Hours-long conversations with an AI "companion" character.
Secrecy about which AI apps your child uses.
Strong emotional reactions when AI access is paused.
Disclosures of self-harm or relationship problems made only to a chatbot.

Review in 30 days

Come back to these questions

→Look at the chat history together if available.
→Discuss anything the AI got wrong or made up.
→Check no new AI "friend" apps have been installed.
→Talk about how AI was used for school work this month.
→Adjust which tools are allowed based on the conversation.

Frequently Asked Questions

Last reviewed: 2026-05-16Next review: 2026-11-16Reviewed against: UK safeguarding practice

This is practical educational content to support families. For case-specific concerns about a child's safety, contact the NSPCC helpline on 0808 800 5000 or your local safeguarding team.