Subscribe
,

How to set up safe AI settings for your children and teens

AIooze Editorial

In one minute

  • You will be able to find and switch on the parental controls for the main AI chatbots, and agree a few family rules.
  • You need: your own account with each service your child uses, and your child’s account or device.
  • It takes about 20 to 30 minutes, ideally sitting down together with your child.
  • Controls change often. Each step links to the maker’s own help page so you can check the latest version.

Why it is worth doing

Children and teens are using AI chatbots for homework, questions they would not ask an adult, and company. Most of the big makers now offer some parental controls, but they are scattered across different apps and settings, and they only work if a parent switches them on. Schools are changing too: Google said last week it will turn off its AI tools, and Search with them, for school student accounts until administrators switch them on (AIooze brief).

The steps

  1. Check the age rules first. They differ a lot:
    • Claude: Anthropic says on its help centre: “We require all users to be at least 18 years old to create and use a Claude account.”
    • Microsoft Copilot: at least 13, or older where local rules require it (Microsoft support).
    • Google Gemini: children under 13 can use the Gemini app only if a parent turns it on through Family Link (Gemini help). Google says this is not available for supervised accounts in the UK, the EEA or Switzerland.
    • ChatGPT: OpenAI’s parental controls are built for teen accounts linked to a parent or guardian (OpenAI help).
  2. Talk before you tick boxes. Google’s own advice to parents is a good starting line: “It’s important for you to help your child understand that it’s an AI tool, not a person to confide in.” Explain that chatbots can be wrong, and agree what your child does if something upsetting comes up (stop, and tell you).
  3. ChatGPT: link your accounts. According to OpenAI’s help page, as of 10 Nov 2026, either the parent or the teen can send the invite. In ChatGPT, open Settings, then Parental controls, then Add family member (a teen chooses Invite a parent instead). Once linked, protections that reduce sensitive content are on by default. You can then set quiet hours when ChatGPT cannot be used, turn off voice mode, saved memories and image generation, and stop your teen’s chats being used to improve the models. Linking does not let you read your teen’s conversations, but OpenAI says it may notify you in limited situations involving a serious concern about your teen’s safety.
  4. Gemini: use Family Link. Google’s help page, as of 9 Nov 2026, gives these steps: go to familylink.google.com or open the Family Link app, select your child, then tap Controls, Gemini, Gemini Apps, and switch it on or off. Google emails you when your child first uses it, and you can change your mind at any time.
  5. Copilot: use your family settings. Microsoft says on its support page that parents can use Microsoft Family Safety, or Apple or Google device settings, to block Copilot, set time limits, and choose which features are available, such as image generation, voice mode and AI companions. It also says: “If you’re under 18, your chats aren’t used to train AI models.”
  6. Instagram and other Meta apps: check AI characters and time limits. Meta said in January 2026 that it had paused teens’ access to its AI characters while it builds a new version, and that new parental controls (turning off chats with AI characters, blocking specific ones) will come with it. Teens can still use the Meta AI assistant. Today you can already set daily time limits for the app, which include time with AI.
  7. Ask the school. If your child has a school account, ask which AI tools are switched on and what rules apply. The school, not you, usually controls the settings on a school account.

A worked example

Priya is 14 and uses ChatGPT for homework. Her dad sends her an invite from Parental controls in his own ChatGPT account, and she accepts it on her phone. He sets quiet hours from 9pm to 7am, turns off image generation and stops her chats being used to improve the models, and leaves saved memories on because she finds them useful for revision. They agree three house rules: no full name, address or school name in chats; check anything important with a teacher or a second source; and tell a parent if a chat gets weird or upsetting. It takes 15 minutes.

Watch out for

  • Controls only cover the signed-in account. A child using a friend’s device, a different app or a logged-out browser is not covered. Device-level settings (Family Link, Microsoft Family Safety, Apple’s Screen Time) help close some gaps.
  • Filters are not perfect. Google says so plainly: “Our filters try to limit inappropriate content from appearing, but they’re not perfect.” Keep talking, not just blocking.
  • Chatbots make mistakes. They can state wrong facts confidently. Teach your child to check homework answers and never rely on a chatbot for health, legal or money decisions.
  • Personal details stay out. Agree that names, addresses, school, photos of themselves and friends, and passwords never go into a chatbot.
  • This is general guidance, not professional advice. If you are worried about your child’s wellbeing or how they are using AI, talk to their school, your GP or another professional you trust.
  • Settings, menus and age rules change often; if a step looks different, follow the maker’s current help page.

Go deeper

Sources

Related