AI for parents: age limits and parental controls for ChatGPT, Gemini and more

From what age children may use which AI

Each provider sets its own age limits, and they differ considerably.

ChatGPT: According to OpenAI's European terms of use, you must be at least 13. Anyone under 18 needs their parents' permission. The German consumer advice centres point out that age is usually not verified.

Claude: Anthropic requires all users to be at least 18.

Gemini: According to Google, you may manage your own Google account in Germany from 16. For younger children, parents can create a supervised account via Family Link. However, Gemini apps are not available for such accounts in the European Economic Area. Anyone who wants to use Gemini with a paid Google AI subscription must be at least 18 in the EU.

Snapchat with its built-in chatbot My AI: According to Snapchat, the minimum age is generally 13. According to the parents' guide Medienkindersicher, children under 16 in Germany need their parents' consent under the General Data Protection Regulation.

Meta AI: According to Meta, the assistant has been built into WhatsApp, Instagram, Facebook and Messenger in Europe since March 2025. The consumer advice centres state that Meta, like ChatGPT, can be used from 13 with parental consent.

Parental controls in ChatGPT

OpenAI offers parental controls. Parents and teens link their accounts: one of them sends an invitation by email or text message from the parental controls section of the settings, and the other accepts it.

Afterwards, according to OpenAI, parents can among other things reduce sensitive content, set quiet hours, switch off voice mode and image generation, turn off memory of earlier conversations, stop chats being used for training and set study hours in which new chats start in study mode.

Important: parents see no conversations, no history and no live activity. Notifications are only sent in a few serious cases, for example when specially trained reviewers identify a serious risk of self-harm. Even then they contain no conversation transcripts.

Also note: according to OpenAI, your child can unlink the accounts at any time. You are notified, but afterwards you can no longer manage their settings and no longer receive safety notifications for that account.

If ChatGPT concludes from account details, an age check or automatic age prediction that someone is under 18, it automatically switches on a teen version with age-appropriate protections and break reminders, according to OpenAI.

Snapchat, Meta and Gemini: what parents can set

Snapchat: In Family Center, parents can see who their child is friends with and who they have messaged in the last seven days, but not what was said. According to Snapchat, this is also where you can stop My AI from replying to your child. Parents need their own Snapchat account, and the child must accept the invitation. According to Medienkindersicher, Snapchat also has a chatbot from Perplexity, which you can block in Family Center as well.

Meta: In January 2026, Meta temporarily stopped teens' access to its AI characters worldwide in order to build a new version with parental controls. According to Meta, teens can still use the general Meta assistant, with age-appropriate safeguards. Since April 2026, according to Meta, parents supervising a Teen Account can see which topics their child has asked Meta AI about in the last seven days, but not the conversations themselves. It launched first in the US, the UK, Australia, Canada and Brazil, and Meta announced a global rollout. Since 5 October 2026, according to Meta, parental supervision is also available worldwide in the standalone Meta AI app. There, parents can additionally choose a stricter setting that further limits what their teen can talk about.

Gemini: Because Gemini is not enabled for supervised child accounts in the European Economic Area, there are no Gemini parental settings for them here either. Google itself writes that its filters are not perfect and children may come across inappropriate content.

What else to talk about with your child

Technology alone is not enough. The German online safety initiative klicksafe published a parents' guide on AI chatbots and so-called AI companions in early 2026. Its key message: parents should actively get to know chatbots and try them out together with their children. Children should learn to name their feelings and, when worried, seek help from trusted people rather than from an AI.

Google gives parents similar advice: explain to your child that a chatbot is not a person and not a friend to share secrets with. They should not enter their address, school name or family information. And AI can be wrong and present false things as facts.

Our assessment: age limits are usually only checked by self-declaration. An open conversation about what your child uses AI for therefore often protects more than any setting.