Kids use ChatGPT, Character.AI and Snapchat AI — here's how to guide them

What your child actually uses

Teenagers rarely say "artificial intelligence". They say ChatGPT like it is a classmate, "CAI" for roleplay, "the bot on Snapchat" for late-night chats. Most AI products students touch run on large language models — the same technology that powers OpenAI, Google and Anthropic.

All four are available in EU countries. ChatGPT runs on a free tier and a paid ChatGPT Plus plan for around €20 a month. Google's equivalent AI plan costs about the same. Character.AI is free with an optional paid tier; Snapchat's extra features cost a few euros a month. In practice, most children use the free versions.

Meta AI is another piece of the puzzle: it is slowly arriving inside Instagram and WhatsApp in Europe after a delayed start over regulatory questions. That means your teenager will soon meet an AI assistant inside apps where they already share photos and messages.

What parents actually worry about

Ask any group of European parents and three concerns come up repeatedly. Homework outsourcing: the child copies, pastes and stops thinking. Emotional dependence: a bot that is always kind becomes the friend that never disappoints. And exposure: chatbots can confidently invent facts, repeat stereotypes or produce sexual and violent content despite filters.

These worries are not theoretical. In October 2024, a US family sued Character.AI after a 14-year-old's death, alleging the platform encouraged him to harm himself. The company responded with additional safety features for minors in 2025, including stricter filters on romantic and self-harm topics. Those events changed how the industry treats teenage users — but built-in caution still has a long way to go.

What European law says about children and chatbots

Europe is a special case for one practical reason: the age at which a child can legally agree to data processing in an app is not the same everywhere. Under GDPR Article 8, EU member states chose a threshold between 13 and 16:

What this means in practice: the same app may legally need your consent for a 14-year-old in Germany or the Netherlands, while a 14-year-old in Spain, Austria or Italy can consent alone — and in Belgium, Denmark or Poland the line sits even lower. If your child uses an AI service, it is worth checking the provider's policy for your country. If the company gets it wrong, it faces GDPR fines, not just bad PR.

Two more EU rules matter. Since 1 August 2024 the AI Act is in force; from 2 February 2025 it bans AI systems that exploit a person's vulnerability due to age. A "therapist bot" designed to press a depressed teenager into emotional dependence would sit squarely in that danger zone. From August 2026 chatbots will also be obliged to clearly tell users they are interacting with a machine. And the AI Act's AI literacy duty applies to schools that use the technology — as a parent, you can ask your child's school how it teaches AI safety.

One underused European right: GDPR data access. As a parent (or holder of parental responsibility), you can ask an AI provider for a copy of all data related to your child's account — including chat history in many cases. Companies must answer within one month, usually free of charge.

What you can set up this weekend

• Check the age rule for your country against the table above and the provider's terms.
• Set the phone to switch off at night. Google Family Link and Apple Screen Time both allow a "bedtime" during which apps shut down. Most problem chatbot use happens between 11 pm and 2 am.
• Open Snapchat Family Center — Snapchat's guide explains how to see who your teen talks to and when (message content stays private).
• Turn off ChatGPT history in the app settings, or use the account together on a shared family device.
• Enable Character.AI's stricter safety mode if your child uses it; under-18 accounts should already have the tighter defaults.
• Keep devices in common rooms and make the rule about the behaviour ("no secret conversations after lights out"), not about AI as such.

If something goes wrong, use the app's report/block function, keep a screenshot and contact the provider. Your national Safer Internet Centre offers free advice in your language.

How to talk about it without losing the room

Banning chatbots rarely works: your child will just use a friend's account. It is more effective to make curiosity an everyday habit. Try these openers:

• "Show me something your bot told you this week — anything, good or weird."
• "What do you think happens to what you type there?" — a natural entry into privacy.
• "Would you tell me if a bot said something that scared you?"
• "A chatbot isn't a friend, but it is designed to sound like one. How do you tell the difference?"

Two points deserve repeating, even to children who have heard them before. First, a chatbot does not know you; it predicts the next word. It can be wrong, biased and self-assured at the same time. Second, a chatbot will never punish you for quitting — real relationships need a human on the other side. If your child ever reads advice that suggests hurting themselves or others, the rule is simple: close the app, tell an adult. The European child helpline 116 111 is free in most EU countries and works in local languages.

Watching for serious signs matters too: late-night secrecy, a sudden drop in real-world friendships, a "friend" without a name, or heavy reliance on a bot for emotional support. None of these make you a bad parent. They are an invitation to talk, not to confiscate.

Three habits that help more than any app setting

First, use a chatbot in front of your child occasionally. When you ask a dumb question and get a confident wrong answer, you have just taught the most important lesson: AI is useful, and AI is not an authority. Second, agree on one family rule together — for example, "no AI in exams or homework unless the teacher allows it" — and write it down. Third, if your child is already ahead of you, let them teach you. The conversation that starts with "show me how it works" is the one that stays open.

If you want to build your own AI awareness before the school meeting, our AI courses for absolute beginners walk through the same tools your children use — the honest way.