Meta AI – The AI assistant in WhatsApp, Instagram and Facebook

Meta AI is a new digital assistant from Meta, the company behind Facebook, Instagram and WhatsApp. Without a separate app, young people can now access artificial intelligence directly via chat or search – for learning, chatting or collecting ideas. The question for parents is: how does it work – and is it safe?

In brief

• AI directly in popular apps(WhatsApp, Instagram, Facebook)
• Minimum age according to GTC: 13 years (without age check)
• Problematic: data protection, emotional attachment, misinformation
• Free of charge, but Meta uses data for training, personalization and advertising

How does Meta AI work?

Meta AI has been officially available in Germany since March 2025. A blue circle with a sparkling pattern signals access to AI-supported chats – directly in WhatsApp, Instagram, Facebook or Messenger. A separate app is not required. The text-based assistant responds to questions, gives tips and suggests content. Meta AI is also integrated into the search bar on Instagram and Facebook. This can lead to users interacting with the AI unintentionally.

• WhatsApp and Messenger: With “@MetaAI”, the AI can be activated in individual or group chats. It helps with planning, researching, texting and improving.
• Instagram and Facebook: Questions can be asked directly via the search bar or as a direct message. The AI then suggests posts, hashtags or content – based on existing data and probabilities.
• Restricted in Europe: Functions such as image generation or creative design tools are deactivated in the EU – due to the stricter requirements of the European AI Act.
• No EU training data according to Meta: Meta assures that no private chats or personal data of minors from Europe will be used for AI training – provided that they correctly state their age and are recognized as minors.

What fascinates children and young people about it?

Many young people experience Meta AI as practical support in everyday life. The AI is available exactly where they are anyway – on WhatsApp or Instagram. It provides quick answers to school questions, helps with translations, writes texts or makes suggestions for posts and content. The AI is friendly, approachable and helpful, almost like a conversation partner.

Especially in comparison to more complex AI offerings such as ChatGPT or Perplexity AI – which specify sources or require special user knowledge – Meta AI has a much lower-threshold and more familiar effect because it appears directly in the familiar Allatgs apps. Many young people also try out Meta AI because they talk about it with their friends or share content.

What can be problematic?

• MisinformationMeta AI often sounds convincing, but can give factually incorrect or one-sided answers. The tool is based on training data that contains biases or does not cover certain topics at all. In such cases, the AI may simply invent (“hallucinate”) content.
• Data protectionMeta uses publicly available content from Facebook and Instagram as well as user behavior to improve AI – also in Europe.
• PrivacyAccording to Meta, private chats and data of minors should not be used for AI training. It remains unclear what other data is actually used.
• Opt-out required: If you do not want your own public posts to be used for AI training, you must actively object – via a web form on Facebook or Instagram. Important: The opt-out only applies to future content. Data that has already been used cannot be deleted retrospectively.
• Emotional closeness: The AI imitates human conversations. Some young people might mistake them for real friends.
• Problematic content: Despite protective filters, Meta AI may address sensitive or problematic content, for example on sexuality, eating behavior or psychological problems.
• Cultural differences: Meta AI was predominantly trained with English-language content. Some answers do not fit well with the German or European context.
• Data protection in the EU: data protection experts accuse Meta of circumventing European laws. Complaints to supervisory authorities are under review.
• Personalisierte Werbung: Laut Meta können Interaktionen mit der KI auch zur Personalisierung von Werbung genutzt werden.

What does the provider think?

Meta emphasizes that no private messages or data from minors in Europe are used for training. According to Meta, public content is accessed within the framework of applicable data protection laws. In the EU, users must actively object if they do not want their content to be used for AI training. However, data protectionists criticize the fact that there is no active consent.

What should parents pay attention to?

Parents can best protect and empower their children if they stay in conversation and reflect together on how AI works.

• Start a conversation: Ask your child openly whether they use Meta AI, what they do with it and what they like about it. Show interest without being controlling.

A minimum age of 13 years is required to use Meta AI, or even 16 depending on the app, but there is no automatic age check. Talk to your child about these age limits too

• Strengthen critical thinkingEncourage your child to question statements made by AI. Review content together and talk about uncertainties, mistakes and the right way to use it, i.e. also about do’s and don’ts when dealing with AI.
• Check privacy settings together: Go through the settings with your child. Also think about the objection to AI training.
• Deactivate or hide Meta AI: In some apps, the Meta AI icon can be muted or hidden. This can prevent unintentional or accidental use.
• Emphasize human contact: Remind your child that AI does not replace real relationships. Feelings, worries or important topics belong in real conversations with family, friends, confidants or professional counselors.