Skip to content

Guide for parents

How to Teach Your Kids to Spot Deepfakes and AI Fakes

Updated

By . Founder of ThinkaBit: a dad based in the Netherlands who has worked in tech for 25 years and now works with AI.

A few years ago, “don’t believe everything you read online” covered most of it. Today a child scrolling through videos can meet a cartoon character singing a song it never sang, a “news clip” of a flood that never happened, a photo of a celebrity at a party they never went to, and a voice message that sounds exactly like a friend. Some of it is harmless fun. Some of it is made to fool people, sell them something or upset them.

This guide brings together everything we’ve written about helping kids aged 7–12 spot deepfakes and other AI fakes. It’s meant as a starting point: read it once for the big picture, then follow the links to the posts that go deeper on each part.

What kids need to understand first

Before any clue or trick, children need one simple idea: pictures, videos and voices can now be made by a computer, and they can look and sound real. A deepfake is a video or recording where someone’s face or voice has been swapped or made up by AI. An AI-generated image is a picture that no camera ever took.

You don’t need to explain the technology in detail. It’s enough to say that AI learned from millions of real photos and recordings, so it can make new ones that look like them. If your child already understands that AI is a pattern-guesser (our guide on how to explain AI to a child shows how to get there), this step is easy: the same guessing that finishes a sentence can also “finish” a face.

Two things are worth saying out loud, because children often assume the opposite:

  • Looking real is not proof. A clear, sharp, professional-looking video can still be fake.
  • Being fooled is normal. Adults get fooled too, including people who work in tech. The goal is not to never be fooled; it’s to check before believing or sharing.

The clues: useful now, less useful later

There are visual clues that still catch a lot of fakes, and kids enjoy hunting for them. In faces and videos: odd blinking, eyes that don’t line up, blurry or shimmering edges where the face meets the hair or glasses, and lips that don’t match the words. In still images: hands with the wrong number of fingers, writing that turns into nonsense, backgrounds that melt, lighting and shadows that don’t agree, and skin that’s too smooth.

How to teach your kid to spot a deepfake covers the five clues for video, and how to spot AI-generated images has seven for pictures, with a “Real or Rendered?” game you can play at home.

Be honest with your child about one thing, though: the clues are getting weaker. Each new generation of AI tools fixes some of the mistakes, and some fakes have none of them. If a child learns only a checklist, they’ll eventually meet a fake that passes it and feel fooled by the checklist too. That’s why the clues are the fun part, not the whole lesson.

The habit that keeps working: pause and check

What lasts is a habit, not a list. Teach your child to pause when something online makes them feel a strong reaction (shock, excitement, anger, fear) and ask three questions:

  1. Who made this? A person, a channel, a company? Can I tell?
  2. Why did they make it? To inform, to entertain, to sell, to get clicks, to trick?
  3. Is there proof from somewhere I trust? Does a known news source, an official page or a grown-up I trust say the same?

Strong feelings are a signal, not a reason. Fakes are often designed to make people share before they think, so the moment a child feels “I have to show everyone this!” is exactly the moment to slow down.

The best tool to pair with the pause is reverse image search: searching with a picture instead of words to find where it first appeared. A “photo from today” that turns up on websites from ten years ago is solved in seconds, and kids love feeling like detectives. Both posts linked above show how to do it.

Talking about it without scaring them

Children take their cue from us. If every fake becomes a lecture about how dangerous the internet is, many kids stop telling their parents what they see. A calmer approach works better: start with what your child already watches, ask questions instead of giving answers, and treat a fake you find together as an interesting puzzle rather than a disaster.

How to talk to your child about AI-generated videos gives you a short script for that first conversation: what to say about how these videos are made, three questions to teach instead of a checklist, and what to do when a video is about someone they actually know, like a classmate or a teacher.

If your child has already been fooled by something, the most useful response is curiosity, not blame. “What made it look real?” teaches more than “You should have known.” The goal is that next time, they come to you.

When fakes are used to scam people

Not every fake is a funny video. AI makes scams more convincing: a voice message that sounds like a family member asking for help, a “friend” in a game asking for a password, a message saying they’ve won a prize. For children, the same rules apply as with any fake, plus a few that are specific to scams: urgency is a warning sign, a request to keep something secret from parents is a red flag, and personal information (passwords, addresses, school names, photos) is never something to give to someone online.

One of the most practical steps a family can take is to agree on a code word: a word only your family knows, which anyone calling or messaging “from the family” in an emergency must say. Teaching kids online safety in the age of AI scams explains how to set one up and covers the five rules that still work against AI-powered scams.

Make it a habit, not a one-off talk

One conversation won’t do it. What works is small and often: a minute at breakfast with a picture that might be fake, a “would you share this?” question in the car, a pause together when something surprising pops up on the TV. Over weeks, the pause stops being your rule and becomes your child’s own habit.

A few ideas to try:

  • Make it a game. Take turns showing each other pictures and guessing “real or AI?”. Score points for explaining why, not just for being right.
  • Say it when you check. When you look something up before believing it, say so: “That sounds surprising, let me see who else is reporting it.” Children copy what they see.
  • Praise the pause. When your child stops to ask “is this real?”, that’s the win, even if the answer turns out to be “yes”.

How ThinkaBit helps

ThinkaBit is an app for kids aged 7–12 that turns this into short games. In the Deepfake Den world, children practise spotting made-up pictures and messages, and the daily Fake or Fact game keeps the habit going. Glitch, the app’s trickster, slips fakes in on purpose, so children get used to checking in a safe place before they meet the real thing. There are no ads, no tracking, and children never chat with an AI in the app.

You can also try our free two-minute Spot the Fake quiz together, right in the browser, with no sign-up.

Where to start this week

If you only do three things:

  1. Have the first talk, using the script in how to talk to your child about AI-generated videos.
  2. Teach the pause: who made it, why, and is there proof?
  3. Agree on a family code word for calls and messages that claim to be urgent.

Then come back to the posts below whenever you want to go deeper. For trusted organisations near you that help families with online safety, see our resources page.

Practise it with ThinkaBit

Short, playful lessons for kids aged 7–12 on how AI works, spotting fakes and staying safe online. No ads, no tracking, free to start.

Share: