Skip to content
← All posts

· 6 min read

Why YouTube Keeps Showing Your Kid the Same Videos

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

Share:

Your child watches one video about slime. Then another one starts by itself. Twenty minutes later, every single thumbnail on the screen is slime: bigger slime, crunchier slime, slime that glows in the dark.

Nobody at YouTube sat down and decided your kid likes slime. A computer did, in a fraction of a second, and it will do it again tomorrow with whatever they watch next. That’s worth understanding, both for you and for them, because the “next video” is where a lot of your child’s screen time actually comes from.

Here’s how it works, in plain words, and how to explain it at the kitchen table.

The “next video” is a guess

Every time your child finishes a video (or scrolls past one), the app has to answer one question: what should we show next?

It answers by guessing. It looks at what your child just watched, how long they watched it, what they skipped, what they tapped on and what other people who watched the same thing went on to watch. Then it picks the videos most likely to keep them watching.

The set of steps a computer follows to make that guess has a name: an algorithm. An algorithm is just a list of instructions, like a recipe. “If they watched slime to the end, show more slime” is a tiny, made-up example of one. The real ones are far bigger and are changed all the time, but the idea is the same.

The important part for kids: the app isn’t picking what’s good for you. It’s picking what you’re likely to keep watching. Those are often not the same thing.

How to explain it to a 7-year-old

Kids get this quickly if you make it physical. Try this:

“Imagine a helper who stands next to you while you watch. Every time you watch something to the end, the helper writes it down. Then the helper runs off and brings back more of the same, because they want you to stay as long as possible. The helper never asks ‘Is this good for you?’ They only ask ‘Will you keep watching?’”

Then ask your child: “What would the helper bring you if you watched three cat videos?” They’ll laugh and say “more cats”. That’s the whole lesson. They’ve just explained a recommendation system.

For older kids (9–12), you can add the next step: the helper also learns from everyone else. If lots of kids who watched slime then watched a “giant slime challenge”, the helper guesses your child will want it too, even if they’ve never searched for it.

Why it matters more than it looks

A guess that keeps showing more of the same isn’t dangerous by itself. Most of the time it just means a lot of slime. But it has a few side effects worth knowing about.

It narrows things. Children’s interests are wide and change fast. A feed that only serves “more of what you watched” can make a world feel smaller than it is.

It rewards whatever grabs attention. Videos with loud thumbnails, shocking titles or “you won’t believe this” hooks tend to get watched, so they tend to get recommended. That’s also exactly how AI-made fake videos and scams spread: they’re built to make you click first and think later.

It makes “stop” harder. When the next video starts by itself and it’s picked to be hard to resist, stopping takes effort. That isn’t a failure of willpower in your child. It’s what the design is for.

It doesn’t know your child. It knows what they did, not what they meant. One curious click on something odd can tilt the next twenty suggestions. That’s why a child can end up somewhere strange without ever searching for it.

None of this means the app is evil or that YouTube must go. It means your child should know there’s a guessing machine on the other side of the screen, so they can notice when it’s steering.

A 5-minute game: “Be the algorithm”

This works at dinner, in the car, anywhere.

  1. You name three videos you “watched”: a dog on a skateboard, a dog catching a frisbee, a cat falling off a sofa.
  2. Your child plays the algorithm and picks what to show you next, and why.
  3. Swap roles. Now they name three, and you guess.
  4. Ask the twist question: “What would you show me if you wanted me to watch for as long as possible?”

Kids usually discover on their own that the “keep watching” answer is the most exciting, most surprising video, not the most useful one. That moment, when they see the choice the machine is making, is worth more than any rule you could give them.

Settings that put you back in charge

You don’t need to understand the algorithm to tilt it in your favour. A few settings help a lot:

  • Turn off autoplay. In both YouTube and YouTube Kids you can stop the next video from starting by itself. That one change turns “what’s next?” back into a choice.
  • Pause or clear watch history. History is what the guesses are built on. Pausing it means new videos don’t feed the recommendations; clearing it gives the feed a fresh start after a strange run.
  • In YouTube Kids, use the parent controls. Pick the age setting that fits your child, consider turning off search, or allow only videos and channels you’ve approved.
  • Watch the first few minutes together of any new channel your child likes. You’ll learn quickly what the feed is serving.
  • Use the “not interested” option with your child when something odd shows up, and say out loud what you’re doing: “I’m telling the helper not to bring this again.”

Settings move around between app versions, so if you can’t find one, search the app’s help pages for its name.

Teach the question, not just the rule

Rules like “no more than an hour” are fine, but they stay yours. The skill that lasts is a question your child asks on their own: “Why is this video being shown to me?”

A child who asks that can notice when they’ve been watching the same thing for too long, can spot a thumbnail that’s built to hook them, and can recognise when something unfamiliar has crept into the feed. It’s the same habit that helps them spot fakes and understand what AI is: look behind the screen, and ask who made this and why.

How ThinkaBit teaches it

This is the idea behind Algorithm Alley, the first world in ThinkaBit. Children aged 7–12 start with what an algorithm is (steps in order, like a recipe) and work up to how apps use examples to guess what you’ll like. They learn it by teaching Sparky, our robot, who only knows what your child shows it: give Sparky lopsided examples, and Sparky makes lopsided guesses. It’s the recommendation feed, seen from the inside.

There’s no chatbot and no generative AI in the app, no ads and no tracking, and lessons take a few minutes a day. ThinkaBit is free to download on the App Store and Google Play, and the first unit of every world is free.

The next time the screen fills up with slime, your child might be the one who says it first: “The helper thinks I only like slime.” That’s a kid who knows who’s choosing.

Share: