📺 Recommendation Algorithms
4th Grade 🤖 Artificial Intelligence
When you open a video or music app, a “Recommended for you” list is waiting for you. That list is picked by a recommendation algorithm — a program that predicts what you’ll probably want to watch or listen to, based on what you did before and what people like you did.
What is a recommendation algorithm?
An algorithm is a set of instructions a computer follows. A recommendation algorithm takes in information about you and about the content on a site, and out of millions of options picks just a few to show you. Without it, you’d have to search on your own through an ocean of videos, songs, games and products.
Recommendations are almost everywhere: on video sites, music services, online stores that suggest “People who bought this also bought…,” social networks and even game apps. Today, many of these algorithms use artificial intelligence.
How does it know what you like?
The algorithm collects clues from everything you do: what you clicked on, how long you watched, whether you watched to the end or skipped after two seconds, what you “liked,” what you searched for and even what time of day you watch. All of these are data.
One method is to look for similar content: if you watched lots of soccer videos, you’ll probably like more soccer videos. Another method is to look for similar people: if thousands of kids who loved the same three songs as you also loved a fourth song, the algorithm will predict you’ll love it too. Most systems combine several methods.
The algorithm doesn’t “know” you the way a friend does. It doesn’t know how you feel or why you watched a video — it only spots patterns in numbers.
Maps and navigation apps sometimes use a similar idea: an app that recommends a restaurant or a nearby park chooses based on places many people liked, on ratings and on what you searched for before. Here too, what shows up first on the list isn’t necessarily the best — it’s what the algorithm predicted you’d choose.
Why is it so hard to stop?
Many companies have a clear goal: to keep you in the app as long as possible, because that’s how they make money — for example, from ads. So the algorithm learns what keeps you watching and serves you more and more of it. Autoplaying the next video and endless scrolling make stopping even harder.
That doesn’t mean you’re weak — it means the system is built very well to grab attention. Once you know that, it’s actually easier to notice and stop.
The filter bubble
If the algorithm shows you only what you already liked, you might get stuck inside a “bubble” — seeing the same kind of content and the same opinions all the time, and missing new and interesting things. This is called a filter bubble.
Also, the algorithm doesn’t always know whether a video is true or right for your age — it mostly focuses on whether people watch it. Companies have rules and filters, but they don’t catch everything. So sometimes scary, untrue or inappropriate things show up in your recommendations. If you see something like that, don’t keep watching — tell an adult, and you can also click “Not interested” or report it.
How to take control
A few simple tips: decide ahead of time how long you’ll watch, and turn off autoplay if you can. Search for new topics on your own now and then, to break out of the bubble. Use “Not interested” when you’re shown something you didn’t want. And ask yourself every so often: “Am I choosing what to watch, or is the app choosing for me?”
Recommendations can be really useful — that’s how you discover a new song you love or a great series. The goal isn’t to fear them or give them up, but to use them wisely: notice when they help you and when they’re just pulling you into scrolling some more.
Key facts
- A recommendation algorithm predicts what you’ll want to watch or listen to.
- It learns from clicks, watch time, skips and “likes.”
- One method is to recommend what people similar to you liked.
- Many apps are built to keep you in them as long as possible.
- A filter bubble is when you see only content similar to what you already liked.
Quick quiz
What does a recommendation algorithm do?
Which information helps the algorithm know what you like?
What does “People who liked this also liked…” mean?
Why do many companies want you to keep watching?
What is a filter bubble?
Does a recommendation algorithm always check that a video is true and right for your age?
What should you do when a scary or inappropriate video shows up in your recommendations?
What helps you get out of a filter bubble?
💭 Think and talk about it
- Look at the “Recommended for you” list in an app you know. What do you think the algorithm “thinks” about you?
- When did a recommendation help you discover something good, and when did it make you waste time?
- What rules would you set for yourself so that you choose what to watch, and not the app?
Questions kids ask
Is the app listening to me through the microphone to make recommendations?
Most recommendations are built from what you do in the app — clicks, views and searches — and from data about similar people. That’s already enough to predict very well. Either way, it’s a good idea to check with an adult which permissions you’ve given each app.
Why does my brother see different recommendations in the same app?
Because the algorithm tailors recommendations to each user based on what they watched and clicked. Two different people get different lists.
Can I reset my recommendations?
In many apps you can delete your watch history or click “Not interested,” and the recommendations change. Ask an adult to help you find the settings.