Your child opens a video, watches for a few seconds, pauses, maybe rewinds, and the feed starts changing. A few swipes later, everything feels strangely on point — same vibe, same topics, same kind of content.
That's not luck. That's the algorithm doing its job.
Most social media platforms run on recommendation systems that learn from behaviour in real time. What your child watches, how long they stay, what they skip — every small action feeds into what shows up next. These systems aren't trying to show what's true or even what's useful. They're trying to show what will keep someone watching.
Research and platform analyses have consistently found that these systems optimise for engagement above all else. The Center for Humane Technology has documented how recommendation engines are designed to maximise time spent, often by leaning into emotionally charged or familiar content. Similarly, the Pew Research Center has shown that teens' online experiences are heavily shaped by algorithmic feeds rather than active search.
Over time, this creates what researchers call a feedback loop.
Are Feedback Loops Dangerous?
The more a child interacts with a certain type of content, the more the platform feeds it back to them. Work by researchers affiliated with institutions like Columbia University has highlighted how recommendation systems can gradually narrow what users see, sometimes pushing them toward more extreme or repetitive content patterns. For adults, this can be subtle. For children, it can be formative.
Instead of exploring broadly, they end up in a tighter loop of similar ideas, trends, and viewpoints. A few clicks can quietly shape what feels normal, interesting, or even true. Over time, this doesn't just influence preferences — it can shape how children process information and form opinions.
Gurdit Singh Chhabda, a digital wellbeing educator who works with Indian parents and schools, explains what is actually happening under the surface:
“The algorithm is running experiments on your child's phone in the background, constantly. It is trying to profile them — what they love, what makes them anxious, what keeps them watching. And then it starts playing on their insecurities.”
He illustrates this with an example that will feel uncomfortably familiar to many parents. If a child repeatedly visits the profile of someone they're infatuated with — or watches content related to something they're privately anxious about — the algorithm registers that pattern and amplifies it. It isn't malicious. It's just doing what it was built to do.
“Is it the algorithm's fault? No. It's a programme designed to show you what you respond to. But that means the responsibility lies with us — to make sure our children are aware of what they're reacting to.”
Enter: Short Format Content
There's also a second layer to this: attention. Short-form content is built for quick hits, where every swipe delivers something new. That constant novelty can make slower tasks feel harder to stick with. Research cited by the American Psychological Association suggests that frequent switching between stimuli can weaken sustained attention, especially in younger users.
This is where everyday concerns around focus start to connect with what's happening online. Difficulty concentrating, restlessness during study time, or a constant pull toward the phone are often brushed off as habits — but they're also shaped by the environments children spend time in.
Is It All Gloom and Doom?
At the same time, it's not all negative. Algorithms can help children discover things they genuinely care about — from useful tutorials to niche interests, and even the occasional study resource that actually helps. The same system that keeps someone scrolling can also introduce them to something meaningful.
The problem is that this system is largely invisible. Unlike screen time, which you can measure or limit, algorithmic influence is harder to spot. A feed feels neutral — like a window into the world — when in reality it's a filtered stream shaped by past behaviour. Research from the OECD shows that many young users don't fully understand how personalised their feeds are.
Teaching Children to Recognise the Loop
Gurdit has developed a practical exercise he runs with students specifically to build algorithmic literacy — an understanding of how their own behaviour shapes what they see, and what they can do about it.
“Give children six or seven healthy topics and ask them to find the relevant hashtags, search the content, find the high-performing posts. Children are already tracking likes and shares naturally. Then the homework is: after three days, what is your feed showing you?”
The point of the exercise isn't to shame children for what they watch. It's to make the invisible visible. Once a child understands that their reactions are literally teaching the algorithm what to serve them next, they start to engage with their feed more deliberately.
He extends this in a more surprising direction: teaching children to create the very content they're being manipulated by.
“We asked children to design two versions of the same reel — one using images relevant to the topic, one using a popular face. The popular face always performs better, and children know it. If a child can create clickbait, they are much less likely to fall for it. Once you understand how the machine works, you can step back from it.”
Every major platform also gives users the ability to tell the algorithm what they don't want to see. Teaching children to actively use that option — to go to a post and select 'don't show me this' — is another tool for moving from passive consumption to intentional engagement.
The Way Forward
That's where digital wellbeing needs to shift. It's not just about how much time children spend on their phones, but what that time is doing to their attention, their interests, and their sense of reality. Helping children recognise that their feed is curated — even lightly — can change how they engage with it.
There are also simple ways to push back against the loop. Encouraging kids to explore different types of content helps widen what the algorithm shows them, while watching together and talking about what they see adds context. Using a parental control app or basic parental control settings can help guide exposure, especially for younger users, without turning everything into a restriction.
Even small changes matter. Screen-free time before bed protects sleep, and breaks during long sessions help reset attention. Over time, these habits support better focus and reduce the pull of constant scrolling.
Algorithms don't just respond to behaviour — they shape it. Left on autopilot, they can quietly narrow what a child sees, what they engage with, and how they think. And when that happens, something designed to personalise content can start to influence beliefs and attention in ways that are much harder to notice, and much harder to reverse.
