Your child doesn’t have to tell a social-media platform what they like.
They show it.
They watch one video twice.
They swipe away from another after two seconds.
They pause on a photograph.
They search for a celebrity.
They follow a fitness account.
They like a video about dieting.
They read the comments.
They watch another.
Individually, those actions can seem meaningless.
Together, they can tell a platform an enormous amount about what captures a child’s attention.
And then something important happens.
The platform can use what it has learned to decide what the child sees next.
That is the basic idea behind many recommendation algorithms. They help transform social media from a library of content a user chooses to visit into a personalized stream of content selected specifically for that user.
For parents, understanding that distinction is increasingly important.
Your child’s social-media feed isn’t simply showing them the internet.
It may be showing them their internet.
What Is an Algorithm?
The word “algorithm” makes this subject sound much more complicated than it needs to be.
At its simplest, an algorithm is a set of instructions a computer uses to make a decision or solve a problem.
Social-media companies can use algorithms for many different purposes.
One might help identify spam.
Another might determine which advertisements to display.
Another might decide which posts appear first.
And recommendation algorithms can help determine what content a user is most likely to watch, click, like, share or otherwise engage with.
Think of it this way:
Imagine someone watched your child use social media for a month.
They wrote down every video your child finished.
Every video they skipped.
Every account they followed.
Every topic they searched.
Every post they liked.
Every creator they returned to.
Every time they stopped scrolling.
At the end of the month, that person would probably know quite a bit about what interests your child.
Technology can perform versions of that observation continuously and at enormous scale.
Your Child Doesn’t Have to Click “Like”
This is one of the most important things for families to understand.
Children may believe that a platform learns about them only when they actively provide information.
I told the app my birthday.
I followed this account.
I liked this post.
But behavior itself can provide information.
A platform may potentially learn from signals such as:
- what content a user watches;
- how long they watch it;
- what they skip;
- what they replay;
- what they search for;
- which accounts they follow;
- what they like or share;
- what they comment on;
- what they click;
- and how they interact with recommendations.
A child doesn’t necessarily need to press a button saying:
“Show me more of this.”
Their behavior can communicate that preference.
Why Do Platforms Personalize Content?
Because personalization can make a product much more compelling.
Suppose you opened a social-media application and half the content involved subjects you found completely boring.
You probably wouldn’t stay long.
Now imagine nearly every few swipes produced something connected to your interests.
A tennis highlight.
A comedian you like.
A recipe you might actually make.
A musician you follow.
A news story about something you searched yesterday.
That experience is considerably harder to leave.
Personalization can be genuinely useful.
It helps people discover communities, creators, entertainment and information they enjoy.
The concern isn’t that personalization exists.
The concern is what can happen when powerful recommendation systems are applied to children—and particularly when the signals they detect involve a child’s vulnerabilities.
One Video Can Become a Pattern
Imagine a 13-year-old watches a video about getting in shape.
There is nothing inherently concerning about that.
The child watches another.
Then a “what I eat in a day” video.
Then a transformation video.
Then an extreme diet.
Then content focused heavily on body weight.
The child didn’t sit down one afternoon and consciously decide:
I would like to spend the next two hours consuming increasingly intense content about weight and food.
The progression can happen one recommendation at a time.
Not every platform will make that exact sequence of recommendations, and not every child who watches fitness content will encounter harmful material.
But the example illustrates why recommendation systems matter.
The question isn’t only:
What did the child search for?
It can also be:
What did the platform decide to show the child after that search?
A Recommendation Can Feel Like a Choice
This distinction is subtle.
When a child actively searches for something, parents intuitively understand that the child made a choice.
Recommendation feeds blur that line.
A video appears.
The child watches it.
Another related video appears.
The child watches that.
Soon, the user may be traveling through a subject largely by choosing among options the platform has already placed in front of them.
The child is still making choices.
But the platform may be choosing the menu.
That is why discussions about children’s social-media use increasingly involve not just content moderation, but content recommendation.
The Algorithm Doesn’t Need to Understand Why
There is another important limitation.
A recommendation system may be very good at identifying what holds someone’s attention without understanding why that content is holding it.
A child might repeatedly watch videos about body image because the videos make them feel worse about themselves.
A teenager might linger on frightening content because it makes them anxious.
Someone may repeatedly view posts connected to sadness, loneliness or self-harm because they are struggling emotionally.
Attention does not necessarily mean enjoyment.
Engagement does not necessarily mean something is healthy.
A system optimized to predict what someone will continue watching may detect the behavior without understanding the human experience behind it.
That distinction becomes particularly important when the user is a child.
Injured and Have Questions?
Call us or request a case review by our experts.
When Curiosity Becomes a Rabbit Hole
Children are curious.
That is usually a wonderful thing.
But online curiosity can move quickly.
A teenager wonders whether they are overweight.
They search.
A child is nervous about something happening at school.
They search.
A teenager has a bad day and looks for content about depression.
They search.
The first results may be relatively ordinary.
But recommendation systems can potentially continue feeding a user’s demonstrated interest.
This is where parents sometimes describe a “rabbit hole.”
The child does not necessarily encounter one shocking piece of content.
Instead, their online environment gradually changes.
A feed that once contained dancing, soccer and jokes may begin containing large amounts of material related to one emotionally charged subject.
That shift can be difficult for a parent to see because the parent’s feed looks completely different.
Your Feed Is Not Your Child’s Feed
This may be one of the biggest misunderstandings in modern parenting.
A parent downloads the same app their child uses.
They scroll through it.
Everything looks relatively harmless.
Recipes.
Dogs.
Sports.
Comedy.
Travel.
So the parent thinks:
What’s everyone worried about?
But recommendation systems personalize experiences.
A 45-year-old parent’s feed may tell you very little about what appears on a 14-year-old child’s screen.
Even two teenagers in the same classroom can have radically different feeds.
One may see basketball highlights and comedy.
Another may see makeup tutorials and music.
Another may receive repeated content about weight loss.
Another may encounter increasingly dark material involving depression or self-harm.
They are technically using the same product.
They may not be experiencing the same product.
Body Image and Eating-Disorder Content
Adolescence has always involved insecurity about appearance.
Social media can add several new dimensions to that experience.
Children may see carefully selected photographs, edited images, filters, fitness transformations, dieting content and endless comparisons with other users.
Recommendation systems can potentially magnify that experience by learning that appearance-related content captures a particular user’s attention.
The issue isn’t simply that an unhealthy post exists somewhere online.
It is whether a vulnerable young user can be repeatedly exposed to related content after demonstrating interest in the subject.
Parents should pay attention when a child’s feed becomes unusually dominated by:
- extreme dieting;
- rapid weight-loss claims;
- obsessive calorie counting;
- “thinspiration” or similar appearance-focused content;
- repeated body comparisons;
- extreme exercise;
- or content that appears to glorify disordered eating.
The presence of one video isn’t necessarily meaningful.
Patterns matter.
Self-Harm and Emotionally Distressing Content
The same concern can arise with content involving sadness, depression, self-harm or suicide.
Young people sometimes seek online communities because they are struggling and want to know that someone else understands what they are experiencing.
Supportive communities can be valuable.
But not every community or piece of content is supportive.
Material may romanticize suffering, normalize dangerous behavior or expose a vulnerable child to increasingly disturbing content.
Parents should avoid assuming that simply seeing a particular subject means a child is engaging in dangerous behavior.
But repeated exposure deserves attention.
If you discover concerning material, the first question should not automatically be:
“Why are you looking at this?”
Try:
“I’ve noticed you’re seeing a lot of this. How does it make you feel?”
That may open a very different conversation.
The Algorithm Can Learn Faster Than a Parent
Parents often know their children extraordinarily well.
But children don’t tell parents everything.
They don’t announce every insecurity.
They don’t report every crush.
They don’t describe every argument at school.
They may not even have words for something they are beginning to feel.
Online behavior can reveal interests and vulnerabilities before a child ever discusses them aloud.
A recommendation system may detect that the child suddenly watches a certain category of content much longer than before.
That does not mean a computer “knows” your child better than you do.
It means it has access to a different category of information.
Parents know the person.
Platforms can know the behavior.
Those are very different forms of knowledge.
What Can Parents Actually Do?
Parents cannot realistically inspect every video a teenager watches.
Nor should every family try to turn a child’s digital life into constant surveillance.
There are more useful approaches.
Ask Your Child to Show You Their Feed
Don’t begin with an interrogation.
Ask them to scroll with you.
“What does this app think you’re interested in?”
You may learn a lot.
Make it an exercise in curiosity rather than a surprise inspection.
Teach Children That Recommendations Are Chosen
Young users can easily experience a feed as simply “what’s on TikTok” or “what Instagram is showing today.”
Teach them that what they see may be selected partly because of what they have previously done.
That creates an important moment of awareness:
This isn’t just appearing. Something is deciding to put it here.
Use “Not Interested” and Similar Controls
Many platforms allow users to tell recommendation systems that they do not want certain content.
Teach children to use those controls deliberately rather than simply scrolling past disturbing material.
Reset When Necessary
Some platforms provide ways to reset or influence recommendations, manage interests or review content preferences.
If a child’s feed has become dominated by content they no longer want, explore the platform’s current controls together.
Don’t Punish Curiosity
If your child searched for something concerning, try to understand why.
A harsh reaction can teach a child to hide future searches rather than stop making them.
Pay Attention to Sudden Changes
If a child’s online environment changes alongside major changes in mood, sleep, eating, school performance or offline activities, don’t dismiss that combination.
The feed may not have caused the change.
But it may tell you something about what your child is experiencing.
Talk to Kids About the Business Model
Older children can understand a surprisingly sophisticated idea:
Your attention has value.
Many digital products make money when people spend time using them, interact with content or see advertising.
That does not automatically make the product bad.
But it means the company and the user may not always have exactly the same objective.
Your child may want to watch three videos and go to sleep.
The platform may benefit if they watch thirty.
Helping teenagers understand that tension turns digital safety into something more empowering than a list of restrictions.
They can begin asking:
Why am I seeing this?
Why did this notification arrive now?
Why does another video start immediately?
Why is this app so difficult to close?
Those are excellent questions.
Ask a Better Question Than “How Much Screen Time?”
Parents understandably worry about hours.
How long were you on TikTok?
How much time did you spend scrolling?
But time is only part of the picture.
Two children can spend the same hour online and experience entirely different things.
One watches comedy and soccer.
Another spends an hour being shown content that intensifies an insecurity they were already struggling with.
The clock says both children had sixty minutes of screen time.
That number tells us almost nothing about what happened during those sixty minutes.
As technology becomes more personalized, parents need to ask a better question:
What is the technology learning about my child—and what is it doing with what it learns?
Understanding that relationship may be one of the most important pieces of digital literacy families can develop.
When Algorithmic Recommendations Contribute to Serious Harm
For some families, concerns about recommendation algorithms are not theoretical.
A child may have been repeatedly shown content involving self-harm, eating disorders, dangerous challenges, extreme body-image material or other subjects associated with serious physical or emotional harm. Parents may discover only afterward that the child’s online experience looked dramatically different from what they believed the platform contained.
When that happens, important questions may arise.
What information did the platform collect about the young user?
What signals did its recommendation system identify?
What content did it repeatedly recommend?
Did the company know that certain recommendation patterns could create risks for children?
What safeguards existed for young users, and did they work as intended?
Those questions are increasingly relevant in litigation involving social-media platforms and harm to children.
At Rafferty Domnick Cunningham & Yaffa, our attorneys represent children and families in cases involving social media, online platforms and serious preventable harm. Our work examines not only individual pieces of content, but how digital products are designed to learn from young users, personalize their experiences and keep them engaged.
If you believe algorithmic recommendations or other features of a social-media platform contributed to serious harm to your child, contact Rafferty Domnick Cunningham & Yaffa to discuss what happened and learn more about your family’s legal options.
