Log in to track your lesson progress and completion.
Log inLearning Goal: Understand how recommendation algorithms build a profile of your child within minutes and shape what they see.
Your child opens TikTok. Within 40 minutes, the algorithm has figured out what makes them tick. A Mozilla Foundation study found that after just 40 minutes of interaction, TikTok locks in on a user's niche — and from that point, 80% of the videos in their feed reinforce that same theme. Even if the user stops engaging with that content, the algorithm keeps serving it.
Let that sink in. Your child is not choosing what they see. An algorithm is choosing for them — based on a profile built from their behavior, assembled faster than most people can finish a TV episode.
How does this work?
Every app with a recommendation engine — TikTok, YouTube, Instagram, Snapchat, Netflix — collects behavioral data about your child. Not just what they click on. How long they pause on each video. Whether they scroll past quickly or watch again. What time of day they are most active. Whether they share content or save it. Their typing patterns. Their browsing behavior.
This data feeds a machine learning model that builds a personalized profile. The model's goal is simple: predict what will keep your child engaged for the longest possible time. It does not care about their wellbeing, their sleep, or their homework. It cares about engagement, because engagement is what generates ad revenue.
The Center for Humane Technology describes ByteDance — TikTok's parent company — not as a social media company, but as a "persuasive AI company." The algorithm is the product. Your child's attention is what is being sold.
Here is why this matters for parents: your child is not in a neutral information environment. They are in a highly optimized persuasion machine. The content they see is not random. It is curated specifically to keep them scrolling. And it gets better at this job every day, because the more data it collects, the more accurately it predicts what will hook them.
By age 13, the average U.S. child has accumulated approximately 72 million data points that shape the ads, content, and recommendations they receive. Those data points follow them across platforms and into adulthood.
This is not about paranoia. It is about literacy. You cannot help your child navigate a system you do not understand.
Exercise: Open the "For You" or recommendation feed on one of your child's apps (or your own). Scroll for five minutes. Notice how the content shifts based on what you pause on. Try pausing on a topic you would not normally watch — then see how the feed changes. This is the algorithm learning in real time.
Key Takeaway: Recommendation algorithms build a profile of your child in minutes and serve content designed to maximize engagement — not wellbeing. Understanding this system is the first step to helping your child navigate it.