If you have a child who uses social media, you have probably had the conversation about content: what's appropriate, what to avoid, which accounts to unfollow. It's a reasonable conversation. But if it's the only conversation you're having about digital life, you are misunderstanding where most of the risk actually lives.
The most consequential thing shaping your child's online experience is not any individual post. It is the recommendation system determining which posts they see, in which order, how many times, and what comes after. A harmful loop—a child caught in an escalating spiral of body image content, extremist videos, or anxiety-inducing material—is almost never caused by one piece of content. It is caused by the system feeding one piece of content into another, learning from each engagement signal, and optimising continuously for more.
You cannot solve an algorithmic problem with content advice alone. Filtering a specific video doesn't change the system that served it. Blocking an account doesn't retrain the recommendation engine. Only addressing the layer that created the problem can actually solve the problem.
Practical Actions for the Algorithmic Layer
Helping your child understand that they are training the system—not just passively receiving it—is foundational. "Every tap teaches the system. So teach it on purpose." Practically: use the "not interested" function deliberately and consistently. Clear watch history periodically. Unfollow accounts that pull toward content that distorts mood or confidence. Deliberately follow content that is useful, joyful, or skill-building—not just to see that content, but to retrain the model. These small acts accumulate into a different feed over time.
The goal is not to remove your child from the system. The goal is to teach them that the system is something they can act on—not just something that acts on them. Moving from passive subject to active participant is the most meaningful shift available to any of us online.
The Conversation to Have
Instead of asking "what are you watching?" (content), try "why do you think this kept appearing?" (algorithm). Instead of "that account is bad for you" (content), try "what would happen to your feed if you clicked on that kind of content more often?" (system). These questions build a different kind of awareness—one that transfers across platforms and scales with age in a way that content restrictions cannot.
The aim is not to raise children who are afraid of digital systems. It is to raise children who understand those systems well enough to make deliberate choices within them. Understanding is the precondition for agency. And agency—not restriction—is the most durable protection available.
