When you first notice it, you may be in a kitchen, watching a four-year-old use the serene muscle memory of a person twice their age to swipe up on a tablet. Not a single show is being watched by them. In the traditional sense, they aren’t even observing. A glitchy animation, a fast-paced cartoon, and an odd little man dancing on a beach while wearing a coffee cup helmet are all being shown to them. The child swipes, laughs, and then laughs once more. A server somewhere logs the interaction and determines what happens next.
Adults today still find it difficult to identify this aspect of the internet. It’s not quite TV. It’s not quite advertising, at least not in the sense that we were raised with it. Stranger still, an increasing amount of data indicates that some of its most dependable clients are hardly out of diapers.
A pattern that seems almost too obvious to be true has been identified by researchers looking into youth media exposure. On the same platforms that their parents and older siblings use, children as young as four are receiving algorithmically optimized content, which consists of brief, loud, emotionally charged videos designed to keep them watching. The pipeline is nearly the same. The viewer’s age and possibly their capacity for resistance are the only variables that alter.
This argument has been made for years by Leah Plunkett, a Harvard Law School lecturer who specializes in youth digital privacy, but lately she seems less scholarly about it. According to her, social media behemoths like YouTube and TikTok are more akin to Hollywood than a neighborhood bulletin board, but they lack the labor regulations, content standards, and gatekeepers that Hollywood eventually had to adopt. There’s a feeling that the regulations just never materialized.
Since the term “persuasive algorithm” has become so widely used that it has lost its impact, it is worthwhile to take some time to understand what it truly means in this context. These systems do more than just make suggestions. They send alerts, nudge, autoplay, and discreetly track how long a tiny face remains focused on the screen before turning away. The threshold for what is considered engaging drastically decreases when the face is four years old. A color that flashes is effective. It works to use a nonsense word. It’s embarrassing how well an AI-generated ballerina with a cappuccino cup head performs in a repeating loop.
The hard sell backfires, as content marketers have long realized. A viewer’s defenses rise as soon as they feel duped. Therefore, the current strategy, particularly on social media, is to hide the pitch behind something amusing, strange, or sentimental. This is already unethical when it comes to adults. When it comes to preschoolers, it ceases to be marketing and becomes something completely different because they lack the cognitive skills necessary to distinguish between a sponsorship and a story. As of yet, there isn’t a clear term for it.
This is often discovered by parents in brief, confusing moments. A young child humming a tune that no one in the household has ever heard. A preschooler requests a specific brand. When the tablet is taken away, the child becomes agitated in a manner that differs from how they responded to books or blocks. For some time now, pediatricians have been discreetly comparing notes on this. It’s difficult to ignore the fact that the discussions sound more like worries about behavior than screen time.
It may take years for the legal framework to catch up. AI-generated content is growing more quickly than any regulator can keep up with, and Section 230 continues to shield platforms from most accountability for what their users upload. Although states are gradually enacting child labor laws for child influencers, the audience side—the four-year-old being observed, assisted, and guided—remains largely unaddressed.
As all of this is happening, a small, unsettling question remains. We debated what kids should and shouldn’t watch on television for a century. The TV now observes, learns, and makes decisions. Furthermore, it’s unclear exactly who should be in charge.⁖※
