top of page
Homepage

Behavioral Economics

Algorithms are designed to hold your attention.

Every social media platform runs on the same basic business model: it doesn't sell you a product, it sells your attention to advertisers. The longer you stay on the app, the more ads you see, and the more money the platform makes. 

Social media companies prioritize profit over children's wellbeing. 
 

The business behind your feed

Social platforms are, at their core, advertising companies. Advertisers pay for "impressions," meaning each time an ad is shown to a user, and they seek out platforms that can promise them as many impressions as possible (Georgetown Law Technology Review, 2017). That creates a direct, built-in incentive: the platform's success is measured by how "engaged" users are, where engagement means viewing, liking, commenting, sharing, and saving content, not by whether users feel good afterward.

​

This is the core tension of the attention economy: a platform's revenue goal (keep people scrolling) and a user's wellbeing goal (use the platform in a healthy, balanced way) are not the same goal, and when they conflict, the business model is built to win.

To maximize engagement, the algorithm has to predict, for each individual user, exactly what will make them stop scrolling and interact. It does this by constantly analyzing data most users never think about, including likes, comments, and shares, but also passive signals like how long a post stays on someone's screen before they scroll past it, whose profiles someone visits, what they search for, and even who they message (Georgetown Law Technology Review, 2017). The system then ranks content by predicted appeal and prioritizes the posts most likely to keep that specific user engaged.

​

Layered on top of that prediction engine are design features borrowed directly from behavioral psychology: infinite scroll removes any natural stopping point, autoplay decides for you what comes next, and notifications and unpredictable "rewards" (e.g., a surprising like, a viral comment) work on the same variable-reward principle that makes slot machines hard to walk away from. None of this is accidental. It is engineered, tested, and refined specifically to extend time on the app.

How the algorithm learns what will keep you there

Children pay the price

Because the algorithm's job is to maximize engagement, not wellbeing, there is no incentive to distinguish between content that helps a child and content that harms one, as long as both keep them scrolling. Internal Facebook/Instagram research made public by whistleblower Frances Haugen in 2021 found that 13.5% of teen girls said Instagram made their suicidal thoughts worse, and 17% said it worsened their eating disorders. Haugen told Congress that company leadership "know how to make Facebook and Instagram safer but refuse to because they have put their astronomical profits before people" (NPR, 2021).

​

The U.S. Surgeon General's 2023 advisory on social media and youth mental health backs this up at a population level. By 2022, 95% of teens ages 13 to 17 reported using social media, over a third said they use it "almost constantly," and nearly 40% of children ages 8 to 12 were on social media despite minimum age requirements. Teens who spend more than three hours a day on social media face roughly double the risk of depression and anxiety symptoms, with adolescent girls and teens already struggling with their mental health especially vulnerable. The advisory also points to something specific to childhood: ages 10 to 19 are a highly sensitive period for brain development, when identity, emotional regulation, and impulse control are still forming, which makes frequent, algorithm-driven use potentially more disruptive for kids than adults (Yale Medicine, 2023).

​

Lawmakers have started responding directly to the algorithm itself, not just to screen time. New York's SAFE for Kids Act, finalized in 2026, requires platforms to turn off algorithmically personalized feeds for minors by default, showing them only chronological posts from accounts they already follow, unless a parent consents otherwise. The state's rationale is explicit: these feeds are "designed to encourage a user to continue to use and return to a platform," often by surfacing content from accounts the user never chose to follow (New York Attorney General, 2026).

​

In August 2026, a court case revealed that Meta knew its own features, like infinite scroll and "like" counts, worked against users' well-being, but kept them anyway because they boosted engagement. The case, backed by 29 states, ended with Meta agreeing to pay up to $18 billion and change how it builds products for kids. It's proof that the patterns this site talks about aren't accidental. They're deliberate choices, made by real companies weighing profit against people's mental health, and they're finally starting to have real consequences.

​

Engagement-driven algorithms also change what kind of world a user sees, not just how much time they spend seeing it. Because the system's core strategy is to show people more of what they've already responded to, it narrows rather than broadens what reaches them over time. Researchers Cinelli et al. found that recommendation algorithms limit users' "selection processes" by constantly suggesting content similar to what they've already engaged with; a user who shows interest in conservative sources gets pushed further right, and a user who shows interest in liberal sources gets pushed further left, gradually shifting the range of content each person sees away from the center (Cinelli et al., 2021, PNAS; explained further at the YIP Institute). For children still forming their understanding of the world, growing up inside a feed that is quietly and continuously narrowing is a very different experience than growing up with broad, diverse exposure to people and ideas.

A narrower view of the world

Who's trying to fix this

The people writing these algorithms work for the companies profiting from them, which raises an uncomfortable question: can a system be trusted to regulate itself when its incentives point the other way? As one legal analysis put it, "so long as algorithms are written by humans bound by contractual obligation to companies, flaws and biases will remain" (Georgetown Law Technology Review, 2017). That's part of why momentum has been building around outside oversight, including calls for algorithmic transparency and independent audits, laws like New York's SAFE for Kids Act and the federal Kids Online Safety Act, and ongoing debate over whether the fix requires more regulation, more independent (even AI-assisted) auditing of these systems, or, most likely, both.

​

Edit/update: Companies cannot be held responsible for regulating themselves. government now assuming the role to regulate--prioritizing children's wellbeing.

Sources
 

What This Means for You

bottom of page