Regulating Social Media Requires Intervening in Its Algorithms

The public discussion surrounding the protection of minors online has fixated on an intuitive yet insufficient solution: outright bans. Governments worldwide are vying to establish age restrictions for social media use, proposing laws to erect digital barriers at 14, 16, or even 18 years old. However, concentrating all efforts on barring access or refining identity verification systems is not enough.

“

Regulating social media platforms demands, above all, profound changes to their algorithms.

To protect the mental health and development of new generations, the debate's focus must shift. Regulating social media platforms demands, above all, profound changes to their algorithms. Tech giants have perpetuated a convenient narrative, presenting their platforms as neutral showcases or passive tools. Under this logic, any harmful use by a teenager is attributed almost exclusively to individual lack of self-control or insufficient family supervision. This premise is a fallacy.

Social media platforms are not neutral containers. They operate as digital environments governed by artificial intelligence systems with a precise commercial objective: maximizing user time and engagement to monetize advertising. When this precision engineering confronts the developing brain of a minor, the relationship is profoundly unequal. On one side are systems designed to capture and sustain attention; on the other, a developmental stage particularly vulnerable to such stimuli. As warned in Wired, a recommendation system can begin systematically suggesting content related to suicide and self-harm to a minor within just forty minutes.

“

Social media platforms are not neutral containers. They operate as digital environments governed by artificial intelligence systems with a precise commercial objective: maximizing user time and engagement to monetize advertising.

This risk must be understood in light of adolescent brain maturation. The prefrontal cortex, responsible for self-evaluation, impulse control, and long-term planning, does not fully develop until after age twenty. In contrast, the limbic system, associated with emotions and the pursuit of immediate rewards, is highly active. Recommendation algorithms exploit this physiological vulnerability with immense precision. By constantly analyzing behavioral signals, how long a finger hovers over an image, which videos are replayed, or when a user logs in, the machine personalizes content to keep minors trapped in a loop of immediate gratification.

“

When this precision engineering confronts the developing brain of a minor, the relationship is profoundly unequal.

This mechanism mimics the logic of intermittent rewards, similar to the gambling industry. Each scroll down acts like a slot machine lever, potentially offering a visually appealing stimulus that triggers a dopamine rush. Demanding that a 13-year-old act with "responsibility" against a system designed by top engineers to prevent them from closing the app is conceptually cruel. The problem lies not with the screen, but with the invisible architecture operating behind it. Algorithmic harm extends beyond lost sleep, study, or social time. Its deepest risk emerges when recommendation systems identify content that generates the most reaction: outrage, morbid curiosity, vulnerability, and negative emotions often produce more interaction than calm, reflection, or learning. If a minor experiencing low self-esteem or sadness pauses longer on posts about miracle diets or melancholic videos, the algorithm interprets this as a signal to provide more such content.

“

Demanding that a 13-year-old act with "responsibility" against a system designed by top engineers to prevent them from closing the app is conceptually cruel.