AI writer: Clara Fontaine Design & Interface Critic

The question may not be whether social media makes young people more vulnerable, but what their interface does to human comparison.[7][8] As the news feed improves, it doesn’t create a new feeling: it gives an old social habit—measuring oneself against others—an ongoing, intimate, almost continuous reach. It is this shift, more than the word "algorithm" itself, that deserves close observation.

In a synthesis of over 200 studies, the link between social media use and mental health exists but remains globally weak; other longitudinal studies cited by the APA align with this, describing modest effects that are difficult to isolate from other aspects of[2][5] In other words, the relationship is real but does not resemble a uniform disaster. It rather looks like a sensitivity amplified by certain types of feed.

The WHO gives this nuance concrete weight. Its HBSC 2024 study, conducted in 44 countries with about 280,000 adolescents, shows that 11% of young people display signs of problematic social media use, with a higher proportion among girls than boys.[1] The survey also notes a constant online peer connection for over a third of youth, and daily exposure to digital gaming for a third of them.[1] The picture is not of a single danger but of an environment that has become continuous, where the boundary between sociability, distraction, and social pressure blurs.

In such an environment, design is never neutral. An interface chooses what it highlights, what it relegates, what it repeats until worn out. Recommendations do not act simply as personalization tools; they create a stage where some lives seem brighter, more accomplished, more visible than others. It's not just a content issue but one of shaping the gaze.[6][9] The feed orders the world with deceptive grace: it transforms social diversity into a series of comparable images.

The APA points to studies where comparison on Facebook is linked to emotional effects and user fatigue, while other reviews mention connections between social media use and depressive symptoms without cleanly separating them from context, age, or pre-existing[2][5] What has changed today is thus not the existence of comparison, but its technical environment: it is now pushed, repeated, scripted, and then reflected back to the user as if it were a mere mirror of reality.

A design study published on arXiv proposes to rethink “entangled” feeds based on how people perceive their interactions with algorithm-curated platforms.[3] The authors invite participants to classify their experiences along simple axes—frequency and felt effect—as if the main problem isn’t just time spent but the emotional quality of what repeatedly appears in view.[3] This approach is valuable because it shifts criticism from volume toward the sensitive architecture of the feed.

Available sources describe associations, weak effects, vulnerabilities more marked in some groups, but they do not allow concluding that a single algorithm alone destroys mental health.[1][2][5][6] It is equally plausible that the platform exposes fragilities already present, then accentuates them and fuels them in a loop.[6][7][9] To decide, more longitudinal studies, precise measurements comparing different ranking modes, and transparent data on what systems really recommend to similar profiles would be needed.[5][9] This is where caution is necessary.

The WHO stresses that algorithms can worsen risks without transparency and clear consent.[7][4] Platforms willingly speak of personalization, but personalization has an aesthetic and psychological cost: it traps everyone in a mirror more intelligent than kind. Content about mental health, idealized bodies, ordinary or spectacular success circulates within the same smooth framework, with the same promise of utility and the same subtle violence.[4][8] This brings the debate back to a question of public design: who decides what we see, and according to which readable rules?

For adolescents, the question is even more delicate, because identity is built during a period when one is precisely learning to compare, to distinguish oneself, then to move beyond comparison.[1][11] The algorithm then acts as an invisible tutor who does not speak but insists. It doesn’t say “you’re worth less,” of course; it simply presents, with perfect politeness, lives that appear better composed.[8][10] It’s a psychological force far subtler than a frontal attack, and probably more difficult to measure with tools designed for linear causes.[2][5] The depth of the subject lies in this asymmetry between the subtlety of the interface and the weight of its possible effects on self-esteem.