In 2017, Facebook's own data scientists found that Instagram made teenage girls feel worse about their bodies. Thirty-two percent of teen girls said so directly. The company shelved the report. Four years later, Frances Haugen walked out with the documents.
Jonathan Haidt published *The Anxious Generation* in 2024 and delivered the curves: adolescent depression up 145 percent among girls since 2010, self-harm ER visits for ten-to-fourteen-year-olds nearly tripled, loneliness rising in every country with smartphone penetration above 80 percent. The camp read those numbers and saw a generation being poisoned by platforms that monetize insecurity.
Then Andrew Przybylski and Amy Orben at Oxford and Cambridge pooled 2.4 million adolescents across forty-two countries. The association between social media use and well-being was negative, significant, and roughly the same magnitude as the association between well-being and eating potatoes. The waved the forest plots and asked: if this is a generational toxin, why does the signal shrink every time you tighten the controls?
Meanwhile a sixteen-year-old in rural Arkansas who had never met another openly queer person found a Tumblr community that told her she was ordinary. She is twenty-six now. She is alive. The do not dispute the self-harm curves. They dispute a framing that treats an entire medium as pathology while erasing the kids it kept breathing.
The watch this argument cycle and point at the thing nobody wants to name: the business model. Attention is the product being sold. The platforms don't cause harm because screens are toxic. They cause harm because the ad model rewards engagement, and engagement runs on outrage, anxiety, and comparison. Fix the incentive and the pathology resolves. Ban the screen and you've treated the thermometer.
The empirical crux has a clock. Several school districts banned smartphones in 2024. Australia restricted social media for under-sixteens in 2025. If those cohorts show measurable improvement by 2028, the camp wins something stronger than correlation. If the curves keep climbing without the phones, the question becomes: what is actually breaking, and why did we spend a decade blaming the screen?
Where do you stand?
Join the conversation →