Social Media Is Radicalizing Young People. Left and Right.
May 22, 2026 · Frisian News
New research from the University of Michigan shows that algorithmic feeds push teenagers toward extreme political content, regardless of where they start. The platforms profit from engagement, not truth.
The study tracked teenagers aged 13 to 17 for six months, following their social media consumption across TikTok, Instagram, and YouTube. Researchers found that most of them, whether they began with moderate left-wing or right-wing content, ended up viewing material at the political extremes within weeks. The algorithm showed them more of what kept them watching. No one sat still. The platforms made money either way.
TikTok's recommendation engine proved most aggressive. Teenagers who watched a single video about climate activism found themselves seeing accelerationist green content, then eco-fascist material, before long. The same happened to kids who watched a mildly conservative economics video. The algorithm did not care about their starting point. It cared about their time on screen.
The research contradicts the common claim that social media merely reflects existing political divides. The platforms actively widen them. One participant told researchers that she started watching videos about body positivity and ended up in communities that blamed capitalism, the government, and men for nearly every problem she faced. She had not sought out extreme content. The algorithm delivered it to her.
Industry representatives claim that removing the profit motive would kill innovation, but that argument ignores what we know: these companies employ thousands of engineers whose sole job is to maximize watch time. They have the tools to slow radicalization. They choose not to use them because slower feeds mean lower ad revenue. Meta, TikTok, and Google have known about this for years. Internal documents show it.
Parents and schools bear some blame for ignoring the problem, but the real power sits with the platforms. They could limit recommendations to content outside a user's current filter bubble. They could slow feeds. They could be transparent about their algorithms. None of these steps happen because the business model depends on keeping young people angry and engaged. Until that changes, expect more teenagers sliding from curiosity into conviction.
Published May 22, 2026 · Frisian News