A test TikTok feed went from 13% to 56% misogynistic content in 5 days

A controlled UCL/Kent audit found a TikTok feed set up as a teenage boy climbed from 13% to 56% misogynistic recommended videos within five days, evidence that the recommender amplifies the content rather than the viewer seeking it.

The number

In the UCL/Kent/ASCL "Safer Scrolling" study, researchers built TikTok accounts around archetypes of teenage boys and watched the For You page for seven days. Recommendations that started by matching the stated interests shifted toward anger and blame directed at women, and after five days the share of recommended videos carrying misogynistic content — objectification, sexual harassment, discrediting women — had quadrupled, rising from 13% to 56%. ^[1]

What it does and doesn't establish

This is a feed audit: it measures what the recommender serves to accounts coded as teenage boys, proving the content is amplified and delivered without the viewer seeking it out. It does not measure what real viewers come to believe, so on its own it cannot show that watching causes misogynistic attitudes — that causal link is not established by any study on the source page. For the full, graded picture of amplification, association and the open causal question, see Does watching Andrew Tate content cause misogyny?

Sources

1 source · all archived at publication

[1]Social media algorithms amplify misogynistic content to teens · University College London (announcement of the UCL/Kent/ASCL report) · 2024-02-05 archived ↗

By The Fieldnote Desk · How we verify · Corrections