UK children’s feeds become gendered within minutes, new social-media study finds

Algorithms rapidly divided content by gender

Social-media feeds created for teenage users began serving distinctly gendered material within 20 minutes, according to research released by Plan International UK and reported by Sky News. The organisation created six test TikTok accounts representing users aged 13 to 16. After less than two hours of scrolling, the feeds were directing the accounts towards increasingly specific themes associated with gender, including appearance and luxury content for girls and messages about wealth, physical strength and emotional isolation for boys. The findings concern the behaviour of recommendation systems rather than material deliberately requested by children.

Children described a wider pattern

The project also surveyed 1,440 people aged seven to 17. Seventy percent of girls and 67 percent of boys said they regularly encountered posts telling them how they should look, dress or behave, or what careers and interests they should pursue. Sixty-two percent wanted social-media companies to remove more harmful material. In one test, an account representing a 13-year-old girl interested in drama, music and dance was eventually shown material concerning sadness, self-harm, suicide and substance abuse. Researchers reported the content during testing, but TikTok determined that it did not violate its rules.

Why the findings matter

The study adds evidence to the argument that child-safety policy must address the design of recommender systems, not only individual prohibited posts. A platform may remove obviously illegal content while still creating a pathway that repeatedly narrows a young user’s feed towards harmful or stereotyped themes. The distinction has become central to regulatory debates in Britain, the European Union and the United States, where governments and courts are examining whether engagement-driven product design exposes minors to foreseeable risks.

Britain is preparing broader safeguards

The UK government announced on September 9 that it would prepare legislation requiring stronger default protections on devices and apps used by children. Its proposed principles include blocking users under 18 from taking, viewing or sending nude images unless robust age assurance establishes that the user is an adult. The government has also said continuously personalised feeds and autoplay functions will be disabled by default for 16- and 17-year-olds under separate measures. Those proposals do not independently validate Plan International’s experiment, but they confirm that algorithmic and device-level safeguards are active government policy questions.

The next test is implementation

The study does not establish how every child’s feed behaves, and its six experimental accounts represent a limited test rather than a full audit of TikTok’s system. What follows will be more consequential: independent replication, disclosure of the researchers’ methodology, the platform’s response and evidence showing whether reported material changes future recommendations. For ministers, the challenge is translating broad safety commitments into enforceable technical standards without relying on parents or children to configure protections themselves. Regulators will also need methods for testing recommendation pathways repeatedly, because a compliant individual post can still contribute to a harmful sequence.