Plan International UK's research finds that recommendation algorithms regularly expose children to prescriptive, gendered content. A survey of 1,440 UK children and tests of six simulated TikTok accounts showed feeds for 13–16 year-olds became gendered within 20 minutes and escalated to more harmful material in under two hours. Girls were often shown unrealistic beauty and luxury content, while boys saw messages glorifying wealth, strength and misogyny. The charity urges platforms to do more to remove damaging material.
Algorithms Push Gendered Messages to Children — UK Study Warns of Harmful TikTok Feeds
Almost a third of girls and nearly a quarter of boys say content they view online is stopping them acting, dressing or behaving in ways they want, research shows. Annette Riedl/dpa
New research from Plan International UK warns that online recommendation algorithms are rapidly steering children toward gendered and potentially harmful content. The charity's combined survey and platform testing found that algorithm-driven feeds can normalise narrow gender stereotypes — and in some cases escalate to content about self-harm and substance abuse.
Key Findings
Plan's representative survey of 1,440 seven- to 17-year-olds across the UK (conducted in February–March) found that more than two-thirds of children regularly encounter messages online telling them how they should look, behave or what careers and interests to pursue. Specifically, 70% of girls and 67% of boys said they regularly saw prescriptive content, and 62% of young people want platforms to do more to remove harmful material.
The charity also created six simulated TikTok accounts to observe recommendation behaviour. Feeds aimed at 13–16 year-olds became clearly gendered within 20 minutes of use and escalated to more narrowly targeted and harmful material in under two hours.
What the Feeds Showed
Profiles representing girls were predominantly served videos promoting unrealistic beauty standards and luxury lifestyles. Accounts built to represent boys were shown content that glorified wealth, physical strength and emotional isolation, alongside misogynistic narratives.
Researchers noted a particularly alarming progression on one simulated account for a 13-year-old girl: a profile that began with interests in music, dance and drama was later shown content about feeling sad and mental health, then videos related to suicide, self-harm and substance abuse.
Plan reported the problematic material during testing, but TikTok determined that the items did not breach its community guidelines. TikTok has been contacted for comment.
Kathleen Spencer-Chapman, Director of Influencing and External Affairs at Plan International UK, said the research suggested misogyny 'is not confined to the darkest corners of the internet, it is woven into the everyday content that children are fed without asking for it.' She added that the algorithms can trap boys in an echo chamber that normalises narrow, dominance-centred ideas of masculinity.
Why This Matters
Recommendation algorithms are designed to maximise engagement by filtering and promoting content based on profiles and past behaviour. That design can rapidly narrow the range of material young people see, reinforcing stereotypes and exposing them to increasingly specific and harmful narratives. These attitudes can shape self-image, peer relationships and expectations about gender roles.
Plan is calling on social media companies to improve detection and removal of harmful gendered content and for regulators to scrutinise how recommendation systems affect children. Parents, educators and policymakers are urged to press platforms for greater transparency and safer default settings for minors.
Help us improve.
