The Algorithm Experiment
The Big Tech’s Little Victims Algorithm Experiment set out to answer a simple but urgent question…what are children actually shown by social media platforms when they sign up at 13, the current age of access?
Our researchers created four fictional profiles of British 13 year olds – two girls and two boys – across TikTok, Snapchat, YouTube and Instagram and scrolled for up to 30 minutes a day for seven days. Dates of birth were entered when we created the accounts, and a predetermined list of influencers were followed aligning to typical teenage interests like gaming, beauty, football and friendship. Randomised scrolling mimicked real-life online behaviour and no comments or messages were shared.
What the experiment revealed
Within just three minutes, on average, the profiles were served concerning content.
On average, one piece of harmful or inappropriate content was shown for every minute spent scrolling.
In one session on Snapchat, 86 pieces of concerning content were flagged in 30 minutes of scrolling.
Content included material glamourising guns and knives, explicit references to sex and pornography, promotion of extreme fitness regimes and diets, and content encouraging misogyny, isolation, self‑harm and even suicide.
Mental health distress content appeared in 50% of sessions, rising to 74% across TikTok, Snapchat and YouTube.
Algorithmic content loops made it difficult – and sometimes impossible – to escape escalating harm.
How algorithms target girls and boys differently
The experiment revealed clear and worrying patterns..
Girls were disproportionately served extreme body‑focused and sexualised content.
TikTok served the girl profiles content on extreme health, fitness or diet in 92% of the sessions, and sexualised content in 83% of the sessions.
Boys were funnelled towards violence, misogyny, racist and anti‑immigration narratives, and figures linked to extremist or conspiratorial views - like Tommy Robinson, Andrew Tate, Joe Rogan and Jordan Peterson.
Across TikTok and YouTube, boys’ profiles were served content featuring hate speech or racist narratives in 77% of sessions.
On TikTok, misogynistic content appeared in 85% of boys’ sessions, compared with 13% of girls’ sessions.
All sessions that flagged content involving weapons or violence were associated with male profiles.
What needs to change
The Algorithm Experiment adds to the growing evidence that social media platforms cannot be trusted to protect children, even when they know their age.
Big Tech’s Little Victims is calling for the age of social media access to be raised to 16.
Have your say before it’s too late
The UK Government has launched a national consultation on potential measures to limit the influence of social media platforms on the lives of young people. The consultation closes on 26 May. We need your help to show the Government that children cannot afford delay – they must act now.
Respond to the government consultation here, or read our guidance on how to take part here.
Contact
- info@bigtechlittlevictims.org
Big Tech’s Little Victims is powered by the National Education Union