Recommending toxicity: The role of algorithmic recommender functions on YouTube Shorts and TikTok in promoting male supremacist influencers

This report by the Anti Bullying Centre at Dublin City University investigates how TikTok’s recommender system exposes young male users to misogynistic, anti-feminist, and far-right content. Using experimental "bot" accounts simulating teenage boys with varied initial interests, researchers tracked algorithm recommendations over consecutive viewing sessions. Findings demonstrate that engaging with baseline male-oriented content (e.g., gaming, fitness) rapidly triggers algorithmic pipelines serving toxic material, including extreme gender hostility and conspiracy theories. The study highlights the speed with which platform algorithms can radicalise or introduce harmful tropes to young users, calling for stronger algorithmic accountability, safer platform design by default, and targeted digital media literacy interventions for adolescents. Key takeaways include the finding that whether or not young boy accounts sought ‘manosphere’ or generic masculine content, they were served toxic content across both platforms; in the final phase of data collection, three-quarters of the content being served was of this category.

Methodology

Algorithmic audit methodology using newly created automated "bot" accounts representing teenage male profiles. Data collection involved tracked automated watch sequences on TikTok’s "For You" page over controlled sessions, logging video metadata, watch durations, and content classifications to observe how algorithmic recommendations evolved following specific watch habits.

Country or region of researched population

Ireland

Citation

Anti Bullying Centre. (2025). Recommending toxicity: The role of algorithmic recommender functions on YouTube Shorts and TikTok in promoting male supremacist influencers. Dublin City University. https://www.dcu.ie/sites/default/files/advanced_processing_technology_editor/2025-12/1.-recommending-toxicity_the-role-of-algorithmic-recommender-functions-on-youtube-shorts-and-tiktok-in-promoting-male-supremacist-influencers-2.pdf

Record created:

04 August 2026

Go to publication
2024
"Prof Debbie Ging, Dr Maja Brandt Andreasen, Dr Catherine Baker, Anti-Bullying Centre.
Report
English
research, policy and decision makers
BIK+ strategy pillar 1 - protection, BIK+ strategy pillar 2 - empowerment
potentially harmful content

This report by the Anti Bullying Centre at Dublin City University investigates how TikTok’s recommender system exposes young male users to misogynistic, anti-feminist, and far-right content. Using experimental "bot" accounts simulating teenage boys with varied initial interests, researchers tracked algorithm recommendations over consecutive viewing sessions. Findings demonstrate that engaging with baseline male-oriented content (e.g., gaming, fitness) rapidly triggers algorithmic pipelines serving toxic material, including extreme gender hostility and conspiracy theories. The study highlights the speed with which platform algorithms can radicalise or introduce harmful tropes to young users, calling for stronger algorithmic accountability, safer platform design by default, and targeted digital media literacy interventions for adolescents. Key takeaways include the finding that whether or not young boy accounts sought ‘manosphere’ or generic masculine content, they were served toxic content across both platforms; in the final phase of data collection, three-quarters of the content being served was of this category.