Over 1,000 Child Abuse Videos Found on TikTok — Spain May Now Shut Down the Platform

A 35-day investigation by Spanish fact-checking outlet Maldita.es, conducted between September 11 and October 15, 2026, documented 1,182 public posts containing child sexual abuse material published from 231 TikTok profiles, with a combined view count exceeding 48 million. Researchers systematically trained TikTok's recommendation algorithm by mimicking predator behaviour, using the platform's native search function and entering livestreams, and found that TikTok's own infrastructure actively directed them toward illegal material at every stage. During the research window, investigators witnessed the live sexual assault of an infant broadcast directly through the platform. TikTok removed fewer than one in ten formally reported videos.

Pattern-based detection at the point of contact addresses the structural vulnerability the Maldita.es research exposed: the moment between publication and virality, when abusive material is first shared or a livestream begins. Guardii operates across Instagram, Snapchat, Discord, Roblox and other platforms, applying anti-CSAM and anti-grooming filters that flag suspicious content-sharing behaviour, requests for images, or the transmission of material matching known abuse indicators, before that material reaches a child or is amplified by an algorithm. The technology reports the shape and seriousness of the exchange rather than handing anyone a feed of what was said, enabling rapid escalation to the appropriate authority in an emergency. Where a platform's native moderation relies on post-publication review and user reporting, real-time detection intercepts the harm as the conversation unfolds, identifying a child in crisis and surfacing that child to a parent, school or professional before millions of views accumulate or a perpetrator moves to the next target.