The Internet Watch Foundation identified 3,440 AI-generated child sexual abuse videos in 2025, marking a 26,362% increase from just 13 videos in 2024. Over half of the detected videos meet the UK's Category A classification, the most severe category depicting penetrative sexual activity, sadism, or bestiality. The surge reflects the rapid advancement of AI video generation tools, which have lowered technical barriers and enabled offenders to produce increasingly sophisticated abuse content with minimal expertise. IWF analysts noted that AI-generated material now appears deliberately imperfect to mimic amateur footage, making it nearly indistinguishable from photographic CSAM except to trained specialists familiar with known victims.
Guardii's AI-driven anti-CSAM detection module — designed to intercept synthetic and deepfake abuse material before it circulates — could have flagged these threats at the point of contact across Instagram, Snapchat, Discord, Roblox and other monitored platforms. As a Meta Business Partner backed by Startmate, Guardii applies pattern-recognition intelligence rather than mass surveillance, blocking hostile CSAM sharing in direct messages before it reaches the child, surfacing acute risk to parents or safeguarding professionals, and enabling rapid escalation to law enforcement. The IWF data underscores the operational gap Guardii closes: detection and intervention before synthetic abuse material normalizes predatory demand or re-traumatizes identifiable victims.
