New research shows AI-generated child sexual abuse images rising sharply

A report published by the Internet Watch Foundation documents a significant increase in the volume of AI-generated and AI-manipulated child sexual abuse material circulating online, with analysts identifying entire forums dedicated to creating and trading synthetic imagery. The research found that offenders are using generative AI tools to produce photorealistic images and to alter existing photographs of children, lowering the technical barrier to creating illegal content and enabling the proliferation of material that can be used for grooming or distributed within offender networks. The Foundation warned that the realism of AI-generated imagery complicates both detection and victim identification.

Synthetic and manipulated child sexual abuse material shares a distribution signature with contact-offence imagery: it is produced, collected and weaponised within direct-message exchanges, peer networks and coercive grooming conversations. Guardii's CSAM detection module, trained to identify AI-generated and deepfake content as well as contact-offence material, scans image and video attachments exchanged in children's conversations across platforms including Instagram, Discord and Snapchat, flagging explicit or exploitative content for review and enabling immediate reporting to law enforcement or platform trust and safety teams. Because the system operates at the conversation layer, it intercepts material at the point it is shared with a child, whether as part of a grooming sequence or as a tool of coercion, surfacing the incident to a safeguarding professional before further harm occurs. The approach is platform-agnostic and treats AI-generated content with the same operational urgency as traditional CSAM.