UN report documents surge in AI-generated child sexual abuse material
Guardii Analysis
The United Nations Office on Drugs and Crime has reported a sharp increase in AI-generated and AI-manipulated child sexual abuse material, with offenders using widely available generative AI models to create photorealistic imagery and manipulate existing photographs of children. The report notes that synthetic CSAM complicates victim identification, circumvents traditional hash-matching detection, and enables offenders to produce material at scale without direct contact offending, while deepfake imagery inflicts renewed trauma on victims whose likenesses are manipulated. Law enforcement agencies report that AI tools are lowering the technical barrier to CSAM production and distribution, and that international legal frameworks have not kept pace with the technology.
Synthetic and manipulated imagery evades signature-based detection but retains the conversational and distribution patterns characteristic of CSAM offending. Guardii's AI platform detects child sexual abuse material, including AI-generated and deepfake content, by analysing the context and behaviour surrounding image sharing rather than relying solely on known-image hashes, intercepting distribution in direct messages across Instagram, Snapchat, Discord and Roblox before material reaches a child or is further disseminated. The system identifies the coercive request, the escalation to image exchange, and the markers of abuse material in real time, surfacing the conversation to law enforcement and enabling intervention before harm compounds.