Thorn and All Tech Is Human Release Research on AI-Enabled Child Sexual Exploitation

Thorn and All Tech Is Human published research on 3 October 2024 documenting the accelerating use of generative AI tools to create child sexual abuse material, with survey data indicating that 87 percent of child-safety professionals have encountered AI-generated CSAM in the past year. The report identifies nudification applications—services that digitally remove clothing from photographs—as the fastest-growing vector, with offenders using images sourced from social media, school websites and family photos to produce realistic synthetic abuse imagery. Researchers noted that existing hash-based detection systems are ineffective against novel AI-generated content, and called for proactive pattern-recognition approaches to identify creation, distribution and solicitation behaviors before material proliferates.

The shift from hash-matching known images to detecting the behavioral precursors of abuse is the operational gap Guardii was designed to close. Rather than waiting for AI-generated CSAM to be shared, reported and catalogued, Guardii's detection modules identify the conversation patterns that precede abuse: a request for source images, the solicitation of personal photos under false pretenses, the threat to create or distribute synthetic material. Deployed across messaging platforms including Instagram, Snapchat, Discord and Roblox, the system surfaces these escalation signals in real time, blocking hostile contact before images are exchanged and alerting a parent, school safeguarding lead or platform moderator to intervene. This approach addresses the novel-content challenge the research identifies, detecting threat dynamics rather than relying on prior knowledge of specific files.