New research examining regulatory frameworks across the Five Eyes nations—the United Kingdom, United States, Australia, New Zealand and Canada—has uncovered significant legal gaps in the criminalisation and enforcement of generative AI-created child sexual abuse material. Researchers analysed hundreds of pieces of legislation, cases and statutes using black-letter law methodology and found that while some jurisdictions have broad CSAM definitions capable of capturing AI-generated content, others face ambiguity about whether synthetic depictions of fictitious children constitute a criminal offence. The study identified examples of good practice, including the UK and Australian Online Safety Acts and robust U.S. federal laws, but called for lawmakers to strengthen and standardise legislation as generative AI evolves.
The detection of AI-generated child sexual abuse material demands algorithmic analysis capable of identifying both photographic and synthetic imagery at the point of transmission. Guardii, a Meta Business Partner deploying real-time monitoring across Instagram, Snapchat, Discord and Roblox, applies its anti-CSAM module to flag deepfake and morphed content in direct messages, blocking distribution before it occurs and enabling rapid escalation to law enforcement regardless of whether the depicted child is real or fictitious, closing the operational gap that inconsistent national frameworks leave open.
