Australia publishes draft online safety codes mandating AI-generated CSAM detection

The Australian eSafety Commissioner has released draft industry codes that would require social media platforms, messaging services, and generative AI providers to detect and report AI-generated and manipulated child sexual abuse material. The proposed codes, open for consultation until November 2026, place obligations on services operating in Australia to deploy hash-matching and classifier-based detection systems capable of identifying synthetic and deepfake CSAM, and to report findings to the Australian Centre to Counter Child Exploitation within twenty-four hours. The Commissioner cited evidence that generative AI tools are being used to create sexualised images of real children from innocuous source photographs.

Conventional hash-matching cannot identify novel AI-generated imagery, and platform-native classifiers trained on historical datasets struggle with synthetic variations that lack known signatures. Guardii's AI-generated CSAM detection module combines behavioral analysis of sharing patterns with classifier models updated to recognize deepfake and algorithmically manipulated content, flagging distribution in direct messages across platforms before material proliferates. The world-leading system, backed by Startmate and operating as a Meta Business Partner, extends coverage to environments where generative tools and social interaction intersect, providing the detection capability the proposed Australian codes anticipate.