UK Online Safety Act: Ofcom Issues First Guidance on AI-Generated Child Abuse Content

Ofcom published its first statutory guidance under the Online Safety Act on September 10, requiring platforms to proactively detect and remove AI-generated child sexual abuse material with the same urgency as photographic CSAM, and mandating risk assessments specifically addressing deepfake and synthetic content. The guidance establishes that synthetic CSAM carries equivalent legal status to photographic material, sets a 24-hour removal standard, and requires platforms serving UK users to deploy detection technologies capable of identifying algorithmically generated abuse imagery, with non-compliance attracting fines of up to ten percent of global revenue.

Compliance frameworks that mandate detection of synthetic abuse material create demand for systems capable of distinguishing harmful content within the context of a conversation rather than relying solely on hash-matching or known-image databases, which are ineffective against novel AI-generated material. As a Meta Business Partner backed by Startmate, Guardii operates at the message level, where its anti-CSAM module evaluates imagery in real time alongside conversational signals—such as coercion, requests for secrecy, or escalation—that distinguish distribution of abuse material from benign image sharing. Statutory obligations to detect synthetic CSAM are most effectively discharged by pattern-recognition systems that assess the exchange itself, not just the content in isolation, enabling platforms to meet regulatory standards without imposing blanket surveillance on all users or requiring universal age verification that would displace risk rather than intercept it.