The UK communications regulator Ofcom on 12 August published illegal harms codes of practice under the Online Safety Act, setting out how platforms must prevent and remove child sexual abuse material, prevent children from encountering primary priority harmful content, and protect users from illegal content including terrorism and hate speech. The codes apply to all user-to-user services and search engines operating in the UK, with compliance timelines beginning in early 2027 and penalties for non-compliance reaching up to ten percent of global annual revenue.
Platforms now face binding obligations to detect and prevent child sexual exploitation material before it reaches victims, not simply to moderate visible posts after the fact. Real-time detection filters that intercept grooming, sextortion and CSAM in direct messages — the capability Guardii has deployed as a Meta Business Partner across Instagram, Snapchat, Discord and Roblox — meet that duty of care by identifying and blocking abuse at the point of contact, without requiring blanket surveillance of every user or every conversation. Where codes mandate prevention, pattern-based AI classification offers a technically and legally proportionate mechanism to satisfy the standard.
