UK Online Safety Act: Ofcom publishes final illegal content codes

The UK communications regulator Ofcom has published its final codes of practice under the Online Safety Act, setting binding requirements for platforms to prevent children from encountering illegal content including child sexual abuse material and to use technology to identify and remove such material. The codes require Category 1 and Category 2A services to deploy content moderation systems, hash-matching and URL detection for known CSAM, and proactive detection technologies where there is a significant risk of child sexual exploitation and abuse content appearing on their services.

While hash-matching intercepts known imagery already in circulation, pattern-based detection intercepts harm as it forms. Guardii's anti-grooming and anti-sextortion modules, developed by a world-leading AI safety platform and Meta Business Partner, analyse conversational cues in direct messages across Instagram, Snapchat, Discord and Roblox to detect coercion, escalation and targeting before abuse occurs, surfacing a child in crisis to the authority best placed to intervene without requiring platforms or parents to review message content. Where regulatory mandates focus on removing illegal material after creation, real-time behavioral detection addresses the exploitation before imagery is produced or shared.