UK regulator Ofcom publishes draft child safety codes under Online Safety Act

Ofcom released draft codes of practice requiring social media and messaging platforms to assess and mitigate risks of child sexual abuse, grooming, and other harms under the UK's Online Safety Act. The codes mandate proactive detection of harmful content and conduct, age assurance measures, and transparent reporting on safety interventions. Platforms face significant fines for non-compliance, with the regulator emphasising that companies must deploy technology capable of identifying child sexual exploitation material, including AI-generated imagery, and patterns of predatory behaviour in private messaging.

Regulatory mandates that require platforms to detect abuse in encrypted and ephemeral messaging shift the compliance burden toward systems that recognise harmful patterns rather than keyword surveillance of message content. Guardii's approach addresses this requirement directly: its AI detection modules for grooming, sextortion and child sexual abuse material, including deepfakes, analyse the dynamics of an exchange—coercion, escalation, repeated targeting—without building a searchable archive of what children say. Backed by Startmate and operating as a Meta Business Partner, the platform's real-time pattern recognition surfaces a conversation that poses serious risk to a parent, school or authority while preserving the confidentiality of routine communication, enabling platforms to meet their duty of care without constructing the infrastructure of mass message reading that privacy frameworks aim to prevent.