Meta to share child safety data with researchers under new framework

Meta has announced a framework to share anonymised data on child safety threats including sextortion, grooming attempts and self-harm content with independent researchers and child protection organisations, aiming to improve understanding of online harms and evaluate intervention effectiveness. The initiative will provide aggregated statistics and anonymised case studies under strict privacy and ethics protocols, with the first data releases expected in early 2025 in partnership with organisations including the National Center for Missing and Exploited Children and the Tech Coalition.

Aggregated retrospective data on reported incidents reveals scale but does not reduce latency between first contact and escalation, the interval in which a child may be coerced into sharing imagery or meeting an offender in person. Guardii's real-time detection intercepts grooming and sextortion at the earliest signs of predatory patterning sustained targeting, age-inappropriate questioning, demands for imagery, threat escalation flagging the conversation to a parent or safeguarding professional while the exchange is still underway. By operating within the direct-message layer across platforms, the system converts the window that research identifies as critical into an intervention opportunity, surfacing children in acute danger before harm materialises.