Are child safety measures changing outcomes or only processes?
Singapore has explained how it evaluates online child-protection measures. The Infocomm Media Development Authority assesses annual platform reports and publishes service-level results, while the Screen Smart programme tracks public awareness, parental confidence and the adoption of healthy digital practices.
What it could improve
The approach moves evaluation beyond the existence of rules. Singapore reported identifying serious weaknesses in the detection and removal of child sexual exploitation and abuse material on X and terrorism-related content on TikTok. It issued letters of caution, placed the services under enhanced supervision and required progress updates. This demonstrates that reported compliance can lead to scrutiny and follow-up action.
What remains unresolved
Process measures, platform reports and public-awareness indicators are useful, but they do not by themselves establish whether children are safer across the services and networks they actually use.
- Platform reporting needs independent validation and comparable outcome measures, including exposure, contact, reporting and recovery indicators.
- An improvement on one named platform may be offset if children move to other services, private channels, games or unmanaged applications.
- Parental confidence should not become a substitute for duties on services, app stores, networks and device providers.
- Evaluation should distinguish different ages and developmental needs rather than treating all children as one group.
- Authorities need privacy-preserving evidence about whether safeguards reach children without creating new identity databases or cross-service tracking.
- Healthy-practice campaigns should be assessed alongside changes to persuasive design, recommendations, contact permissions and commercial profiling.
The Child-Friendly Internet perspective
Singapore’s emphasis on evaluation is valuable because child safety policy should be judged by outcomes, not announcements. A Child-Friendly Internet measurement framework would examine whether appropriate safeguards activate consistently across participating services and connection environments, while collecting only aggregated or purpose-limited evidence needed for accountability.
- Measure whether children encounter fewer harmful contacts, recommendations and commercial pressures.
- Compare outcomes across services and network settings so risk is not merely displaced elsewhere.
- Use independent evaluation and child participation to test lived experience, not only provider reports.
- Preserve privacy by separating safety measurement from persistent identification of individual children.
- Track digital wellbeing, beneficial participation and growing autonomy alongside harm reduction.
The central question
How can regulators demonstrate that protection follows children across their digital lives, rather than measuring each platform and campaign in isolation?