A Large-Scale Analysis of Privacy Labels in Child-Accessible Mobile Applications
摘要
Mobile apps are a primary medium for children’s digital interaction, making transparent and responsible data practices essential for child protection and regulatory compliance. Both Apple and Google have implemented privacy labeling systems requiring developers to disclose data collection. However, the extent to which these labels reflect actual app behavior in safeguarding children’s data remains unclear. In this work, we conduct a large-scale analysis of privacy disclosures in child-accessible mobile apps on Apple App Store (AAS) and Google Play Store (GPS). We construct a dataset of 5,662 pairs of popular child-accessible apps that offer both Apple and Android versions. Our analysis reveals significant disparities in disclosing data handling behaviors, where GPS apps more frequently report collecting sensitive data for advertising and personalization, whereas AAS apps do so at comparatively lower rates. We also analyze how app-level factors such as category, pricing model, and platform influence disclosure behaviors. These findings highlight the need for improved enforcement and standardization of privacy labeling mechanisms in child-accessible mobile ecosystems.