AI-driven Content Recommendation Systems for Children: Privacy Protection and Algorithm Governance in TikTok

This is the authors' abstract. We don't add key points for this paper.

This paper examines the dual framework of privacy protection and algorithm governance in AI-driven content recommendation systems targeting children on short video platforms, with a specific focus on TikTok. As the platform increasingly shapes children's digital experiences through personalized content recommendations, significant concerns regarding children's privacy rights, safety, and developmental wellbeing have emerged. Through an in-depth analysis of TikTok's "For You" algorithm, this research investigates the regulatory landscape, ethical considerations, and technical implementations. The paper analyzes TikTok's privacy protection measures and algorithm design choices, evaluates their effectiveness through the lens of regulatory compliance and child safety, and examines key regulatory actions including the landmark COPPA enforcement cases of 2019 and 2024. The findings suggest that effective protection requires a multi-stakeholder approach combining technical safeguards, industry self-regulation, enhanced regulatory oversight, and algorithm literacy development. This paper contributes to ongoing discourse on ethical AI development in recommender systems for children and provides actionable recommendations to ensure that AI-driven recommendation systems respect children's privacy while delivering age-appropriate content.