Ethical Controversies in AI-Driven Risk Control and Content Moderation on Social Platforms—A Case Study of Xiaohongshu

The use of artificial intelligence (AI) to identify and manage content risks on social platforms is growing. It raises ethical questions about uncertain classifications, proportionate intervention and opportunities to challenge decisions. This paper examines Xiaohongshu through a qualitative documentary case study comprising a technical publication, a regulatory document and three original news reports. The analysis distinguishes source identification, harmfulness assessment and moderation action, and examines the relationship between internal explanations and user-facing remedies. The selected documents separately describe evidence for reviewers, information for audiences and appeal access, leaving their integration unevaluated. This study argues that technical explanations alone do not establish user-facing procedural protection, understandable reasons and accessible correction are also necessary. Source labels can provide transparency without establishing harm, while stronger restrictions require an additional justification appropriate to their consequences. Differentiated review, bounded disclosure and traceable appeals are proposed as design requirements whose effectiveness remains to be evaluated. The contribution emphasizes linking content-safety objectives with user protections in platform governance.