CONTENT MODERATION AND FREE SPEECH: LEGAL LIMITS OF AUTOMATED TAKEDOWN SYSTEMS ON SOCIAL PLATFORMS

This paper examines automated content moderation as a system of private speech governance and evaluates its compatibility with foundational principles of freedom of expression. Moving beyond technical accounts of algorithmic error, the study conceptualizes platforms as private governors exercising regulatory power through automated, risk-driven enforcement mechanisms. Drawing on legal theory, empirical scholarship,and human rights law, the paper demonstrates how automation restructures moderation into a system marked by procedural failure, substantive arbitrariness, and weak accountability. It argues that opaque notice regimes, ineffective appeals, and biased outcomes are not accidental flaws but structural consequences of scale-oriented automation.Focusing on copyright-based takedown systems as a prototype, the analysis traces the expansion of automation from intellectual property enforcement to general speech control. The paper further evaluates human rights and EU regulatory responses, concluding that transparency and compliance obligations mitigate but do not resolve core legitimacy deficits. Ultimately, it argues that automated moderation faces normative limits and must remain subordinate to law, fairness, and contextual human judgment in speech governance.