Algorithmic Design, Product Liability, and the Protection of Children Online: K.G.M. v. Meta Platforms, Inc., et al. and Lessons for Brazil's Digital ECA

Abstract

This article examines the emerging legal paradigm of algorithmic accountability through a comparative analysis of the U.S. litigation K.G.M. v. Meta Platforms, Inc., et al. and Brazil’s Digital Statute for Children and Adolescents (Law No. 15,211/2025). It argues that both developments signal a structural shift in the legal treatment of digital platforms, from neutral intermediaries protected by broad immunity regimes to designers of behavioral environments subject to duties of care, safety, and risk mitigation. The analysis focuses on how plaintiffs in K.G.M. advance a conduct-based theory of liability that targets platform architecture such as algorithmic recommendation systems, infinite scroll, and engagement-maximizing features rather than third-party content, thereby challenging the traditional scope of Section 230 immunity. The article adopts a doctrinal and interdisciplinary approach, combining legal analysis of intermediary liability, product liability, and constitutional constraints with a review of contested empirical literature on social media and adolescent mental health. It examines the role of corporate knowledge, internal documents, and foreseeability in establishing liability, while also addressing the limits of analogies to prior public health litigation, particularly the tobacco cases. The findings suggest that, although scientific evidence on social media harms remains heterogeneous, the legal standard of foreseeability does not require consensus, and internal corporate awareness plays a decisive role in bridging evidentiary gaps. In the comparative section, the article analyzes the Digital ECA as a proactive regulatory framework that operationalizes similar concerns through ex ante obligations, including safety-by-design requirements, risk assessment, and restrictions on exploitative practices. It argues that the Brazilian model represents a legislative consolidation of theories that are still being tested through litigation in the United States. The article concludes that these parallel developments reflect a broader transnational convergence toward recognizing platform design as a source of legally cognizable harm.