From Safe Harbors to Duties of Care: Rethinking Platform Liability for Illegal Content in Comparative Perspective

For much of the Internet’s legal history, intermediary liability has been structured around a defensive question: when should an online service provider be protected from liability for unlawful content created by someone else? That question remains indispensable, but its dominance has produced a conceptual blind spot. Legal debates too often move directly from the proposition that intermediaries should not automatically bear liability for third-party content to the much broader assumption that they should bear no affirmative responsibility for identified illegality. The result is a false choice between expansive immunity and generalized platform monitoring. This Article challenges that binary. It argues that protection against derivative liability and the existence of affirmative duties are analytically distinct questions. Safe harbors determine when another person’s wrongdoing should not be attributed to an intermediary; they do not, without more, determine whether the intermediary’s own knowledge, function, control, and capacity may generate independently justified obligations. Conversely, recognizing such obligations does not require transforming platforms into universal guarantors of online legality. Drawing comparatively on U.S. intermediary law, the European Union’s Digital Services Act, and relevant approaches in South Korean and Chinese law, this Article develops a differentiated and layered model of intermediary responsibility. Under this model, an affirmative duty depends on six interrelated variables: trigger, legal clarity, function, control, capacity, and proportionality. These variables determine not only whether a duty arises, but also what form of intervention can reasonably be required. The framework accordingly distinguishes among duties to receive and assess notices, restrict access, preserve information, cooperate with public authorities, and prevent repetition in narrowly defined circumstances, while preserving the prohibition on generalized monitoring and safeguards against over-removal and privatized legal adjudication. The Article ultimately argues that intermediary law should move beyond the binary of immunity versus liability. The central question is not whether “platforms” as a category should be responsible for users’ speech, but when a particular intermediary’s legally relevant relationship to an identified risk justifies a particular obligation. Reconstructing intermediary responsibility in this function-sensitive, trigger-based, and proportionate manner preserves the essential protection supplied by safe harbors while making room for carefully bounded duties arising from intermediaries’ own role in the governance of illegal content.