Abstract
For over two decades, there has been a heated debate among legal scholars, activists, judges, and others about the scope of Section 230 of the Communications Decency Act. A persistent theme in those debates has been hyperbolic claims about the necessity of immunity from state laws for digital tech platforms and fearmongering that anything less than maximum immunity will destroy the Internet.
This Article argues that states retain considerable discretion to regulate digital platforms’ design and engineering decisions. We argue that manipulation, engineered behavior, and even habituation and normalization of engineered behavior by digital platforms are content-agnostic harms attributable to defective designs, and neither Section 230 nor the First Amendment categorically immunizes platforms for causing such harms.
We develop a typology that distinguishes between direct harms perpetuated by the platforms’ own design decisions, secondary harms caused by user-generated content and the platforms’ content-moderation failures, and further attenuated tertiary harms. Our approach enables and empowers policymakers and judges to distinguish the platforms’ regulable conduct from their expressive decisions immunized by Section 230 and the First Amendment.
We focus on personal-data-driven algorithmic targeting, a manipulative design feature of social media systems that directly causes primary harm. Other design features—such as infinite scrolling and other dark patterns—also directly cause primary harm, and this form of manipulation also includes hijacking attention, disabling or overriding user autonomy, undermining self-governance, scripting behaviors, and engineering addiction. Of course, personal-data-driven algorithmic targeting and other manipulative design features may exacerbate content-based secondary and tertiary harms, but those are properly understood as analytically distinct from the primary harm of manipulation.
We show how this approach is consistent with current law and is already surfacing in recent state regulations and in litigation over social media platforms’ tortious design. This Article illustrates how careful delineation of the relevant actors, actions, causal relationships, effects, and harms is crucial for understanding the proper scope of Section 230 and the First Amendment.
Recommended Citation
Brett Frischmann & Peter Ormerod,
Regulating Manipulative Design Is Not Preempted by CDA 230 or the First Amendment,
75
Emory L. J.
1101
(2026).
Available at:
https://scholarlycommons.law.emory.edu/elj/vol75/iss5/1