Abstract
This Comment examines Section 230(c)(1)’s immunity for online platforms that use machine-learning algorithms to recommend third party content and proposes a narrow exception for truly “bad-actor” cases. It reviews Section 230’s text, as well as its early cases, showing that Congress intended broad protection for editorial functions such as content organization. It then traces how recommendation algorithms evolved and how courts treat algorithmic sorting as a neutral tool. This Comment addresses conflicting interpretations and defends broad immunity for routine curation. This Comment argues that immunity should be stripped only after a balancing test under a three-factor “bad-actor” framework—one that evaluates the factors in the totality of the circumstances: (1) the platform’s knowledge or reckless disregard that the content category is illegal or inherently dangerous, (2) an intentional configuration of algorithms to prioritize that content, and (3) the foreseeability of extreme harm. Finally, this Comment recommends that courts adopt this framework in jury instructions, that Congress codify the framework, and that industries use specific tools to manage risk. This balanced approach preserves free expression and innovation while holding “bad-actor” platforms accountable for knowingly amplifying destructive content.
Included in
Computer Law Commons, Internet Law Commons, Legislation Commons, Torts Commons