2025 Theses Master's

Feed-om of Speech: The Technical Design and Legal Interpretation of Social Media Recommender Systems

Gerger, Dessa

Social media platforms have become the modern American town square, spaces where public discourse unfolds dynamically and ideas propagate outward. The Founding Fathers could not have imagined that, instead of cobblestones, machine learning models would form the foundations of personal expression in this new square. This work explores the technical foundations of personalized recommender systems that curate content and shape user experience on social media. By unpacking the architectural and algorithmic principles behind such systems, it examines the intersection of technical design and the legal frameworks governing speech and liability. Central to this inquiry is the question of how to legally classify recommender systems, whether as First Amendment-protected speech or neutral conduct. The paper argues that this dichotomy oversimplifies the issue, and instead advocates for a nuanced understanding that accounts for algorithmic intervention, editorial control, and financial interests. Bridging the gap between technical and legal perspectives is essential for developing informed regulatory and cultural responses in an increasingly social media-dependent society.

Keywords: social media, Section 230, recommender systems, free speech online, information ecosystem, attention economy