Computational Approaches to Enhance the Integrity of Social Media: From Detection to Intervention
MCML Authors
Abstract
Abstract
This dissertation advances social media integrity through computational methods for detecting harmful content, auditing platform algorithms, and testing interventions. Using case studies, it identifies QAnon users on Parler, analyzes rumor diffusion, audits political ad delivery on Meta, and evaluates AI-generated counterspeech against online hate. Combining computer science and social science, the work reveals key vulnerabilities in social media systems and proposes data-driven strategies for transparency, accountability, and safer digital environments (Shortened).
BibTeXKey: Bae25