Essays on Content Moderation, Misinformation, and Fraud
Date
2024-07-29
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In the digital age, ensuring online safety has become paramount. This dissertation explores three crucial dimensions of this challenge: fraud detection, misinformation management, and the control of harmful and extreme content. First, we introduce a novel framework for detecting fraudulent websites by analyzing the interactions between websites and third-party vendors. This approach transcends the limitations of traditional content-based methods, offering a more robust and adaptable solution that leverages the underlying infrastructure of websites. Our findings demonstrate that the structure of third-party requests can effectively differentiate between fraudulent and legitimate websites across diverse categories, enhancing the accuracy and efficiency of online fraud detection. Second, we empirically investigate the indirect effects of content moderation policies on social media platforms, specifically examining the unintended consequences of a policy change on Reddit designed to prevent misinformation. Through a natural experiment, we uncover that even seemingly minor policy adjustments can have far-reaching effects on user behavior, content sentiment, and overall platform dynamics. This study underscores the need for a nuanced understanding of the indirect effects of moderation policies and their broader implications for social media platforms, users, and regulators. Finally, we develop a theoretical model to investigate the economic incentives and welfare implications of different content moderation strategies, with a particular emphasis on shadowbanning. We compare shadowbanning with content removal, the most prevalent moderation strategy, to illuminate the economic motivations behind platform choices and their impact on consumer surplus and social welfare. Our analysis reveals that shadowbanning can be economically advantageous for platforms under certain conditions, but it also raises concerns about potential negative consequences for users and overall social welfare.
Description
Keywords
Content Moderation, Fraud Detection, Misinformation, Social Media, User generated content
Citation
Hojati, A. (2024). Essays on content moderation, misinformation, and fraud (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.