RAPID: Investigating Social Influence and Mitigating Disinformation Campaigns in Non-English Social Media

University Of Southern California, Los Angeles CA

Investigators

Abstract

Non-English Twitter and other social media have served as public spaces for activists in non-Western countries to express opinions and exchange information during public protests. However, cyber-attacks by different actors have been used to distort the news, malevolently distract audiences, interrupt collective action, and foster affective polarization and outgroup hate. The project team will be studying how cyber-attacks and communication in this online world are correlated with offline protests, violence, and state action. The project team is collecting a large representative corpus of Non-English social media. The team is building tools to identify misinformation and its sources in order to provide a more secure and trustworthy cyberspace to everyday users who rely on social media for receiving news. Given the scientific significance of the corpus being collected, along with the potential security concerns of this domain, the project team is working on a data sharing agreement for researchers interested in studying this corpus while providing a level of security and anonymity to users. Part of the data is being annotated for bots, misinformation, and different types of rhetoric. The data will be used to design and optimize models of bot and troll detection for this domain that will be incorporated in browser plug-ins to help users detect problematic information and accounts. The interdisciplinary project will advance our basic understanding of the role of online communication and offline violence in an understudied cultural setting, adding to the emerging line of work on online behaviors outside Western democracies. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.