Data Analysis of Russian Disinformation Supply Chains

• In Event: Global Developments in Digital Mis-/Disinformation and Potential Solutions

Abstract

This paper uses artificial intelligence to identify Russian disinformation narratives and track them to their original sources online. This project, using online content that is localized and categorized by the VAST OSINT system, identifies and analyzes domains and websites associated with Russian propaganda to leverage both hybrid media theory and strategic narrative insights. By using human open-source intelligence to identify known propaganda conspiracy narratives, we are able to linguistically tag and track how Russian disinformation narratives move through the online media ecosystem in real time.

One of the most significant challenges in political communication is replicating human content analysis at speed and scale by using artificial intelligence. This would allow us to operationalize and test the way online information flows are theorized by Chadwick (2017). At the same time, we are faced with the challenge of operationalizing the promising concept of strategic narratives, which are described by Miskimmon, O’Loughlin, and Roselle as “tools that political actors employ to promote their interests, values, and aspirations for international order by managing expectations and altering the discursive environment” (2017). Our approach that combines network analytics with natural language processing holds promise in bridging this gap between theory and application.

Continuing developments in natural language processing and network analytics allow us to not only map how propaganda narratives travel, but which of these narratives gain resonance and prominence within the hybrid media system. We use Russian propaganda, particularly strategic narratives surrounding the threatened Russian invasion of Ukraine in 2021-22, to demonstrate the way messages originate, spread, and affect different segments in the media ecosystem. This methodology that combines human open-source intelligence with artificial intelligence has significant promise for supporting the real-time identification of what enables propaganda narratives to reach wide audiences. While this particular approach is useful for testing theories of hybrid media and strategic narrative, it also has practical applications in the real-time detection and deterrence of propaganda narratives. In an era of hybrid warfare in which states combine their military movements with propaganda campaigns, this type of instrument is particularly useful.

References

Chadwick, Andrew. 2017. The Hybrid Media System: Politics and Power. 2nd Edition. New York, NY: Oxford University Press.

Miskimmon, Alister, Ben O’Loughlin, and Laura Roselle (eds.) 2017. Forging the World: Strategic Narratives and International Relations. Ann Arbor, MI: University of Michigan Press.

Author

• Sarah A. Oates, University of Maryland