Influencing Communities and Narratives: New Techniques for Understanding Adversarial Influence Online
Abstract
China is leveraging influence operations to coerce neighboring countries to reorder the Indo-Pacific region to their advantage.itwill continue to pursue a military modernization program that seeks Indo-Pacific regional hegemony in the near-term and displacement of the United States to achieve global preeminence in the future." -- National Defense Strategy, 2018.The participatory affordances of online social platforms provide novel opportunities for adversaries to engage in large-scale manipulation of our digital environment, such as the one witnessed soon after the deadly clash between China and India in their disputed Himalayan border. More recently, over 2,500 accounts on Facebook, Twitter, and YouTube were found to spread Chinese propaganda and disinformation in India. Theseexamples underscore the following aspects of adversarial influence, all of which inform the objectives of this proposal: 1) First, the Indian Ocean Region (IOR), despite being a hotbed of influence operations, often fomented by its neighboring countries, has not been studied. 2) Adversaries leverage both traditional and new media to run online influence. 3) They engage in two parallel efforts: i) alter the online narrative and ii) alter the community. 4) Additionally, atforms often serve as vectors of adversarial influence in the IOR. Yet systematic algorithmic audits for inauthentic/adversarial content is missing. 5) Finally, adversaries in the IOR are adept at using multiple social media platforms when waging a deceptive information operation, which makes cross platform investigation a necessity in the IOR. This project addresses these gaps in current adversarial influence investigations.The work will proceed along three symbiotic lines of inquiry. Goal 1 will map the new and traditional media landscape by first developing a novel empirical measure of social media based audience similarity, followed by mapping thenews landscape on this similarity space, and then characterizing influence operations launched within this ecosystem of news and social media. Goal 2 will investigate another influence vectoralgorithms and their popularity mechanismsan area that has received vero investigate search, recommendation, and trending topicsall popularity mechanismsas avenues for surfacing adversarial content. While our audits will be conducted in the IOR context, the outcome from this goal will result in novel methodologies and audit frameworks for investigating algorithmically-enabled influence operations worldwide. Goal 3 will expand on the capabilities of existing techniques for detecting narrative and community maneuvers in a way that works in the novel context of IOR adversarial influence. Specifically, our techniques will have multi-language and cross-platform capabilities to determine who is conducting the attack, what narrative and community maneuvers are used to impact whom. By leveraging advances in natural language processing and through rigorous temporal statistical modeling of information flow, we will invent methods to track the dynamic trajectories of adversarial participation, maneuver used, detect their emergent fine-grained roles, and their variations over time. PI is well versed in multiple IOR languages and has the technical skills to bring these goals to fruition.Upon successful completion, our work will have direct applicatioctors: algorithm, community, narrative, news and social media landscape, with actionable knowledge to counter their effects. Moreover, this work is situated in the context of the Indian Ocean regionan area of strategic importance to the U.S. naval forces.(Approved for Public Release)
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 05, 2021
- Source ID
- N000142112748
Entities
People
- Tanushree Mitra
Organizations
- Office of Naval Research
- United States Navy
- University of Washington