Assessing the Influence of Chinas Belt and Road Initiative in Local Communities at Global Scale
Abstract
The 2022 National Defense Strategy highlights the threat posed, by the rising influence of PRC activities across the spectrum of strategic competition, and the necessity of new approaches to deterrence that combine diverse elements of national power, including economic, informational, and diplomatic influence, especially in forging stronger ties to allied and partner nations.In recent years, Chinese influence operations have increasingly been designed to disrupt these ties through soft power narratives emphasizing economic development, while attempting to obscure gray network connections to elites in weak-state environments that lie outside the reach of standard approaches to monitoring and assessment. A central component of these efforts has been the PRC#s Belt and Road Initiative (BRI), an unprecedented multi-trillion dollar investment campaign spanning over 140 countries.-While the BRI is the subject of growing concern for U.S. policy makers, little is known about the soft-power effects generated amongst the local communities most directly impacted by BRI investment-projects. A key barrier to understanding these dynamics is that BRI investment projects are often-conducted in difficult-to-access environments. A second, related challenge is that the computational tools traditionally used for assessment of the information environment are often not directly applicable in such settings, because they typically involve methods focused exclusively on English-language sources or the use of low quality automated machine translations that obscure key nuances embedded in local dialects and emergent discourse communities. This leaves the U.S. national defense community largely in the dark regarding the local impact of PRC-influence operations in much of the world.-We aim to transform this situation via the construction of computational social science methods-and tool sets that allow researchers and domain experts to co-construct an understanding of the local-impacts of BRI investments in these difficult-to-access environments. First, we leverage tools and data we have previously developed to construct a large-scale, geo-tagged archive of social media messages that are sent in close proximity to known BRI investment projects across the globe. Our data are unique both in the scale of the core database from which we start, and in our novel approach to geo-tagging these messages across languages. Second, we develop two new language-agnostic computational approaches to identify distinct social groups to which individuals posting on social media are aligned. Critically, we propose two approaches that are complementary in both their level of automation and their alignment with different socio-theoretic models of group formation and identity, while also developing new approach es to human-machine teaming that allow models to be iteratively retrained through the inputs of domain experts. Third, and finally, we build on these socio-political group identification models to train a language-agnostic deep neural network to classify group-level shifts in discourse occurring following the announcement of new BRI investment projects, allowing flexible discovery of ideational themes operating in support and opposition to these projects amongst local communities. Doing so allows us both to further validate our proposed tools for group identity detection and shifts in discourse within them, and to construct a statistical model that leverages these features to generate forward-looking predictions of the success and failure of the implementation of announced BRI investment projects. Taken together, our new methods and tools aim to transform the current opacity of PRC gray zone activities, providing the defense community with radically enhanced capabilities for the assessment of the strengths and weaknesses of Chinese soft power amongst local communities in difficult-to-access environments.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 11, 2023
- Source ID
- N000142312723
Entities
People
- Kenneth Joseph
Organizations
- Office of Naval Research
- Research Foundation for the State University of New York
- United States Navy