Navigating Online Narratives on Israel-Hamas War with LLMs

Abstract

In recent years, international conflicts have seen a steady rise, accompanied by the increased use of technological advancements to achieve strategic and military objectives. One such conflict, the ongoing hostilities between Israel and Hamas, exemplifies the complex interplay of political, ideological, social, and technological factors shaping contemporary warfare. While extensive analysis is devoted to analysing strategic, historical, and economic dimensions, still limited attention is directed through research efforts towards the vast troves of information available on global social media platforms such as YouTube. These platforms serve as conduits for a wide range of narratives, opinions, and representations related to this war, by providing valuable insights into public discourse and sentiments. Given the capability of automatically and systematically analysing vast amount of content, this research tackles the above stated knowledge gap by developing a novel explorative LLMs (Large Language Models)-based modelling approach to efficiently extract, categorize, and interpret nuanced discussions surrounding the Israel-Palestinian war. It does that following the Data Science methodological approach and focusing on YouTube data aiming at not only contributing to enriching understanding of contemporary conflicts through social media platforms, but also to reflect on the evolving role that state-of-the-art AI technological developments play in shaping the global security landscape.