Fake Accounts and Political Manipulation: Topic and Sentiment Analysis in Indonesia

ABSTRACTFake accounts have evolved from mere tools of online deception into strategic infrastructures of political manipulation that systematically shape public opinion and intensify polarization. This study aims to uncover how fake accounts operate as coordinated actors that exploit platform architectures to amplify divisive narratives in Indonesia’s digital public sphere. Using a Natural Language Processing (NLP) pipeline—data collection, preprocessing, topic modeling, sentiment analysis, and interpretive mapping—this research analyzed 8,102 posts generated by 100 identified fake accounts across X (Twitter) and Instagram. The findings reveal that negative, morally charged narratives—although constituting a small fraction of all content—disproportionately drive engagement and virality, particularly on X, whose open architecture facilitates mass diffusion. In contrast, Instagram’s private network design confines manipulative content within insulated circles, suggesting differentiated strategies of disinformation. The study contributes new empirical evidence on how inauthentic actors leverage platform logics to influence collective sentiment. Beyond theoretical implications, it recommends multi-level interventions through digital literacy education, platform governance reforms, and regulatory oversight to mitigate the systemic risks of coordinated inauthentic behavior.