The paper will apply sentiment analysis to examine specific forms of verbal aggression and hate speech characteristic of far-right online discourse. Drawing on recent advances in artificial intelligence and large language models (LLMs), the study moves beyond conventional polarity-based sentiment detection to capture the nuanced emotional and semantic dimensions of hostility, exclusion, and resentment. LLM-supported sentiment analysis will be applied to large-scale social media corpora to identify patterns of affective intensity, gendered or racialized hate speech, and the linguistic strategies through which aggression and polarization are articulated. Particular attention will be given to markers of far-right narratives, such as nationalism, xenophobia, and anti-immigrant rhetoric. By integrating AI-driven modeling with computational text analysis, we aim to develop a scalable and theoretically grounded approach to understanding contemporary manifestations of far-right communication.
