AUTOMATED CRITICAL DISCOURSE ANALYSIS: HOW AI IDENTIFIES IDEOLOGY IN POLITICAL SPEECHES

Abstract

Critical Discourse Analysis (CDA) has long been employed to uncover hidden ideologies, power relations, and dominance embedded in political discourse. However, traditional CDA relies heavily on manual qualitative analysis, which is time-consuming, subjective, and limited when dealing with large political corpora. With the rapid advancement of Artificial Intelligence (AI) and Natural Language Processing (NLP), new possibilities have emerged for automating discourse analysis. This study investigates how AI-based tools can be applied to conduct Automated Critical Discourse Analysis in order to identify ideological patterns in political speeches. Adopting a mixed-method, corpus-based research design, the study analyzes a selected corpus of English political speeches using NLP techniques such as keyword extraction, sentiment analysis, and topic modeling. The analysis is theoretically grounded in Fairclough’s three-dimensional model of CDA, Van Dijk’s socio-cognitive approach, and Halliday’s Systemic Functional Linguistics. The findings reveal that AI effectively identifies key ideological markers, including pronoun usage, modality, evaluative language, and framing strategies, which align closely with traditional CDA interpretations. The study concludes that automated CDA enhances analytical efficiency and objectivity while retaining critical linguistic insight when combined with human interpretation. This research contributes to the growing interdisciplinary field of computational linguistics by demonstrating the potential of AI as a supportive tool in critical discourse studies