DOI: 10.22215/cjers.v19i1.5214

Abstract

This article provides a pilot bimodal analysis of 20 texts in Russian and their accompanying AI-generated images (RuAI) from the RIA-Novosti news channel, totalling 16,650 words. The major goal of the study was to determine whether ChatGPT can provide this analysis of highly ideologically and culturally diverse texts and images under human supervision. The selected texts and images, spanning from April to May 2024, focused on the war in Ukraine and the involvement of Western countries. The qualitative analysis was conducted using ChatGPT 4.0 for discourse analysis tasks, including theme and subtheme identification and symbolism interpretation in the images vis-à-vis the texts they illustrate. The results indicate that ChatGPT 4.0 can successfully serve as a qualitative discourse analysis tool of texts in languages other than English. However, some limitations of ChatGPT analysis have been noted, such as a generic rather than specific identification of themes and subthemes. While ChatGPT successfully identified symbolism in AI-generated images, some minor details were overlooked due to insufficient exposure to Russian cultural and historical contexts. The article also suggests the emergence of a new bimodal genre in political propaganda—articles accompanied by AI-generated images that enhance the message. Although there are issues with the ambiguity and repetitiveness of some imagery, AI has the potential to be a significant tool in political propaganda as well as in discourse analysis.

Citation format

MAKAROVA, Veronika. AI as a political weapon and a research tool for discourse analysis of russian propaganda. Canadian Journal of European and Russian Studies, 2026.