LINGO-PSYCHOLOGICAL ANALYSIS OF THE POSSIBILITIES OFACTUAL GENERATIVE LANGUAGE MODELS IN POLITICAL DISCURSIVE PRACTICES (ON THE EXAMPLE OF CHAT GPT)

Irina N. Ziryanova,Alexander S. Chernavskiy

DOI: https://doi.org/10.47475/2070-0695-2023-50-4-29-39

2023-12-21

Sign problematic field in mediaeducation

Abstract:This article is a theoretical and practical generalization based on the results of an experiment conducted by the authors in the spring of 2023 as part of a joint project of Moscow Pedagogical State University and Baikal State University. This project was carried out by the Department of Political Science at Moscow State Pedagogical University and the Department of Theoreticaland Applied Linguistics at BSU, which carried out scientifi c research on the topic of research work “Language and Man”. Despite the emphasis on relevant applied results in this publication, the authors also provide a brief theoretical overview to understand the main historical and theoretical-methodological aspects of the research topic. An interdisciplinary approach to the use of modern generative language models in current political discourse is the basis of this work. Neural network generative language algorithms already provide unique opportunities for creating a variety of content, including the political sphere. In addition, the authors offer a brief analytical overview of modern psychological and linguistic techniques of manipulation, which continue to develop and play a signifi cant role in modern political discourse. The article also contains an overview of contemporary scientifi c knowledge related to the stated research topic. At the same time, the authors see the main goal of the work in a theoretical and applied study of the current potential of the leading publicly available generative language model Chat GPT (4.0) in the fi eld of political discourse using suggestive techniques and manipulations. The basis of this study was the author’s corresponding bilingual requests for the generation of three texts of a manifestational and ideological nature, which then became the object of psychological and linguistic analysis. At the end of the experiment, brief fi nal analytical statements were formulated in the context of the stated topic. The authors come to the conclusion that generative language models can already be widely used today to create a wide variety of media content in different forms of communication, while quite complex linguistic constructions can be implemented according to appropriate requests.