Analysis of the features of representation of candidates in the 2024 US presidential elections in the application of generative artificial intelligence

Modern artificial intelligence applications are increasingly covering various areas of society. They are beginning to have a noticeable impact on politics in general and political communication in particular. Of considerable interest is the analysis of the role of AI applications in the electoral process. The purpose is to identify the content and semantic features of the presentation of texts generated by one of the most popular online services (neural networks) ChatGPT for key text queries related to the two main candidates in the 2024 US presidential election - Donald Trump and Kamala Harris. The main empirical methods of the study were quantitative content analysis of an undirected type and frame analysis. The information array consisted of the results obtained from 41 text queries. The total volume of generated texts was 3503 words. As the result, it was found that the texts of the largest volume are generated by the artificial intelligence application based on queries related to the possible role of the two specified politicians in solving certain current problems of American society, and not to their personal and professional traits. In the course of the research, at least two popular frames found in the generated texts were identified - “efficient politician” and “ineffective politician”. Both the overall frequency of occurrence of the designated frames in texts dedicated to the two political leaders and the number of key expressions-markers responsible for the manifestation of both frames were determined. This made it possible to establish in relation to which topics a particular frame is more common. The general conclusion was made that in the texts generated by ChatGPT in the texts about Kamala Harris, the frame “efficient politician” is more often presented than in similar texts dedicated to Donald Trump.