Metlushko I.V. Representation of historical figures in AI-mediated memory discourse

The article investigates AI-mediated memory discourse, i.e. a type of memory discourse, in which the creation, storage, systematization, transmission, and forgetting of individual and collectively shared representations of the past are distributed between humans and large language models. The empirical corpus incorporates 100 “revived stories” published in social media in 2025–2026. The study examines the ways historical figures are represented. The structural and semantic organization of the “revived stories,” along with the use of first- and third-person narration and their combination as a means of retaining audience attention are considered. The mechanisms of nationalizing memory in the corpus under analysis, as well as the frequent address to contemporaries through avatars of heroes of the Great Patriotic War are analysed. The technical solutions used by neurocreators and the subordination of “revived stories” to social media requirements concerning the predominance of shared elements in the verbal component of videos and of divergent elements in the non-verbal component are studied. The conclusion is made that AI-mediated memory discourse shapes a new mode of dealing with the past, contributing to the nationalization of memory and the redistribution of what is subject to digitalization and transmission in social media and, we assume, will influence the selectivity of remembering and the forgetting of the collective past by the mass addressee.