TRANSFORMATION OF JOURNALISTIC TEXTS IN MULTILINGUAL MEDIA SYSTEMS: FROM EQUIVALENCE TO ALGORITHMIC RECONSTRUCTION

This research investigates how multilingual news circulation produces fundamentally new textual formations rather than reproducing the original message across languages. By tracing a single news event through three media environments – KCNA’s propagandistic source text, BBC’s internationally oriented reframing, and Ukrinform’s domesticated Ukrainian interpretation – the study demonstrates that news organizations act as autonomous meaning-making systems. Their outputs differ not because of linguistic variance alone, but due to editorial selectivity, ideological positioning, and contextual supplementation. The MTJ (Multilingual Transformation of Journalism) model developed in the article identifies four layers of transformation: equivalent reproduction within a unified ideological context, intertextual rewriting characteristic of global media, combined multilevel transformation emerging from sequential editorial processing, and AI-mediated modification shaped by neural machine translation and large language models. Special attention is devoted to algorithmic influences that restructure syntax, alter evaluative cues, and introduce statistically generated patterns that may shift narrative framing. These findings reveal that multilingual journalism no longer produces translations but transformants – communicative entities that preserve thematic continuity while acquiring new pragmatic roles and ideological configurations defined by the media ecosystem in which they operate.