The use of large language models in translating large-scale official documents
Abstract
The rapid growth of artificial intelligence, particularly large language models, has shaped public opinion that the translation profession is gradually becoming obsolete, with its core functions already being largely handed over to large language models. The present article demonstrates the fallacy of this viewpoint by analyzing the machine translation of a large official text using widely known tools. It addresses the problems and limitations of current solutions — linguistic, stylistic, and technical — and provides strong arguments for the impossibility of fully replacing human translators with machines. The role of the human translator remains particularly indispensable when translating lengthy texts where the standardization of terms and other concepts is crucial.
Received: 11/13/2025
Accepted: 11/30/2025
Accepted date: 15.01.2026
Keywords: artifi cial intelligence, large language models, AI, LLM, machine translation, chat bots, neural networks
DOI: 10.55959/MSU2074-6636-22-2025-18-4-66-89
Available in the on-line version with: 02.04.2025
-
To cite this article:
This work is licensed under a Сreative Commons Atribiution - NonCommercial 4.0 International (CC BY-NC 4.0)