Neural Machine Translation in the Military Sphere: Evaluation of Efficiency, Limitations, and Prospects of Use

Aim. To evaluate the effectiveness of neural machine translation (NMT) for military texts of various styles in English and Vietnamese, identifying its advantages and limitations. Methodology. An experiment was conducted with NMT systems (Google Translate, ChatGPT, Gemini) and a comparative analysis of their translations against those of professional translators. Quality was assessed based on content accuracy, terminological precision, stylistic consistency, and error rates. Results . The study shows that NMT ensures high speed and acceptable quality for scientific and official texts but encounters difficulties with literary and conversational texts requiring stylistic expressiveness. Research implications. This research clarifies the applicability of NMT, provides recommendations for its use in military translation, and highlights the need for domestic offline solutions to enhance translation quality and ensure data security.