Artificial intelligence in translating phraseological units of a literary text

The article reveals the essence and specifics of applying artificial intelligence technologies in the translation into English of Russian phraseological units used in the context of poetic works (based on the poems of S. Yesenin). The experiment tests the translation engines of statistical, neural and hybrid types represented by the DeepL API, Google Neural Machine Translation and PROMT platforms, respectively. The article presents a comparative analysis of the translations made with the help of each of the engines under consideration and the literary translation performed by people. The engines that exist at the moment are not perfect, which determines their popularity at the pre-translation stage, while professional literary translation still remains the prerogative of relevant specialists, implying the need for careful editing of the machine translated product. Special attention in the article is paid to the pragmatic component, which, when working with machine translation, is one of the key characteristics of phraseological constructions, in particular, the possibilities of preserving it in a context adequate to the original. The research was carried out due to the significance of the topic, the active introduction of artificial intelligence technologies into the life of a modern person and the rapid growth of popularity of digital translation, which has confidently defined its position in global translation practice and relevant theoretical developments.