Modern Machine Translation in the Communicative and Pragmatic Dimension: A Study of Presentational Internet Discourse

The article introduces research results obtained through a comparative analysis of the quality of machine translation and the translation carried out by the artificial intelligence (AI) systems and focuses on preserving the communicative and pragmatic characteristics of presentational Internet texts. The relevance of the study is determined by two key factors: firstly, the development of modern translation studies within the framework of the communicative-discursive approach, which emphasizes the pragmatic potential of a text: its importance and means of its preservation in translation; secondly, the widespread integration of various translation technologies into practice, necessitating an evaluation of their ability to recognize contextual and discourse-related nuances, as well as the development of effective post-editing strategies. The research material comprises Russian-language promotional texts and their translated versions in English and German, produced by MT systems (Yandex.Translate, Promt, DeepL) and the AI-based program DeepSeek. The difference in developers' approaches to transmitting the communicative potential of the original text is revealed. The translation analysis of source and target texts revealed both ingenious solutions that preserve the pragmatics of the original text and errors that hinder the creation of a communicatively equivalent translation. It has been established that traditional MT systems demonstrate a limited ability to work with text as a whole unit, which leads to literal translations, violation of grammatical norms and inaccuracies in terminology. AI chat DeepSeek shows utterly different results and performs structural adaptation of the text, takes into account the stylistic and pragmatic aspects of creating a translated text by analyzing the context. The results prove that modern AI systems are superior to traditional MT in communicating pragmatic potential, but depending on the language pair, they require improvement in the field of specialized terminology and conveying stylistic characteristics.