Abstract:

With the rapid and inevitable launch of AI in different fields, the demand for human intellectual power seems to be rapidly decreasing. This is also relevant in the field of translation. The concern is whether AI translation tools will substitute human intelligence. An even more significant concern is if the AI translation is reliable enough. The research aims to reveal how the quality of translations provided by translation tools and post-edited by students and professionals differs. The article focuses on the advantages and disadvantages of the translation tools used by students to translate LinkedIn articles and the factors such as the linguistic nuances of the source and target languages that cause potential challenges in achieving reliable translations.

Methodology:
The research methods applied were the analysis of theoretical sources that helped to review the advantages and disadvantages of human and machine translation theoretically and contrastive descriptive linguistic analysis that helped to identify the difference in the quality of translations practically. The article focuses on how well the translation tools DeepL and Google Translate transfer the meaning from English as a source into Lithuanian as a target language and how students accept or dismiss the suggestions provided by Artificial Intelligence. The added value of the contrastive linguistic analysis is that students’ translations have been revised by a professional translator, and students have to substantiate the improvement of the quality of the translation provided.

Findings:
The practical analysis revealed that both Google Translate and DeepL translation tools had remarkable resemblances, giving similar suggestions and making similar mistakes. However, in some cases, the suggestions differed. The students’ translations occasionally outperformed the machine tools, revealing a better grasp of the context and linguistic sense. The professional postediting proved that both Artificial Intelligence and Human Intelligence find it difficult to translate idiomatic language. In their translations, students lacked consistent lexis, clear sentence structure, proper punctuation, and thorough proofreading. To conclude, Artificial Intelligence wins in the time of producing a translation, and the students win in the human awareness of the language. However, to produce a highly reliable translation, three components are needed, i.e., intelligent translation tools, an intelligent translator, and an intelligent editor.

Keywords:

AI, translation tools, human translation, reliability, the Lithuanian language.