Can AI-based Translation replace Human Translation within Psychology Settings? A Corpus-based Research

Abstract

As we all know, AI has irrupted in our lives, and its adoption has led to substantial job losses in recent years.  Consequently, the field of translation studies and professional translation renderings is also at risk. Hence, within this research paper we are aimed at unveiling whether professional translation could be the exception to the rule by providing the target readership with more accurate translation renderings (as compared to the ones provided by AI translation). Accordingly, we focus on a specific field featured by the use of literal translations, anglicisms and lexical calques, i.e., the field of Psychology. Subsequently, we selected a highly demanded field, which is the field of emotional states. In order to fulfil our goal, we selected 15 lexical items from the aforementioned field, provided a definition, and contrasted their translation renderings (AI translation with DeepL and Google Translate versus human translation). The research outcomes are striking in a sweeping majority of cases from our corpus, IA-based translation tools cannot replace human translation renderings, since most of these translations were inaccurate, unreliable and too vague.