AI-Mediated Translation and Interference in Arabic: Revisiting Toury’s Law in the Age of Digital Language

Abstract

Thanks to artificial intelligence and machine translation technologies, the volume and speed of translation production have increased. However, these technologies also raise questions about the impact of dominant digital languages on less resourced ones. Many languages remain underrepresented in multilingual modelling due to the limited availability of digital language resources. This results in an uneven linguistic representation in natural language processing and AI models, which, in turn, leads to an asymmetry in linguistic influence, as evidenced by the literature. This study investigates linguistic interference in Arabic translation within digital environments, drawing on Toury’s (2012) Law of Interference. It addresses a critical gap by examining, from a qualitative perspective, how uneven language representation in automated translation environments may shape the quality of Arabic translations. It argues that the unequal distribution of digital language resources may contribute to the transfer of source-language features into Arabic translations. Methodologically, the study adopts an exploratory qualitative approach, analysing a small corpus of 100 Arabic translations from English and French, including human and machine-generated translations, focusing on three dimensions of interference: lexical, syntactic and discursive. The findings show repeating patterns of source-oriented elements, including the incorrect use of cataphoric construction and the coordinator واو, the excessive use of periphrastic constructions and the reduction of stylistic adaptation. Similar patterns were observed in machine translation outputs and online Arabic content, suggesting that AI systems reproduce linguistic tendencies already present in digitally circulating texts. The study highlights the need for greater attention to online high-quality Arabic resources.