Machine Translation and Linguistic Changes: How Automated Translation Affects English Usage

Abstract

This study investigates the linguistic and stylistic implications of machine translation (MT) from Arabic into English across three distinct genres: literary, journalistic, and academic writing. It examines how neural machine translation (NMT) tools—particularly Google Translate and DeepL—affect lexical choices, syntactic structures, pragmatic nuances, and culturally embedded expressions in the target language. Through a comparative analysis of machine-generated translations and human-produced equivalents, the research identifies consistent patterns introduced by MT, including lexical flattening, syntactic over-regularization, and the loss of metaphorical or culturally specific content. Using representative texts—a short story (Heart of Glass by Ashti Kamal), a political news article from Reuters (June 20, 2025), and an academic paper (The Science of Language in the Era of Generative AI by Levy et al., 2025)—the study demonstrates both the strengths and limitations of MT in preserving textual fidelity. The findings indicate that while MT performs well in terms of surface fluency and grammatical accuracy, it frequently fails to convey rhetorical nuance and contextual depth. This raises important concerns about its cumulative influence on English usage among non-native speakers. The study contributes to the fields of translation studies, digital linguistics, and multilingual pedagogy by underscoring the need for critical engagement with machine-generated texts and by advocating for hybrid translation models that integrate computational efficiency with human interpretive expertise