Comparative Analysis of Arabic Translation Results Between ChatGPT and Deepl

Abstract

The rapid advancement of artificial intelligence (AI) has significantly transformed translation studies, yet challenges persist in achieving semantic precision and syntactic fidelity in Arabic translation. While ChatGPT and DeepL are among the most widely used AI translation systems, comparative linguistic analyses of their Arabic translation performance remain underexplored. This study investigates both systems' morphological, syntactic, and semantic accuracy through a descriptive–comparative library research design. Data were drawn from Arabic academic texts in Qirā’ah al-Nuṣūṣ, analyzed using a back-translation technique and linguistic equivalence framework. The findings show that ChatGPT tends to generate more communicative and contextually adaptive outputs aligned with dynamic equivalence. In contrast, DeepL demonstrates more substantial formal and lexical precision consistent with formal equivalence principles. These results suggest that both systems offer complementary strengths that can enhance Arabic translation pedagogy and computational linguistics research. The study introduces a back-translation-based linguistic evaluation model that bridges Arabic linguistic complexity with computational precision, filling a notable gap in AI-assisted Arabic translation research.

Ähnliche Arbeiten