ARTIFICIAL INTELLIGENCE AND HUMAN TRANSLATION OF SELECTED QUR’ANIC VERSES (ĀYĀT): A CONTRASTIVE STUDY OF SEMANTIC AND INTERPRETATIVE ACCURACY

Abstract

Advanced Neural Machine Translation (NMT) has revolutionized cross-language communication. However, its extension to religious translation, particularly the Holy Qur'ān, involves challenges beyond linguistic transfer, encompassing deeper theological, cultural, and exegetical nuances. This study conducts a systematic contrastive analysis of AI-generated and human translations of nine verses (Āyāt) containing ten semantically complex and interpretively challenging terms precisely selected Qur'anic verses (Āyāt), focusing on their lexical richness and semantic ambiguity. By comparing the ChatGPT and DeepL outputs with Dr. Mustafa Khattab's acclaimed human translation (The Clear Qur'an, 2001) and adopting Tafsīr Ibn Kathir as an interpretive reference point, this study measures translations along the lines of semantic fidelity, contextual richness, and interpretive depth. The findings of this study reveal that the capacity of AI tools to create linguistically competent outputs is matched by systemically weaker performance in expressing interpretive depth and theological intent and a tendency to fall back on literal or fallacious translations. Human translation, informed by learned tradition, outperforms AI in managing the complexity of Qur'ānic discourse. The study concludes by affirming the need for any future application of AI to religious text translation to adopt a domain-specific exegetic corpus with expert oversight.