On Assessing the Accuracy of Arabic-English Translation by Machine and Human
Syazwan Naim Ibrahim · Advances in computational intelligence and robotics book series · 2025
The ubiquity of artificial intelligence (AI) has significantly influenced research and practice in translation studies. As an integral part of the globalized world, AI's impact continues to grow. But translation accuracy can be compromised if machine translations are not supported by post-editing. This study evaluates the accuracy of machine and human translation by comparing their outputs using a qualitative descriptive approach. Machine translation, while efficient, struggles with polysemous terms, historical references, and culturally embedded expressions, leading to semantic distortions and omissions. The indispensability of human expertise also has been highlighted in preserving the integrity of Arabic scholarly works in cross-cultural academic discourse. This study underscores the importance of hybrid approaches, combining machine translation's speed with human post-editing to enhance translation quality. Recommendations offered include improving machine translation systems through domain-specific training and contextual awareness algorithms.