The rapid proliferation of Artificial Intelligence (AI) in Arabic-Indonesian translation has created an urgent need to evaluate the capabilities of emerging platforms. This study analyzes and compares the translation quality of two latest-generation AI platforms, Dola AI and Grok AI, using the accuracy, acceptability, and readability parameters of the Nababan et al. (2012) evaluation model within Nida's dynamic equivalence framework. A qualitative descriptive-comparative method was employed, with 20 Arabic texts selected through purposive sampling across four genres—basic structure, religious studies, journalistic, and literary—validated by domain experts using an identical prompt. Results reveal that Grok AI consistently outperforms Dola AI with an overall mean score of 2.57 versus 2.33, with the largest gap in religious texts (0.54 points) due to Dola AI's critical terminological confusion in fiqh vocabulary. Both platforms failed equally on figurative language in literary texts. This study concludes that Grok AI is more reliable for academic Arabic translation, though human verification remains essential for Islamic legal and literary texts.
Copyrights © 2026