This paper aims to assess the accuracy of ChatGPT Translation Quality Assessment to see its practicality, competence, and functionality. Larson’s criterion model (Larson, 1984) is adopted to evaluate translation accuracy errors according to accuracy, clarity, and naturalness criteria. Through this (accuracy) sub-category error analysis approach, the study aims to reveal the affecting factors of the ChatGPT translation service and uncover them by identifying the recurring types of errors in MT output by applying Larson’s Model (1984) (Accuracy Criterion). This research study hypothesizes that providing incorrect information is the main source of the lack of accuracy criterion in translating English journalistic texts into Arabic. In addition, translating and assessing journalistic texts is less challenging with prior linguistic and extra-linguistic knowledge of journalistic writing norms. Results show that ChatGPT translation’s output still needs human intervention to modify and adjust the errors, although MT has improved the quality of its translations. Based on the accuracy error analysis, conveying incorrect information or unjustified omission and addition errors seems to be problematic in ChatGPT translations of journalistic texts from English into Arabic, implying that proper guidelines are crucial in post-editing so that post-editors can be aware of the potential repeated errors.
Messri et al. (Sun,) studied this question.