This study investigates the translation quality produced by ChatGPT and Google Translate in translating texts from English to Bahasa Indonesia. It compares the performance of both AI-assisted translation tools across five aspects of Translation Quality Assessment (TQA), consisting of Accuracy, Equivalence, Translation Skill, Text Function, and Grammar and Style. The comparison was conducted across three text types, including informative, persuasive, and instructional texts. A questionnaire of nineteen items using a Likert-scale was used to evaluate the translation results, assessed by three raters. A quantitative analysis was employed using the paired-samples t-test and Cohen’s d effect size. In addition, a qualitative examination by the raters was also conducted to identify translation issues generated by the translation tools. The findings reveal that ChatGPT obtained higher scores than Google Translate across all evaluated aspects. Further statistical analysis indicates that the differences are significant and meaningful in practical contexts. ChatGPT showed better performance in producing natural, contextual, and communicative translation results. However, it was also found that ChatGPT generates additional information not explicitly present in the source text. Meanwhile, Google Translate showed the tendency to produce literal translation that sounds awkward and unnatural translation results. It highlights that revision and post-editing efforts are required to improve the quality of translation results by both AI-assisted translation tools. This study concludes that both ChatGPT and Google Translate are more suitable as translation aids rather than complete substitutes for professional human translators.
Copyrights © 2026