This study investigates the maintenance of Culture-Specific Items (CSIs), translation strategies, cultural meanings, and contextual nuances in Indonesian folklore Asal Usul Nama Cirebon, by comparing translations generated by ChatGPT and a human translation. Employing a qualitative descriptive approach with comparative translation analysis, the study examined CSIs using Newmark’s (1999) classification and Baker’s (1992) translation strategies. The research data consisted of clauses containing CSIs derived from the literacy e-book of six legends in Cirebon. Data were collected through documentation and analyzed using Miles et al (2014), such as encompassing data reduction, data display, and conclusion drawing/verification. The findings identified 11 CSI categories into material culture, social organization and social culture, with material culture emerging as the dominant category. Both ChatGPT and human translation predominantly employ translation by a more neutral word, a loan word, and a paraphrase unrelated word. Although both translations generally preserved denotative meanings, the human translation more effectively conveyed contextual and cultural nuances. In contrast, ChatGPT occasionally encountered difficulties in rendering culturally loaded expressions. The study concludes that ChatGPT serves as a valuable supplementary tool; however, human intervention remains essential to preserve cultural identity and contextual meaning in literary translation practices across diverse cultural contexts.
Copyrights © 2026