Argument-Centered Translation Quality Assessment of the Editorial "Save Raja Ampat" Translated by Kompas

Abstract

The integration of artificial intelligence in the translation of editorial texts, which are essentially persuasive and ideologically charged, raises concerns about rhetorical accuracy, especially for content that is sensitive to the environment. This study evaluates the English translation of Save Raja Ampat produced by artificial intelligence from the Kompas daily editorial, reviewing the ecological impact of nickel mining in Raja Ampat, a UNESCO Global Geopark that is currently the center of controversial environmental policy in Indonesia. Through the application of a Translation Quality Assessment framework centered on Williams' (2004) prism based on the Toulmin model, this study analyzes the preservation of the core components of argumentation: claims, reasons, warrants, support, counterexamples, and rebuttals—in the source text and target text through comparative analytical analysis. The findings of this study show that although artificial intelligence translation maintains the basic logical structure and main elements of ideas, such as environmental sustainability as a reason, this translation risks weakening the rhetorical power by softening the tone, with the shift from a command to a suggestion having a detrimental impact on advocacy for an international audience. Although there is a reduction in the claim section, on the other hand, this editorial translation still accommodates other elements such as grounds, warrant, backing, qualifier, and rebuttal. This study concludes, first, that artificial intelligence still has a probability of rhetorical inaccuracy in opinion journalism, even though its structural adequacy can be expected; second, human supervision is needed in translating high-risk persuasive texts where ideological nuances influence policy outcomes.