I or GAI? A Comparative Analysis of Human and Generative Artificial Intelligence Translations
Jasmine Mercado, Clarence Jinwell Ramos, Jon Paolo Ordoñez
Received: 28 February 2026; Revised: 05 May 2026; Accepted: 17 May 2026; Published: 19 May 2026
DOI: https://doi.org/10.66074/HUM44D66F5
Abstract
The primary goal of this study is to conduct a comparative analysis of human translation and Generative Artificial Intelligence (GAI) Translation in terms of Accuracy, Expression, and Elegance. The study used 20 translated texts sourced from both human translators and GAI platforms, which were analyzed using translation techniques such as word-for-word, grammatical, idiomatic, and communicative translation. The results revealed that while GAI translations generally excel in speed and consistency, human translations still surpass in capturing accuracy, expression, and elegance. Human translators most frequently used idiomatic and communicative techniques, while GAI translations tended to favor word-for-word and grammatical approaches. This study was anchored in Stanfield’s ‘The Measurement of Translation Ability’ and Juliane House’s Translation Quality Assessment (TQA) model, providing a theoretical lens for understanding translation choices and their impacts. Findings show that GAI translations are improving rapidly, but cannot fully capture the subtleties of human emotion and cultural adaptation. This analysis led the researchers to recommend improving the integration of GAI tools into translation tasks, emphasizing the need for human post-editing to ensure quality in professional and academic settings.
Keywords: generative artificial intelligence, human translation, translations, translation criteria, translation techniques
Corresponding Author Information: College of Sciences, Technology and Communication, Inc., Philippines; 20221818@cstc.edu.ph
Volume 2, Issue 2, June 2026