Explicitation in Generative AI Translation: A Comparative Analysis of ChatGPT and DeepSeek in Translating a Chinese Policy Text

With the rapid development of generative AI, increasing attention has been paid to its potential in translation practice. This study aims to investigate generative AI’s use of explicitation in translation. It compares the official English translation of an international publicity text on China’s green and low-carbon industrial transition, released by the State Council Information Office, with translations generated by ChatGPT and DeepSeek. Drawing on focus-point analysis, the study first identifies focus points through a comparison between the Chinese source text and its official English translation to examine explicitation use, and subsequently examines how the two generative AI models address these points. The findings reveal that ChatGPT and DeepSeek demonstrate a tendency towards explicitation, with their outputs showing considerable similarity to the official translation. Both models are effective in making implicit linguistic relations explicit, including omitted subjects, possessive relations, and logical connections between linguistic elements. However, generative AI’s explicitation is primarily concentrated at the linguistic level. Compared with the official translation, the two models show limitations in preserving the stylistic features of the international publicity text and providing culture-specific background information. These findings suggest that while generative AI may serve as an effective tool for initial translation drafting, human involvement remains important for stylistic refinement and culturally informed decision-making. The study contributes to a better understanding of the capabilities and limitations of generative AI in Chinese-English translation and offers implications for the broader use of generative AI in international communication.