The EMPLOYMENT OF CHATGPT-4 AND GOOGLE TRANSLATE IN RENDERING ARABIC FORMAL CORRESPONDENCE AT GOVERMENTAL INSTITUTIONS INTO ENGLISH: A CASE STUDY OF THE PALESTINIAN MINISTRY OF CULTURE
No Thumbnail Available
Date
2025-11-12
Authors
Tamara Abdalhameed Abdallah Ismail
Journal Title
Journal ISSN
Volume Title
Publisher
جامعة النجاح الوطنية
Abstract
Governmental formal correspondence is challenging when rendered between two different languages and cultures. Due to the widespread proliferation of generative AI such as ChatGPT-4o along with MT such as GT in the translation domain, this thesis examines Arabic formal correspondence which has been rendered to English by utilizing GT and ChatGPT-4o. These samples of correspondence are extracted from Palestinian Ministry of Culture through obtaining official permission from the relevant authority, compromising up to 10 samples. Each sample is composed of 2-11 lines which include various topics such as inquiries, requests, invitations and instructions. The researcher seeks to evaluate the selected data by adopting main concepts such as lexical choices, level of formality and the cultural filter along with employing register variables (field, tenor).
The researcher has employed Juliane House’s (2015) model Translation Quality Assessment (TQA). This model concentrates on evaluation of ST and TT which provide a significant challenge in rendering them. The purpose of this thesis is to examine the translations of GT and ChatGPT-4o through rendering institutional Arabic formal texts to English in relation of register variables regardless of mode which is fixed in data analysis. The findings demonstrate that Arabic formal correspondence is required a high level of formality and accuracy which GT achieves and preserves it more than ChatGPT-4o.
ChatGPT-4o tends to use the emotional tone partially rather than the formal one as in the translation of GT. In addition, the translation of ChatGPT-4o forms mismatches in field and tenor variables which lead to shift the subject matter and the social role relationship of the ST between the addressor and the addressee in the TT. These mismatches distort the original meaning which cause a dilemma in the TT and TL. Therefore, the researcher recommends conducting more comparative analysis of institutional formal correspondence translation into other languages to provide a more comprehensive understanding of the influence of languages by utilizing GT and ChatGPT-4o. Moreover, she recommends translators and employees who utilize MT and generative AI, particularly GT and ChatGPT-4o to acquire enough awareness in rendering formal texts. She also recommends applying House’s model (TQA) to other kind of texts such as legal texts.