An An Assessment of English-Arabic Translation Using ChatGPT

Ahmad Al-Harahsheh; Suha Tayseer Abualkarasanneh; Rafat Al Rousan

doi:10.47012/jjmll.17.4.8

Authors

Ahmad Al-Harahsheh Translation Department Yarmouk University, Jordan https://orcid.org/0000-0001-9381-641X
Suha Tayseer Abualkarasanneh Translation Department Yarmouk University, Jordan https://orcid.org/0009-0001-1587-9914
Rafat Al Rousan Translation Department Yarmouk University, Jordan https://orcid.org/0000-0001-7009-1321

DOI:

https://doi.org/10.47012/jjmll.17.4.8

Keywords:

ChatGPT, Artificial Intelligence, Political texts, English

Abstract

This study aims to assess the quality of translating political texts from English to Arabic using the artificial intelligence tool ChatGPT. It adopts the updated version of the Multidimensional Quality Metrics (MQM) framework to evaluate the quality of translations produced by ChatGPT. The data includes 30 political articles selected from BBC World and CNN World in English and their translations into Arabic. The study identified several linguistic errors encountered by translators using ChatGPT, such as fluency and accuracy. Translations lacking these important aspects often lead to misunderstandings and confuse the users. Despite these issues, the ChatGPT tool demonstrated good translation performance in political news, with an overall translation quality rating of 93%.

Highlights

ChatGPT provide spontaneous translation of different texts with nearly high accuracy.
However, human intervention is crucial to rectify this translation, as human translators are more capable to recognize the cultural or social context.
The expertise of human translators is essential for ensuring the building of effective and culturally suitable translation systems at the time AI grows quickly. AI technologies or tools can be developed and improved to address different language and cultural demands.
Furthermore, investing in collaboration and knowledge exchange between AI researchers and human translators would encourage advancements and push translation technology forward. This will lead to more effective precise and accessible translation technologies that will help both people and communities.