Breaking language barriers with ChatGPT: enhancing low-resource machine translation between algerian arabic and MSA

Neural machine translation (NMT) has become an essential tool for breaking down language barriers and facilitating communication between different cultures and communities. However, NMT’s potential impact is limited in low-resource language settings, where the availability of training data and resources is scarce. In this context, this article presents an experiment that explores the effectiveness of using a pre-trained language model, ChatGPT, to improve translation performance between the Algerian Arabic dialect and Modern Standard Arabic (MSA). The results of the experiment highlight the potential of using large pre-trained language models to improve translation quality in low-resource language pairs, contributing to the promotion and preservation of minority languages. In addition to the improved translation quality, another noteworthy aspect of the experiment is that ChatGPT was able to capture the intention of the researchers, even when dealing with a low-resource language pair. This means that the model was able to accurately translate the meaning of the sentences, rather than simply producing a word-for-word translation.