Autoethnographic Journey of Academic Writers as Multilingual Learners in Neural Machine Translation: Human-AI Assistance or Flawed-AI Tool?

In an era where global academic and educational settings are increasingly mediated by technology, integrating Neural Machine Translation (NMT) tools such as DeepL Translator has become indispensable for multilingual learners.This study bridges an empirical gap by exploring the considerations and dilemmas that arise for multilingual learners when incorporating NMT, specifically DeepL into their academic writing through a collaborative autoethnography (CAE) as a qualitative method.Through the personal narratives of two siblings from Yemeni and Indonesian backgrounds influenced by their multicultural upbringing and educational journey, this study revealed key themes, including ethical considerations (e.g., cultural sensitivity and gender bias), educational considerations (e.g., learning dependency, balancing assistance with autonomy, and importance of feedback and revision), and linguistic considerations (e.g., ambiguity and local language variations).This study contributes to establishing a foundation for refining NMT techniques and developing strategies to support multilingual learners, providing practical guidance to navigate the complexities of NMTassisted academic writing while ensuring academic integrity and language proficiency.