• I Putu Eka Indrawan

  Author

• I Made Citra Wibawa

  Author

Translation support systems driven by artificial intelligence are becoming a means of achieving intercultural communication in many areas, including education, business, and public services, that cannot simply substitute words, but help convey meaning and tone in a culturally sensitive way. The present article investigates the rationale behind the design, the architecture, and the outcomes of an AI-based translation system that was created to help instructors working in English with students of different languages in a higher education context. In order to reach this goal, a quasi-experimental design with mixed methods was employed. In a period of nine weeks, twenty instructors made use of the translation system to organize communication with students, while another group of twenty instructors continued to use conventional translation support. The data were collected via a number of means, including surveys, communication logs, feedback forms, and semi-structured interviews, and were analyzed. The findings showed that the group of instructors that used the AI-powered system managed to communicate better than the second group of instructors at the same time. Moreover, students expressed satisfaction with the system in terms of the tone and cultural adaptation of the messages received. The provided architecture comprises a linguistic translation engine, a layer of cultural contextual adaptation, a system for human review, and a feedback loop. The article outlines various practical implications as well, including the risk of oversimplifying culture, the need for human monitoring of sensitive translations, and differences in the reliability of tone detection in various languages. The article concludes by adding that AI-supported translation systems can enhance intercultural communication if cultural calibration and human monitoring are treated as vital factors during the process of development, along with a further need for increased coverage and longitudinal evaluation of the systems.

• PDF