Machine translation pedagogy in the Arabic-English classroom: Insights from a UAE-based data-informed approach

Abstract

This study investigates the development of machine-translation literacy (MT literacy) among Arabic-speaking translation students in the United Arab Emirates, combining quantitative learning analytics with qualitative reflection. Using EduApp, a process-logging platform designed by the author, the research analyzed post-editing behavior across five bidirectional Arabic-English/English-Arabic translation and post-editing tasks. The descriptive results suggest that directionality, diglossia, and educational bilingualism are associated with differences in post-editing efficiency and confidence. Students required more time and produced denser edits when revising English-to-Arabic MT output, often treating stylistic variation as machine error. Content analysis of 160 reflective statements, aggregated at the student level, identified an attitudinal continuum—from skeptical rewriters to pragmatic revisers and strategic users—that coincided with higher throughput and lower edit density. Over the semester, participants displayed growing ethical awareness, shifting from instrumental reliance on MT to reflective, accountable use. The findings highlight that MT literacy is not merely technical but sociocultural, shaped by linguistic diversity, digital experience, and professional ethics. The study proposes a glocalized framework—that is, one that merges global MT-literacy principles with locally attuned pedagogies—for translator education, integrating analytics-based reflection, culturally responsive pedagogy, and data-informed curriculum design to connect global technology with regional linguistic realities.