Post-Editing Skills in a Tech-Savvy World: A Case in an Undergraduate Translation Classroom

Abstract

This article reports on a small-scale qualitative study of how undergraduate translation students work with machine translation (MT) output in a post-editing task. The project grew out of concern that translation classes often still rely on traditional, teacher-centered methods and do not always integrate current AI- and MT-based tools. The study is framed by models of translation and post-editing competence, with a particular focus on how students spot, classify, and correct errors. Data were collected from five students using a short survey, a post-editing task based on an MT version of a general English text used in class, and a structured written interview consisting of 18 questions. The analysis adopts a content-analytic approach and is organized around five types of translation problems, or rich points: lexical problems (including specialized and culture-bound items), morphosyntactic issues (such as sentence complexity and voice), textual problems (coherence and tone), extralinguistic knowledge (cultural and domain references), and intentionality/reader adaptation. Sixteen rich points were pre-identified in the source text and examined through the students' post-edited Indonesian outputs and their own explanations of the choices they made. The findings shed light on how students actually use MT in classroom tasks, which problems they notice and which they tend to overlook, as well as how they view MT and AI tools, and the future of translation work. Based on these findings, the study argues for a curriculum design that uses MT as a starting point for training in error handling, critical judgment, and other core competencies in human translation, rather than treating MT as a shortcut for translation.