<![CDATA[Artificial Intelligence (AI) is now embedded in translation practice, yet frequent use does not necessarily ensure critical or self-regulated translation. Recent studies have examined GenAI performance, student perceptions, prompting behavior, and assessment policy, but fewer studies connect students’ observed translation processes with instructor needs and curriculum requirements to derive process-oriented instructional design. This study identifies the needs of students and instructors and formulates design requirements for a Metacognition- and AI-Integrated Translation Practice Module (MPP-AIM). It reports the Define phase of the 4-D Research and Development framework using a convergent mixed-methods needs-analysis design. Questionnaire data were collected from 15 English Literature students and two translation instructors; the instructors also participated in semi-structured interviews, 11 students were observed during AI-assisted translation tasks, and the Semester Learning Plan (RPS) was analyzed. Students reported very high AI familiarity (M = 4.9) and use for translation assignments (M = 4.8), but lower confidence in independently performing translation tasks (M = 2.9). Observation showed that 9 of 11 students began without systematic planning and 7 performed no post-editing. Instructors were cautious about AI-generated translation quality (M = 2.5) while strongly supporting curriculum revision and professional development (M = 5.0 each). Integrated findings indicate a gap between technological adoption and metacognitive regulation, curricular alignment, and process-oriented assessment. The study contributes an evidence-to-design framework specifying planning prompts, structured prompting, post-editing checklists, reflective activities, ethical-use guidance, process-oriented rubrics, peer review, and instructor support for critical human-AI collaboration.]]>
Copyrights © 2026
