Beyond Agentic Translation: A Design-Based Framework for Human-AI Collaboration in Translator Education

Agentic translation systems, autonomous multi-agent architectures capable of planning, tool orchestration, and end-to-end translation, increasingly perform routine translation tasks that translator education programmes were historically designed to teach. Without reform, programmes risk producing graduates for occupational categories that are projected to be increasingly automated. This design protocol paper proposes a framework for human-AI collaboration in translator education, redirecting training from routine translation skills toward four competencies: cross-disciplinary knowledge, AI orchestration competence, higher-order strategic judgement, and critical AI evaluation. The framework is grounded in distributed cognition, task-based automation theory, human-AI collaboration theory, AI literacy, and the EMT competence framework. We specify (a) a translation-task displacement inventory mapping 18 translation sub-tasks to current AI capability and remaining human role, (b) the four-dimensional competence framework with draft assessment rubrics, and (c) a 12-month design-based research (DBR) protocol with a 5-person co-design team and five separately recruited pilot participants, with pre-registered evaluation criteria on the Open Science Framework (OSF). The study contributes (a) a theoretically grounded four-dimensional competence framework, (b) a reproducible small-team co-design methodology, and (c) a transferable design protocol for institutions facing similar redesign challenges. Positioned within the Special Issue’s three pillars of new epistemologies, pedagogies, and research methodologies, the framework complements existing Special Issue contributions on AI curriculum design and educator-AI workflow design by addressing the learner-competence layer. The empirical evidence on the framework’s feasibility, acceptability, and educational effects will be reported in a subsequent paper following the pilot’s completion; the present paper’s contributions are the framework (whose operationalisation is developed through a co-design process rather than the framework structure itself), the reproducible co-design methodology, and the transferable design protocol.