Students' competency gaps in ai-assisted translation and the transformation of translation teaching — an empirical study based on the MQM framework
AI辅助翻译下学生能力短板与教学转向——基于MQM框架的实证研究

摘要

本研究以某高校翻译专业58名三年级学生为研究对象，采用有无AI辅助交叉实验设计，通过基于多维质量评估模型（MQM）的深度文本对比分析与半结构式访谈，系统探究AI辅助翻译下学生能力短板和AI使用问题。研究结果显示：两种实验条件下，学生在语境识别、句式调整及语用风格处理等方面均存在明显局限；学生对AI存在不同程度的过度依赖，缺乏对AI译文的批判性审校意识与能力。翻译教师应据此调整教学重点，聚焦高阶翻译能力，提升学生人文素养，最终实现人机协同翻译的最优效果。

Abstract

This study recruited 58 third-year translation majors and adopted a crossover experimental design. Using in-depth textual comparison based on the Multidimensional Quality Metrics (MQM) model and semi-structured interviews, it systematically explored students' competency gaps and AI usage issues in AI-assisted translation.
Results indicate that under both conditions, students exhibit obvious limitations in context recognition, syntactic adjustment, and pragmatic style handling. They also show varying degrees of over-reliance on AI, lacking the awareness and ability to critically review AI-generated translations. Translation teachers should thus adjust teaching priorities, focus on high-order translation competencies, and enhance students' humanistic literacy to achieve optimal human-AI collaborative translation.