Exploring the Effectiveness of Pedagogical Orientation of Generative AI Models on Enhancing University Students' Translation Skills: An Experimental Study

Abstract

This study attempted to bridge the research gap in AI-driven pedagogy for translation training in the Arab context, focusing on the potential of generative AI models to improve the translation proficiency of university translation majors. The research explored the effectiveness of pedagogically oriented generative AI tools in enhancing students’ skills across linguistic, cultural, and text-level dimensions in English-Arabic translation, using a true experimental pre-test-post-test control group design. While both groups used identical training materials, the experimental group received AI-guided training, and the control group was taught through traditional instruction. Through a random sampling (n = 37 per group), participants were recruited from four universities in Yemen and Oman, ensuring a comparable educational background. The findings revealed that the experimental group outperformed the control group in translation achievement in all targeted translation skills due to the impact of guided integration of AI. The study underscored the multifaceted pedagogical applications of AI in translation education when grounded in a systematic pedagogical framework under instructor guidance. Through highlighting practical pedagogical implications and offering an evidence-based framework for integrating AI into translation programs, the research opens new avenues for innovative practices in AI-assisted translation pedagogy for instructors and curriculum designers.