Evaluating the Effectiveness of Google Gemini in Enhancing Arabic Lexical Competence and Contextual Translation Skills among Second-Year Arabic Language Undergraduates at Sultan Idris Education University (UPSI), Malaysia

Abstract

This study evaluates the effectiveness of Google Gemini in enhancing Arabic lexical competence and contextual translation skills among undergraduate learners of Arabic as a Second Language (ASL) at Sultan Idris Education University (UPSI), Malaysia. Employing a mixed-methods descriptive-analytical design, the research involved 73 second-year students enrolled in the Arabic Language with Education programme. Quantitative data were collected using a validated 30-item questionnaire measuring technological accessibility, lexical support, contextual translation, and learner autonomy, while qualitative data were obtained through contextual translation tasks to examine linguistic performance. The findings demonstrate that Google Gemini significantly improves learners’ ability to select contextually appropriate lexical items, interpret polysemous vocabulary, and produce translations that reflect the semantic, pragmatic, and stylistic conventions of Modern Standard Arabic. The overall mean score of 4.42 (out of 5) and statistically significant results (p < .05) indicate a high level of perceived effectiveness. Furthermore, learners reported an 82% increase in autonomous learning, reflecting a transition from passive dependence on conventional dictionaries to active engagement with AI-generated feedback. Nevertheless, the findings underscore the importance of AI literacy and prompt engineering to maximise pedagogical benefits and ensure critical, responsible use of generative AI. The study concludes that Google Gemini functions as an effective cognitive scaffold for Arabic language learning and offers practical implications for integrating generative AI into higher education curricula to strengthen lexical development, contextual translation, and learner autonomy.