Arabic Language Competencies For Academic And Professional Purposes In Generative AI–Enhanced Learning And Workplace Environments: An Empirical Needs Analysis Of Fiqh And Judiciary Students At Universiti Islam Sultan Sharif Ali, Brunei Darussalam

Abstract

This study investigated the Arabic language competencies required by Fiqh and Judiciary students for academic and professional purposes within generative AI–enhanced learning and workplace environments at Universiti Islam Sultan Sharif Ali, Brunei Darussalam. It employed a quantitatively driven explanatory sequential mixed-methods design (QUAN→qual). All 43 first-year students in the department were invited to participate, yielding 29 complete responses. Quantitative data were collected using a validated questionnaire assessing the perceived importance of the targeted competencies and students’ self-reported proficiency levels. Subsequently, semi-structured interviews were conducted with a purposively diverse sample of students, academics, and practitioners to provide contextual explanations for the quantitative patterns. Quantitative data were analysed using descriptive statistics, a weighted discrepancy index, and the Wilcoxon signed-rank test, while qualitative data underwent reflexive thematic analysis. The findings were subsequently integrated through a joint explanatory matrix. The results revealed substantial language-learning needs, with an overall mean of 4.31 out of 5, a relative weight of 86.3%, and an aggregate agreement rate of 90.2%. Academic discussion and recognition of Arabic as foundational to academic success jointly ranked highest (M = 4.41), followed by the comprehension of Sharia-related texts and research writing (M = 4.38). Specialized academic writing recorded the highest level of agreement (96.6%). The relatively narrow range of standard deviations (0.54–0.83) further indicated consistency across participants’ assessments. The study recommends developing a task-based curriculum that integrates receptive and productive competencies, terminological precision, and generative AI literacy—particularly prompt formulation, source verification, and the detection of hallucinations and algorithmic bias. The generalizability of the findings is nevertheless constrained by the modest sample size, single-institution setting, and reliance on self-reported data.