Speaker
Description
The widespread adoption of generative artificial intelligence (GenAI) has challenged conventional language assessment practices in English-Medium Instruction (EMI) programs, where students are expected to demonstrate both academic English proficiency and disciplinary knowledge. Existing writing rubrics primarily assess the final product and often fail to account for students' responsible use of AI, critical evaluation of AI-generated content, and transparency in AI-assisted writing, thereby raising concerns about assessment validity and authenticity (Chuang & Yan, 2025; Kane, 2013). This study aims to develop and validate an AI-aware rubric for assessing academic English in Vietnamese EMI programs. A sequential mixed-methods design will be employed. Semi-structured interviews with EMI lecturers and language assessment experts will identify key assessment dimensions, followed by a Delphi study to refine the rubric. The finalized rubric will then be piloted with undergraduate students in EMI courses and evaluated using thematic analysis, exploratory and confirmatory factor analyses, Cronbach's alpha, and inter-rater reliability measures. The study is expected to produce a multidimensional rubric integrating academic language proficiency, disciplinary knowledge, critical thinking, responsible AI use, and learner reflection. The findings will contribute to the growing field of AI-aware language assessment by providing an empirically validated assessment tool that supports more authentic, transparent, and pedagogically meaningful evaluation practices. The study also offers practical implications for TESOL practitioners and EMI lecturers seeking to redesign language assessment in response to the increasing integration of generative AI in higher education.
Keywords: artificial intelligence, english-medium instruction, language assessment, assessment rubrics, TESOL
