From Syllabus to Assignment Design: A Case Study of GenAI Future Role in University Assessment

The growing integration of generative language models in higher education has prompted renewed attention to their role in supporting instructional design, particularly in developing assessments. This study explores the potential of such tools to assist in creating structured assignments within a university-level STEM curriculum. A systematic methodology was applied to evaluate outputs across key pedagogical dimensions, including alignment with learning objectives, appropriateness of difficulty, and cognitive depth. While the tools effectively generated technically accurate and syllabus-aligned content, persistent limitations were identified in their ability to produce higher-order reasoning tasks and multi-layered assessment items. These constraints were especially evident in advanced or design-based coursework. The findings suggest that generative models can enhance instructional efficiency and provide a valuable starting point for educators, but their outputs require ongoing refinement and professional oversight. Their optimal use lies in supporting, not replacing, the instructor, enhancing pedagogical expertise in meaningful assessment.