Harnessing Large Language Models for Scalable and Effective Formative Assessment in Higher Education: A Review

Formative assessment is an integral component of higher education, fostering student learning through feedback, reflection, and iterative improvement. However, despite its pedagogical importance, widespread adoption of formative assessment is often hindered by time constraints, resource limitations, and scalability challenges. The objective of this study is to examine how large language models (LLMs) offer a potential solution to support and enhance formative assessment in higher education across diverse educational contexts by enabling automated, personalized, and scalable feedback that is sustainable and accessible. In this review, we comprehensively examine cutting-edge research and applications of LLMs in various components of formative assessment, including feedback generation, student self-assessment, peer review, and instructor support within the context of higher education. We explore the opportunities LLMs present in enhancing learning outcomes associated with formative assessments and current research gaps while critically discussing the challenges in practical implementations of integrating LLM-driven formative assessments in real-world classrooms. By synthesizing current advancements, this review provides educators and researchers with insights into the transformative potential and responsible implementation of LLM-driven formative assessments in higher education.