Formative assessment is a cornerstone of effective teaching and learning, providing students with feedback to guide their learning. While there has been an exponential growth in the application of generative AI in scaling various aspects of formative assessment, ranging from automatic question generation to intelligent tutoring systems and personalized feedback, few have directly addressed the core pedagogical principles of formative assessment. Here, we critically examined how generative AI, especially large-language models (LLMs) such as ChatGPT, can support key components of formative assessment: helping students, teachers, and peers understand "where learners are going," "where learners currently are," and "how to move learners forward" in the learning process. With the rapid emergence of new prompting techniques and LLM capabilities, we also provide guiding principles for educators to effectively leverage cost-free LLMs in formative assessments while remaining grounded in pedagogical best practices. Furthermore, we reviewed the role of LLMs in generating feedback, highlighting limitations in current evaluation metrics that inadequately capture the nuances of formative feedback, such as distinguishing feedback at the task, process, and self-regulatory levels. Finally, we offer practical guidelines for educators and researchers, including concrete classroom strategies and future directions such as developing robust metrics to assess LLM-generated feedback, leveraging LLMs to overcome systemic and cultural barriers to formative assessment, and designing AI-aware assessment strategies that promote transferable skills while mitigating overreliance on LLM-generated responses. By structuring the discussion within an established formative assessment framework, this review provides a comprehensive foundation for integrating LLMs into formative assessment in a pedagogically informed manner.
Prompiengchai et al. (Thu,) studied this question.