Resignifying Assessment through Knowledge Building: A Grounded Theory on Students’ Perceived Value of Reflective Evaluation with KBAT and AI

Abstract

PREPRINT. This study explores how postgraduate education students reconfigure the value of assessment when participating in an educational innovation based on Knowledge Building (KB). The intervention included the use of Knowledge Building Analytics Technology (KBAT), a reflective self-assessment tool that provides graphical visualizations and indicators, along with scaffolded interpretations and improvement- oriented suggestions generated through generative AI (ChatGPT), showing students' progress in relation to core Knowledge Building principles. The study followed a Grounded Theory approach to analyze students' perceptions and develop an explanatory model of how they understand and carry out assessment within a Knowledge Building implementation. Results show that students moved from viewing assessment as a mechanism of external control to understanding it as a formative, empowering, and meaningful practice. They used KBAT to reflect on their learning, justify their self- assigned grades, and engage more actively in improving shared knowledge. The theory developed offers insights into how epistemically grounded technologies can support student agency and transform the role of assessment in learning.