Beyond Clickers: An AI-Powered Student Response System for Capturing Mathematical Reasoning in Indonesian Mathematics Classrooms

Source abstract

This study aims to examine the design and pedagogical logic of the Si•Xiang•Hui AI Interactive Classroom, an AI-powered student response system developed at Zhejiang Normal University, China, and to explore its potential implications for Indonesian mathematics education. The study employed a qualitative descriptive-analytic case study design. The case involved a hands-on demonstration of the system during a seminar for education administrators from Southeast Asia held at Zhejiang Normal University in September 2026. The system developer served as the key informant, while 34 adult participants, including the two authors, participated as participant observers. Data were collected through participant observation, an audio recording of the session, the developer’s presentation slides, the delegation’s activity report, and relevant policy documents. The data were analyzed using reflexive thematic analysis with a deductive orientation informed by the ICAP framework, formative assessment, and knowledge building, with triangulation conducted across data sources. The findings indicate that the system extends beyond conventional clickers in four ways: it pairs answer choices with spoken explanations; enables every student to pose questions that can be addressed through human–AI collaboration; assigns higher scores to explanations based on their quality (up to ten points, compared with one point for an answer choice); and provides analytics intended to classify responses according to Bloom’s Taxonomy and SOLO taxonomy levels. Two areas for further development were identified: enabling students to revise their recorded arguments and making teacher verification of AI-generated answers and scores an explicit step in the assessment process. The findings suggest that the pedagogical logic underlying the system, rather than the technology itself, can be adapted to Indonesian mathematics classrooms using low-cost tools, provided that teachers receive adequate preparation, scoring criteria incorporate mathematical correctness, and student data are appropriately protected.