Faculty
Dr. Trang Phan, Department of Curriculum and Instruction (CI), Kremen School of Education and Human Development
Abstract
Despite the growing exploration of artificial intelligence (AI) in educational assessment, research comparing AI-generated feedback with peer feedback for student-generated questions (SGQs) remains limited. This study investigated the alignment and divergence between human and AI evaluation of SGQs within an undergraduate teacher education course. Participants included 19 preservice teachers, with data derived from their 533 self-generated questions, peer feedback, and survey responses. Data was analyzed using descriptive statistics and thematic analysis. Findings revealed that AI ratings consistently exceeded human ratings regarding the cognitive levels of the questions. Furthermore, AI-generated feedback was characterized as more standardized and detailed, whereas human peer feedback offered greater holistic judgment and interpretive nuance. The results highlight the potential for hybrid feedback systems that synergize the efficiency of AI coaching with the pedagogical presence and social context provided by peer evaluators. This study also offers implications for designing scaffolded assessment environments in teacher preparation programs.