Pre-service elementary teachers’ generative AI-based mathematics lesson design and evaluation: A mixed-methods study using the extended UTAUT framework

다국어 초록 (Multilingual Abstract)

This mixed-methods study investigated how pre-service elementary teachers engaged with generative AI during mathematics lesson design and how their perceptions—conceptualized through an extended Unified Theory of Acceptance and Use of Technology (UTAUT)—were associated with instructional practices. The study addressed three research questions: (1) What factors underlie pre-service elementary teachers’ acceptance and behavioral intention toward generative AI, based on an extended UTAUT framework? (2) In the context of AI-supported lesson design, what types of questions do pre-service teachers pose, and how do they evaluate AI-generated responses? (3) How do pre-service teachers’ acceptance factors and behavioral intention levels relate to the types of questions they pose to generative AI and their evaluations of AI-generated responses? Fifty-three pre-service teachers completed a survey measuring seven UTAUT-related factors—performance expectancy, effort expectancy, social influence, trust, perceived risk, moral obligation, and behavioral intention—and participated in AI-supported lesson design tasks. Qualitative analyses of prompts and evaluations were combined with quantitative comparisons across factor levels. Findings revealed distinct engagement patterns: higher performance expectancy and trust were linked to student-centered, purpose-driven prompts and greater recognition of diverse instructional resources in AI outputs, whereas lower behavioral intention corresponded to textbook-based prompts and less critical engagement. Across groups, participants typically articulated both positive and improvable aspects of AI-generated content, indicating evaluative processes grounded in pedagogical reasoning rather than simple acceptance or rejection. The study underscores the need for AI literacy in teacher education—encompassing not only technical proficiency but also the capacity to critically interpret, adapt, and align AI-generated content with curricular goals, learner needs, and contextual realities. While situated in the specific context of Korean first-grade mathematics, the findings offer transferable insights for broader subject areas and educational settings, highlighting the importance of context sensitive and pedagogically informed AI integration.