|
 |
Building on the authors’ previous research, which proposed a typological framework integrating the three-layer definitional structure of education-specific AI based on the OECD Digital Education Outlook 2026 (DEO 2026) with a five-stage Teacher–AI teaming model, this study addresses a remaining limitation: the framework did not sufficiently account for situations in which learners interact directly with AI. To address this issue, the study introduces the concept of “Student–AI teaming” by extending Cukurova’s five-stage teaming model to learner–AI interactions. It further conceptualizes the phenomenon whereby learners bypass the scaffolding provided by education-specific AI and directly obtain answers from general-purpose AI as “Scaffolding Bypass,” and examines its relationship to teaming depth. Finally, based on vendor interviews with two educational AI services with distinct characteristics, the study illustrates asymmetries in teaming depth and differences in bypass risk within the integrated model.
Research papers (research meetings, symposium materials and others)
