Learner–AI cognitive collaboration for creativity: a divergence–convergence model in elementary project-based learning

Kim, Kyungri, Kim, Dongho
Asia Pacific Education Review · Springer Nature · 2026-11-21 · DOI: 10.1007/s12564-026-10155-5

Abstract

Preserving learner agency while leveraging generative AI remains a central challenge in AI-integrated creativity education. This design and development study developed principles for learner–AI collaboration (LAC) and an instructional model for elementary project-based learning, and examined how the principles operated in the classroom. Five LAC principles and sixteen guidelines were derived through literature synthesis and two rounds of expert review, reaching content validity index (CVI) and inter-rater agreement (IRA) values of 1.00 in the final round. The principles were organized into a four-stage divergence–convergence instructional model that specifies the roles of learners, AI, and teachers. The classroom study involved 36 sixth-grade students, with 18 in the LAC-based instructional condition and 18 in the control condition. The experimental group reported higher creativity perceptions than the control group ( F = 6.37, p = .017). In the collaborative outputs, fluency and elaboration changed significantly after Holm adjustment, whereas flexibility and originality did not. These findings indicate how the LAC model operates in classroom practice and suggest a need to revisit collaboration within the instructional stages in future design iterations.

Read at publisher →

Similar articles