#1052: Designing AI to Scaffold Learners’ Cognitive Engagement in Generative Learning

Recent advances in AI have raised concerns about students’ overreliance on automated systems that may complete learning tasks without requiring meaningful cognitive effort and engagement. This paper explores how AI can be designed to scaffold and sustain cognitive engagement in generative learning activities. Drawing on generative learning theory and scaffolding, we present a conceptual account of how AI can guide learners through generative processes such as selecting, organizing, and integrating information, while gradually fading support as competence develops. We illustrate this through the design of an AI-assisted system that helps children (aged 8–13) create educational science videos --- a task that embodies generative learning but poses cognitive challenges for young learners. Our work sheds light on how to integrate scaffolding to foster but not replace meaningful engagement with AI in generative learning.