Learning-By-Explaining with Generative AI: A Pilot Implementation in Introductory Biology

Explanation activities promote learning but are difficult to scale. This project applied the Design Implementation Framework (DIF) in a nested Learning Engineering cycle to investigate how generative AI can support learning-by-explaining in introductory biology. In an experimental study, students were randomly assigned to view a brief video about explaining or an extended version that also modeled how to use ChatGPT to craft and improve explanations. The extended instruction resulted in longer exchanges and more causal language in the chats. Findings illustrate DIF’s value for guiding AI integration and instruction that help students to leverage AI as a tool for active learning.