Exploring Pre-service Teachers’ Co-Design of Learning Environments with Generative AI: Effects of AI Support on Instructional Design Quality

This study examines the impact of support provided by generative artificial intelligence (GenAI) on the quality of pre-service teachers' instructional design (ID) outcomes -lesson blueprints.An experiment was conducted with one control and two experimental conditions: 1) producing ID outcomes with a generic GenAI bot (ChatGPT), and 2) producing ID outcomes with a Co-instructional designer bot customized with a single ID model (Systems approach).Participants included 78 pre-service teachers from diverse fields such as behavioral sciences, physics, and architecture.ID outcomes were blindly analyzed by experts in ID and evaluated using a pairwise comparison technique, where two ID outcomes are holistically compared to determine which one is better.The findings reveal a significant improvement in the Co-instructional designer condition compared to the control, suggesting that a GenAI bot can support pre-service teachers in ID when customized with an ID theory.