Learning by Teaching AI (LBT-AI): A case of using Teachable Machine as a cognitive tool

• In Event: AI in K-12 Education

Abstract

Learning by Teaching (LBT) has been widely employed in educational settings to enhance student learning outcomes. This study tested the feasibility of an alternative LBT using Teachable Machine, which is a platform for training AI models developed by Google. With LBT-AI, students teach (or train) an AI model, instead of their human peers. By preparing themselves to be able to train an AI model, the students develop a deeper understanding of the subject matter. The preliminary results of the study showed that the 5th grade LBT-AI students outperformed their counterparts on the categorization of bridges and significantly outperformed their counterpart on the topic of categorization of clouds. The qualitative data also showed the students' overwhelming enthusiasm for LBT-AI instruction.

Authors

• Woei Hung, University of North Dakota

• Jordan Trottier, Surrey Elementary School