The Artificial Intelligence Learning Anxiety and Self-Efficacy of In-Service Teachers Taking AI Training Courses

This research explored the degree of learning self-efficacy of machine learning experience (MLSE) and artificial intelligence learning anxiety (AILA) of elementary and junior high school teachers. The participants were in-service teachers in the technology domain. This research applied the AI2 Robot City, which is a computational thinking board game, to in-service teacher education. The learning content was image classification for AI application. Elementary and junior high school teachers operated the MIT App Inventor(MAI) and Personal Image Classifier(PIC) platform, and trained the model for practicing AI to implement supervised machine learning on the PIC platform. The learners then inserted the model they had trained into the block-based programming environment of MAI. They completed the smart phone app and used the app to recognize the board game cards so as to control the movement of the smart cars on the table map to meet the requirements of the task in the AI2 Robot City board game. In order to understand affective factors such as the self-efficacy and learning anxiety of the elementary and middle school teachers participating in the AI teacher training workshop, the MLSE and AILA scales were administered before and after the classes. A total of 28 samples were collected. The results showed that there was no significant difference between the MLSE of the elementary and junior high school teachers. However, the average AILA degree of the junior high school teachers was significantly higher than that of the elementary school teachers. It was found that AILA was significantly negatively correlated with MLSE. The elementary and junior high school teachers were confident that they could study AI-related courses. However, the AILA of the junior high school teachers was higher than that of the elementary school teachers. Therefore, more teacher training workshops on AI application can be conducted for junior high school teachers to generally improve their familiarity with AI application. Future research can further explore whether teachers will gradually improve their AILA and MLSE with time and as training courses increase.