Computer Vision on Education: Fostering AI Literacy using RBL-STEM with Google Teachable Machine

This study aims to analyze the application of the RBL-STEM learning model using Google Teachable Machine as a computer vision-based learning media to improve AI literacy. The Research Based Learning-STEM (RBL-STEM) learning model is a learning model that integrates research activities in learning using the STEM approach. Convolutional Neural Network (CNN) is a branch of computer vision that uses artificial intelligence algorithms that are very effective in developing AI products to process image-shaped data. This study utilized a mixed methods approach that integrates quantitative and qualitative techniques to explore the improvement of AI literacy. The participants in this study were 139 undergraduate students of science education study program, Faculty of Teacher Training and Education, University of Jember who participated in the study were taking introductory information technology courses for science education, the sample selection method used was purposive sampling. The quantitative method utilized a pre-test and post-test design, which included the analysis of mean scores, standard deviation, and the observed increase in mean scores. The quantitative method used a survey on AI literacy. The pretest mean score was 38.33 with a standard deviation of 13.41, while the posttest mean score was 71.49 with a standard deviation of 9.37 with a Wilcoxon signed rank-test result of -8.468, indicating a significant effect of the RBL-STEM learning model on students' AI literacy. The high standard deviation on the pretest indicates that there is a large variation in the AI literacy level of the students before the learning begins. This is due to students' different backgrounds, experiences and understanding of AI technology. Some students may be familiar with AI, while others have not been exposed to it at all. This inequality causes a wide spread of scores. After the implementation of the RBL-STEM model with Google Teachable Machine, the standard deviation decreased, indicating that this learning not only improved the average AI literacy, but also made the improvement more even. The AI literacy survey results showed an average score of 3.48, indicating that 69% of students showed an understanding of AI literacy. The implementation of the RBL-STEM model of teaching with Google Teachable Machine is able to train students to conduct research integrated in learning activities, the role of Google Teachable machine as an AI-based learning media is able to improve student AI literacy because the use of AI-based learning media creates a new, interactive, and fun learning atmosphere. Based on the findings of the analysis, it can be concluded that the application of the RBL-STEM model has a significant impact in improving students' AI literacy.