Abstract
This paper presents my design and research of a Machine Learning (ML) curriculum for middle school students to critically explore socioscientific issues. My research focuses on how the curriculum supports students’ learning of ML knowledge and practices, as well as their understanding of AI ethics. I implemented the curriculum as a free online afterschool program over three to four weekends and recruited ten middle school students from diverse backgrounds. They were engaged in hands-on projects of solving real-world problems with AI/ML technology, such as collecting photos of themselves wearing and not wearing masks to train their own ML models that can automatically recognize if a person wears mask or not. The curriculum design created a constructionist online learning environment for students to learn through social interactions with ML tool, physical tool, peers, and teachers, in the process of making ML products. Primary data sources include students’ responses to pre and post ML test and class recordings throughout the entire program. As a result of statistical comparison between pre and post-test scores and interaction analysis of learning processes, students showed improved AI/ML knowledge and practices and good understanding of AI ethics at the end of program. The study sheds light on the learning and assessment design of AI/ML education for middle school students.