The emergence and ubiquity of Artificial Intelligence in the form of Machine Learning (ML) systems have revolutionized daily life. However, scant if any attention has been paid to ML in computing education, which continues to teach rule-based programming. A new, promising research field in education consists of acquainting children with ML to foster this much-needed shift from traditional rule-driven thinking to ML-based data-driven thinking. This article presents the development of computational thinking competencies in 12-year-old students who participated in a learning-by-design or a learning-by-teaching ML course. The results, based on a qualitative and quantitative evaluation of the students' achievements, indicate that they demonstrated computational thinking competencies at various levels. The learning by design group evidenced greater development in computational skills, whereas the learning by teaching group improved in terms of computational perspective. These findings are discussed with respect to promoting children's problem-solving competencies within a constructionist approach to ML.
展开▼