Characterizing children’s conceptual knowledge and computational practices in a critical machine learning educational program

Abstract

In this study, we describe the design and implementation of a CML (critical machine learning) education program for children between the ages of 9 and 13 at an after-school center. In this participatory design-based research, we collected learner artifacts, recordings of interactions, and pre/post drawings and written responses to model children’s developing knowledge and practices related to critical machine learning. Drawing from constructionist and critical pedagogical perspectives, our research questions are: (1) How do children develop machine learning knowledge grounded in social, ethical, and political orientations in a CML education program? and (2) What computational practices do children engage in when developing robots for social good in a CML education program? We found that (1) children made more sophisticated connections with socio-political orientations and ML content as they progressed through the program, and (2) they engaged in computational practices, such as experimenting and iterating, testing and debugging, reusing and remixing, and abstracting and modularizing. Further, our findings indicate that a critical lens to ML education can be characterized by posing and answering questions about the roles of AI technologies producers and consumers and identifying how these technologies are designed to apply this knowledge to build applications for marginalized populations. This study suggests that a critical lens is an effective approach towards engaging young children in designing their own machine learning tools in socially responsible ways.