There is a growing consensus on the importance of AI ethics in K-12 education, yet effective teaching remains a challenge. AI ethics requires an interdisciplinary understanding of computer science, philosophy, and the humanities, alongside epistemic insights into how AI systems acquire, process, and apply knowledge differently from humans. To address this challenge, this study presents the design, development, and implementation of three theory-informed activities aimed at fostering epistemic insight and ethical understanding of AI among upper primary school students (ages 10-12). Grounded in constructionism, our pedagogical design leverages hands-on experimentation with guided reflection to concretize complex AI concepts. Students examine rule-based, data-driven, and generative AI systems, employing mathematical reasoning to represent AI decision-making processes and reflect on ethical issues such as fairness, bias, and transparency. The interdisciplinary, constructionist approach encourages learners to discern how AI knowledge construction differs from human cognition, thereby enhancing their ethical reasoning. The findings show that students not only developed a foundational understanding of ethical principles but also gained epistemic insight into AI’s relationship with human knowledge and values. This article provides a practical, theory-informed framework and interdisciplinary teaching resources to advance K-12 AI ethics education and support educators in fostering AI literacy.

Magnetic Resonance Imaging (MRI) is important in clinic to produce high resolution images for diagnosis, but its acquisition time is long for high resolution images. Deep learning based MRI super resolution methods can reduce scan time without complicated sequence programming, but may create additional artifacts due to the discrepancy between training data and testing data. Data consistency layer can improve the deep learning results but needs raw k-space data. In this work, we propose a magnitude-image based data consistency deep learning MRI super resolution method to improve super resolution images' quality without raw k-space data. Our experiments show that the proposed method can improve NRMSE and SSIM of super resolution images compared to the same Convolutional Neural Network (CNN) block without data consistency module.