While most XAI (Explainable AI) methods have been designed primarily to assist AI researchers and engineers, the author argues that they also offer significant potential for AI education. By making complex model behavior more transparent and accessible, XAI techniques can support students in understanding how AI, especially Deep Neural Networks (DNN), learn and make decisions. This paper introduces Project LogosXAI, a framework that integrates both static and dynamic visualizations to enhance the interpretability of DNNs for educational purposes.