From image to decision: developing a mobile AI application as a model of interdisciplinary learning

Abstract

This paper presents a model of interdisciplinary learning based on the development of a mobile AI application for image classification. The proposed approach connects several interdependent layers of an AI system, including image classification as an introduction to the basic concepts of computer vision, the use of both a pre-trained and a custom model, a Python backend as an intermediary communication layer, and a mobile application developed in the MIT App Inventor environment. Particular emphasis is placed on the educational potential of such an approach, which enables students to perceive artificial intelligence not merely as a ready-made tool, but as part of a functional digital system. The paper builds on contemporary approaches that emphasize AI literacy, project-based learning, and the interdisciplinary integration of content, and the proposed model can be regarded as a possible framework for introducing AI-related topics through concrete and functionally connected activities. The paper argues that such an approach has the potential to contribute to the understanding of fundamental AI concepts, the relationships among different components of an AI system, and the integration of knowledge from multiple fields into a coherent educational whole.