Author Information
Christopher Klupak, University of Hamburg, GermanyElmar Dammann, University of Hamburg, Germany
Simon Vock, University of Hamburg, Germany
Jula Brügmann, University of Hamburg, Germany
Julius Groß, University of Hamburg, Germany
Abstract
Engineering programmes are widely regarded as demanding fields of study, with high dropout rates especially in early semesters. Foundational subjects such as Technical Mechanics pose particular challenges, requiring students to integrate conceptual understanding, mathematical reasoning, and structured problem-solving. Meanwhile, generative AI is increasingly used as an informal learning tool, often without sufficient guidance, raising concerns about surface-level learning strategies. The KI4ING project investigates how generative AI can be systematically embedded in digital learning environments to support students in Technical Mechanics. Following a Design-Based Research approach, the project combines iterative system development with empirical insights from expert interviews and accompanying student studies examining learning behaviour, problem-solving processes, and AI use in practice. Key findings reveal persistent difficulties in structuring solution strategies, applying conceptual knowledge, and critically engaging with AI-generated outputs. These insights directly informed the design of a custom AI-supported learning environment integrating a Moodle-based learning management system, a dialogue-based AI tutor, and structured learning pathways. The system's modular architecture enables domain-specific configuration, scaffolding-oriented interaction, and privacy-conscious deployment in higher education. This paper presents the conceptual design and technical implementation of the KI4ING system and discusses how empirically grounded design decisions can foster structured problem-solving and deeper conceptual understanding in engineering education.
Comments
Powered by WP LinkPress