HUMAN AI COLABORATION IN SOFTWARE DEVELOPMENT: A CASE STUDY OF SGU MBKM SYSTEM

Abstract

This study explores the integration of Large Language Models (LLMs), particularly ChatGPT, as co-creators in the software development process of the MBKM system at Swiss German University (SGU). In response to limited IT resources and the need to manage complex student activity data, the development adopted a modified Feature-Driven Development (FDD) approach, with LLMs assisting in key phases such as requirement elicitation, system modelling, UI/UX design, coding, and testing. A structured prompting framework was applied to optimize interactions with the LLM, ensuring more accurate and usable outputs. The final system prototype was evaluated through development artifacts and user feedback, showing that LLM-assisted development improved efficiency by enhancing documentation completeness by 26,7% and reducing development time by 60%. Despite some limitations, such as visual output inconsistencies and contextual limitations, the collaboration between human developers and LLMs proved valuable in accelerating development and enhancing creativity.