Design and development of a web-based prototype for human-AI collaboration : integration an AI participant to enhance creativity

Abstract

In this study, a web-based platform was developed to integrate AI as a participant in the idea generation process, to facilitate remote brainstorming and human-AI synergy using Design Science Research Methodology. It examined how the LLM (Large Language Model) can be specifically useful in the creative process and how it is currently integrated into other platforms. A large part was devoted to the specific development of the tool and its architecture. The existing models were analyzed for a model that supports idea generation, and ChatGPT was finally chosen for further implementation. The web platform was designed and developed using JavaScript and React.js for the frontend and Java with Spring Boot for the backend. The system architecture also utilizes a PostgreSQL database for message storage and a RabbitMQ message broker to facilitate real-time communication in a group chat. The platform was evaluated through a group brainstorming session and a survey assessing participants’ experiences. The results were analyzed using correlation analysis, confirming previously established relationships and revealing a new potential link between trust in AI (Artificial Intelligence) and perceived collaborative performance. This newly identified correlation highlights an area for further investigation.