Human-AI Collaboration in Generating Graphical Museum Descriptions
Juyeon Kim, MyoungHun Han, SeungJun Kim, Jin-Hyuk Hong · Journal on Computing and Cultural Heritage · 2025
This study explores the potential of human-AI collaboration in generating graph-based descriptions of artifacts in museums. We implemented an AI system that automatically transforms textual descriptions into graph-based representations by fine-tuning a general-purpose language model on a museum dataset and designing an ontology for artifacts. A user study conducted with curators as experts and lay users such as visitors demonstrates the quality and user satisfaction of the graphs generated by the collaboration of AI and a human expert. The results of our study demonstrate that AI is highly effective in extracting detailed and reliable information from textual descriptions. However, human experts play a crucial role in refining the AI-generated graphs, thereby enhancing both the accuracy and readability. Human-AI collaboration even promotes greater consistency across graphs designed by different experts, effectively satisfying diverse user preferences. This research presents a scalable and engaging solution for graphical museum artifact descriptions and a deeper understanding of human-AI collaboration, particularly within the domain of cultural heritage information delivery.