Towards human-AI collaborative urban science research enabled by pre-trained large language models

• DOI: 10.1007/s44212-024-00042-y
• Publication Date: 2024-04-29T01:01:21Z
• Authors: Jiayi Fu; Haoying Han; Xing Su; Chao Fan
• Keywords: FOS: Computer and information sciences; 11. Sustainability; Computer Science - Human-Computer Interaction; Human-Computer Interaction (cs.HC)

Abstract Pre-trained large language models (PLMs) have the potential to support urban science research through content creation, information extraction, assisted programming, text classification, and other technical advances. In this research, we explored opportunities, challenges, prospects of PLMs in research. Specifically, discussed applications institution, space, information, citizen behaviors seven examples using ChatGPT. We also examined challenges from both social perspectives. The application were then proposed. found that can effectively aid understanding complex concepts science, facilitate spatial form identification, assist disaster monitoring, sense public sentiment so on. They expanded breadth terms content, increased depth efficiency multi-source big data enhanced interaction between disciplines. At same time, however, face evident threats, such as limitations, security, privacy, bias. development fundamental based on domain knowledge human-AI collaboration may help improve future.