Exploring the Applicability of ChatGPT in the Landscape Architecture Design Process

다국어 초록 (Multilingual Abstract)

(Background and Purpose) The rapid development of Artificial Intelligence (AI) has brought significant changes to various fields, including landscape architecture. This study explores the potential application of ChatGPT in the landscape design process, particularly focusing on its role in site analysis, concept generation, spatial composition, and visualization.
As AI continues to integrate into diverse creative fields, understanding its capabilities and limitations becomes increasingly important. This research aims to evaluate how AI-based tools like ChatGPT, with its wide accessibility, can enhance the early stages of landscape design, specifically in concept development and spatial planning, while identifying the constraints that arise when applying AI in a real-world educational setting and practical design workflows. (Method) To systematically assess the capabilities and limitations of ChatGPT in landscape architecture, an experimental study was conducted. The framework for the study was established through literature review, and Hongik Cultural Park was selected as the test site.
The design process was carried out using ChatGPT, which incorporated site information, generated concepts, developed spatial layouts, and visualized the design. The study also involved analyzing the strengths and limitations of ChatGPT's applicability in a landscape architecture studio based on the generated results. (Results) ChatGPT was found to be highly effective in the early stages of design, particularly in generating text-based ideas, conducting site analysis, and spatial composition. However, several key limitations were identified. These included challenges related to geographical accuracy, scale application, and the integration of natural environmental and socio-cultural factors, which are critical in the contextual adaptation of the design. Moreover, ChatGPT lacked the nuanced understanding required for comprehensive landscape design. Although it could quickly propose initial design solutions and directions based on general analysis, it was not capable of fully replacing the expertise and depth needed to address complex environmental and cultural elements that human designers bring. (Conclusions) This study emphasizes the need to integrate AI as a complementary tool, rather than as a replacement for human input, in landscape architecture. While AI can effectively assist in site analysis, evaluation, and setting directions, it lacks the comprehensive understanding necessary for executing independent designs. The research contributes to ongoing discussions about AI-based design methodologies and underscores the importance of AI-human collaboration to ensure both the efficiency and quality of the design process. Future studies should explore efficient ways of incorporating professional data into landscape design through AI and investigate more accessible approaches to collaborative AI-human models in design processes and educational settings.