Journal of the Architectural Institute of Korea (대한건축학회논문집)

Architectural Institute of Korea (대한건축학회)
권호 표지
DOI QR코드

DOI QR Code

The Role and Utilization of AI: An Integrated Approach with ChatGPT and DALL.E in Architectural Design

AI의 역할과 활용: 건축 디자인에서 ChatGPT와 DALL.E의 통합적 접근

  • Received : 2023.12.18
  • Accepted : 2024.01.27
  • Published : 2024.02.29
⟨ Previous Next ⟩

Abstract

This study delves into the use of generative AI, specifically ChatGPT and DALL.E, in architectural design, focusing on their abilities to comprehend architectural blueprints, interpret sketches, and generate 3D images. It represents pioneering research in integrating ChatGPT into architecture, offering fresh insights. The methodology consists of two phases: the first phase assessed ChatGPT's capacity to interpret and provide feedback on architectural sketches, revealing its potential for creative input but limitations in understanding complex designs; the second phase examined DALL.E's effectiveness in generating 3D images from textual descriptions, where it showed promise in producing contextually relevant images, albeit with some challenges in detailed architectural understanding. This study finds that while these AI tools can enhance initial design stages with innovative suggestions and preliminary analyses, they cannot fully replace the nuanced judgment and expertise of professional architects. However, this research highlights ChatGPT's relevance for specific architectural tasks, enhancing their efficiency. One limitation is the reliance on a single building design, potentially limiting the scope of architectural challenges examined. Future studies should incorporate diverse designs for broader validation. In summary, integrating generative AI into architectural design shows promise for boosting creativity and efficiency, although these tools are not yet equipped for complex architectural planning. They lay a valuable foundation for future advancements in this field.

Keywords

References

  1. Barr, J., & Shaw, P. (2018). Ai application to data analysis, automatic file processing. In 2018 First International Conference on Artificial Intelligence for Industries, AI4I, 100-105. IEEE. 
  2. Grace, K., Salvatier, J., Dafoe, A., ... & Evans, O. (2018). When will AI exceed human performance? Evidence from AI experts. Journal of Artificial Intelligence Research, 62, 729-754. 
  3. Haymaker, J. R. (2011). Opportunities for AI to improve sustainable building design processes. In2011 AAAI Spring Symposium Series.
  4. Hoffmann, J., Borgeaud, S., Mensch, A., Buchatskaya, E., Cai, T., Rutherford, E., ... & Sifre, L. (2022). Training compute-optimal large language models, 2203.15556. 
  5. Joshi, G., Walambe, R., & Kotecha, K. (2021). A review on explainability in multimodal deep neural nets. IEEE Access, 9, 59800-59821. 
  6. Marcus, G., Davis, E., & Aaronson, S. (2022). A very preliminary analysis of DALL-E 2. 2204.13807. 
  7. George, A. S., & George, A. H. (2023). A review of ChatGPT AI's impact on several business sectors. Partners Universal International Innovation Journal, 1(1), 9-23. 
  8. Radford, A., Wu, J., Child, R., Luan, D., ... & Sutskever, I. (2019). Language models are unsupervised multitask learners. OpenAI blog, 1(8), 9. 
  9. Ray, P. P. (2023). ChatGPT: A comprehensive review on background, applications, key challenges, bias, ethics, limitations and future scope. Internet of Things and Cyber-Physical Systems.
  10. Ruiz, N., Li, Y., Jampani, V., Pritch, Y., ... & Aberman, K. (2023). Dreambooth: Fine tuning text-to-image diffusion models for subject-driven generation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 22500-22510. 
  11. Schwartz, R., Dodge, J., Smith, N. A., & Etzioni, O. (2020). Green ai. Communications of the ACM, 63(12), 54-63. 
  12. Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M. A., Lacroix, T., ... & Lample, G. (2023). Llama: Open and efficient foundation language models. 2302.13971. 
  13. Van Den Oord, A., & Vinyals, O. (2017). Neural discrete representation learning. Advances in neural information processing systems, 30. 
  14. Van Gerven, M. A. J., & Bohte, S. M. (2017). Artificial neural networks as models of neural information processing.
  15. Zhang, H., Song, H., Li, S., Zhou, M., & Song, D. (2023). A survey of controllable text generation using transformer-based pre-trained language models. ACM Computing Surveys, 56(3), 1-37. 
  16. Zhao, L., Zhang, L., Wu, Z., Chen, Y., Dai, H., Yu, X., ... & Liu, T. (2023). When brain-inspired ai meets agi. Meta-Radiology, 100005. 
  17. Zylinska, J. (2020). AI art: machine visions and warped dreams, Open Humanities Press, 181