Abstract

This research systematically reviews the latest research progress and application challenges of artificial intelligence ( AI) technology, especially generative large models, in the field of engineering construction. Research shows that generative AI, centered on diffusion models and Transformer architectures, has become the mainstream technical foundation, achieving end-to-end generation from “text-image-3D models” and significantly enhancing design efficiency. Current research focuses on the integration of multi-objective optimization and interpretability, as well as the fusion of cognitive interaction and sensibility engineering. However, it faces problems such as fragmented data copyrights and insufficient interpretability of generation results. In response to these issues, this research proposes that enterprises should balance model scale and business needs, build high-quality knowledge assets, and optimize the workflow of intelligent agents; individuals should transform into intelligent agent commanders, strictly control AI hallucinations, and cultivate deep decision-making and innovation capabilities. This research provides a systematic strategic framework for the industry to deal with the productivity reconstruction driven by AI, emphasizing the core development path of “human-machine collaboration” rather than “ machine replacement”.

Download paper here

Download preprint here

Finacial Support: National Key R&D Program (No. 2023YFC3804600) and National Natural Science Foundation of China (No. 52378306)

Financial Sources: