The rapid development of artificial intelligence technology and the accessibility of development tools have opened new chapters for engineering innovation. In this presentation, I will introduce the evolution from AI 1.0 to AI 2.0, discussing the underlying principles behind these technologies. Through two case studies, we will explore how artificial intelligence can assist engineers in achieving higher-quality projects within the construction industry.

In the first case, I will demonstrate how Convolutional Neural Networks (CNN) can be applied to real-time image segmentation, extracting features of rebar assembly at construction sites to aid in rebar inspections. I will also discuss methods to enhance inspection precision and recall rates through synthetic image generation and domain-adaptive learning. Additionally, the potential for automated inspection will be examined by creating true 3D models, identifying features, and comparing them with designed Building Information Models (BIM).

In the second case, I will present how combining Large Language Models (LLM) with domain-specific knowledge can facilitate the traceability and generation of construction inspection forms. This will be achieved through Retrieval-Augmented Generation (RAG), Knowledge Graphs, and Agentic Reasoning.