Artificial Intelligence in Civil Engineering: Current Applications, Engineering Capabilities, and Limitations
TL;DR
This review identifies the most significant AI achievements from 2023 to early 2026 and presents a task-oriented roadmap that highlights how deep learning, graph models, physically-based learning, and surrogate modeling are making complex engineering workflows more efficient.
Abstract
In civil engineering, Artificial Intelligence (AI) is a significant engineering tool that accelerates analysis, improves monitoring reliability, and enables more effective solutions for engineering decision-making throughout a building’s life cycle, particularly under conditions of nonlinear behavior and complex data. This review identifies the most significant AI achievements from 2023 to early 2026 and presents a task-oriented roadmap. The roadmap covers structural analysis, structural condition monitoring, damage detection, design optimization, and project management. It highlights how deep learning, graph models, physically-based learning, and surrogate modeling are making complex engineering workflows more efficient. Additionally, the review explores the relationship between optimization practices and multi-objective search using surrogate models, as well as the roles of large models in risk triage and evidence management. While summarizing these achievements, the article also identifies several implementation challenges, including data heterogeneity, limited generalizability of results across contexts, limited interpretability, and inconsistencies in standardization and regulatory requirements. Furthermore, priority areas for additional research and implementation are discussed.