Artificial intelligence-driven predictive modeling in civil engineering: a comprehensive review

Abstract Artificial intelligence (AI) has become an essential force in modern civil engineering, reshaping conventional methods through intelligent data processing, predictive modeling, and automated decision-making. This review presents a comprehensive synthesis of AI applications across eight major civil engineering domains: structural analysis and design, concrete technology, geotechnical engineering, hydraulic and water systems, transportation and pavement engineering, construction management, building information modeling (BIM), and green infrastructure with sustainability assessment. The study systematically analyzes findings from 106 peer-reviewed sources, highlighting how AI models, especially machine learning (ML), deep learning (DL), and hybrid combinations such as ANN-PSO and CNN-XGBoost, have significantly improved performance in complex, nonlinear, and data-intensive scenarios. The review reveals that AI-based systems have enhanced the precision of strength predictions, classification of soil types, analysis of fluid dynamics, optimization of transport networks, and the integration of BIM environments with real-time data for advanced infrastructure management. In construction, AI supports safety monitoring, resource planning, and risk forecasting, while in sustainability, it aids in carbon footprint modeling and energy efficiency analysis. Moreover, the incorporation of Internet of Things (IoT) devices and sensor-based systems with AI has enabled real-time feedback and intelligent decision support in several infrastructure systems. Despite its transformative impact, AI integration in civil engineering still faces limitations. These include data scarcity, model transparency challenges, high computational costs, and lack of standardized benchmarks. To address these issues, the review outlines key research recommendations, including the use of physics-informed neural networks (PINNs), the adoption of transfer learning for small datasets, explainable AI (XAI) frameworks, and AI-enabled digital twins for infrastructure lifecycle management. This paper serves as a consolidated reference for researchers and professionals aiming to harness AI in the planning, design, operation, and sustainability of civil infrastructure systems. It not only maps the current state of AI in civil engineering but also offers a roadmap for future advancements toward smarter and more resilient infrastructure.