Artificial intelligence in construction risk management: a decade of developments, challenges, and integration pathways

The increasing complexity and uncertainty of construction projects have driven a paradigm shift in risk management, with Artificial Intelligence (AI) emerging as a transformative tool. This systematic review examines 87 peer-reviewed studies published between 2014 and 2024, assessing the role of AI technologies—including machine learning (ML), natural language processing (NLP), knowledge-based reasoning (KBR), optimisation algorithms (OAs), and computer vision (CV)—across the construction risk management lifecycle. Using PRISMA guidelines and NVivo-based mixed-method analysis, the review identifies AI’s core functions in risk identification, assessment, response, and monitoring, while uncovering key barriers such as fragmented implementation, limited adaptability, and socio-organisational resistance. ML and CV dominate predictive and visual analysis, while NLP and KBR support compliance and rule-based decision-making; OAs optimise resource trade-offs. Despite increasing research, critical gaps persist, including fragmented AI integration, limited attention to implementation barriers, and underrepresentation of adaptive response and monitoring stages. This paper makes four specific contributions: (1) it offers a functional taxonomy that maps AI methods to distinct risk management tasks and phases; (2) it synthesizes technical, organisational, and ethical challenges hindering real-world AI adoption; (3) it introduces an integrated AI–risk management lifecycle framework that aligns AI capabilities with construction project needs; and (4) it extends classical risk epistemologies by embedding Beck’s, Power’s, and Aven’s perspectives into the discourse on AI governance. These contributions provide a structured foundation for advancing research and guiding practitioners toward responsible, resilient, and explainable AI deployment in high-stakes construction environments.