Buildings and construction are among the most resource-intensive activities on the planet, consuming vast quantities of concrete, steel and aluminium and accounting for a substantial share of global energy use and energy-related carbon emissions. As extreme weather events intensify, the sector is under growing pressure both to reduce its climate impact and to become more resilient to it.
In a study published in the KeAi journal Risk Sciences , an international team of researchers from Australia, Finland, Morocco and Thailand examined how artificial intelligence (AI) could help. Following PRISMA guidelines, they systematically reviewed research published over the past decade and combined bibliometric and content analyses to map where AI is being applied, how mature those applications are, and what stands in the way of adoption.
"AI should not be treated as an isolated technological application," says corresponding author Fahim Ullah of Victoria University, Australia. "It is a lifecycle-wide enabling technology. It can support predictive decision-making, optimise energy and carbon, improve resource efficiency, automate processes and monitoring, and it works best when integrated with complementary tools such as building information modelling, digital twins and the Internet of Things."
The review found that AI can contribute across the whole construction project lifecycle: enabling lower-carbon and more resilient materials, greener procurement and supply chains, energy-efficient and climate-resilient design, optimised equipment and scheduling on site, smarter building operation and maintenance, and better recycling and waste recovery at end of life. The team distilled these contributions into a conceptual framework that identifies three principal pathways: cleaner and more resource-efficient energy use, decarbonisation of the built environment, and responsible production and consumption of resources.
Notably, the evidence is unevenly distributed. Most studies focus on mitigation, such as cutting emissions and energy use, while far fewer explore adaptation, such as climate-resilient infrastructure or extreme-weather preparedness. Applications like predictive maintenance, building energy optimisation and computer-vision monitoring are relatively mature, whereas generative climate-resilient design and fully integrated digital twins remain largely at the pilot stage.
The authors also identify key barriers, including poor data quality, interoperability problems, high implementation costs, organisational readiness, regulatory uncertainty and cybersecurity concerns, and they stress that AI is not environmentally neutral.
"The presence of AI capability does not automatically produce a net-positive climate outcome," Ullah adds. "Training and operating these models requires computational resources and electricity. The climate benefits of AI must be evaluated against its own environmental footprint. AI can make a major contribution to a low-carbon, climate-resilient construction industry, but it is neither a standalone solution nor inherently sustainable."
Contact the author: Fahim Ullah, Institute for Sustainable Industries & Liveable Cities, College of Sport, Health and Engineering, Victoria University, Footscray Campus, Melbourne, VIC 3011, Australia; School of Science, Engineering and Digital Technologies, University of Southern Queensland, Springfield Campus, Australia. Email: Fahim.ullah1@vu.edu.au
The publisher KeAi was established by Elsevier and China Science Publishing & Media Ltd to unfold quality research globally. In 2013, our focus shifted to open access publishing. We now proudly publish more than 300 world-class, open access, English language journals, spanning all scientific disciplines. Many of these are titles we publish in partnership with prestigious societies and academic institutions, such as the National Natural Science Foundation of China (NSFC).
Risk Sciences
Systematic review
Not applicable
Artificial intelligence for climate change adaptation and mitigation across the construction project lifecycle
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.