Skip to content
‹ Back to IndustriesEngineering & Construction

206 construction AI studies show far more work on emissions than on extreme weather

A review of 206 construction AI studies spanning a decade finds AI can support climate change mitigation and adaptation across every project stage, but the evidence is lopsided. Tools that run assets more efficiently are rated relatively mature, while generative design for climate resilience and fully integrated digital twins are still largely in pilots.

This story was produced through MarketScale. See how Engineering & Construction teams put it to work with Partner & Channel Enablement.

By MarketScale Newsroom · Victoria UniversityRisk SciencesAi in ConstructionClimate Adaptation
Share

Key takeaways

01

After 206 studies spanning a decade, the construction AI rated relatively mature is mostly about running assets more efficiently, while AI that would design buildings to ride out extreme weather is still largely in pilots.

02

Any AI pitched as a climate win should be judged net of the compute and electricity it consumes to train and run.

Free workspace

Turn your Engineering & Construction expertise into content.

Record interviews, organize footage, and write with AI on a free trial of the MarketScale platform for qualifying companies. No demo required, no credit card.

Try it Free

Researchers from Australia, Finland, Morocco and Thailand, working as an international team, looked at how artificial intelligence could help construction cut its climate impact and better withstand climate change. They charted the areas where AI is being used, how far along those uses are, and which barriers are holding back adoption. Their finding was that AI can support both climate change mitigation and adaptation at every stage of the construction project lifecycle, from material production through end-of-life management. The evidence, however, is spread unevenly.

The review was published in the KeAi journal Risk Sciences. It sorts the research three ways: by project stage, by how mature each application is, and by what blocks adoption. Most of the studies deal with mitigation, meaning lower emissions and energy use. Far fewer deal with adaptation, such as climate-resilient infrastructure and readiness for extreme weather. That split could matter to anyone choosing construction technology this year.

The stakes are physical. According to the release, few activities on Earth demand as many resources as building and construction. The sector uses enormous amounts of concrete, steel and aluminium, and it is responsible for a large portion of the world's energy consumption and of carbon emissions tied to energy. The paper appears in a Risk Sciences special issue titled "Building Resilience Against Climate Risks," which makes its main finding a little ironic: in the evidence it maps, resilience is the weaker side.

A decade of studies sorted by project stage

The corresponding author is Fahim Ullah of Victoria University, Australia. The team paired bibliometric analysis of the publication record with a detailed reading of the papers' content. According to the paper's abstract on ScienceDirect, the review covers six lifecycle stages: material production, procurement, planning and design, construction operations, use and maintenance, and end-of-life management.

The eight authors followed PRISMA guidelines, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses, a standard protocol for how a literature review finds, screens and reports the studies it includes. That matters for the maturity ratings below, because it makes the sample a defined, documented set rather than a curated reading list.

The review in numbers

206
Articles in the evidence base
2015 to 2025
Publication window, so the map reflects a decade of research
3
Pathways in the authors' framework linking AI to climate outcomes
4
Countries represented on the author team

Risk Sciences (KeAi); EurekAlert!

Across those stages, the release summarizes what AI can contribute in roughly this order:

  • Material production: lower-carbon and more resilient materials
  • Procurement: greener procurement and supply chains
  • Planning and design: energy-efficient and climate-resilient design
  • Construction operations: optimised equipment and scheduling on site
  • Use and maintenance: smarter building operation and maintenance
  • End of life: better recycling and waste recovery

Read the list for the word "resilient" and it attaches to only two stages, materials and design. Everything else is framed as efficiency or waste reduction. That pattern lines up with the study count: adaptation shows up mostly at the front of the lifecycle, where a building's ability to handle heat, flood or storm gets decided, while the efficiency work spreads across every stage after it.

The authors condense all of this into a conceptual framework with three pathways: cleaner and more resource-efficient energy use, decarbonization of the built environment, and responsible production and consumption of resources. The framework links AI interventions across the project lifecycle to climate outcomes, implementation feasibility, technological maturity, barriers and potential adoption trade-offs. The paper also highlights potential contributions to UN Sustainable Development Goals 11, 12 and 13.

A systematic review counts and classifies published research. It shows where studies cluster and how far applications have progressed, but it does not measure emissions cut or costs saved on any particular project. A maturity rating here describes the state of the evidence, not a vendor's product.

Mature tools run buildings; resilience tools are still in pilots

After 206 studies spanning a decade, the construction AI that researchers rate as relatively mature is mostly about running assets more efficiently, while the AI that would design buildings to ride out extreme weather is still largely in pilots.

Relatively mature

The review names three applications as relatively mature: predictive maintenance, building energy optimization and computer-vision monitoring. All three live in the construction-operations and use-and-maintenance stages, and all three are about doing existing work with less energy, less downtime or fewer surprises.

They also map closely onto the capabilities Ullah lists in the release: predictive decision-making, optimizing energy and carbon, improving resource efficiency, and automating processes and monitoring. That overlap suggests the mature end of the field is mature for a reason. These tools act on assets that already exist and produce data, so their results are easier to measure and publish.

Largely at pilot stage

Fully integrated digital twins and generative design for climate resilience are still largely in the pilot stage.

Generative climate-resilient design sits in the planning-and-design stage, which the stage list above identifies as one of only two places where resilience is explicitly in scope. The digital-twin finding carries its own twist. Digital twins are one of the complementary tools Ullah says AI works best with, so the fully integrated version of the setup he recommends is itself still largely a pilot.

Ullah's main point in the release is about how the tools work best. He describes AI as an enabling technology that runs across the whole lifecycle rather than a standalone application, and says it performs best alongside building information modeling, digital twins and the Internet of Things. For firms whose BIM models and building sensors already share data, that is an argument to build on what they have. For firms without those connections, the barrier list below suggests the data work has to come first.

The resilience gap meets a resilience agenda

The review landed as governments were talking about exactly the thin side of its evidence. The U.N. COP31 climate summit ran Nov. 9-20 in Antalya, Turkey, with Turkey as official host and Australia leading the negotiations, according to Associated Press reporting published by PBS News. COP31 president Murat Kurum, Turkey's environment minister, announced an "Action Agenda" that includes building climate-resilient cities, alongside reducing methane emissions and raising the share of electricity use from 21% to 35% by 2035.

Adaptation money was contentious from day one. Delegates were expected to take up the question of how poorer countries could pay for adapting to higher temperatures and extreme weather, and negotiators from nearly 200 countries were already arguing over funding for climate adaptation and for recovery after climate disasters. In opening remarks, Kurum said this was the first COP at which there is no reasonable doubt the world will exceed the 1.5-degree target.

The paper's own introduction frames the same pressure through events: the Black Summer fires in Australia, Cyclone Gabrielle in New Zealand, Cyclone Idai in Africa, the European drought and the Henan flash floods in China, along with IPCC, WMO and UNDRR assessments pointing to more frequent and intense climate hazards. Put that next to the summit agenda and the gap reads differently. If policy attention and adaptation funding move toward climate-resilient cities, the construction AI meant to serve that goal would be drawing on the field's least-documented evidence.

Owners and design firms writing specifications for resilience work could reasonably treat AI resilience features as pilots, with their own evaluation criteria, instead of assuming the maturity of an energy-optimization tool carries over.

The strongest objection to reading too much into the gap is that it measures where researchers have published, not where tools fail. Fewer adaptation studies means fewer documented results to check a vendor's claim against. It does not show that the claim is wrong. The framework's own structure also cuts against a hard line: a tool that makes a building more efficient to operate may also help it cope with stress, and the review does not sort every application neatly into one bucket.

Barriers, and the model's own power bill

The authors also identify key barriers: poor data quality, interoperability problems, high implementation costs, organizational readiness, regulatory uncertainty and cybersecurity concerns. They stress that AI is not environmentally neutral.

The footprint point should change how projects are scored. In the release, Ullah said that simply having AI capability does not by itself guarantee a net-positive climate outcome. He noted that training and running models use computing resources and electricity, so any climate benefit has to be weighed against that cost. He concluded that AI can contribute a great deal to a low-carbon, climate-resilient construction industry, but it is not sustainable by default and cannot fix the problem on its own.

Turned into procurement questions, which is our reading and not a requirement set by the authors, the barrier list gives a buyer three things to check:

  1. What data quality does the tool assume, and does the project or portfolio actually have it?
  2. Does it exchange data with the BIM and sensor systems already in place, or does it need a parallel pipeline?
  3. What electricity does it draw to run, compared with the savings it claims?

When a vendor quotes energy or carbon savings, ask for the number net of the tool's own compute and electricity use. The review's authors say that is the comparison that decides whether the climate outcome is positive.

What would move the pilot column

Adaptation AI has been flagged as unfinished before. In a 2022 paper in Results in Engineering, Beril Sirmacek and Ricardo Vinuesa proposed combining AI and simulation to extract climate-adaptation indicators for cities from satellite images and to project how those indicators might change. They described the area as "an open field and an on-going area of research," and said such models could help decision makers and early responders once they become more robust and are used in real-life applications.

Their reasoning also touches the data-quality barrier. They noted that not all cities can collect large volumes of data from distributed sensors, and argued that satellite-derived indicators could support predictive models usable in any city. That is one route around a sensor gap. The new review's finding that adaptation remains the thinner side of the literature suggests proposals like theirs have not yet turned into a large body of documented field results.

The read changes when published field results for generative climate-resilient design or fully integrated digital twins start to accumulate. Until then, the operating-phase tools have the stronger evidence, and the design-phase resilience tools are where a careful buyer builds in a trial period, a baseline and a way to stop.

Featured companies

Your experts belong here

Every story in MarketScale Engineering & Construction starts with a company putting its project engineers, superintendents, and estimators on the record. Buyers are already reading this topic. The only question is whose experts they find.

Owners shortlist firms they already trust, and your field leaders become the reason your name is on that list.

Book DemoSee how it works15 minutes, straight to a calendar.

About the author

MarketScale Newsroom
MarketScale NewsroomEditorial Team, MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

B2B Weekly

The week in Engineering & Construction, and sixteen other industries, every Monday.

Ten stories, one-line takes, five minutes. Free.

Engineering & Construction: are you visible to AI?

Before they reach out, Engineering & Construction buyers ask AI engines which vendors to trust. Explore how your experts, customers, and partners can become useful content for buyers and AI search.

Free Trial

You just read one Engineering & Construction expert. Your company is full of them.

This article was produced through MarketScale. The same platform turns your project engineers, superintendents, and estimators into the articles, video, and social content Engineering & Construction buyers are searching for. Start a free trial and see it with your own people. For qualifying companies, no credit card, no demo required.

NPS +73 · 1,000+ creators · 38+ countries

What your free trial includes

Hands-on access to the MarketScale platform
Media requests to your crowd, remote recording, AI writing tools
No demo required. No credit card.
For qualifying companies. Company confirmation required.

More Engineering & Construction Insights

Why Lighthouse Electric joins data center jobs before the drawings exist

Why Lighthouse Electric joins data center jobs before the drawings exist

Lighthouse Electric is brought onto data center projects before any drawings exist, contributing labor capacity, material lead times and logistics knowledge instead of installation. That moves much of the risk into version control. Every handoff from the modeling team to the fab floor to the field crew has to stay on the current file.

  • 01Lighthouse Electric gets hired onto data center jobs before the drawings exist, which means the first thing it sells is what it knows about labor, lead times and logistics.
  • 02ABC’s nationwide backlog indicator stood at 8.7 months in September, with most of the growth attributed to data centers. That demand, plus rising tariff-affected inputs like copper, makes crew availability and lead times a key constraint to surface before owners lock in systems.
  • 03One signal to watch: ETASV wants to add an exam and issue a Bluebeam certificate that contractors could request by name, according to Bluebeam. Contractors calling the union hall can already ask for workers certified in confined space or high voltage.

Nov 20, 2026

Blended cement is ready to sell, and public tenders decide how fast it spreads

Blended cement is ready to sell, and public tenders decide how fast it spreads

A campaign backed by sponsors including the Global Cement and Concrete Association wants public tenders to specify blended cement, arguing government buyers make up a large share of the infrastructure market. Some blended cements can cut carbon by up to half. The mixes already exist, so public-works specifications, building standards and contractor schedules help determine how quickly they scale.

  • 01Blended cement that cuts carbon by up to half is already for sale, and what moves it at scale is a public tender that asks for it.
  • 02Before switching to a high-replacement mix, ask the supplier how fast it reaches strength. A delay of a few days changes a pour schedule, and some newer admixtures are designed to close that gap.

Nov 19, 2026

Jacobs and TYLin get up to $150M to bolster MWAA project oversight

Jacobs and TYLin get up to $150M to bolster MWAA project oversight

The Metropolitan Washington Airports Authority hired Jacobs and TYLin under task-order contracts capped at $150 million over five years, expandable to $300 million over 10. The authority is paying for extra staff to handle overlapping work. Dulles' Package T tunnels must fit with three other construction packages, while Reagan National's Terminal 1 procurement runs in parallel.

  • 01MWAA lined up up to $150 million in outside program management support while Package T was still at the request-for-information stage, as the authority prepares for multiple projects and overlapping procurements at Dulles and Reagan National.
  • 02Even at its full 10-year, $300 million ceiling, the oversight pool serves a Dulles redevelopment valued at more than $20 billion, and it also covers Reagan National.
  • 03The first task orders will show whether Jacobs mainly leads Package T or provides staff across both airports. Package T responses are due Nov. 23 and Reagan National construction manager at-risk proposals are due Dec. 7.

Nov 19, 2026

Explore More Engineering & Construction Insights

Read more expert perspectives from across Engineering & Construction.

Browse Engineering & Construction Hub

About the Expert

MarketScale Newsroom
MarketScale Newsroom

MarketScale Newsroom

MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

For B2B teams

Your experts could be publishing here

Stories like this one run on content MarketScale captures from real practitioners. See how your team's expertise becomes coverage in Engineering & Construction and beyond.

Book a Demo

Or call us. No forms required. We pick up. 214-945-2512