Will the Robots Dance?

The Pending Use of AI in Construction

By Stephen Carter

My 6am morning alarm is now the incessant beeping of a moveable crane reversing to collect and hoist building materials onto a third level platform. I now live less that 200 feet from a construction site of a multiuse complex. This is a relatively new location as we downsized from a leafy neighborhood to an emerging urban village at the convergence of two rivers.

Observing construction, even as I write this, can be mesmerizing as the logistics of a postage stamp-sized laydown area is navigated and re-navigated. While material storage and access may be extremely well-planned, on some days, time seems to mostly to involve multiple moves to accomplish the same task. While construction sites are always more of a punk rock concert than a Handel symphony, artificial intelligence (AI) has begun to significantly alter the chaos.

Just one example. In my youth, I spent my summers working on a construction site and as a non-skilled laborer, one of my jobs was to fetch tools and materials. When this wasn’t some cruel initiation stunt (“get me the lumber-stretchers”), I truly did spend a lot of my minimum wage hours hunting for stuff.  Thanks to emerging AI applications involving site cameras and monitors, IoT (Internet of Things) sensors, and a project data base, finding and confirming a tool can be a nanosecond undertaking.

I have previously written about AI advances in the design and engineering fields in this column. These professions were quick to recognize and incorporate certain aspects of AI in the creative and productive process. The next generation in the design community seems to be generative design which is extremely expansive and open to much anticipated discussion on whether human designers will become an endangered species, but we’ll save that for later.

Back to the construction site, for us AI dummies, the vision of project delivery is mostly limited to robots substituting for sweating laborers. No doubt they already populate some sites and especially in factories where modular or component products are being produced. And, clearly, this will expand and demonstrate time, safety, and cost efficiencies. A pre-pandemic Trends Report on the Digital Future of the Construction Industry reported a year-on-year increased usage of AI at 243%.

Writing on the future of AI in Canada’s construction industry, Simion Kronenfeld, a well-recognized real estate developer, noted that the reduction of risks, prevention of cost overruns, and improvements in job safety and labor shortages would eventually accelerate experimentation with various forms of AI in construction. Kronenfeld suggests that construction applications of AI have the potential to reduce costs by 20%.

While the most visible and politically controversial, robots are a small part of the changing construction delivery scene. Roland Berger, a Canadian-based construction management consulting company with offices in the US, takes a wholistic perspective regarding AI suggesting five areas where introducing the use could improve capacities and outcomes: 1) Design and Planning; 2) Construction; 3) Supply and Facility Management; 4) Material Production; and 5) Marketing and Sales.

In an article on use cases for AI in construction, Kai-Stefan Schober quotes a Roland Berger report that suggests that “productivity output of the whole construction industry has been improving marginally or not at all for decades” and suggests that a major reason is logistical inefficiencies. While I did not think of this while having a morning coffee watching the same laborers move the same material back and forth, logistics was at the heart of the inefficiency problem.

At the most basic level, our ability to accumulate, analyze, and incorporate data is the beginning of efficiency and cost reduction. Schober uses the example of combining an inventory database, site cameras, drone-supplied images, a mobile phone, and IoT sensors to find a hot saw in less time than Brady can find an open receiver. In a better time (before supply-chain and material shortages), labor was at least 50% of the construction cost and any improvement in job-site efficiency is bound to reduce costs and accidents.

Turning briefly to the correctional construction market: as more owners are considering the use of modular or factory-assembled building components, the opportunities for generative design and AI-based construction approaches increases. As with the operational concepts of direct supervision, dynamic security, and normative design, the key to success will be education. We need to talk more about what it is; why it should be considered; and how results will accrue through every step of the project delivery process. And, we have to promote planners, designers, engineers, contractors, operators, and maintainers taking deep dives together into how this future can be managed.

Finally, I just want to acknowledge that there are a host of political and ethical issues to be addressed.  Let’s be honest; much of America’s infrastructure was and continues to be constructed by immigrants. I can’t imagine (yet) a day when a prison is constructed exclusively through AI, but I can easily envision a day when the typical on-site labor requirement is halved. This opens complicated discussions on immigration policy and, frankly, financial survival for millions of construction workers.

And then there are the ethical issues of what happens after jobs disappear; how do we distribute wealth created by machines; how will buildings created through artificial intelligence affect our behavior; how will we design-out bias in robots (I wrote about this in an earlier Trends article); and how do we literally prevent evil (eg.: potential expanded use of drones) from corrupting our outcomes? There is a lot of thinking, debating, and planning to be done.

Maybe at my age and stage, I should just eat my toast and drink my morning coffee in gladness and not worry too much about the construction site noise. A building will result (and with the better views of the skyline and river may be a new home for me); laborers will feed their families; the future restaurant will promote civil conversation over dinner; and this small, incorporated village will feel a noticeable economic lift. The next generation will sort out the dance of the robots.

Stephen Carter, AICP, is the executive vice president and global strategic development officer for Miami-based CGL.

Editor’s Note: This feature originally appeared in the July/August issue of Correctional News.