Word to the Wise: How AI Will Shape Tomorrow’s Construction IndustryPosted on October 19, 2021 by Ceco Building Systems
You know the type. They’ve been in the business since before you were born. They’ve seen it all. They’ve been through it all. They’ve solved every problem, adapted to every trend and rode every paradigm shift that came their way. At Ceco, we’ve known many like this – both within our organization and outside it.
This typically older, wise person is utterly unflappable and collected no matter what the situation. Everyone goes to them with their questions and crises and comes back with an unexpected answer or solution. Maybe you’ve been lucky enough to work with someone like this. Maybe in your organization, you are this person. Either way, it’s unfortunate that once they retire or move on, all that knowledge goes with them. But it doesn’t have to be that way.
You’ve probably seen artificial intelligence (AI) and deep learning represented in a variety of spectacular ways, as intelligent android servants, super soldiers or omnipotent machine brains. But in reality, the best use of AI for the construction world is much closer to the old-timer who’s seen it all. And, nearly literally, that’s how AI and deep learning work. The more raw, historical data and outcomes you feed them, the better their intelligence and predictions.
Intelligence Vs. Learning
Though they’re often used interchangeably, AI and deep learning are slightly different. Artificial intelligence is a computer or program which mimics the cognitive functions of humans. This could be problem solving, recognizing patterns, understanding images or learning from previous data and results.
Deep learning (or “machine learning”) is a subcategory within AI, and uses statistical techniques to “learn” from data, without any explicit programming. It can quickly digest millions of data inputs, spot trends within them, create insights and make predictions based on them. Deep learning can learn to do anything from analyze financial reports and project future outcomes to find the fastest way to beat Super Mario Brothers. The more data and more attempts it’s allowed, the better the outcome. Whereas AI is comparable to the research and raw data end of the spectrum, deep learning is more like the interpretation of that data – and execution based on that interpretation.
Intelligence That Never Retires
Instead of requiring a lifetime of experience, deep learning holds the potential to be deeply wise in the space of a few months – maybe even less. And there’s no reason for it to ever retire. You’ve already seen this type of learning put to use; you just may not realize it. For instance, do you remember when you had to sort through all your own emails? Now, smart servers automatically prioritize messages likely to be important and categorize the rest into social, promotional, spam and other categories. Or, have you noticed a search engine automatically filling in the rest of your search? These are relatively commonplace applications of smart machine learning, but for the construction industry, the possibilities are much more exciting. Better yet, each year, the barriers to entry get steadily lower.
Intelligent Foresight & Intelligence for Sites
The first place we’re seeing introduction of deep learning is in budgeting, cost and project management. Giving machines years and years of past project data, including costs and impacts of rework, delay, errors, and other crucial data points can give them the information they need to find patterns and outliers, preventing costly delays. On small projects, it might be easy to predict what could happen if a permit gets held up. But on multi-year, large scale projects with many moving parts – or for companies managing many simultaneous projects, trades and vendors – these impacts are impossible to foresee. AI-empowered management software can run through both “what if?” scenarios and contingency planning. Unexpected situations are no longer unexpected. They can be anticipated and planned for.
We’re Only Human
Human error is a reality we can’t ignore, and unfortunately the more labor-intensive a task is, the more vulnerable it is to human error. That can be incredibly expensive. In 2018, rework cost the industry more than $31 billion. AI can spot and catch potential problems earlier than humans, provide a real-time view of a project’s financial health, and even provide intelligent options for streamlining the process. This prevents costly delays and downtime and leads to better-managed schedules, wider margins and higher returns.
Have Your Cake & Eat it Too
AI can provide similar benefits in the design process. Using a synergistic integration of BIM technology and deep learning, it’s possible to explore countless variations of a model to identify the best option. Called ‘generative design’, this technology allows you to experiment with tens of thousands design changes – from minor to major – then use the best outcomes for a more stable, safer, economical or aesthetically pleasing design. Time-intensive exploration like this – that would take a human months – can be executed by AI in a few hours.
Get Your Piece of the AI Pie
As promising as these potential breakthroughs are, they’re just the beginning. Around the world, more than $10 trillion is spent each year on construction. That figure is expected to grow by 4.2% each year, and with creative use of AI, you can secure more and more of that growth. A McKinsey report from 2017 pointed out that productivity could be boosted by as much as 50 percent through real-time analysis of data. But it can also be put to use on-site.
Such Great Heights
AI can help you see around corners – both figuratively and literally. Drones may have become commonplace, allowing surveying, overhead images and project views that were either impossible or very expensive in the past, but this only scratches the surface of what drones could do. Combined with AI that can train drones to ‘understand’ what they’re seeing, drones could monitor sites for adherence to drawings and plans, create records of progress and even police workers for safety protocols such as wearing helmets or practicing social distancing.
We’re even beginning to see the use of robotic construction equipment. Though the days of a worksite full of autonomous erectors, bricklayers and concrete finishers is still in the distant future, today it’s not unheard of for remote diggers to work 24/7 clearing and leveling a site. There are even robots capable of pouring concrete, installing drywall and laying brick. Though these technologies are still in their infancy, recent decades have taught us the technology grows up quickly.
Eye on the Prize
With expensive equipment and suddenly valuable steel and lumber sitting unmonitored on a job site, these locations can be a thief’s dream. But smart sensors can monitor everything from temperature and application of materials to equipment condition and worksite security. And with cameras and sensors hooked up to AI, you can keep an intelligent eye on your worksites, monitoring not only for suspicious activity, but also to give you a heads up when heavy equipment needs maintenance.
There are numerous businesses trying to train AI systems, but a shortage of the kinds of data scientists needed to conduct this training has made for extremely high salaries. Competition is fierce to attract top AI talent, meaning the biggest firms tend to win out. But these same high salaries and intense competition mean many of today’s students will be eager to enter the field, bringing tomorrow’s supply up.
The idea of AI that can think, plan and manage better than us, that can monitor worksites and even perform construction trade functions might be more frightening for some than it is exciting. It leads to wondering, “If we don’t need people to do the work, we don’t need buildings. If we don’t need buildings, we don’t need construction. Even if we did, robots could build for us. How long until we’re all obsolete?”
AI experts assure us these worries are in vain, because AI just helps us do our jobs better by freeing up massive amounts of time that would’ve been required to do tasks by hand. It’s like a secret weapon, but it needs human intelligence, and allows us to focus on the more important, strategic concerns. That means more human minds doing what humans do best – dreaming up new ideas and innovative solutions. So, those among the first to embrace AI will likely come out on top.