Like all industries, construction is also being automated, but it is one of the slowest that experiences this. This is because the real life variables that come into play in construction are very high, and real life variables are very hard to predict. In other words, it is harder for AI to manage construction projects, than manufacturing cars in an assembly line, making diagnostics in healthcare, analyzing stock trends, cleaning floors, check credit card fraud, profile airline passengers, play chess, because all these tasks can be performed in highly controlled environments, with limited number of variables, which are all highly defined, given to the computer.

Let’s think about one of the most straightforward construction projects, a single story office building. Even a simple project as that, still involves hundreds of real life variables everyday. Material deliveries, quality of the material and installation, coordinating trades and workers… all multiplied by the number of trades. In addition, making and coordinating designs such as architectural to structural to mechanical to electrical… weather effects, unforeseen site conditions such as underground, accessibility to construction site being affected by many outside effects that can happen anytime, access and coordination of all trades inside the site, coordination with neighbors who live or work around the site, administration effects, finance effects, safety related concerns, schedule and critical path of schedule related concerns and many, many more…

So, the number of variables for a construction project are far, far more greater in number and complexity than many other industries that are already experiencing automation to higher degrees. This is the main reason why, for construction, automation is extremely slow. It will be one of the last fields, to resist to ever increasing power of AI and automation. However as we said in the beginning of this post, all industries are being automated, and even construction is not an exception. Why? Because given enough time, the amount of information in EVERYTHING approaches to 100%. Let’s explain what this means…

Let’s think about manufacturing a car 100 years ago. Our knowledge about building cars was low and we needed to spend lot of effort in doing it as we also were learning how to better do it. So the most part of making a car 100 years ago, was the effort, with little knowledge. Now think about 50 years ago. Our knowledge was more. We had to spend less effort, because we knew more. We had more information. The percentage of information in making that was more. Now think about today. We have a lot of knowledge in making cars. Yes workers in factories still spend effort on some tasks, but to a great extent, making a car is now automated. The information percentage in it, is now even greater and effort keeps getting less and less. Now think about making a car 200 years from now. May be all we will need to do is to say “make me a car with this and this characteristics” to a commercial 3d printer in one of the 3d printing stores on the street. And the car will be manufactured quickly. Now the information in it is almost 100%, with no effort spent. So as we can see from this example, the amount of information in everything converges to 100%, given enough time. It is just that for some fields, it takes less time, as they are less complicated, and for some fields, it takes more time, as they are more complicated. This is the only difference. Having said all these, it doesn’t mean that in construction no efficiencies or automation is happening. This is absolutely not true. Already we have seen many, many signs and improvements.

Let’s give some examples below:

• When I started in construction industry almost 20 years ago, the project plans to contractors were distributed manually by mail. It was a lot of mailing and manual work. When I remember now, I still feel it in my brain, how the edge of those plans could cut hand and cause it to bleed, if I tried to sort them quickly. Now, everything is being distributed online, in a much more efficient and organized way, without the very costly and tiring mailing work. You see, it is sufficient to distribute the “information” of the plans, and not the plans themselves, because the parties who receive that information can print it themselves or use it even without printing anything on paper.

• Now we have design and coordination software that we can lay everything out in 3d, which allows us to put everything dimensionally and locationwise, together in a virtual environment, before they get installed in the field. This way we can identify potential conflicts beforehand.

• Engineering design software can make very complicated engineering analyses, once we put into the model everything correctly. The software can make the very complex calculations in only a fraction of the time and more accurately. In fact, it can use methods such as the finite element method, to analyse with greater precision, which then it compares to traditional engineering methods that were done by hand. The use of finite element method by humans is simply not possible as it would take millions of years due to large matrices involved, which is something computers can solve in minutes.

• Now we can inspect things from outside with drones which especially becomes important for high structures, instead of having to rent a helicopter or going near the problem areas one by one with a crane.

• Now we have remotely operated small excavators, which do not need a human operator on it, and can operate in narrow environments effectively.

• As the robotics technologies improve, we will start to see robotic workers on sites, more and more and in increasing sophistication every year. I would say that the field operation will be greatly automatized by robotic workers, in the next few decades, as the robots are now learning to work in undefined real environments. The robotic workers will exceed humans in their particular tasks for speed and accuracy.

The design and coordination part of construction, which is one of the last forts among jobs that humanity will hold against AI, will also eventually fall and be automatized. Bottom line is that, automation is inevitable, and it is only a matter of time, based on the complexity of the industry, the real life variables involved and the tasks in hand.

Here is an example about robots in construction:

http://spectrum.ieee.org/automaton/robotics/industrial-robots/robots-do-construction-with-brick-and-concrete