In an industry that trades on the physical and the tangible, artificial intelligence remains somewhat elusive. Its applications are not so obvious, but its potential impact is broad, and still quite nascent.
But before we analyze the possibilities, let’s clarify what AI is:
Type of AI
- Aautomation: Process things on a screen and perform everyday tasks such as processing invoices.
- Image and data analysis: with sensory cameras. We now use it in workplaces to flag safety issues by identifying who is wearing safety glasses or appropriate PPE.
- Great language model(LLM): Recognizes and generates text. This didn’t exist until ChatGPT launched in 2023. It has gone from infancy to being an “intern” in the last year. In its early days, you had to guide it to produce answers; now, it has more robust capabilities, but still requires a human to check its work.
LLM is the most widely discussed type of AI, and some misconceptions about it are worth clearing up.
What is and is not an LLM?
LLM is a database of all human knowledge, all human experience in one place. It’s essentially an algorithm that allows you to consult human experience and get an answer. A common question is “why is this a breakthrough? Isn’t it just Google?” Yes, but no: Unlike Google or other Internet search engine query results, LLMs understand common language more fully and are better suited to more complex requests. Each LLM has an entire database from which to extract and synthesize data, rather than publishing it via individual links.
Byrtus
A common misconception about an LLM is that it speaks a language, when in fact it processes patterns. Each sentence is a pattern, and it responds in the same way through patterns. It takes the patterns surrounding your query and produces a likely answer based on all human experience for that pattern. For example, the steps required to construct a table transcend language. Regardless of the language the carpenters speak, their steps remain the same: they are a pattern. LLMs sound pretty amazing, don’t they? But what about its limitations?
Well, an LLM has no reasoning ability, and no memory. It does not “learn” in the sense that it cannot refer to a previous question to lead to a new one. The query above will increase the probability that the answer is correct for that pattern, but it won’t “remember” what you just asked. He also has no spatial understanding. If a table is next to a chair, the LLM doesn’t understand what that means. He doesn’t understand where things should be in a given space, i.e. a table goes next to a chair, nor does he understand how to practically lay out the room.
He understands that there is a pattern where chairs are paired with tables, but he doesn’t spatially understand where the chair fits in relation to the table. It does not use visual analysis in its processing, only pattern recognition.
AI capabilities within construction
Construction as an industry faces a looming labor shortage and with it a skills gap as many employees age out of the workforce and not enough people enter the pipeline to replace them. AI can help ease this transition, specifically with its ability to take over more tedious tasks and automate them. You can send a change order; send notifications about security issues, such as a water leak; guiding people to think about problems in a workplace; autofill contract procedures; do complex math; replace structural engineers by automating formulas in a plan (engineers will still need to review and stamp plans); engage in basic architecture (nothing too aesthetic and artistic, but it can process requests to provide a new design based on existing models, but again, it doesn’t understand space, so human input still is essential). Unlike some industries where AI is seen as a threat to the workforce, construction will increasingly rely on AI to ensure it can continue to operate efficiently amid a shortage of human labor.
This means AI can go a long way in getting people out of the trailer and into work. At the moment, the administrator has a lot of weight: there is a lot of paperwork and legal work under construction. Many project engineers spend most of their time in a trailer. With the help of AI, they can be free to focus on building.
A recent use case: One of our competitors had a federal project that required a 17-step process for all change orders, all of which had to be reviewed by every member of a 70-person department. They couldn’t keep up and were weeks behind. With the help of AI, they could program a bot to process these cases within minutes. This AI application is still new, but the opportunity is huge.
The future of AI + Construction
robotics
At the moment, the use of robots in construction is quite limited. Robots can draw lines on a wall and help with some basic tasks if they’re programmed with knowledge, but they don’t yet operate at an AI level. The actual construction of a building is still very much a human activity, from building aesthetics to concrete pouring/placing to electrical installation. It is still an industry focused on human labor.
There is a general fear that humans will be replaced by robots, but this is only applicable to some of the simpler or more dangerous tasks. With an AI engine, you could talk to a robot that would far surpass Siri. They require a lot of high-speed bandwidth and their spatial capabilities are limited, but once you implement visual sensors, their capabilities and understanding increase, and with it, their productivity.
How far are we from this kind of AI robotics in construction? I’m looking at the autonomous car industry, how its spatial component reflects the demands of construction. When they get to level 5 automated driving and figure out how to plug in AI engines, we’ll be close. It is always difficult to put a time frame on these technological developments, but it is likely within the next 10 years. In Europe they can appear even faster.
data
AI has a bright future when it comes to data, where it is currently underutilized. A lot of data exists, but it is far from fully optimized to apply it to increase inefficiencies or distill new learnings from it. Simple tasks are the most accessible: the employee handbook, perhaps, could have an AI engine that allows users to submit a query and receive a response. Other use cases include document summaries, location information, and other data-driven tasks that deliver knowledge quickly and efficiently. Construction is lagging in this application of AI and I expect we will see growth in this area in the coming years.
Quantum computing
One of the challenges of AI is its heavy use of energy and storage. There is never enough internet bandwidth in the workplace and power is always an issue. But as quantum computing becomes more powerful and compact, it will allow us to use AI in smaller, more accessible packages, which means implementing it in better and more efficient robots, computers, and algorithmic engines. Quantum computing will also accelerate computing power, further increasing efficiency.
A word of caution
We don’t have to look very far to find examples of well-intentioned technological advances that quickly go dark. So what could possibly go wrong with AI in construction?
If you don’t use an enterprise tool and just use open AI on the Internet, all entries become public domain. This is a risk because the AI often “hallucinates” as it doesn’t always give the right answer. It’s a probability engine, and there’s always the chance that it will give you a “bad” answer, which could lead to safety issues, contractual liability, and lawsuits. This means that the AI is only as good as the person checking the work. Eventually we’ll have an AI that checks the AI’s response, but not yet. It is still in its “internal” phase.
Despite the challenges, AI will accelerate our ability to execute projects and solve problems. It will also help people focus on becoming better builders, rather than simply learning systems. And more time spent building is a win for everyone.
Matt Byrtus is the vice president of business technology for XL Construction based in Milipitas, California.
