Artificial Intelligence isn’t just a pipedream in the civil infrastructure sector – it’s already on-site, rolling up its cyber sleeves and getting stuck in helping the team with what it does best.
But will it take your job? That’s unlikely, at least according to Matt Ensor, Kia Ora AI Director and Chair of the AI Forum of New Zealand’s working group on generative AI.
Matt says the reality on worksites is AI augmenting and supporting workers by providing a new and powerful tool that makes jobs safer, easier to master and more rewarding.
“AI won’t take your job – it’ll come with it. I’ve yet to find a job that it could fully replace.”
For people considering careers in civil infrastructure, this presents an exciting opportunity to enter an industry that’s embracing cutting-edge technology to solve age-old challenges.
“AI will become part of people’s jobs and I’m hoping quite a natural part, you won’t have to understand how it works to use it.”
Smart support means less admin
Matt says the first trend is AI taking over the tedious administrative tasks that nobody enjoys, like writing reports, summarising documents and creating routine emails, with Microsoft Copilot, ChatGPT and Claude popular.
“AI helps you do all the stuff you don’t want to do, the admin stuff.”
That’s good news for those who like working on the field. With AI handling admin, people can spend more time on the tools, doing site management, building relationships and problem solving rather than tackling paperwork.
Less time spent on admin allows newcomers to the industry to develop core industry skills faster as well, he says.
AI as an expert consultant
Matt says more exciting is AI’s ability to help with more complex tasks, by supplying high-level knowledge people may not possess.
“Look at those tasks that you do which require a higher level of skill than you have. That’s where AI is really making a difference.”
Existing technology lets people use their smartphones to access AI-powered insights on-site, for example to identify safety risks or make engineering calculations, Matt says.
“Everyone now has access to using AI through a phone’s camera and AI is pretty good at construction.
“What we’re looking at is how do you create what we call ‘agents’, so they can effectively give you information or assist you in ways that are beyond your expertise.”
An AI ‘agent’ is a specialised AI that knows your job inside-out and can give you expert advice in real-time.
For example, a junior engineer can use an agent to access top-level insights on advanced problems by taking a photo of concrete cracking and asking the agent what might have caused it and what they should do next, accelerating their learning and improving project outcomes.
Instant access to decades of experience
One of Matt’s favourite applications draws on his 35 years as a civil engineer, where knowing which senior colleague to choose to ask was important.
“One of the best skills to have was knowing who to ask about something because there’s nothing new in most engineering projects – someone would have come across that problem before, or someone had almost made that mistake before.”
Modern AI tools can delve into your company’s files and access exactly what you’re looking for, he says.
“You can ask some very specific questions and it will search everything that your company’s ever recorded and come up with an answer. It won’t give up.”
For people new to the industry, this is like having instant access to the wisdom and experience of every senior colleague in the company – past and present.
AI-powered safety and risk management
Safety is paramount in civil construction, and AI is revolutionising how risks are identified and managed. Matt demonstrates this with a practical example.
“Hold up your phone while someone else is driving and ask ChatGPT, ‘What are the risks? What might go wrong? What should the driver be worried about right now?’. You’ll see that it has this uncanny ability to detect risk and overlay it with human behaviour.”
This technology could be integrated into helmets, other safety equipment or site-monitoring systems, providing real-time risk assessment and preventing accidents before they happen, he says.
“AI can sit there watching the site for health and safety issues all day while you’re doing something else.”
Quality control: AI as a second set of eyes
Contrary to popular belief, AI is better at reviewing work than generating it from scratch, Matt says.
“Most people think because generative AI is called generative AI, it’s good at generating, which is entirely wrong. It’s good at review.”
This is particularly valuable when dealing with complex specifications, client requirements and regulations.
“You can get the AI to worry about all that stuff and just say: ‘Hey, what do you think about what I’ve done and review this against everything you know about this project’. It does a really good job.”
For new entrants developing their skills in the sector, this provides an invaluable development tool.
“If you just get AI to generate everything for you, you learn nothing, but if you get AI to review everything that you do, you actually accelerate your learning because you’ve got access to its expertise on your phone at any moment.
“You can also ask those embarrassing questions that you might be a bit shy about asking a real person.”
This review capability extends beyond documents to visual inspection through machine learning. On construction sites, AI can analyse photos taken by drones or smartphones to spot potential issues that human eyes might miss.
For example, AI can review images of bolt installations and immediately flag if the wrong number of bolts have been used or if they’re incorrectly positioned. By training on databases of correct and incorrect installations from similar projects, it can identify patterns and anomalies that even experienced workers might overlook.
Staying in control
While AI offers tremendous benefits, Matt emphasises the importance of maintaining human oversight.
“Don’t ever let AI take over a process without a human still being fully accountable for that process.”
He says the key is treating AI like any other tool that requires proper risk management.
“You can’t just say, ‘Oh, the risk is too high’. You need to work out how to manage those risks with accountability.”
Privacy is another key consideration, with ChatGPT remembering everything you’ve ever said to it, he says.
Some users accept this as a trade-off for the convenience of their software knowing them inside out, and delivering them perfectly tailored content, while others aren’t comfortable with an AI knowing all their business.
An industry that evolves
Looking to the future, the civil construction industry is continually evolving to become more efficient, safer and innovative, with AI just one piece of the puzzle, Matt says.
“I think the sector will get more exciting because things will happen more quickly, we’ll make fewer mistakes and hopefully it will make people happier.
“People shouldn’t be fearful about AI taking their jobs, it’s not going to change the need for people in civil construction.”
