AI in construction: addressing the first steps and the fears

That’s one of the key takeaways from the Digital Construction Summit webinar focused on AI in the wake of the recent publication of the Chartered Institute of Building’s AI Playbook.

Vicki Reynolds, chief technology officer at Catalyst, emphasised the need for quality data and a back-to-basics approach. “First things first, any data you use has to be accurate, relevant, and trustworthy. And it will be on a case-by-case basis that you determine what that means for you, and the problem you’re trying to solve.

“Ultimately, it comes back to that first question: ‘what’s my problem and how am I going to use this tool to make life easier, greener, safer?’ That then allows you to determine the most appropriate data set for you. The great thing about it is, it is just all of the same practicalities and processes around good project management.”

Stefan Mordue, senior manager for education programs and partnerships at Bentley Systems, noted: “One of the things we advocate [in the Playbook], particularly with the use case templates, is to think about establishing an internal AI working group in your organisation of like-minded people, where you get together and you start to think about these big questions. But start small, think about a test. Use test cases, evaluate them and what you’re going to get out of it; use trial-and-error projects as well. It can be very daunting at first, but try to break it down.

“The more we get comfortable with using AI in our own personal lives, the better we’ll get at things like asking the right questions [of AI in our professional lives]. I think we’ll become a little bit more comfortable and we’ll have more instinct about whether we trust this data or not.”

A webinar attendee asked whether any current standards can help inform best practice use of AI and whether the UK government will follow the EU AI Act? May Winfield, global director of commercial, legal and digital risks at Buro Happold, answered: “One thing to bear in mind is that there’s a lack of standardisation globally on how AI is being approached. The UK has taken what it calls a pro-innovation light touch, but in the King’s Speech they said there would be laws regarding AI. What will they look like? The intention is for them not to be as comprehensive as the EU AI Act.

“The EU AI Act, in turn, separates things, depending on how risky they are. The higher the risk, the more regulations there are. It’s [a case of] watch this space. As with all things, the law and regulations tend to play catch-up with digital and technology.”

David Philp, chair of the CIOB’s innovation advisory group, suggested the Alan Turing Institute’s AI Standards Hub is a good resource for those seeking best practice guidelines.

Responding to an audience question about whether clients should ask their supply chains about AI use, Winfield said: “I think it’s sensible for a client to be asking, ‘are you using AI?’ You want to be comfortable that your supply chain is using tools that you are happy with. For example, what tools are being used to record meeting minutes?”

She reminded the audience: “At the very basic level, you’re liable for what you do. If you have used AI, and you have relied on it, it’s still your work in the contract. AI isn’t a living person: liability cannot attach to it.

“There’s an interesting aspect in terms of copyright. In the US, there are some cases that say if something has been created by AI without sufficient human input, then no copyright attaches to it. And that’s a problem if you create something for a project where you think you have copyright, but actually it’s free for all to use. And your client may also be a bit upset with that. What did they pay for? Think of AI as a tool – like you would a calculator.”

The future impact on skills and jobs was inevitably a part of the panel debate. Karoliina Torttila, director of AI at Trimble, highlighted the challenge of upskilling experienced construction workers to use AI and training inexperienced new entrants on AI. “The interesting thing for universities to figure out is, as we upskill our labour force on AI, how do we do it in a way where we still retain the core knowledge and core skills instead of just automating things to a level where nobody understands how the physics work or how the legal details work?”

Musing on the fear factor, Mordue added: “I think it’s one that we all ponder, and I think it’s one we can’t say with any certainty. However, I think what we can do is look at history. We can look at how these things play out. I think a lot of the repetitive and mundane tasks will be automated. But you still need somebody to make sense of that information. I think it will be a very slow change.

“My advice would be to broaden your skill sets. We need people that are digitally literate. Prompt engineering is a new, emerging role. Ten years ago, app developers weren’t really around, but now that’s an in-demand job.”

Torttila said it’s more productive to engage early with how AI will benefit your role rather than to default to the defensive position of ‘what will AI take away that makes me redundant?’.

“Take estimators, for example. Anybody who has to count electrical symbols on a 2D drawing time and time again knows that this is not the most fun thing to do on a day-by-day basis,” she said. “What’s really great in developing this technology is when people who are doing take-offs today, and are struggling with how much manual work there is, engage with us early in the development process. It’s really inspiring to hear how they think their roles are evolving; an estimator’s job is much more than just counting symbols or executing something without much thought.”

Watch the webinar on demand.

The webinar was part of the Digital Construction Summit, organised by BIMplus and Construction Management, in association with the CIOB, Atvero, Ideagen Mail Manager and Trimble.

Don’t miss out on BIM and digital construction news: sign up to receive the BIMplus newsletter.