How Can AI Be Used in the Construction Industry?
AI is an exciting field of development in the modern world. Over the coming years, an estimated $1 trillion are expected to be spent on CapEx, including significant investment in data centres, chips and the power grid. As companies try to be amongst the first to successfully implement AI into their practices to improve productivity and efficiency, the construction industry must ask itself if there is a place for AI to help create more sustainable designs in the future.
The first question to ask is where would engineers and architects be able to use AI effectively? Certain architectural firms have begun to use AI software at the conceptual design stage, allowing them to create more iterations of the design at an early stage that helps mould the most efficient structure possible, and steer the project in the correct direction. By inputting information about the site, an AI programme can generate structural forms that optimise variables such as cost, carbon emissions, light gains and energy consumption, which may result in more sustainable designs later on.
While AI may be useful at a conceptual stage, it is unlikely to be useful during later design stages. Much of AI in its current form relies on a huge log of historic data that it can use to create a new, possibly innovative solution that aligns with what has been done in the past. However, it doesn’t inherently understand the output it gives, which allows for a relatively large amount of error. Human input is still needed to check the solution an AI software may give, meaning that AI would likely impede work at this time, leaving designers to prefer designing with a pencil rather than a computer at these latter stages.
Alternatively, AI could be used more for monitoring and maintenance during and after construction. Similar to the medical industry where AI is used on MRI scans to detect lesions on the brain far before a human would be able to, engineers and surveyors could utilise this technology to monitor buildings. Data taken from site could be analysed to assess building elements (such as the structural or mechanical systems) and predict where maintenance may be needed. Maintenance costs could then be minimised by hitting a balance between carrying out a larger number of cheaper repairs very regularly, or carrying out fewer, larger repairs once systems have started to fail further.
AI-powered cameras and sensors could also be used on site to assess the quality of work being carried out and alerting workers when something comes up. Tighter building control could allow engineers to design to smaller safety factors which would reduce the volume of material needed for a project, and may have huge carbon savings if used across the industry.
In conclusion, the development of AI may offer tools that the construction industry could use to move towards more sustainable design. However, in its current state, it’s unlikely that AI will be used widely at later design stages. With these latter tasks requiring a dialogue between practices, and an understanding of the brief and how buildings work, companies would likely have to develop a “general AI” – a currently theoretical type of AI that can understand, learn and apply its intelligence across a range of tasks. This leads into a whole other debate on whether this is possible, with experts arguing both sides of the case. We must also ask ourselves if the use of AI is sustainable. Vast amounts of energy are required to power AI, with studies suggesting that the AI industry could use 85-134 terawatt-hours of electricity by 2027. With current infrastructure, this would produce huge amounts of CO2, and though the electric grid may decarbonise in the future, the large amounts of energy used could create shortages elsewhere.
Moving forward, designers must be wary of using AI simply because it is fun toy, and only utilise it where necessary, knowing that it will give a correct result that improves the efficiency of the project, rather than undermine it. However, if used effectively, AI could be an exciting tool that may be the next step towards reaching the industry’s carbon goals.
By Joe Spillane
References
85-134 terawatt: https://www.bbc.co.uk/news/technology-67053139
$1 trillion: https://www.goldmansachs.com/insights/top-of-mind/gen-ai-too-much-spend-too-little-benefit
Uses of AI in construction: https://www.constructconnect.com/blog/ai-in-construction-has-landed
AI in medical: https://www.sciencedirect.com/science/article/pii/S0720048X24003541#:~:text=5.,up%20images%20must%20be%20compared.