AI has an unrivalled ability to explore variations of a design model, a concept known as generative design. For companies involved with building information modelling (BIM), it offers a solution to a number of challenges they face, such as optimising space or making a building more sustainable.
Effectively, AI could take a BIM model and explore hundreds of different designs that would make the building safer, more stable, cheaper to build or better for workers. What would take AI a matter of hours would take a human weeks or even months to achieve.
Using generative design principles also allows for better planning by taking into account the requirements of architects, engineers, mechanics, electricians and plumbers - and the sequence in which work needs to be carried out. AI has the capacity to identify and mitigate clashes between the various trades, per model. This has a significant bearing on costs and delivery time.
Planning and cost management
Approximately US$10trn a year is spent on construction projects, meaning it’s an expensive business. It’s common for projects to overrun and for costs to spiral out of control because of unforeseen circumstances, which is something AI can help to mitigate.
Artificial neural networks can be used to predict costs based on the project size, quality of contractor and the type of contract. By using historical data (such as start and end dates), predictive models are also able to indicate accurate and realistic timelines, which helps to plan projects far more effectively.
On the planning front, drones that are able to capture video footage and feed it into Computer Vision AI models are also proving highly effective in the construction industry. They’re able to track, monitor and review the progress of sub-projects within a build, apply data learnt from other projects and suggest the best path forward.
Building Information Modelling (BIM)
Building Information Modelling (BIM) is the collection of data on every aspect of a building – from initial construction through to the finished article. It draws on information collected collaboratively at various stages of the build.
BIM isn’t itself new (it became the norm in the early 2000s), however when combined with AI, it can explore possibilities for each aspect of a construction project far quicker than a human can. BIM collects tonnes of data, data that AI models and algorithms can learn from. In doing so, AI-powered BIM can make predictions that explore opportunities for modifications and improvements, assess resource-efficient solutions, and even create execution plans that minimise the risk of loss
Risk Mitigation & health and safety
Risk is an inherent part of construction. It happens daily and evolves. There are thousands of open issues, hundreds of contractors and subcontractors on site, and the construction itself changes daily. Managing those factors is a very big task, and to put it in context, there are on average 1000 fatalities on US constructions sites a year.
However, AI offers a couple of solutions to both prevent fatalities and mitigate the circumstances that cause them to arise. Most issues that arise are recorded, usually on a phone. Most sites are also under CCTV surveillance – both of which provide a rich data source through which to train and categorise incidents, certain types of behaviour and different scenarios.
Take for example a worker climbing a ladder. By analysing behaviour patterns and incidents, the AI can be trained to detect whether said worker is in a high-risk situation. So, for example, it could be trained to detect whether the worker is wearing hi-vis gear, if the ladder is harnessed correctly or if the hammer left overhanging scaffolding is likely to fall. The system can then flag the risk and alert co-workers to rectify the situation before it becomes a hazard.
This is the area of greatest impact for Computer Vision AI in the construction industry. Incidents and injuries are devastating for individuals and for workforces, but they have a major impact on the bottom line too. In the UK alone, it’s estimated that workforce injuries and accidents cost the construction industry in the region of £6bn each year, with many being preventable.
The knock-on effect for accidents and injuries in construction are increased insurance premiums, delays to work that result in contractual obligations being met, fines and penalties. These not only damages the brand but the organisation’s reputation too – which has significant impact on the company’s future.
To that end, the ability to effectively manage and mitigate the risk of accident and injury is not only a duty of care but also one of operational resilience. Without the capability to monitor, review, refine and improve procedures, businesses cannot realistically expect to grow or to excel. That is especially true where health, safety and risk are concerned.
Though AI-use cases are still relatively nascent in the construction industry, they are prescient. The need to effectively bridge the gap between risk and resilience is increasingly urgent; it will become a competitive advantage for those that successfully do – and for those organisations, Computer Vision AI will have a significant role to play.