In the first instalment of our construction x technology series, we explore the potential of integrating AI with Building Information Modelling (BIM) to revolutionise the construction industry.
BIM has been improving the construction sector for years, as it allows for models of construction projects and buildings to be created as a digital representation, often in three-dimensional virtual structures. When used in combination with artificial intelligence (AI), BIM has the opportunity to progress the digitised automation of construction projects, from design to construction and ultimately to the ongoing maintenance of the building – reducing the pressures on project teams, mitigating construction risks and reducing costs in the long term.
How AI and BIM can assist the construction industry
Whilst developments are at their earliest stages, the potential for integrating AI with BIM technology is clear and the possible technological advancements will ultimately revolutionise the construction process moving forward. For example:
- Automation: Artificially intelligent machines trained on vast data sets of tried and tested designs can automate the design process within the parameters set by the user (generative AI). Rather than humans completing the design and then modelling it within the BIM software, the AI-powered BIM software could itself generate the design (e.g. floor plans, site footprints, and more). In theory, having been trained on the strengths (and defects) of previous designs, the AI-powered software should produce the best possible models – avoiding human error and accidental repetition of past design shortcomings.
Further, given that the data on the relationships between the BIM models’ virtual objects are all interconnected in the database, the AI software could generate the necessary consequential adjustments to the models where a human user has made an alteration to part of the design. This could significantly reduce the risk of design clashes in BIM models (whether those are hard, soft or false clashes), which can lead to significant costs, particularly if hard or soft clashes are not identified until the construction stage.
- Virtual reality: AI-powered immersive visualisations of the BIM models allow stakeholders to experience and evaluate the design before it has been physically constructed. This technology would enable project participants to assess the feasibility and overall buildability of the designed structure, helping to determine project viability. Additionally, it would allow design teams to view their creations in a more true-to-life form, facilitating the resolution of any potential issues prior to construction.
- Simulations: AI can enable the use of sophisticated simulations within BIM. Such simulations could replicate requested scenarios and predict possible outcomes. For example, the software could simulate the impact of light or temperature on the degradation of the construction materials in the models. The use of AI-generated simulations in conjunction with BIM allows the project team to make informed construction and design choices by using data-driven insights on possible outcomes and alternatives.
- Project management: AI-powered BIM software can also enhance project management during construction. The BIM platform itself serves as a digital database containing detailed and intricate model information. AI software can utilise that information in combination with historical project data to aid in economical resource allocation, the preparation of realistic and achievable programmes (including by considering factors such as weather and site conditions), the evaluation of potential risks, and the accurate forecasting of project costs.
- Human interactivity: AI could enhance the user interface of existing BIM software, making it more user-friendly and therefore allowing novice users to utilise BIM. The artificially intelligent software combined with generative pre-trained transformer technologies and natural language processing could create a form of ‘virtual assistant’ within the software. This ‘virtual assistant’ could answer user questions on the building models, request simulations or projections for costs and timing, or even allow human users to set parameters for a new model to be created by the software itself.
- Underground construction: Mapping out underground sites can be difficult and create safety hazards for workers where, for example, there are tight tunnels or unstable environments. Information and data available on the underground environment are also generally less accurate than above ground information, and in less abundance. Therefore, small AI-powered autonomous vehicles which are fitted with sensors and cameras to map out pathways and analyse the surrounding environment could in theory transmit data into the BIM database directly. This could confirm site suitability for construction or enable accurate modelling for subsurface structures such as pipes and cable networks.
Risk factors and considerations
Whilst there are many benefits of integrating AI and BIM systems, the risks must also be considered. For example:
- Initial high costs for implementation of AI-integrated software: AI has many cost benefits in the long-run, given that it can drive efficiency and reduce manpower where it is not essential. However, AI technology is still developing and remains a relatively new technology, therefore the start-up costs for integrating AI into BIM software can be prohibitive. This includes the cost of the software, but also the cost of the hardware (e.g. processors) running the software and project staff responsible for the AI-powered BIM.
- Discipline-specific models: BIM models are often described as being ’discipline-specific,’ in that they are often built to represent the semantics or expertise of one single professional view (e.g., an architectural perspective or structural engineering perspective, etc). In fact, many BIM databases operate using federations of separate models, which frustrates the idea of a multidisciplinary and collaborative model that would be the intended outcome for artificially intelligent BIM software. This could mean there are too few multidisciplinary models available to effectively train the AI-powered software, resulting in the automated model being ineffective.
- Human intervention: As mentioned above, whilst AI is making new advancements day by day, there is still the requirement for human monitoring and intervention. AI is by no means perfect as it stands, and it has the capability (and sometimes propensity) to make mistakes and therefore will need to be closely monitored by any project team. Too heavy a reliance on the capabilities of AI could lead to detrimental effects on the construction project.
- Human-learned bias and constraints: Given that AI is ultimately programmed by humans, there is always the potential for algorithms to inadvertently incorporate human-learned biases and to be limited by the quality of data on which they rely. It is therefore imperative to AI-integrated BIM platforms that there is rigorous testing of the software and ongoing monitoring of data set quality to ensure that any AI-generated BIM models are as effective and accurate as possible.
- Legal considerations: The relationship between AI and the law is an ongoing and complex matter that continues to evolve. There are many legal implications for the role of AI in construction projects that will need to be thoroughly considered at the outset and prior to the implementation of AI into any project. These include issues relating to intellectual property and ownership of AI-generated models, licensing for using the technology, data protection and cybersecurity concerns, as well as considerations of liability if things go wrong. For further information on the legal considerations for AI see a selection of our blog posts here.
Conclusion
The intersection between BIM and AI can produce significant developments for the construction industry. It can automate the design process, assist with project and costs management, mitigate safety risks and much more. However, as with the implementation of any new technology, it is important to consider the risks to pre-empt where things can go wrong and to mitigate those risks as much as possible. But whoever masters these risks and can utilise AI-powered BIM will have a competitive advantage and so it is worth consideration.