With ChatGPT gaining mass attention recently, Dr Andrew Pidgeon, global head of strategic innovation at RSBG Infrastructure Technologies, examines the opportunities and challenges for artificial intelligence in design, construction and consultation
Since the start of 2023, tools that deliver heightened artificial intelligence in construction and deeper understanding in response to complex problems have been shown to generate both great excitement and promise.
They have attracted a new and formerly sceptical audience and widened their horizons.
They have replaced complex programming requirements with simple interfaces and platforms exported via plain language outputs.
The initial excitement has been mostly found to be within the Large Language Model (LLM) area.
ChatGPT has given more to the hands of a diverse audience whereby intrigue can be pointed towards the machine; outputs can be counter-queried, amended and delivered in a rapid and cost-effective manner by non-technical personnel.
19% of resource workloads will have at least 50% of their tasks supported by AI
Adding to the potential benefits and impacts, research undertaken by OpenAI, Pennsylvania University and OpenResearch initially suggests that 19% of resource workload will have at least 50% of their tasks supported or complemented through the evolution of LLM.
PwC provided further insights, stating that 30% of occupations could directly be affected by automation by the year 2030, enhancing skillsets but, more importantly, noting that the actual impacts will be as part of a phased approach in terms of saturation across sectors.
What is foreseen is not a total role replacement but an opportunity to advance skillsets and increase the quality of outputs through wider data integrations for enhanced decision-making and complex factors.
Factors towards the drivers and enablers of artificial intelligence in construction
The advancement of artificial intelligence in construction brings an opportunity to harness this technology to counter legacy issues and apply a focused benefit towards the challenges we face today and in the future.
Sir John Egan in 1998, led the production of the UK-based report Rethinking Construction, whereby conclusions were made on the state of the construction sectors.
The report found that disconnected teams, alongside the under-utilisation of performance data metrics to advance decisions, require improvement in order to reduce the barriers of inefficient manual process management.
We also have new challenges, such as how we make positive steps toward a sustainable future within our infrastructure sector. The construction industry also needs to address the resource crisis, which creates restrictions on the speed of delivery and, ultimately, the ability to complete projects on time and within budget.
There are also societal impacts, digital transformation pressures and revenue generation gravities which affect people within organisations, economies and regions.
Will AI in construction save the day?
Although the implementation of artificial intelligence in construction has potential, attention needs to shift to its ability to incorporate intelligence directly into deliverable tasks.
Internal process development, enhancement and automation for day-to-day business operation improvements must also be focused on.
By itself, AI and thusly LLM will add a benefit, but its benefit may soon be lost when it comes to finding a purpose, which is where the technological fatigue may well kick in and the power seen as more of a complicated burden.
As we venture into this exciting new realm, opportunities to complement existing roles with a heightened awareness and opportunities to gain enhanced outcomes will become more apparent.
A fundamental understanding of artificial intelligence is needed to reap the benefits
As stated within Blackbox Thinking by Matthew Syed, what appears to go on inside of complex systems appears to be aesthetically clean, crisp and simple, but what really occurs inside the box is actually more difficult to understand than the original situation.
To harness the greatest benefit from AI, fundamental understandings are required to enable the validation of outputs.
I like to use the analogy of teaching mathematics to children to understand this. Why do we bother if we have calculators?
It is the fundamental understandings and the validation of the data which supports the complexity and benefits of the “Blackbox” of AI. This promotes intervention for the discussive points on what to deem as ill-fitting, what matches the purpose and how we decide the right approach for the next steps, given what the algorithm produces.
What factors are key to enabling the success of AI in the construction sector?
Stated below are some of the fundamental enablers steering towards these beneficial outcomes, which I believe are key to the successful incorporation of AI methodologies:
- Skillsets will need to be supportively developed in terms of foundations alongside, and not in replacement of, traditional roles towards the computational and scientific domain.
- A fundamental understanding for validation purposes, and not understanding of what exactly goes on inside the Blackbox is core to the successful development of personnel.
- Focus on more than storytelling with AI will advance the ability to improve the complexity of a more widespread decision-making approach, truly utilising “big data” and countering complexities in plain language.
- Using the principles of Oswald Wald Evaluating Dynamic Data, there is an ability to gain additional information from the gaps in data we don’t have, which holds great value and justification of the implementation and thusly, investment in this field.
- Potential to reduce the impacts of resource shortages due to doing more with less.
- Ability to create new, value-added services built around automation of processes, sustainability and carbon impacts, operational optimisation of assets and querying of decision making (rapid ‘what if’ analysis across varying situations and stages).
Although there is potential, I believe that we are only at the beginning of our journey with artificial intelligence in construction.
This fresh perspective needs extra thought about what long-term benefits this application of AI can bring to sectors and societies, such as the construction industry, including security and commercial compliance impacts.
I personally see this phase of development as being similarly referenced to the discovery that the world wasn’t, in fact, flat.
This then led to entirely new areas of exploration, technology development and bold choices, such as curiosity to explore unknown regions and the connecting of societies that were able to accomplish greater things and achieve even greater benefits.
Construction must provide organisations with the correct training to understand AI
Applying a solution-first approach doesn’t add as much benefit as understanding the nut we are trying to crack. As an industry, we need to create some grounding to benefit from AI.
We need to not ‘rely on the AI’ by throwing it questions to solve and accepting the answers.
People in organisations, peer networks and focus groups must be provided with the opportunity to develop their capabilities in a stepped yet supportive approach to facilitate the development of their understanding of the fundamentals of AI.
This will enable them to develop and advance by aiming real-world problems hand-in-hand with the artificial enhancements and not solely alone, keeping the human in the loop via validation of the cycle, output compliance and intelligence analysis.
Dr Andrew Pidgeon
Global head of strategic innovation
RSBG Infrastructure Technologies
