Development of new technology drives progression within Civil Engineering, such as BIM, photo-voltaic glaze, modular construction, CAD, kinetic roads and walkways, to name a few. The diffusion of artificial intelligence (AI) is no different and is another string civil engineers can add to their bow.. sooner than you might expect.
Discussion of AI revolutionising civil engineering may seem a far-fetched concept to some, with AI having connotations of being sci-fi technology - it’s not uncommon. Many people’s understandings and expectations of AI, are futuristic and originate from movies. For industry-innovators grappling at the new and exciting opportunities AI presents for construction, the rate of AI changing civil engineering might not be fast enough.
Just how advanced is AI in civil engineering (types of AI)
Although AI can’t yet fully replicate human intelligence, it would be wrong to assume that a human would be superior at competing against AI to do certain tasks. Notably, productivity, loss of attention, speed to calculate, data overwhelm and bias are things which can be eradicated when delegating tasks to artificial intelligence, rather than relying on human intelligence.
Use-cases of AI are mostly configured to execute narrow tasks, specific to the setting (domain) in which the technology is being applied. Input of information from various technology sources is collected and processed to create a valuable output which solves a task or task-set. Completing tasks to a high-standard or exceeding human ability to complete tasks is known as Narrow AI and is where we see most development in the civil engineering sector in the short term.
Not only this, but AI models can learn to improve their understanding and efficiency overtime with reinforcement from AI-users, and from their own data and experience so an AI process would be expected to improve over time.
More-advanced AI which has the ability to mimic the way our mind works, and be able to think and understand, is known as Strong-AI, which is yet to exist, and is not expected to be commercially-viable or adopted until 2050-2070. These will of course present new opportunities to AI, but not for several years to come.
Is AI a threat to civil engineering jobs?
No. Not in this generation at least. Narrow AI looks to enhance and develop the ability of engineers by assisting with a narrow set of assigned tasks. As AI becomes more intelligent and widely-adopted, the role of a Civil Engineer will likely develop to utilise more systems, technology and achieve a better outcome, faster. AI looks to build efficiency and value, by multi-tasking, mitigating risks and faster communication.
Don’t be put off by Domestic AI
Smart speakers in domestic settings, self-driving cars and highly-targeted advertising are often people’s first known interactions with AI technology, and it can be difficult to envision practical applications in a civil engineering setting, if these are the best use-cases which spring to mind.
The reliability of devices such as Alexa and Google Home may not seem ground-breaking when used for tasks such as playing songs and setting timers, but what makes them intelligent is less their ability to complete such basic tasks, but their ability to learn spoken dialect and language. This is known as Natural Language Processing (NLP) and when this is applied to live commercial problem-solving is when the ground-breaking AI becomes front and centre.
Computer sight is a prime example of how computers have learnt to understand what they see, increasing the useful opportunities of AI. Computer Vision is where a lot of development is occurring in the civil engineering sector, in the short-term.
Early-innovations in Narrow-AI have a lot to offer.
Health & Safety in Civil Engineering
Civil engineering contractor Winvic are front-runners in implementing AI in a civil engineering setting, after securing a Innovation Grant in partner with other organisations.
Computer-vision through live site footage, supported by GPS technology, 5G communication and wearable technology provide an AI system the information required to map a construction site, autonomously identify health and safety concerns and communicate the findings in a way that can significantly increase the site safety of contractors.
By always observing and mapping the site surroundings, this AI system can create ‘safe-zones’ around large or moving machinery and can directly alert contractors on site of personal safety concerns, through wearable tech. Instances of missing PPE will be easily identified and resolved, by flagging issues directly to Project Managers or Health & Safety manager.
One way this model can develop further is the use of drones to capture footage and investigate black-spots created by fixed cameras being obscured. Additionally, drones can travel faster than a person on foot, which can reduce time spent by Project Managers travelling to explore the site.
Plug and play - AI software for construction.
Leading construction software firm Autodesk introduced plans for new AI technology in a report from 2018, detailing the applications for AI in construction.
Autodesk announced their intentions to allow their future data platforms (BIM i360) to connect with third-party AI vendors used throughout the construction industry. Linking data systems is a vital preparatory step, to achieve the best outcomes from an AI project. Connecting up systems already used by firms makes the roll-out of AI in construction one-step easier and more achievable.
To build bespoke AI requires immense skill, vast development and data preparation so being able to purchase pre-packaged AI models from leading construction firms makes AI technology more accessible to construction firms. Permission for third-party vendors to connect informational systems makes the process of implementing AI less arduous than the roll-out of other information systems which require migration from one system one to the next. More firms who are becoming BIM ready or BIM compliant will already be on-track to innovate with this new tech.
Pre-construction stages of projects can also benefit from AI by machine-supported design (Generative Design), intelligent clash detections and more which can prevent costly mistakes or problem-solving on-site.
Other applications in Civil Engineering
An AI report from the Institute of Civil Engineers, predicted applications of AI that can think and do in construction. The use-cases are detailed in this table form a 2017 survey of Built Environment professionals.
|
Process |
Percentage of respondents (>150) |
|
Design and optioneering |
71% |
|
Construction operations |
69% |
|
Risk and crisis management |
58% |
|
Project communications |
55% |
|
Business case planning |
40% |
|
Procurement |
37% |
|
Dispute management |
23% |
|
Consenting and approval |
22% |
Table 1: Built environment processes that artificial intelligence technologies will improve
Here, we can start to envision other areas where AI can add real value to projects, by being tasked with set tasks to support Engineers and Managers on site.
What is and what isn’t AI?
The definition and classification of what is, and what isn’t AI, changes based on our understanding of technologies. A highly simplified definition of AI might be:
“Informational technology that mimics human intelligence.”
In a civil engineering context, the opportunity AI presents relates to handling tasks previously only conducted by humans due to the level of intelligence required to execute that task reliably, or to a suitable standard.
If AI can be proven to not be exhibiting a replication of human intelligence, then it is no longer considered AI. What was considered AI yesterday, might not be AI today. And what is considered AI today, might not be considered AI tomorrow – something known as the AI-effect.
Interesting fact: Back in the 50s during the War, Alan Turing famously developed one of the first computers to crack the unbreakable enigma code by building one of the first computers. Turing’s invention was hailed as one of the most notable early instances of AI and paved the way for modern day computing, but we wouldn’t consider the everyday computer to be AI today, despite modern computing being far superior in terms of technology and capability. Technology is discounted as AI when researchers can prove that the technology is not exhibiting true intelligence.
AI in civil engineering is a hot topic for us and something we’re interested in hearing more about. Get in touch to share your ideas, experiences and ambitions of AI and be involved in future content about AI in civil engineering.
