Construction – Call the next robotic witness please
11th March 2021, 8:33 am
Specialist construction lawyer Peter McHugh, a partner with national law firm Clarke Willmott LLP, examines the advantages and disadvantages of using AI intelligence to gather evidence and assist in the success of a construction project.
Today the risk of something going wrong on a construction project hangs over the heads of the contractor and the professional team like the ‘Sword of Damocles’ but the tools for spotting problems and gathering evidence is getting more sophisticated, including with the use of artificial intelligence or AI.
Collection of evidence
Recent and ongoing developments in technology have allowed the construction industry to start recognising the potential and current uses of artificial intelligence to streamline processes and overcome some of the industry’s challenges such as delay, cost, changes and safety. Whilst many of the potential uses of AI are not yet being used daily or on every site, there are several devices and processes coming to the market that can assist with many elements of the construction process, from design up until practical completion. Moreover, the significant collation and storage of huge amounts of data in a logical and easy to locate manner has the potential to revolutionise a lawyer’s ability to collect evidence that can be relied upon in court, adjudication or simply in making / defending a claim on behalf of employers, contractors and consultants.
Mitigating delay and hazards
AI is also being used to prevent or mitigate delays, monitor evolving designs, identify any health and safety concerns, forecast “what if” scenarios and plan contingencies, and monitor conditions for weather or other hazards.
Reducing the need for changes
Currently, design firms are adopting the use of “generative design” which streamlines the design process and checks any design against set parameters and requirements, such as for planning, or any functional, spatial, performance, cost or material requirements. This may reduce the need for changes, instructions and subsequent variations, one of the main causes of dispute between parties to a construction contract.
BIM seeks to resolve delay and identify defects
By way of example, Building Information Management (BIM) creates a “digital twin” of the building against a physical representation that is monitored by robots and drones or other image recording devices. This allows artificial intelligence to monitor the build, identify any issues or deviations from plans and significantly reduce the risk of defects. The AI would also track, plan and change any programming accordingly. As well as providing cogent evidence for any delay claims (either as a claim or defence to a claim), the use of BIM would be the epitome of “delay mitigation”, as it would seek to resolve and reduce any delay caused. Furthermore, BIM could be used to identify any potential defects which are considered deviations from the plan and which could potentially prevent significant issues arising before the end of construction, where it becomes invariably more difficult to rectify the same.
Electronic site diaries and drones
Site diaries are commonly completed by the onsite manager by hand. This relies on the site manager correctly inputting the information and having the time to do so. Devices to assist with digital site diaries are potentially coming to the market and could be used to collate evidence automatically such as through a fingerprint or facial recognition software to identify who is on site and when, and which deliveries have come in with materials. Drones are already being used to tag and locate people, vehicles, material and equipment on large sites, as well as being used to monitor tricky to reach places for surveys or during construction. Clearly for any delay claim, all evidence collected could be invaluable, as an accurate site diary will assist any delay expert in accurately and properly programming the project to identify where the delay has occurred. This is of course not an all-inclusive solution, and in many circumstances, issues in dispute will be fact dependent and will still likely require the input of those closely involved in the project to provide their input on potential causes of delay.
In projects where modular and offsite production methods are being used alongside traditional methods of onsite construction, contracts must contain detailed provisions to allow for design co-ordination of the two construction methods, and cohesive programming and planning to ensure a finalised fully integrated project. These types of contracts are being termed “smart contracts”. Combined with the use of BIM, smart contracts can potentially be read by AI, and used to monitor actual performance of the construction to ensure the project is being delivered in accordance with it, for example, are the payment provisions of the contract being properly complied with as AI could potentially highlight payment clause deadlines to help prevent “notified sum” claims. In addition, AI used in this manner could provide any required “as built” information upon conclusion of the project which would also assist in creating an accurate record and fulfilling any “as built” contractual obligations.
Whilst the use of AI in the construction industry could provide a useful means of evolving designs, monitoring the process, mitigating risk, and reprogramming a schedule, to name but a few of the many ways it might be used, the potential legal implications and risk allocation in contract documents does need to be considered. Parties to a construction contract need to make sure it is clearly agreed and established who takes on the risk and the degree of liability of that risk, particularly regarding smart contracts where elements are being built offsite. The parties will ultimately have to agree on what terms the information and data collected by AI forms part of the contract, to include considering the extent to which that information provides a definitive picture of events and to what extent that position can be rebutted. Another element that needs to be carefully considered is who will own the data that is being collected, and any contract provisions will need to make sure data protection protocols are observed.
Who bears the cost of AI?
Due consideration may also need to be given to who bears the cost of any AI hardware / software and who maintains the technology; is it the contractor or employer, does it form part of overheads or need to be defined in any Bill of Quantities? This may also help to decide where any theoretical risk lies. It may be easier for any AI to be provided under a Subcontract, though depending on the type of AI used, leased or purchased, a contract may contain non-construction provisions or no longer fall under the statutory definition of a “construction contract” at all.
Risk with AI and bias in algorithms
One of the inherent risks of AI, as we have seen in other industries recently, is bias in the algorithm, (for example the Google facial recognition patterning system where people of colour were disproportionately underrepresented in the dataset and therefore misidentified). Who obtains the datasets and who writes the algorithm (the input) will have some impact on the results (the output) and may create significant issues where bias creeps in and nullifies the outcomes.
Whilst the benefits of using AI are numerous, and the potential capacity for recording and processing huge amounts of data is undoubtedly useful to the construction industry, AI is not completely risk free. One of the features of AI is its ‘training, learning and development’; how the algorithms are trained to recognise patterns and anomalies from data that is currently available. Imagine a lot of data about accidents on construction sites; AI could easily take these datasets and be trained to look for patterns and anomalies. However, a bias in types of information available may influence the output, for example where only specific types of accidents happened until a change in construction process made it safer; we may not see those accidents as much, but a machine would not know that just from interpreting the data it had analysed.
AI is still a very new and continually developing technology that is only starting to be applied to the construction industry. Some firms involved in construction are beginning to explore how the technology currently available could be used to help deliver construction projects on time, in budget and safely; or to start collecting data that could be used to improve AI for future use. We are still some way away from a “robotic witness” in the courtroom, however if you are starting to use technology to enhance and improve delivery of your construction project, consider taking legal advice to make sure you are fully complying with your contractual and legal obligations. Our experienced and approachable team are always on hand to provide swift and thorough advice.
Clarke Willmott is a national law firm with offices in Birmingham, Bristol, Cardiff, London, Manchester, Southampton and Taunton.
For more information visit www.clarkewillmott.com