ARTIFICIAL INTELLIGENCE

ThinksARTIFICIALINTELLIGENCE –SHAPING THE FUTUREOF THE BUILTENVIRONMENT

Artificial Intelligenceshaping the future of thebuilt environmentThe ability of computers is transforming our livesat an increasing rate. The prospect of machinesthat can think, rather than just do, is somethingwe are beginning to take for granted. Thetransformative power of artificial intelligence (AI)to change the infrastructure sector is only justbeginning, but now is the time to assess andanalyse how we can best take advantage of it.All stakeholders with an interest in tomorrow’s infrastructure needto come together to identify the challenges and opportunitiespresented by this impending AI revolution. This will ensure weare better equipped to develop the engineers of tomorrow,harness the transformative power of AI and ensure that futureinfrastructure technologies are delivering the best outcomes forpeople around the world.ICE and the National Infrastructure Commission recognisedthis need and convened a workshop on AI in the builtenvironment to address four key areas:1. PERFORMANCE;2. TRANSITION;3. G OVERNANCE AND DATA; AND4. SKILLS AND ETHICS.The workshop was attended by representatives from the techsector, civil engineering consultancies and contractors, the legalprofession, academia and world-leading research centres.The drafting of this paper was also supported by Eleanor Earl, civilengineer at Arup and one of the 2017 ICE President’s Apprentices.02Artificial Intelligence –shaping the future of thebuilt environment

1. PerformanceHow can AIimprove theperformanceof the builtenvironmentsector in the UK?Nearly 80% of built environmentprofessionals think AI will have apositive impact on the sectorOver 60% think AI will helpimprove productivityIt is anticipated that AI will increasethe efficiency of infrastructuredesign, delivery and maintenance. Less than 25% thinkAI will improve disputemanagement, consentingand project approvalComputer systems that can analyse largevolumes of data to highlight patternsin the performance and use of existinginfrastructure assets will enable betterdecisions to be made on the type ofinfrastructure that is required and how itcan best be delivered.ICE surveyed more than 150 builtenvironment professionals1 to understandwhere current thinking on AI sits. Nearly80% of those surveyed felt that AIwould have a positive impact on the builtenvironment sector.1ice.org.ukTable 1 shows that 71% of respondentsbelieve that AI will improve designand optioneering, while 69% think itwill improve construction operations.Conversely, less than 25% of respondentsthink that AI will improve disputemanagement, consenting and projectapproval.ProcessPercentage ofrespondents( 150)Design and optioneering71%Construction operations69%Risk and crisis management58%Project communications55%Business case planning40%Procurement37%Dispute management23%Consenting and approval22%Table 1: Built environment processes thatartificial intelligence technologies will improveThe survey also revealed that over60% of respondents think that AI willhelp to increase productivity in theconstruction industry, while enablingbetter connectivity and integration ofinfrastructure networks.Respondents to the survey were built environment professionals working in boththe UK and internationally. The survey was conducted by the ICE in July 2017.03

What would bethe benefits of adigital twin of thebuilt environmentand what are thesteps we need totake to get there?The data and information we createabout an asset for its whole life-timecan become a rich source of analysis.This concept has become known as –the digital twin.The digital twin of a physical assethelps us to monitor and improve itsperformance. By realising the valueof our data resources through digitaltwins, we can deliver more productiveinfrastructure, and optimise ourexisting assets.Digital twins reduce costand riskDigital twins should be predictiveand adaptive, meaning they canreflect changes that occur in thephysical environment and respondto them providing greatly improvedmodelling for assets. The benefitsinclude: harmonisation of operationsto deliver optimal user outcomes;clash identification and automatedremediation; and reduced costsand risk.Managing and analysing the volumeof data required presents a significantchallenge. Most industries analyselittle of the data that they gather.However, the benefits are clear.Google realised savings of 40% onits own data centres using DeepMindanalysis on its assets. 2To realise these savings and benefitsinfrastructure delivery needs toharness available knowledge. Thebuilt environment sector needs tofully integrate data expertise intoits teams, as well as working moreclosely with the technology sector.This means going beyond traditionalalliancing models.The opportunities of dataare not yet understoodClients need to start seeing digitalintegration as standard, whilstdrawing on other sectors experience,like social sciences. This will enablethem to avoid replicating existingproblems that have already beenovercome elsewhere such as designbiases from the physical environmentinto the digital environment.Before the true benefits of digitaltwins can be unlocked there areblockades that need to be overcome.There is a need to identify andaddress the policy and regulatoryincompatibilities that are currentlypreventing the realisation ofdigital twins.Managing the common dataenvironment is important but farmore pressing is agreeing the datagovernance, risks and liability atthe outset of each project. It is alsonecessary to acknowledge that datasecurity is unlikely to ever be 100%,so risk management - rather than arisk prevention mind-set - is requiredto ensure innovation is not inhibited.Making digital integrationstandard will help enableall these benefitsThey improve coordinationof operations to deliverbetter outcomes for people042DeepMInd (2017) DeepMind AI reduces Google Data Centre Colling Bill by 40%Artificial Intelligence –shaping the future of thebuilt environment

2. TransitionHow and whenwill we movefrom automatingexisting practicesto using AI todrive radicalchanges?AI is already being used on constructionsites to automate tasks like bricklayingand concrete pouring,3 but it is difficultto predict when their application acrossmore complex design and delivery taskswill become the norm.Machine learning and AI technologies,together with robotics and virtual reality,could transform the built environment.4However, each of these technologiesis at a different stage of developmentand maturity. As a result, the point atwhich they will each become ready forcommercial use is also likely to differ.The size of construction andinfrastructure companies has a significantimpact on the speed at which they areable to embrace new technologies andways of working. Large client bodieswith dated asset bases and challengingbusiness models are also not primed forrapid change.Procurement in the built environmentsector is typically driven by lowest cost,rather than the long-term value that aninfrastructure owner will gain from aninvestment.5 The higher costs associatedwith technologies like AI – whetheractual or perceived – can therefore actas a barrier.ice.org.ukDigital data is a prerequisite for theeffective use of AI in any industry;it underpins the development oftechnologies and their practicalapplication. In construction, this includesdata on the design and operation ofinfrastructure assets, constructionproducts, logistics, health and safetyand the like. At present this data eitherdoesn’t exist or isn’t being shared.AI has huge potential totransform the supply chainHowever the technologiesare all at different stages andthe benefits are not yet fullyunderstood ariation in terms ofVcompany sizes affect abilityto embrace new tech quicklyIt is clear, then, that there are earlysteps that need to be taken if the builtenvironment sector is to fully harness thenew powers of AI.L eadership and top-downorganisational change is keyto successHow do we move from our currentway of doing things to practicallyusing AI in the built environment?Successful piloting of practical uses forAI is required in order to demonstrateto infrastructure owners, contractorsand the whole supply chain the positiveimpact on their businesses.Organisational cultures that championdisruptive innovation are needed to drivefresh approaches to the way in whichthe design and delivery of infrastructuretakes place. Large companies mustaccept that in a digital age, currentproject management methods needupdating, and that developing theircapabilities in technologies like AI willimprove the effectiveness of the servicesthat they offer.Strong leadership is necessary to bringabout this shift in organisational culture.Creating a vision and instilling this acrossan organisation is something that mustbe led from the top-down. It is notpossible to change the operating modelof an organisation without the buy-in ofits employees.Developing and maximising therelationship between AI and BuildingInformation Modelling (BIM) will speedup the adoption of AI in the builtenvironment sector. The widespread useof BIM in business case development,optioneering and construction operationswill create the quality data that isnecessary to inform the development ofAI technologies and their practical use.It is important to acknowledge thatthere isn’t necessarily a universal desirethroughout the supply chain for AI.The value will not be shared equitablyand as the large contractors developtheir capabilities in AI there is a risk tothe future role of smaller constructionbusinesses. There is a need to supportthese organisations as the use of AIbecomes more common.3AEM (2016) How Artificial Intelligence Could Revolutionize Construction4NIC (2017) Driving innovation in infrastructure through artificial intelligence5ICE (2017) From Transactions to Enterprises (Project 13)05