Justice Philip Jeyaretnam: Lecture given at the 2024 Seah Choo Meng Public Lecture

2024 Seah Choo Meng Public Lecture

Organised by SUSS School of Science & Technology

20 June 2024

COLLABORATIVE CONTRACTING AND DISPUTE RESOLUTION FOR THE CONSTRUCTION INDUSTRY IN THE AGE OF AI

Emerging solutions to the drawbacks of traditional contracting and adversarial
dispute resolution and the role that AI can play

The Honourable Justice Philip Jeyaretnam
Judge
Supreme Court of Singapore
President, Singapore International Commercial Court

I. Introduction

1               I am honoured to be invited to deliver this lecture which bears the name of one of the most distinguished doyens of the Singapore construction industry. I congratulate the SUSS School of Science & Technology who are the organisers of the series. At a time of rapid change, annual lectures such as this one offer an opportunity for reflection on how best to address its impact.

2               My intention today is to spark discussion concerning possible reform and innovation in relation to conflict avoidance, conflict management and dispute resolution procedures for the benefit of the construction industry. Having been a construction lawyer for more than thirty years now, I offer my legal perspective on how construction is organised and managed through contracts, how project communications are managed and how a project is documented.

II. Thesis

3               People have an innate tendency to complicate things. The thesis of this lecture is that this tendency has been magnified and exacerbated by two effects that the swift advances of technology have had. First, many enterprises, including the business of constructing buildings and infrastructure, have been made increasingly complex – in short, a smart building involves a lot more than a plain building. Secondly, methods of communication have been revolutionised, with instantaneous written communication largely replacing oral interaction. This change has spawned thundering torrents of words and images that threaten to swamp human powers of comprehension and assimilation. As recently as when I started my career, we bemoaned the lack of records, and hence the lack of information to understand retrospectively how a project had proceeded and who was at fault for any shortcomings or failures. Today, we suffer from an over-abundance of information – there’s often just too much material to realistically get our heads around. Technical and evidentiary complexity compound each other.

4               The rise of artificial intelligence or AI has given rise to the spectre of AI taking over human management and decision-making functions. This fear may be exaggerated, but there are in any case positives to the rise of AI. I would suggest that AI is best seen as a supplement to human intelligence, and not its replacement. AI offers us the development of tools to help humans navigate technical and evidentiary complexity. If the past twenty years has seen a flood of information inundating us, AI offers us ways to erect levees and dig channels that will enable us to take control and use knowledge and information to irrigate thought and human flourishing, rather than drown it.⁽¹⁾

5                Further, as part of its response to complexity, I believe that the construction industry may develop and expand appropriate methods to achieve quick “good enough” outcomes with the support of commercial courts. My argument is that fast “good enough” methods of dispute resolution are indeed practical, useful and important for the construction industry. However, they must be anchored by a fair and efficient court system that offers litigants the possibility of seeking and obtaining justice, if necessary.

6               Moreover, it is in the context of “good enough” methods of dispute resolution that AI has the most potential to quickly make a positive difference, in terms of costs, time and general efficiency.

III. Complexity

7               Let me outline four elements of complexity.⁽²⁾

8               First, there is increasing technical complexity. While technology gives us new capabilities and makes our lives easier and more convenient as consumers, it has also helped drive the increasing specialisation of skills and compartmentalisation of knowledge. To be an expert today means to know more and more about less and less. This increased specialisation is true of construction as well.

9               Large building projects have always depended on planning and organisation.Clear delineation of roles and responsibilities promotes efficiency and accountability. I will take as my example the construction of the tallest functional buildings in existence prior to the use of structural steel in the nineteenth century. I am referring to the monumental cathedrals built in northern Europe from the late twelfth century onwards. Their design was inspired by the religion of their builders. Their soaring height and windowed light seemed to render the divine visible. A strong grasp of mathematics, including advances made by Arab mathematicians, was essential. But at a practical level it was good planning, clear instructions and division of labour into specialties that enabled master masons to build these transcendental cathedrals. Unlike other impressive stone structures such as Angkor Wat whose relief and tracery was carved in-situ after the stones were in place, stone for these cathedrals was shaped and carved by masons first on the ground and then fitted into place. Doing it like this was the only way when the stones would eventually be at such great heights. Moreover, it was more efficient to carve stone in workshops at ground level. But to do it in this more efficient manner required a greater degree of planning and organisation, breaking down the design into its constituent parts and giving clear instructions to multiple masons working separately from each other. Shop drawings were etched on plaster-laid tracing floors for the instruction and direction of the masons working under the master. A new layer of plaster would be laid when there were fresh details to be traced, worked out and explained. Modern construction methods and organisation of the construction site can be traced to these stone behemoths of Gothic architecture.

10               With the advent of the digital age, the past few decades has seen a sudden mushrooming of the number of trades, sub-trades, disciplines, and sub-disciplines required to successfully complete a large project. This is because our built environment involves more and more interconnected networks which themselves have become more complex – district cooling systems, building access systems, wind, and temperature sensors – the list goes on and on. Buildings have become vastly more complex. They integrate numerous mechanical and electrical services and must take account of new codes, including environmental standards. To add yet another layer of complexity, new buildings have been required to incorporate a living element in the form of green spaces and plants.⁽³⁾

11               Secondly, with the increase in the number of interconnected elements in our built environment, organisational complexity has continued to increase. Construction is a complex exercise, which necessarily involves numerous skills and trades. These must all be coordinated in time and space. Traditionally, this coordination has been achieved through breaking tasks down into linear paths for each trade, with bilateral contracts between a single main contractor or project manager and each of the subcontractors responsible for a particular series of tasks. This has begun to change with the rise of collaborative contracting.

12               Collaborative contracting has been a response to the perceived misalignment of incentives within traditionally organised construction projects. Among the hallmarks of collaborative contracting is a multilateral contract, with elements of partnership, or at least references to the use of best endeavours and good faith, and the adoption of mechanisms for profit and loss-sharing of some description. Even without bringing into consideration collaborative multilateral contracting, the mapping of bilateral contracts onto the performance of a project does not eliminate the interplay of the different skills and trades in the real world. Almost every delay claim, for example, will involve assertions about the performance of other players on the building site beyond the contracting parties. I will say more about this when I turn to collaborative contracting.

13               The third complexity is the explosion of recorded communications: first emails and now direct messaging platforms. In the past, the universe of documents would be limited to daily records of specific activities, minutes of weekly meetings, and monthly progress reports. Not only were there fewer documents but these documents were the product of contemporaneous processing. All these records would involve summaries, ie, note-taking that would facilitate later comprehension. While some such summaries would be done unilaterally, others were the product of a joint process with contemporaneous review by other parties. Today, by contrast, the art of contemporaneous summarising has been eroded. Instead, there is a vast and ever-growing thicket of electronic communications that largely remains in a raw uncondensed state right up until the time a dispute erupts. In short, more information and unprocessed information.

14               The fourth complexity is increasing legal complexity. The number of statutes and regulations has greatly increased. Whereas in the past only truly significant judgments would be reported, today every judgment is readily accessible at the click of a search button. This means counsel cite more cases. Judgments themselves have become longer, and each judgment refers to many more cases than would have been typical in the past. Transcripts of witness evidence relieve judges and arbitrators of the need to keep their own record of the oral evidence but in turn creates reams of additional material. Verbatim records of spoken language are harder to read than they were to listen to. This is because spoken language is different in structure and syntax from written language. When one listens to someone speak in one’s presence there are a host of accompanying cues that transmit meaning. This context is lost once the words take the form of a written transcript.

15               The combination of these complexities leads to an information overload on decision-makers. This includes judges and arbitrators faced with determining legal rights and wrongs when disputes arise from a project. Both interim adjudication (in its statutory and contractual forms) and collaborative contracting are responses to this challenge of complexification.

16               I will now provide a brief sketch of these two responses, beginning with interim adjudication.

IV. Good Enough Justice: Statutory and contractual adjudication

17               I should start by clarifying that the word “adjudication” has both a broad meaning and a narrow meaning. Its broad meaning refers to any formal judgment or legal decision on a disputed matter. Its narrow meaning refers to statutory adjudication under national statutes such as Singapore’s Building and Construction Industry Security of Payment Act 2004 (2020 Rev Ed). For the purpose of this lecture, I am focused on interim adjudication, subdivided into statutory and contractual.

18               Statutory adjudication is at its core a fast and inexpensive method of enforcing payments for work done or related goods and/or services supplied in the construction industry on a provisional basis, with the full merits of the dispute deferred to arbitration or court process. Contractors use statutory adjudication to obtain a quick, rough and ready answer to the question of how much they are due, an answer that is temporary and subject to further review in arbitration or court. In the meantime, it facilitates payment such that the contractor, rather than the owner, has use of the money pending final determination of the merits. Its impetus was concern about blockage of cash flow for contractors and subcontractors when owners or contractors higher up within the chain withheld or delayed payments for work done. Strikingly, however, the temporary answer given by statutory adjudication is often accepted by parties as a “good enough” outcome for everyone. People often adopt this “good enough” answer and move on without the time, effort and trouble of a full trial. This fact has demonstrated that a quick rough and ready answer given within a few months may in fact be more useful to businesses in the construction industry than an in-depth and forensically meticulous answer achieved only years later.

19               This preference in some circumstances for a quick and inexpensive good-enough answer drives the quest for equivalent contractual mechanisms for contracts or claims that are not within the scope of statutory adjudication. Indeed, such contractual mechanisms pre-date statutory adjudication. One example is that of the dispute board, which first gained popularity with the growth of international construction projects. In such projects, participants were typically from different countries. Projects were large in scale. They required years to complete. Demand grew for ways to address issues as they arose quickly so as to forestall their escalation into project-threatening fights. A successful example is that of the dispute boards that are such an important feature of FIDIC contracts. In our context, one may refer to the Singapore Infrastructure Dispute Management Protocol (“SIDP”) which is incorporated into Optional Module E to the Public Sector Standard Conditions of Contract.

20               One concern with contractual adjudication is that the enforcement of a determination may be cumbersome and convoluted. Typically, decisions and determinations of a dispute board must first proceed to an arbitration award and only after that to court enforcement. One cannot directly grant judgment for the money that has been contractually adjudicated because that would be a final decision and would raise an issue estoppel if the dispute proceeds for fuller determination on its merits. The conceptual answer is for the court to be asked to grant specific performance of the obligation to comply with the temporary determination of how much should be paid. That would be a final order, but its result is simply that the paying party has performed its obligation to comply. Later, there could be an adjustment that would take account of that compliance in the same way as when the obligation is complied with voluntarily. The contract could include a clause providing for this. For example, it could state that parties are obliged to comply with the outcome of the contractually mandated adjudication process pending a final resolution of the dispute. It should also state that this obligation may be specifically enforced either in a chosen court or in any court of competent jurisdiction. This would be particularly useful in international construction projects where there is no statutory adjudication regime in the country where the construction is taking place. For example, there could be express choice of the statements tracks under the rules of the Singapore International Commercial Court (“SICC”). This could potentially simplify the enforcement of contractual adjudication decisions in international construction projects. My suggestion is of course court-agnostic: what is required is that there be express choice of a court whose powers, rules, and processes fit with the task of summary enforcement of the obligation to comply with contractual adjudication decisions.

V. Collaborative contracting

21               I now turn to collaborative contracting, of which Optional Module E is an example.

22               As the name suggests, collaborative contracting is about putting in place a contractual framework that fosters teamwork, open communication, and mutual trust among owner, designer, contractors, and other stakeholders to achieve project goals. The aim is to create a working environment that prioritises problem-solving over fault-finding.

23               Today, the tasks of design, planning, coordination, and instruction are all aided by computer programs. These programs effectively extend the capacity and improve the efficiency of the human mind. Without them, the construction of complex buildings would be nigh on impossible and would certainly take far longer than they do today. The number of trades and disciplines that have to be coordinated is also far greater than in the twelfth century, with numerous internal networks required for any building to meet the electrical, ventilation, sanitation, and communication needs of its occupants and users.

24               Each of the steps involved in a modern building project, such as design, planning, seeking approvals, procurement, construction, commissioning, handover, and maintenance, subdivides into a myriad of tasks that have to be carried out by different people in coordination with each other. Health and safety obligations must be met. Sustainability requirements must be fulfilled. It is a wonder and a marvel that we build so much and so quickly.

25               Traditional methods of contracting seek to define and demarcate in advance the responsibilities of each participant, with a concomitant allocation of risk. There is little or no financial incentive to do additional or different things within one’s own sphere of responsibility that serve only to aid or assist the work of another in their separate sphere of responsibility. In addition to lateral divisions of responsibility, there are temporal ones too, with the design phase separated from the construction phase, and little or no opportunity for the contractor to influence the buildability of the project during the design phase. If a concern about the design arises during the construction phase, the contractor is typically not rewarded if he finds a practical solution for the owner.

26               Collaborative contracting is an attempt to align individual project participants’ incentives with the interests of the project as a whole.

27               I return now to Optional Module E to the Public Sector Standard Conditions of Contract introduced by Singapore’s Building and Construction Authority (“BCA”) in 2020. It aims to reduce cost, minimise variations, and expedite completion by encouraging teamwork and cooperation among stakeholders. Early sharing of information and open communication are intended to facilitate the early identification of risks as well as prompt solving of problems that occur during construction. Moreover, with the SIDP (to which I have already referred) processes are put in place to help avoid full-blown disputes and encourage early dispute resolution. The SIDP provides for the empanelling of a dispute board that deals with disputes on an interim basis during the course of construction.

28               Clause E2.0 of Module E obliges parties “to use their best endeavours to work together in a collaborative environment and act in a spirit of mutual trust and cooperation”. To this end, key officers, including key officers of subcontractors, must participate in “Partnering Workshops”. This express obligation is an expected feature of long-term relational contracts such as those governing substantial large-scale infrastructure projects.

29               It is important to appreciate that the clause does not use the language of good faith. What it does is require parties to use “best endeavours” to behave in two specified ways: (a) to work together and (b) to act in trust and cooperation. One may anticipate much ink being spilled in the coming years as the parameters and implications of this obligation is worked out. Without pre-empting how the clause will be interpreted, my view is that it provides a clear threshold of required acceptable conduct that is amenable to retrospective assessment by arbitrators or courts. It will be possible to establish that a party has failed to meet this threshold. Accordingly, my view is that from a legal point of view it offers a workable and effective alternative to traditional forms of contracting. Ultimately, it is for the market to determine its utility. For it to find acceptance, there will have to be changes in business culture, including in relation to the procurement and tendering process.⁽⁴⁾

VI. Appropriate dispute resolution

30               I turn now to appropriate dispute resolution methods: how parties may choose the best way to resolve their dispute. It may be better to use the plural – best ways – because different methods may be appropriate for different parts of a dispute.

31               It cannot be over-emphasised how critical it is for parties to think about the right processes and methods for resolution of their dispute. Lawyers have an important part to play in this. They are under an ethical duty to guide their clients towards the most cost-effective and expeditious methods of dispute resolution. They are also ethically obliged to cooperate with each other and with the court or other adjudicator so that a dispute is resolved by appropriate, efficient and economical methods. I would now like to highlight another option that parties can adopt as a model of collaborative dispute resolution. I am referring to having a professional “signalperson” who in the interests of the project directs parties to the appropriate tracks for resolving particular disputes that arise in the course of the project. By this I mean the appointment, at the time of contracting, of a person who does not themselves act as mediator, member of a dispute resolution board, evaluator, or arbitrator, but instead assesses disputes as they arise and channels parties to what in his or her assessment is the most appropriate mode of dispute resolution for that dispute, be it mediation, neutral evaluation, interim adjudication, arbitration, or litigation. In the context of construction projects, such a person could channel bilateral disputes to an arbitrator, while funnelling multiparty disputes to a court that readily accommodates such multiparty disputes such as the SICC. It can be difficult to establish and manage multiparty disputes in arbitration, because at its root arbitration arises from contracts, which are themselves bilateral.

32               This “signalperson” function is precisely what is provided by the new INTEGRAF scheme operated by the Singapore Mediation Centre (“SMC”). INTEGRAF stands for Integrated Appropriate Dispute Resolution Framework. The SICC is collaborating with SMC on this scheme, and Model Clauses have been formulated for this purpose.⁽⁵⁾

33               Where international construction projects are concerned, the SICC maintains the Technology, Infrastructure and Construction List (the “TIC List”) for trials of technically complex issues and questions, including construction and engineering disputes. How cases are tried once placed on the TIC List as well as the opportunities for mediation or neutral evaluation, makes it a good fit for projects involving collaborative contracting.

34               It is to the utility of mediation and neutral evaluation (especially for parts of disputes) that I now turn. Courts and other tribunals find their origin in the human desire that injustices be remedied. We speak of doing justice but what we really mean is remedying injustice: stopping one person from infringing the rights of another or compensating the victim for the infringement. Mediation in and of itself has only an indirect relationship to justice because it focuses on the interests of the parties and not merely their rights. On one hand, this makes mediation attractive, with the potential to rescue parties from the zero-sum game of a formal and final adjudication where one party must lose for the other to win. On the other, however, if mediation is unmoored from the legal system and the possibility of redress in court is cut off, it may favour the party with greater bargaining power and so advantage the strong over the weak. Where seeking legal remedies takes too long or is too costly, then mediated outcomes that give up some of one’s rights become more attractive, but not for the correct or proper reasons. Hence, it is necessary to anchor mediation within the broader context of a fair and efficient court system. Parties who know that the legal case can in principle be fully litigated within a reasonable time will be better placed to reach just mediated outcomes, accommodating both parties’ interests and not overriding either party’s desire for justice.

35               For this reason, mediation often works best once a claim has been formulated and responded to. Moreover, some disputes are of such complexity that they are difficult to resolve fairly and efficiently by one mode alone. For such disputes, it is useful to mediate some parts but litigate others. Which one to do first may also be important. A good example would be technically complex construction disputes which include contractual aspects, defects aspects, and delay aspects. To take an example, one aspect of a construction dispute might be numerous defects claims, but what counts as a defect may also involve an element of contractual interpretation. For this reason it will be helpful to have a court decide the correct interpretation before a mediator or a neutral evaluator works through the defects claims. This is what we mean when we speak about “unbundling” large complex disputes.

VII. How AI can help

36               Artificial Intelligence or AI is a general umbrella term first coined in 1956 at Dartmouth in the US for a scientific conference. At that time, the envisaged approach was that “every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it”.⁽⁶⁾ Today, the excitement is about a different approach, one of machine learning, in which a program trained on a large data set is able to recognise patterns. When interrogated by a prompt it finds the most probable next token (and then the next and the next) to generate an output that may be an answer to a question or may be an image that fits a prompted description. Anyone who has used such a program knows how magical it seems. Generative AI is rapidly building a new market in the entertainment sphere where the user co-creates their own imagined world. In that context, what is being created is a fictional world and there is no need for factual accuracy. However, for generative AI to be useful in a business context, factual accuracy and verifiability are critical. The production of what has been termed “AI hallucinations” is a problem that has led to, among other instances of folly, naïve lawyers filing submissions into court relying on case authorities that AI has made up.⁽⁷⁾

37               Nonetheless, AI tools have already demonstrated utility in at least four areas. One is in translating between languages. Another is in transcribing speech. A third is in searching out relevant points within voluminous material. A fourth lies in organising and summarising specified material, including by generating chronologies.

38               The third and fourth points are not limited to language documents. AI-powered search and organisation tools can work with images and video as well.

39               Having reviewed what AI offers already, and keeping in mind that its capability is growing quickly, I believe that it can be adopted to help manage complexity and address the curse of information overload.

40               There are two stages where AI tools can help to address complexity. The first stage is at the time of the project. I spoke earlier about how there is more information collected in the course of a project but there is less processing of it contemporaneously. AI tools can be used not merely to collect information but process it at the same time. This strengthens how management can respond in a timely and effective way as problems arise. To give an example: autonomous drones not only provide fuller and more complete visual records of construction progress but those records can be digested and problems flagged in real time for human operators to follow up on.

41               An important way in which AI tools can be used is to improve the quality of minutes of meetings. I am not talking about verbatim transcripts. Word for word transcription of oral communications is hard to read. Instead, the AI can summarise and rewrite, effectively translating spoken English into written English. This would be using pattern recognition in the same way that it might translate one written language into another written language, say French into English.

42               The second opportunity to use AI tools is when claims are being made or defended. They can be used to review and analyse defined material, to build timelines and organise disparate claims into intelligible categories.

43               I emphasis that none of this means replacing humans altogether. It is about extending our own capabilities, which has always been the function of tools. Humans need to review what AI has produced and analyse it further. Moreover, it will be organisations that must take responsibility for the output, and in doing so organisations will have to rely on real people as well.

44               Moreover, courts will have to rely on parties and counsel to check and verify. One can have recourse to the rebuttable presumption that computer systems are working as intended in the production of output, but the fact that some forms of AI involve greater autonomy than traditional methods of computer processing means that human oversight is essential. AI also gives rise to the possibility of “deepfakes” – fake evidence that deceives human senses – but AI will also offer tools to detect such fake evidence. None of this however can simply happen unsupervised. Humans will have to be involved, so that we make the best of AI without succumbing to the worst of it.

45               On this condition – that humans are involved and take responsibility – AI can help to manage complexity.

VI. Conclusion

46               The construction industry faces many challenges. It is not enough to build quickly and profitably. It is also necessary to do so safely and sustainably. Similar challenges face the management and maintenance of the built environment. Technology continues to advance at an ever-quickening pace. We are on the cusp of significant deployment of AI, at the design and construction stages and both embedded within construction machinery and within buildings being constructed. I have not spoken about that because that is beyond my expertise. My focus has been on how technological advances add to complexity, overloading human capacity. For these reasons and more, the need to adopt flexible and appropriate dispute resolution procedures is heightened and accentuated. Part of the answer is to adopt, adapt, and expand “good enough” methods, such as statutory or contractual adjudication, that offer business users relative speed and affordability. Another important element in addressing complexity and making the dispute resolution process less costly and more efficient is AI. AI can help contemporaneously during the project with the summarising and organising of the morass of information that swells with every single hour and day of a construction project. AI can also help with searching, interrogating and analysing evidence when it comes to making claims or defending claims. Ultimately, however, AI should be a tool that assists real people, extending their capabilities and expanding their capacities, and like any tool it is the human user who must take responsibility for it.

_{(1)       For discussion of AI in the context of the courts, see Sundaresh Menon, Chief Justice, Supreme Court of Singapore, “Judicial Responsibility in the Age of Artificial Intelligence”, Keynote Speech at the Inaugural Singapore-India Conference on technology (13 April 2024), https://www.judiciary.gov.sg/news-and-resources-details/chief-justice-sundaresh-menon--keynote-speech-at-the-inaugural-singapore-india-conference-on-technology (accessed 19 June 2024).
(2)       For discussion of the concept of complexification of disputes see Sundaresh Menon, Chief Justice, Supreme Court of Singapore, “The Complexification of Disputes in the Digital Age”, Speech delivered at Goff Lecture 2021 (9 November 2021), https://www.judiciary.gov.sg/docs/default-source/news-docs/goff-lecture-2021.pdf? (accessed 20 September 2023).
(3)       Urban Redevelopment Authority of Singapore, “Landscaping for Urban Spaces and High-Rises (LUST 2.0 Programme: Landscape Replacement Requirements for Non-Landed Private Residential and Commercial/Mixed-Use/Hotel Developments outside Identified Strategic Areas”, Circular (12 June 2014) https://safe.menlosecurity.com/https://www.ura.gov.sg/corporate/guidelines/circulars/dc14-13 (accessed 20 September 2023).
(4)       For discussion of the barriers to collaborative contracting in the Singapore context, see Na Zhang, Bon-Gang Hwang, Xiaopeng Deng, and Fiona Tay, “Collaborative Contracting in the Singapore Construction Industry: Current Status, Major Barriers and Best Solutions” (2020) 27(10) Engineering, Construction and Architectural Management 3115 at 3124.
(5)       https://www.judiciary.gov.sg/singapore-international-commercial-court/sicc-and-integraf
(6)       Professor John McCarthy, https://home.dartmouth.edu/about/artificial-intelligence-ai-coinded-dartmouth
(7)       See for example the recent decision of Zhang v Chen 2024 BCSC 285 in British Columbia where the lawyer concerned was made personally liable for the costs, to compensate for the additional effort and expense to which opposing counsel had been put in researching and addressing fake cases.}