Mohammed Khamisa KC, Leith Ben Ammar, Johnny Shearman, Greenberg Traurig LLP
This is an extract from the sixth edition of GAR’s The Guide to Construction Arbitration. The whole publication is available here.
This is an Insight article, written by a selected contributor as part of GAR’s co-published content. Read more on Insight
This chapter represents current understanding of rapidly evolving technologies and legal frameworks. Practitioners should consult current guidance and consider engaging specialist advice when implementing AI solutions.
Introduction to construction arbitration in the age of intelligent technologies
The construction industry, like many global industries, is undergoing a profound digital transformation. Artificial intelligence (AI), machine learning algorithms and intelligent automation now pervade almost every aspect of project delivery. The convergence of these technologies with traditional dispute resolution mechanisms presents both unprecedented opportunities and formidable challenges that demand careful consideration by practitioners, arbitrators and institutions alike.
The reality is that the arbitral resolution of construction disputes is evolving and the integration of intelligent technologies into construction arbitration is not a distant future consideration: it is a present reality requiring immediate attention. The 2025 White & Case and Queen Mary University of London International Arbitration Survey (the 2025 QMUL Survey) reveals that 91 per cent of respondents expect to use AI for research and data analytics within five years. This anticipated adoption is driven primarily by time savings (54 per cent of respondents) and cost reduction (44 per cent), factors that are particularly compelling in construction disputes, which are known for their complexity and expense.
It is thus not surprising that the construction sector’s embrace of intelligent technologies has been accelerated by economic pressures and the sector’s chronic productivity challenges. According to recent industry analysis, AI tools have transformative potential to minimise disputes throughout project life cycles while positively affecting profitability. This transformation extends beyond mere operational efficiency; it fundamentally alters how disputes arise, how evidence is gathered and analysed, and how arbitral proceedings are conducted.
The Law Commission for England and Wales’ recent discussion paper, ‘AI and the Law’, underscores the legal system’s growing imperative to engage meaningfully with AI’s expanding influence. The Law Commission identifies three critical themes that resonate deeply within construction arbitration: AI autonomy and adaptiveness; interaction with and reliance on AI systems; and the complex web of issues surrounding AI training and data utilisation. These considerations are particularly acute in construction disputes, where the stakes are invariably high and the technical complexity substantial.
The Chartered Institute of Arbitrators’ ‘Guideline on the Use of AI in Arbitration (2025)’ (the CIArb Guideline) represents the first comprehensive framework for navigating these waters. The Guideline acknowledges that AI technologies can enhance efficiency, improve quality and potentially remedy inequality of arms. However, it equally emphasises the risks: threats to confidentiality, data security concerns, potential bias in algorithmic decision-making and the fundamental challenge of maintaining human oversight over increasingly autonomous systems.
Yet this technological embrace must be tempered by proper understanding of AI’s limitations and risks. The ‘black box’ problem poses a particular challenge in arbitration, where transparency and due process are fundamental. As construction projects become increasingly complex, generating vast data sets and requiring sophisticated analytical approaches, the tension between technological capability and procedural integrity becomes ever more acute.
The way forward requires a nuanced understanding of how intelligent technologies can enhance rather than replace human judgement in construction arbitration. This chapter explores that delicate balance, examining both the transformative potential and the necessary safeguards required to ensure that technological advancement serves the cause of justice rather than undermining it.
Digital tools and technology use cases in the modern construction industry
The construction industry’s digital revolution has transformed how projects are conceived, executed and disputed. Understanding these technological foundations is essential for arbitrators and counsel seeking to navigate the complex landscape of contemporary construction disputes.
For example, Building Information Modelling (BIM) has evolved from a design tool into a comprehensive project management ecosystem. The advancement to 4D and 5D BIM, incorporating time scheduling and cost estimation, respectively, represents an important step in project oversight capabilities. These technologies allow for sophisticated sequencing of works, early detection of clashes between trades and predictive analysis of potential delays and cost overruns. When combined with AI capabilities, 4D and 5D modelling begin to approach the realm of prescient project management, potentially identifying problems before they materialise. The implications for dispute resolution are significant. Traditional delay analysis, long the domain of expert schedulers armed with conventional software, is being transformed by AI-powered analytical tools that can process vast data sets to identify causation patterns and quantify the effects with considerable precision. These tools can automate the identification of critical path changes, acceleration measures and disruption events.
Alongside the increasing use of AI, smart contracts and blockchain technology are beginning to permeate construction procurement and execution. These technologies promise automated compliance monitoring, instantaneous payment triggers and immutable record-keeping. These are all factors that could significantly reduce the incidence of payment disputes that arise in the construction sector. However, their implementation raises novel questions about liability allocation when automated systems malfunction or when smart contract logic produces unintended consequences.
Sensors deployed throughout construction sites (giving rise to an internet of things (IOT)) generate continuous streams of data regarding weather conditions, equipment performance, workforce productivity and safety compliance. Drones, aided by AI, can conduct ongoing safety inspections and progress monitoring. This real-time monitoring capability enables project managers to make adjustments to maintain schedule compliance and quality standards. In a dispute context, this data provides a granular level of evidence regarding project performance, replacing traditional reliance on often incomplete or subjective project records.
In addition, historical project data can be analysed using machine learning algorithms to predict likely disputes and recommend mitigation strategies. For example, predictive analytics may help identify patterns in stakeholder behaviour, specification changes and external factors that historically correlate with project delays and cost overruns.
Going one step further, AI has the potential to transform contractual administration. AI systems can already analyse contractual clauses, highlight potential conflicts and even draft variations based on pre-established templates. Coupling this technology with real-time site data could give rise to dynamic risk mitigation capabilities where corrective actions under an existing contractual framework are recommended instantly. This could shift contract management from a reactive to a proactive discipline and, were it to be deployed across the International Federation of Consulting Engineers (FIDIC) suite of contracts, the construction industry would, in all likelihood, operate very differently.
However, these technological advances potentially create new categories of disputes. The United Kingdom’s Civil Engineering Contractors Association recognises that, while AI implementation can drive significant benefits, it also creates new risks, including reduced privacy, overdependence on AI tools, cybersecurity vulnerabilities and potential loss of intellectual property. Already, we are seeing questions around algorithm accountability, data ownership, system interoperability and cybersecurity breach liability becoming increasingly common in construction disputes.
Therefore, the integration of these technologies requires careful consideration of existing legal frameworks. Under the European Union’s AI Act, high-risk AI systems (which could include many construction applications) must comply with transparency requirements ensuring their ‘operation is sufficiently transparent to enable deployers to interpret a system’s output and use it appropriately’. This creates obligations for providers to design systems with appropriate transparency features.
While England and Wales currently lack direct legislation specific to AI, existing legal obligations -including data protection requirements, professional negligence standards and contractual performance criteria – must be navigated carefully when implementing AI solutions. However, this regulatory gap creates uncertainty for construction professionals seeking to harness AI’s benefits while managing legal exposure. For instance, existing tort law principles, exemplified in Donoghue v. Stevenson and refined in cases such as Caparo Industries plc v. Dickman, must be adapted to address AI-related harms.
Despite this regulatory uncertainty, the transformation of the construction industry through intelligent technologies continues apace. The challenge therefore lies not in avoiding these technologies but in ensuring their integration serves the interests of fairness, efficiency and justice.
Claims, liability and remedies in algorithmic ecosystems
The integration of AI and automated systems into construction projects has the potential to challenge traditional concepts of contractual liability, fault attribution and remedial frameworks. If algorithmic decision-making becomes embedded in every phase of project delivery, from initial design optimisation to real-time construction management, the legal framework for addressing disputes must evolve to accommodate this new reality.
The Law Commission’s analysis of AI autonomy and adaptiveness highlights a critical challenge: determining liability when AI systems make independent decisions that result in project failures or delays. Traditional construction contracts allocate risk based on human agency and decision-making but AI systems have the potential to adapt and learn beyond their original programming parameters. This adaptiveness creates novel questions about whether liability should attach to the AI system developer, the project party deploying the system, or some combination thereof.
The emergence of AI-driven predictive analytics in construction risk management exemplifies these challenges. When machine learning algorithms analyse historical project data to recommend schedule adjustments or resource reallocation and those recommendations prove erroneous, determining liability becomes complex. Is the fault with the underlying data quality, the algorithm’s design, the implementation of the recommendations or the failure to adequately supervise the AI system’s outputs?
Taking AI-driven structural design as an example, an algorithmic error causing project delay might not fit traditional agency, requiring tribunals to extend attribution to system deployers or designers as the ‘best-placed’ risk-bearers. The Privy Council in Meridian Global Funds Management Asia Ltd v. Securities Commission recognised that traditional attribution rules may be inadequate in novel circumstances. Lord Hoffmann held that:
there would be little sense in deeming such a persona ficta to exist unless there were also rules to tell one what acts were to count as acts of the company. It is therefore a necessary part of corporate personality that there should be rules by which acts are attributed to the company.
The court developed ‘special rules of attribution’ for exceptional cases where primary and general rules fail, requiring courts to ‘fashion a special rule of attribution for the particular substantive rule’. This contextual approach becomes crucial for AI systems where algorithmic decisions may not fit conventional agency frameworks.
Comparatively, the EU AI Act imposes strict duties on high-risk AI providers across supply chains, mandating human oversight and liability for systemic failures, applicable to construction tools such as predictive analytics. For instance, Article 14(1) of the EU AI Act requires that ‘[h]igh-risk AI systems shall be designed and developed in such a way, including with appropriate human-machine interface tools, that they can be effectively overseen by natural persons during the period in which they are in use’ and Article 16 establishes that providers of high-risk AI systems mustensure that their systems are compliant with such oversight by natural persons. These provisions create clear lines of human accountability that supplement traditional corporate attribution principles.
Against this background, the concept of ‘smart contracts’ within construction agreements introduces additional complexity to traditional liability frameworks. Self-executing contracts, with terms directly written into code, have the potential to automatically trigger payments, impose penalties or initiate dispute resolution procedures based on predetermined criteria. However, the deterministic nature of smart contracts may not account for the nuanced circumstances that often arise in construction projects, potentially leading to harsh or inequitable outcomes that traditional contract interpretation would avoid.
Accordingly, there is a balancing act that is required where a degree of human oversight and accountability, at least for now, is necessary. However, the practical implementation of this balancing act requires clear contractual frameworks defining the boundaries of algorithmic decision-making authority. Construction contracts may have to address questions that were unimaginable a decade ago:
The allocation of liability for AI system failures presents particular challenges in construction’s multilayered contractual structures. When an AI-powered scheduling system fails to account for weather delays, resulting in project overruns, liability could theoretically attach to the software developer, the systems integrator, the project manager who relied on the system or the contractor who failed to maintain adequate backup planning procedures. Traditional contractual risk allocation mechanisms struggle to address these interconnected technological dependencies.
The emerging doctrine of ‘algorithmic negligence’ requires construction professionals to maintain reasonable competence in supervising AI systems within their domain of responsibility. This standard recognises that, while AI systems can enhance human capabilities, they cannot replace the duty of care owed by construction professionals, who remain at fault in selecting and supervising AI systems. The challenge lies in defining what constitutes reasonable supervision of systems that operate at speeds and scales beyond human cognitive capabilities in the construction context. Further challenges are caused by the ‘black box’ problem, which hinders causation analysis under English law’s ‘but for’ test and foreseeability requirements.
Remedial frameworks must also adapt to address AI-related failures. Traditional delay and disruption claims assume human decision-making timelines and processes. However, AI systems can cascade failures across multiple project systems within milliseconds, creating complex causation chains that challenge conventional delay analysis methodologies.
The quantification of damages in AI-related construction disputes requires advanced understanding of both technological capabilities and limitations. When an AI-powered quality control system fails to identify defective work, the resulting damages may include not only rectification costs but the additional expense of implementing enhanced monitoring procedures and potential loss of confidence in automated quality assurance systems.
Insurance and indemnification provisions in construction contracts will also need to increasingly grapple with AI-related risks. Traditional professional indemnity insurance may not adequately cover liability arising from AI system recommendations. Product liability frameworks may struggle to address the evolving nature of machine learning systems that adapt beyond their original specifications.
Evidence, expertise and the rise of digital forensics
The transformation of evidence gathering, analysis and presentation in construction arbitration represents one of the most significant effects of AI and digital technologies on construction disputes. The traditional paradigms of expert testimony and factual investigation are being revolutionised by tools that can process vast data sets, identify previously hidden patterns and present complex technical information with unprecedented clarity and precision.
The CIArb Guideline acknowledges that AI technologies can support evidence collection, analysis and presentation while emphasising the continued importance of human oversight and validation. This balance between technological capability and human judgement is particularly critical in construction disputes, where expert evidence often determines the outcome of multimillion-pound claims.
Digital forensics in construction disputes now encompasses far more than traditional document review. IOT sensors embedded throughout modern construction sites generate continuous data streams. This real-time data provides arbitrators with an unprecedented level of factual precision, potentially eliminating disputes about basic project conditions that previously relied on witness recollection or incomplete project records.
The application of AI to schedule analysis exemplifies the transformation of expert evidence. Traditional delay analysis required experts to manually analyse thousands of project activities, resource allocations and interdependencies to identify critical path impacts and quantify delays. AI-powered analytical tools can now process this data in minutes rather than months, identifying causation patterns and quantifying the effects with mathematical precision that can exceed human efforts.
However, this technological capability introduces new challenges for expert witnesses and arbitrators alike. The opacity of AI decision-making processes creates tension with fundamental principles of expert evidence. Experts must be able to explain their methodology and reasoning to tribunals but AI systems often operate through neural networks and machine-learning algorithms that are inherently unexplainable, even to their creators.
Therefore, while AI can dramatically enhance the efficiency and accuracy of data analysis, a strong argument can be made at this time that it cannot replace human judgement in interpreting results, understanding context and applying industry knowledge to technical findings. The most effective approach combines AI’s analytical power with human expertise in construction practices, contractual interpretation and dispute resolution.
The emergence of 4D and 5D BIM technology in evidence presentation transforms how arbitrators understand complex construction sequences and cost relationships. Rather than relying on static Gantt charts and cost schedules, parties can now present dynamic, visual representations of how projects unfolded over time, enabling arbitrators to understand causation and effect with unprecedented clarity. These visualisation tools can demonstrate the cascading effects of delays, the effects of design changes and the effectiveness of mitigation measures in ways that traditional expert reports cannot match.
Blockchain technology and smart contracts have the potential to create immutable audit trails that can serve as definitive evidence of contractual performance and compliance with project milestones. This technology potentially eliminates many factual disputes that traditionally consume significant time and expense in arbitration. However, this must be coupled with the questions that arise alongside this technology in terms of data integrity and system security as well as the interpretation of automated contract execution.
The integration of AI into expert report preparation presents both opportunities and risks. AI tools can assist experts with their data analysis, research and even initial draft preparation. However, the CIArb Guideline emphasises the continued need for human oversight and accountability, particularly in ensuring accuracy and avoiding the ‘hallucination’ problems that have been reported in current AI systems.
The 2025 QMUL Survey reveals that although the vast majority of practitioners expect to use AI for research and document review, there remains significant concern about accuracy, bias and confidentiality risks. These concerns are particularly acute when considering expert evidence, where inaccurate AI-generated content could fundamentally undermine a party’s case and potentially expose experts to professional negligence claims.
Cross-examination of AI-assisted expert evidence presents novel challenges for advocates and arbitrators. Traditional approaches to testing expert evidence focus on the expert’s qualifications, methodology and reasoning process. However, when AI systems contribute to evidence generation, cross-examination must extend to understanding the AI’s training data, algorithmic bias and the extent of human oversight applied to AI-generated outputs. In turn, this may give rise to yet more cross-examination at the final hearing stage.
The immediate future of expert evidence in construction arbitration lies not in choosing between human expertise and AI but in developing frameworks that harness AI’s analytical power while maintaining human judgement and industry knowledge.
Procedure, guidelines and best practices
As the underlying construction industry evolves, and as new disputes and causes of action materialise, so the way evidence is stored, gathered and analysed also changes. Somewhat inevitably, the procedural architecture of a construction arbitration is having to adapt. From a procedural standpoint, the challenge for arbitrators, practitioners and institutions lies in harnessing AI and automation and the advantages of these new capabilities while maintaining procedural fairness, transparency and due process, all of which form the foundation of effective arbitration.
The CIArb Guideline represents one of the first comprehensive frameworks for integrating AI into arbitral proceedings. The Guideline adopts a risk-based approach, recognising that different AI applications present varying levels of procedural risk and require correspondingly differentiated oversight mechanisms. This nuanced approach is particularly relevant in construction arbitration, where AI applications range from low-risk administrative tasks to potentially high-stakes evidence analysis and decision support.
The CIArb Guideline establishes four fundamental principles for AI use in arbitration: transparency, human oversight, data security and procedural fairness.⁴ In construction disputes, where technical complexity often engulfs traditional arbitral procedures, these principles provide essential guardrails for technological integration. Transparency requires parties and arbitrators to disclose AI use, explain methodologies and provide access to underlying data and algorithms where feasible. Human oversight mandates that meaningful human control be maintained over AI-assisted processes, particularly those affecting substantive rights or procedural outcomes.
Another example of a procedural framework for working with AI is the Silicon Valley Arbitration & Mediation Center’s ‘Guidelines on the Use of Artificial Intelligence in Arbitration’ (the SVAMC Guidelines), which introduce a principle-based framework to help participants navigate AI in arbitration. The SVAMC Guidelines are specifically for arbitrators. The guidance includes that arbitrators must not delegate decision-making to AI and must independently assess facts, law and evidence. In addition, arbitrators cannot rely on AI-generated material outside the record without disclosure, party comment and source verification.
The London Court of International Arbitration’s 2020 Rules acknowledge tribunals’ authority to make procedural orders regarding technology employment, potentially extending to arbitrators’ own use of technological tools. This broad discretionary power may enable tribunals to craft AI-specific procedures tailored to a particular dispute’s needs and complexity levels. However, it is arguable that this will also require arbitrators to develop a competent, if not sophisticated, understanding of AI capabilities, limitations and risks.
Therefore, best practice development in AI-assisted arbitration must address the tension between technological efficiency and procedural integrity. The 2025 QMUL survey indicates that 54 per cent of practitioners view time savings as the primary driver for AI adoption, while 44 per cent cite cost reduction as motivational. However, these efficiency gains must not compromise the fairness and thoroughness that dispute resolution demands.
When it comes to document discovery and review processes, AI-powered analytical tools can process millions of documents, identify relevant materials and flag potentially privileged communications with great accuracy, in some instances exceeding human reviewers. These capabilities are particularly valuable in construction disputes, which often involve extensive documentary records spanning multi-year project periods. However, procedural orders may need to start to address questions of algorithmic bias, false positive rates and quality control measures to ensure that AI-assisted discovery meets traditional standards of completeness and accuracy. Rarely is a ‘no stone left unturned’ approach adopted in a discovery process. Rather, it is a question of proportionality in terms of the value of the dispute and the cost of the exercise. AI-assisted discovery has the potential to move the dial towards a more thorough review process, but at proportionate cost.
From a case management perspective, the UK Ministry of Justice’s ‘AI Action Plan’ emphasises the potential for AI to reduce court backlogs and improve case management efficiency. Similar principles apply in arbitration, where AI tools can optimise hearing schedules, predict resource requirements and identify potential scheduling conflicts before they arise.
However, automated case management also must preserve essential human oversight and party autonomy. The CIArb Guideline emphasises that although AI can support administrative tasks, substantive procedural decisions must remain under human control. This principle requires careful delineation between appropriate AI assistance and impermissible delegation of arbitral authority.
Accordingly, procedural orders addressing AI use must balance technological flexibility with adequate safeguards. The CIArb Guideline provides model clauses addressing AI disclosure requirements, quality control measures and confidentiality protection. These templates can be adapted for construction arbitration’s specific needs, including technical evidence requirements, expert witness protocols and complex causation analysis procedures.
Training and competency development for arbitrators and counsel must advance to address AI integration challenges. The Society of Computers & Law and similar professional bodies are developing educational programmes addressing AI’s legal implications, but construction-specific expertise remains limited. Arbitral institutions may need to go further by considering whether AI competency should become a qualification criterion for the appointment of arbitrators to technology-intensive construction disputes.
On top of this, quality assurance frameworks for AI-assisted arbitration must address both technological reliability and procedural compliance. This may require the development of new professional standards, validation procedures and accountability mechanisms that ensure AI enhancement serves rather than undermines arbitral justice.
Jurisdiction, awards and enforcement
The digitalisation of construction arbitration extends beyond procedural efficiency to fundamental questions of arbitral authority, award validity and cross-border enforcement. As AI becomes embedded in arbitral decision-making processes and smart contracts automate dispute resolution triggers, traditional concepts of jurisdiction, due process and enforceability face unprecedented challenges.
The Law Commission’s exploration of AI legal personality raises profound questions for arbitration. If AI systems achieve sufficient sophistication to warrant legal recognition, their integration into construction projects could create novel jurisdictional issues. When an autonomous AI system makes decisions affecting construction project performance, determining the appropriate forum for dispute resolution becomes complex. Traditional rules linking jurisdiction to the contracting parties’ domicile or project location may prove inadequate when AI systems operate simultaneously across multiple jurisdictions with distinct laws and conflicting jurisdictional claims. One way of navigating this issue is for the parties to contractually agree on a location or jurisdiction in terms of where the AI systems are deemed to operate.
However, smart contracts embedded in construction agreements present particular challenges for traditional arbitration frameworks. These self-executing contracts can automatically trigger dispute resolution procedures, impose penalties or initiate payment processes based on predetermined algorithmic criteria. The deterministic nature of smart contract execution may conflict with arbitrators’ inherent discretion to consider equity, fairness and unforeseen circumstances that traditional construction contracts acknowledge through force majeure and other relief provisions.
The enforcement of awards generated through AI-assisted proceedings raises novel questions under the Convention on the Recognition and Enforcement of Foreign Arbitral Awards (the New York Convention). While the New York Convention does not explicitly prohibit AI use in arbitral proceedings, enforcement courts may scrutinise such awards for compliance with due process requirements and public policy considerations. In particular, the ‘black box’ problem of AI decision-making creates potential challenges where enforcement courts cannot understand or verify the reasoning of the underlying arbitral decisions.
The integration of AI into arbitral award drafting raises questions about authorship, accountability and enforceability. While AI tools can assist in research, analysis and initial drafting, the CIArb Guideline emphasises that arbitrators must maintain ultimate responsibility for award content and reasoning. This human oversight requirement becomes crucial for enforcement, as courts may be reluctant to enforce awards where AI systems played substantial unsupervised roles in decision-making.
Against this backdrop, procedural transparency and adequate party consent become crucial for ensuring award enforceability.
This equally applies where interim relief is sought and emergency arbitrator procedures are used. The rapid evidence analysis and decision support capabilities of AI, which have already been discussed, could make emergency procedures more effective. However, the speed of AI-assisted decision-making must be balanced against the need for adequate party participation and due process compliance. Emergency procedures that rely too heavily on AI analysis may face enforcement challenges if courts determine that insufficient human consideration was given to complex legal and factual issues, even at an interim stage of proceedings.
An illustration of the courts’ unease where AI decision-making without human accountability risks violating fundamental fairness principles occurred in LaPaglia v. Valve. In this 2025 case, a US district court faced a petition to vacate an award on the ground that the arbitrator allegedly ‘outsourced his adjudicative role to Artificial Intelligence’, raising serious due process concerns about non-human reasoning. LaPaglia v. Valve underscores a tribunal’s duty to avoid an arbitrator performing an alien act (i.e., the delegation to alien actors, human or machine) that may set aside an award under public policy grounds.
Next, blockchain technology’s near immutable record-keeping capabilities offer significant advantages for award enforcement by creating an evidence trail of the arbitral proceedings, award issuance and compliance status. However, blockchain implementation is still not widely adopted. That is in part due to the questions the use of blockchain technology raises in relation to data sovereignty, cross-border data transfer restrictions and the interaction between immutable distributed ledgers and traditional legal modification or correction procedures.
Where construction projects are to be managed through decentralised blockchain protocols with participants across multiple jurisdictions, traditional approaches to determining arbitral seat and applicable law become increasingly complex. This is compounded by the emergence of distributed autonomous organisations and the use of aliases and pseudonyms by blockchain participants. The UK Jurisdiction Taskforce’s Digital Dispute Resolution Rules first attempted to address some of these challenges.
More recently, the Blockchain Expedited Arbitration Rules (the BEAR Rules) facilitated by the London Chamber of Arbitration and Mediation (LCAM) since December 2024 go further in addressing these issues. For example, the BEAR Rules include the option for party anonymity. Where certain conditions are met, the identity of the claimant will not be disclosed to the other parties to the dispute. However, the arbitrator and LCAM can disclose the identity where it is necessary for the fair resolution of the dispute, for the enforcement of any award or order, if required by any law or regulation or court order, or to protect the arbitrator’s own interests. The primary method of enforcement envisaged by the BEAR Rules is on-chain, which would allow for anonymity to be preserved. However, under the BEAR Rules, it is still possible to go to a national court to enforce an award. At that point, it is envisaged that the identity of the anonymous party would be revealed to the party seeking to enforce.
More broadly, the confidentiality obligations inherent in arbitration proceedings face new challenges in AI-assisted contexts. When AI systems require access to confidential project information for analysis, ensuring adequate data protection while enabling effective AI assistance requires sophisticated technical and legal safeguards. Information security teams around the globe are having to assess these risks carefully to ensure confidential data remains secure.
Cross-border data transfer requirements may create additional complications for AI-assisted construction arbitration. The European Union’s General Data Protection Regulation, and various national data sovereignty requirements, may restrict the international flow of information necessary for AI-powered evidence analysis and case management systems. These restrictions have the potential to fragment global arbitration proceedings and complicate the use of largely cloud-based AI services that distribute processing across multiple jurisdictions.
The development of international standards for AI use in arbitration becomes crucial for ensuring consistent enforcement approaches across jurisdictions. The CIArb Guideline represents an important first step, but broader international coordination through organisations such as the United Nations Commission on International Trade Law may be necessary to provide comprehensive frameworks for AI-assisted arbitral proceedings.
Reform and future-proofing construction arbitration
The digital transformation of the construction industry has already begun to reshape the contours of dispute resolution, and arbitration sits at the very heart of this evolution. AI, machine learning, blockchain and intelligent automation are no longer peripheral innovations – they are becoming embedded in every stage of project delivery, contract administration and, ultimately, dispute adjudication. As such, their integration cannot be viewed as an abstract or future concern but as an immediate and pressing reality for practitioners, arbitrators and institutions alike.
This chapter has explored how these technologies are simultaneously creating opportunities and challenges across the arbitral process. From the proactive reduction of disputes through predictive analytics and smart contracts, to the transformation of evidence via digital forensics, 4D and 5D BIM and blockchain audit trails, AI-enabled tools hold enormous potential to enhance efficiency, accuracy and even equality of arms. Yet, these benefits come accompanied by risks that strike at the very foundations of arbitral justice: confidentiality breaches, algorithmic opacity, bias, cybersecurity vulnerabilities and the danger of undue reliance on autonomous systems.
Equally, the rise of algorithmic ecosystems forces a reconsideration of traditional concepts of liability, causation and remedies. The emergence of algorithmic negligence and the increasing complexity of risk allocation in multi-party contractual structures exemplify the extent to which established legal doctrines must adapt. Similarly, the evolution of arbitral procedure, from document discovery to multilingual hearings, requires new competencies and best practices to ensure that efficiency gains do not come at the expense of fairness or transparency. At the jurisdictional and enforcement level, the challenges are no less profound, as questions of AI involvement in decision-making, the rigidity of smart contracts and the enforceability of technology-assisted awards test the resilience of long-standing frameworks such as the New York Convention.
The Law Commission’s forward-looking analysis acknowledges that AI development may accelerate rapidly enough to require fundamental reconsideration of legal frameworks, including the ‘perhaps radical’ option of granting legal personality to sufficiently advanced AI systems. For construction arbitration, this possibility demands preparation for scenarios where AI entities could be parties to disputes, witnesses in proceedings or even serve in quasi-arbitral roles, subject to appropriate safeguards and limitations.
Conclusion
Looking forward, the construction arbitration community must prepare for transformative changes, including quantum computing’s potential effect on cryptographic security, advanced AI systems approaching general intelligence and the possible emergence of autonomous construction projects managed entirely by AI systems. These developments may seem distant, but the rapid pace of technological change demands proactive consideration rather than reactive adaptation.
Regulatory frameworks must strike careful balances between innovation promotion and consumer protection. The United Kingdom’s current approach of avoiding prescriptive AI regulation while maintaining existing legal obligations provides flexibility for arbitral innovation. However, the construction industry may require more specific guidance addressing liability allocation, quality assurance requirements and professional responsibility standards for AI use.
The development of AI-specific insurance and indemnification products becomes crucial for supporting innovation while managing the associated risks. Traditional professional indemnity insurance may not adequately cover liability arising from AI system failures. Specialised insurance products could facilitate AI adoption by providing appropriate risk allocation mechanisms.
What emerges from this analysis is the recognition that AI and digitalisation must be approached not as wholesale replacements for human judgement but as augmentations requiring careful governance. The future of construction arbitration depends on developing principled frameworks that preserve arbitral legitimacy while embracing technological progress. This balance demands close collaboration between lawyers, technologists, industry professionals and arbitral institutions, as well as sustained international dialogue to harmonise standards and manage cross-border complexities.
Ultimately, the arbitral community stands at an inflection point. The task is not to resist innovation, nor to embrace it uncritically, but to chart a responsible course that integrates technological capability with legal integrity. By doing so, arbitration can remain both relevant and resilient – capable of resolving the disputes of tomorrow while upholding the fundamental values of justice, accountability and enforceability that underpin its legitimacy.
Subscribe here for related content, breaking news and market analysis from Global Arbitration Review.
