Charlie Morgan (pictured left) is a partner; Georgia Di Salle is an associate; and James Doe is head of construction and infrastructure disputes at law firm Herbert Smith Freehills Kramer\u00a0<\/i><\/b><\/p>\n
News that the American Arbitration Association and the International Centre for Dispute Resolution have launched an \u2018AI arbitrator\u2019 has put the spotlight on the potential opportunities for artificial intelligence to resolve construction disputes \u2013 at least relatively straightforward, low-value ones.\u00a0<\/span><\/p>\n \u201cIf parties agree to use AI tools, those drafting contracts will need to consider the issue of enforcement carefully\u201d<\/span><\/p><\/blockquote>\n The AI arbitrator is intended for use in two-party documents-only construction disputes and purports to offer \u201cfast, cost-effective and trusted\u201d resolution. Terms such as \u2018fast\u2019 and \u2018cost-effective\u2019 will be familiar to contractors and other users of adjudication which, over the past 30 years, has become a popular form of interim dispute resolution.\u00a0<\/span><\/p>\n However, AI\u2019s rapid progress and the introduction of tools such as the AI arbitrator raise interesting questions. Could AI tools be the next evolutionary step for interim dispute resolution? Might they offer an alternative to established procedures such as adjudication? Could the use of AI adjudicators be possible within existing statutory adjudication regimes? And could AI provide a more practical solution to facilitate cashflow and successful project outcomes?\u00a0\u00a0<\/span><\/p>\n Section 108(2)(b) of the Housing Grants, Construction and Regeneration Act 1996 requires a construction contract to contain provisions for the parties to appoint an adjudicator, failing which the Scheme for Construction Contracts applies. Notably, section 108 says nothing about who (or what) the adjudicator should be, just that a process is agreed to appoint an adjudicator. However, the adjudication provisions of the scheme repeatedly refer to a \u201cperson\u201d being an adjudicator, which would preclude the appointment of an AI as an adjudicator under the scheme.\u00a0\u00a0\u00a0<\/span><\/p>\n Assuming for current purposes that an AI cannot be an adjudicator in UK statutory adjudication, even if parties were to elect to use AI models to resolve their dispute, a dissatisfied party would still be able to exercise its statutory right under the Act to refer a dispute to adjudication at any time. The AI process would therefore be an extra step within the broader dispute-resolution framework. Parties tend to view such extra preliminary steps as an unnecessary waste of time and money.<\/span><\/p>\n The enforceability of AI decisions is perhaps the key issue. Unlike adjudication decisions, which are binding unless, or until, finally determined by litigation or arbitration, AI-generated decisions have no statutory backing.\u00a0<\/span><\/p>\n If parties agree to use AI tools, those drafting the terms of their construction contracts will therefore need to consider the issue of enforcement very carefully. For example, would the AI decisions be temporarily binding, or not binding at all?\u00a0<\/span><\/p>\n Although it seems extremely unlikely (for the time being) that the parties would agree that AI decisions in all disputes should be fully and finally binding, what about decisions in lower-value disputes? There is no apparent reason why, if the parties agreed, that an AI decision could not be enforced in the same way as an expert determination, conclusive engineer\u2019s certificate or undisputed final statement \u2013 ie, by way of summary judgment or accelerated arbitration.\u00a0<\/span><\/p>\n If the parties had contractually agreed that an AI decision was fully and finally binding, presumably any adjudicator, court or arbitrator would, if following the contract provisions, subsequently \u201cenforce\u201d that agreement.<\/span><\/p>\n At the time of writing, it is clear that the promise of AI in dispute resolution as a sophisticated tool for lawyers \u2013 generating significant efficiencies for clients \u2013 is very real. However, in its current form, large language models cannot replace human reasoning, which is so important in the resolution of construction disputes, in particular those that are highly technical in nature.\u00a0<\/span><\/p>\n In our view, while AI continues to evolve rapidly, and those who fail to embrace it will no doubt fall behind, AI must be used responsibly with clear training on its capabilities and limitations. For now, it can support human decision-makers in disputes, but it cannot be a substitute for detailed human analysis.\u00a0<\/span><\/p>\n","protected":false},"excerpt":{"rendered":" Charlie Morgan (pictured left) is a partner; Georgia Di Salle is an associate; and James Doe is head of construction and infrastructure disputes at law firm Herbert Smith Freehills Kramer\u00a0 News that the American Arbitration Association and the International Centre for Dispute Resolution have launched an \u2018AI arbitrator\u2019 has put the spotlight on the potential …<\/p>\n","protected":false},"author":1,"featured_media":535604,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"ep_exclude_from_search":false,"footnotes":""},"categories":[578,579],"tags":[],"class_list":["post-535756","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-legal","category-tech"],"yoast_head":"\nEnforceable outcomes<\/span><\/h3>\n