AI and Arbitration – A Perspective from the UAE
-
Insight Article 19 September 2025 19 September 2025
-
Middle East
-
Tech & AI evolution
-
Dispute Resolution
The UAE is both a leading international arbitration hub and a key investor in artificial intelligence development and technology. In this article, Tom considers the UAE legal and regulatory framework application to AI in arbitration and the prospects for future development.
The United Arab Emirates (“UAE”) is uniquely placed to play a leading role in the development of use cases for Artificial Intelligence (“AI”) in international arbitration. Its key advantages are:
- A government and legal system which are avowedly pro-technology and determined to seize the nation’s place at the cutting edge of AI development and deployment;
- A light-touch approach to AI regulation which emphasizes self-regulation; and
- A diversity of judicial and legal systems.
That last point alludes to the UAE’s complex legal and judicial structure, which provides fertile soil for legal innovation. Though not a large country, the UAE is a federal nation consisting of seven emirates (each with their own laws in addition to federal law), four “onshore” court systems (the courts of Dubai, Abu Dhabi, Ras Al Khaimah, and the federal court system used by the emirates of Sharjah, Fujairah, Umm Al Quwain, and Ajman), and two financial free zones with their own legal systems and common law courts (the Dubai International Financial Centre (“DIFC”) and Abu Dhabi Global Market (“ADGM”)). There are three distinct jurisdictions in terms of arbitral legislation: the “onshore” UAE, which uses the Federal Arbitration Law (No. 6/2018), and the aforementioned DIFC and ADGM, each of which have their own arbitration laws; and two leading regional arbitration centres, namely the DIAC (Dubai International Arbitration Centre) and the ArbitrateAD (Abu Dhabi International Arbitration).
The significance of this pluralist legal system is that it creates “laboratories of innovation”: many opportunities for experimentation and the adoption of distinct approaches, where arbitration laws, rules, and the approaches taken by courts may take different paths until, in time, a degree of consensus may develop on international best practice for the integration of AI into arbitration.
Current Legal and Regulatory Position
The UAE does not currently have standalone laws specifically addressing the regulation of AI. The main supervisory authority is the UAE Artificial Intelligence Office, though several other agencies and government bodies have a remit which may partially embrace AI strategies, initiatives, and regulation. Nor does the UAE have any laws which specifically address the use of AI in the arbitration context.
AI has been deployed in certain use cases in international arbitration for some time. In particular, electronic document disclosure, also known as e-discovery, is a notoriously time-consuming (and therefore expensive) endeavour. The use of AI tools within agreed parameters to assist and streamline document review is well-established in international arbitration (and commercial litigation) and relatively uncontroversial.
Translation has seen major improvements in harnessing AI to produce more accurate and sensitive electronic translations of documents. This is particularly helpful in a jurisdiction like the UAE, where English is the dominant language of business and the language of offshore courts, while Arabic is the official language and the language of onshore courts (which may have supervisory jurisdiction over arbitrations, and/or a role in enforcement of an award), and where many languages are spoken by the expat population, including disputant parties. While certification of translations by a qualified translator is likely to remain a requirement at least for the foreseeable future, the integration of AI into translation workflows can achieve considerable time and cost savings. This is an area where the wider AI technological ecosystem can provide benefits, with the UAE having developed the world’s largest Arabic-language NLP model, NOOR.
More recently, some lawyers have begun to explore the use of AI to perform more substantive and technical legal tasks, such as research and drafting. This is not specifically addressed in the arbitration context by law or regulation in the UAE, but the pitfalls of AI hallucination have been well illustrated by a string of widely publicised cases. In the present state of the technology at least, reputable lawyers will not rely on AI-generated work product and will check its output rigorously, to the extent that it is used. Commonsense parameters are consistent with the DIFC Courts’ Practical Guidance Note No. 2/2023, which provides that practitioners in the DIFC Courts making use of AI should, to the extent possible:
1. Verify the accuracy and reliability of AI-generated content;
2. Disclose their intention to use AI at an early stage in proceedings;
3. Select the most appropriate GCG (Generative Content Generator) for their purposes;
4. Educate their clients; and
5. Protect client confidentiality and comply with legal obligations.
Another relevant consideration is the increased use of AI in the drafting and preparation of contracts, particularly high-volume and standard or commoditised contracts, where it is not economical to engage specialist lawyers to draft or review documentation. Effective AI contract software which is well-trained in dispute resolution clauses is less likely to produce some of the errors often found in individually drafted arbitration clauses, such as the conflation of an arbitral seat with an arbitral institution, or referring to a non-existent institution, which can render an arbitration clause difficult to implement or potentially ineffective. Contract review is not a major part of an arbitration practitioner’s work, but increased use of AI to prepare standard contracts may reduce the incidence of pathological arbitration clauses.
Future Developments
As discussed above, there is in general no requirement to disclose the use of AI in arbitration, unless the tribunal orders otherwise, though it is foreseeable that as use of the technology grows, it will be addressed by more and more tribunals (and arbitral institutions), just as some rulesets and procedural orders now address data protection and privacy after advancing regulation brought these issues to prominence. It seems likely that the use of AI in arbitration in future will be subject to disclosure and transparency requirements, as well as obligations to ensure that data processed by AI tools remains confidential.
The Government of Dubai recently announced the development, by the Dubai Future Foundation, of a scheme of Human-Machine Collaboration (HMC) icons. The icons denote the extent of AI involvement in producing a piece of content – whether it was entirely human, human-led, co-authored by humans and AI, primarily AI-generated and subject to human review, or entirely the product of AI – and the nature of the AI involvement (translation, data analysis, writing, and so on). The intention is that marking content with the HMC icons will be mandatory for the Dubai Government, and those working with the Government. This represents a possible stepping stone and a framework that, if successfully adopted and built upon in the Dubai public sector, could be deployed in the arbitration context.
There is special excitement in some quarters, and consternation in others, about the possibility of an AI programme donning a wig and robes and “sitting” as a judge or arbitrator. With rapid advances in the analytical capability of AI models, a functional AI arbitrator may not be a remote prospect, at least from a technological perspective.
The current UAE legal framework does not specifically contemplate, but in effect does not truly permit, an AI arbitrator. The UAE Federal Arbitration Law requires an arbitrator to be a “natural person” (Federal Arbitration Law, Article 10(1)(a)). The DIFC Arbitration Law and the ADGM Arbitration Regulations do not contain any similar language expressly excluding an AI arbitrator, but clearly contemplate the appointment of a natural person. Interpreting and reconciling those statutes to make sense in the context of an AI arbitrator – with provisions around the nationality of arbitrators, the obligation on them to disclose any impediments to their independence or impartiality, and liability for arbitrators (in extremely narrow circumstances) – would be extremely challenging. The same goes for the DIAC and ArbitrateAD rules. It can be presumed that, at present, the legal framework is not receptive to the use of AI arbitrators. The ethical dimension is also fraught with difficulty: one of the tenets of Smart Dubai’s AI Ethics Principles and Guidelines is that “AI subjects should be able to challenge significant automated decisions concerning them and, where appropriate, be able to opt out of such decisions”. Where an arbitral award almost inevitably disappoints at least one of the parties, there will be significant and unpredictable complications arising from a challenge to an AI-rendered arbitral award, until there is a clear protocol or legislative backing in place for that application of the technology.
Despite the foregoing difficulties, while arbitration conducted by AI is far from imminent, it may possibly loom closer in the UAE than in some more traditionally minded jurisdictions. This article opened with an explanation of why the UAE is well-placed to implement and capitalise on AI developments. The UAE’s recent announcement that it aims to use AI to review, revise, and draft new legislation shows a high degree of willingness to allow AI to shape substantive legal outcomes. If AI decision-making begins to play a part in international arbitration, it would be unsurprising to see the UAE and its institutions at the forefront of the rising trend. It is unlikely that we will see AI adjudicating multi-billion-dollar DIAC arbitrations in construction and infrastructure projects. However, in some relatively straightforward and lower-value sectors where efficiency and cost-effectiveness are paramount, there is potential.
Closing Thoughts
It is a deliberate national strategy to position the UAE at the forefront of AI research and development, and its government is particularly receptive to fostering and harnessing the enormous productive potential of artificial intelligence. The UAE has adopted a general tendency towards self-regulation, and has not so far legislated extensively on AI. AI has already identified some, albeit relatively circumscribed, use cases in arbitration; however, the UAE’s legal structure and forward-thinking strategy may lead to significant advancements.
This article was originally published on Daily Jus on Friday 19th of September, with thanks to Jus Mundi & Jus Connect.
End