In this two-part article, the author explores the impact of artificial intelligence (AI) on arbitration. In this first part, the author expounds on the evolution and functions of AI tools to consider how AI is designed and deployed. Then, the author examines the pros and cons of AI to emphasise how it can be beneficial or threatening for legal practice. Building on that discussion, the author provides a glimpse into the future with the appointment of AI arbitrators. In the conclusion of this article, to be published in an upcoming issue of Dispute Resolution Journal, the author will examine guidelines issued recently by various arbitral institutions setting out ethical boundaries on the use of AI in arbitration. The author will conclude this article by explaining that the arbitration process places considerable confidence in the kaleidoscopic and multifaceted aspects of human intelligence, thereby demanding legal and technical expertise to meticulously study the ramifications of AI arbitration in the future, in order to address concerns that the premature utilisation of AI arbitrators can trigger.
Digitalisation has significantly influenced the legal profession; it is unimaginable to envisage law practice without law-based solutions stimulated by technology. Artificial intelligence (AI) is an automated, machine-based system or computer software capable of making decisions, recommendations, and predictions based on objectives set by humans. AI can perform many tasks that are performable by human beings, and even AI can overdeliver or outperform humans.[1] Arbitration, an alternative dispute resolution mechanism, has also been influenced by technology. AI is being deployed in various arbitral processes as arbitration stakeholders, such as arbitrators, arbitration counsel, arbitral institutions, and experts, increasingly adopt this technology. Arbitration stakeholders are harnessing AI tools to achieve various goals, including summarising documents, editing and formatting text, and analysing legal documents. There exists a range of opinions on whether AI in arbitration is a threat or an opportunity; despite varying perspectives, the AI tools are making room in arbitration legal landscape and the rapid increase in the recognition and deployment of AI in arbitration and the thriving reliance of arbitration stakeholders on AI tools reflect AI’s growing role in the enhancement of the arbitral process and procedure.[2]
To that end, this article explores the impact of AI on arbitration.
First, in broad strokes, this article expounds on the evolution and functions of AI tools to consider how AI is designed and deployed to assist humans and foresee the outcome of a case.
Then, this article examines the pros and cons of AI to emphasise how it can be beneficial or threatening for legal practice. Specifically, this section of the article critically analyses the significant risks associated with AI—including a lack of emotional intelligence, AI bias, hallucinations, and opacity—that impact outcomes.
Building on the discussion, furthermore, this article provides a glimpse into the future with the appointment of AI arbitrators. This article examines whether there is room for AI arbitrators. It explores the indispensable principles and standards of the current arbitration framework to evince how the appointment of AI arbitrators may spark tension with those values and norms, and what the repercussions of such engendered incompatibility might be.
The conclusion of this article, to be published in a future issue of Dispute Resolution Journal, will explain how, in consideration to rapid advancement in putting AI in arbitration practice, various arbitral institutions, like American Arbitration Association®, Chartered Institute of Arbitrators, Silicon Valley Arbitration & Mediation Centre, and Vienna International Arbitral Centre, issued guidelines very recently in 2025 to set out ethical boundaries on the use of AI in arbitration to protect the integrity of arbitration, due process, and maintaining confidence of parties in arbitral outcomes. This article examines these guidelines to determine whether these guidelines provide room for the deployment of AI arbitrators. Given this, this article scrutinises the guidelines to bring the attention of the arbitration community to this contemporary progress, as such guidelines are crucial to ensuring transparency, accountability, and fairness in AI tech-assisted arbitration.
Accentuating legal concerns regarding the application of AI tools in general and AI arbitrators specifically should not be mistaken for skepticism of innovation, a bias against the status quo, or an anti-innovation approach. However, the primary objective of this article is to delineate the creation and operation of AI arbitrators in a manner that is acceptable to courts and arbitration stakeholders, when technically viable. The discussion culminates in a conclusion that the arbitration process places considerable confidence in the kaleidoscopic and multifaceted aspects of human intelligence, thereby demanding legal and technical expertise to meticulously study the ramifications of AI arbitration in the future, in order to address concerns that the premature utilisation of AI arbitrators can trigger.
Demystifying AI
The term “intelligence” refers to brain-related activities, that is, understanding, learning, and reasoning.[3] The intelligence in AI refers to human-like intelligence, as AI stimulates aspects of mental agency. Nevertheless, from a biological perspective, AI is not alive like humans.[4] From a psychological perspective, AI is perceived as emotionless.[5] Physiologically, AI lacks a brain and nervous system; therefore, it lacks consciousness, a characteristic that humans possess.[6]
Outlining the ability of computers to acquire intelligence and establishing a boundary between “natural” or “human” intelligence and “artificial” intelligence is significant in determining whether an AI can effectively act as an arbitrator to protect parties’ legal rights and settle disputes between them. Human desire to create intelligent technology is not new. The constant evolution led to the creation of expert systems, containing the “rule-based expert system,”[7] which is composed of two foremost components: (1) facts vis-à-vis a situation, and (2) rules applicable to such facts. A rule-based expert system requires human expert involvement to specify the procedures and steps that the machine must follow to make a decision or respond to a question. To that end, the decision made is understandable to an expert human because it results from a procedure set by a human expert. That being the case, programmers’ and users’ transparency and integrity are maintained, as humans have the power to review the system’s work and design and ascertain the rules that AI has applied to make a decision.[8] Similarly, the system is transparent, reliable, and cost-effective. Rule-based expert systems require constant updating and the incorporation of new data and information over time. It employs “if” and “then” logic; therefore, it was limited in handling complex realities and subjects. Consequently, in the 1990s, AI was introduced to fill the gaps of the existing systems.[9]
At that stage, more emphasis was placed on machine learning. AI is a machine’s capability to learn in a way that humans call intelligent. The learning capability of technology is the creation of human beings and is an integral element of the digital economy. The intelligence of such technology is not natural but artificial. Machine learning is a subset of AI, whereby controlled data in the form of algorithms is input, and without human intervention, machines learn from the data.[10] It correctly classifies items, performs complex analyses, and automates routine tasks. Information is fed into the machine learning algorithms. In line with the algorithm’s reflexive response to inquiries that shape its operation, the algorithm creates information in the form of output. The more information in the form of data is input via an algorithm, the more it will learn, and the efficacious its responses to inquiries. To that effect, machine learning is capable of recognising connections and rules even though the programmer fails to understand them. For that reason, machine learning is considered a valuable and effective problem-solving tool, particularly in cases where the rules are complex or unclear.[11]
The significant steps involved in designing machine-learning technology include:
The data must be collected and transformed into a form that can be input into the technology.
The applicable model should be selected. For a particular problem, there is a specific algorithm; nevertheless, algorithms address sets of problems simultaneously. Therefore, regarding a particular problem, there can be competition among algorithms, as one may be considered superior to others. Programmers select the most effective algorithms that yield the best results for a specific problem. Various ready-made machine learning models are accessible to everyone.[12]
After selection of the best model, there is a need to train the applicable model. In the present era, resolving minor problems with the assistance of AI often requires extensive training and access to a significant data source.[13] Algorithms require a data source for learning; however, they are capable of learning from data in various ways. Such learning is supervised; therefore, it is attracting legal attention.[14] In the supervisory form of learning, input and output are provided, and the algorithm is required to create a flowchart that follows the input leading to the output. During training, the supervisor’s role is to provide accurate values to initiate learning. The algorithm’s capability to solve problems is increased when it learns from the data, which is characterised by desired responses.[15]
Pros and Cons of AI in Arbitration
Various AI tools are currently assisting in legal drafting and research, including ROSS Intelligence,[16] Lexis+ AI,[17] Harvey AI,[18] vLex,[19] LawGeex,[20] and Luminance.[21] In the segment below, various pros and cons of AI are provided.
Pros
There are several pros to using AI in arbitration, including being time- and cost-efficient, consistent and precise, and capable of making informed predictions. These AI tools automate various time-consuming tasks, including legal research of a specific nature, document review, data management, tracking deadlines, summarising documents, transcription, translation, sending emails, and sending reminders. These tasks may encompass numerous documents, and arbitration counsels or arbitrators may spend hours reviewing such documents. Similarly, complex arbitration involves a meticulous review of bundles of documents to support a claim and evaluate evidence. If AI automates such tasks, it can perform the job efficiently.[22] For instance, law firms are actively adopting AI tools, such as Harvey—an AI-driven legal automation tool—to improve and automate various aspects of legal work, including the preparation of legal memoranda and the enhancement of legal research quality.[23]
As discussed above, one of the unique functions of machine learning (ML) is that it can predict case outcomes by evaluating the likelihood of success in each case. Moreover, it can also predict whether a claim or defence has a higher chance of success. ML can scrutinise former cases and their decisions in order to predict outcomes in future cases.[24] In 2014, Arbitrator Intelligence, an AI-automated tool, was launched that is capable of gathering data and feedback on decision-making patterns of over a thousand arbitrators.[25] Due to this, the parties and their arbitration counsels gain insight into the perspectives of various arbitrators regarding their cases and strategies.[26] This aspect of AI is cost-efficient and beneficial in predicting the weaknesses and strengths of the case and reckoning the success chances. Therefore, arbitration counsels can make timely adjustments to their strategies to ensure a better presentation before the tribunal and achieve more favourable results.[27]
AI has already been deployed to enhance procedural effectiveness. By way of illustration, “Technology-Assisted Review” (TAR) or “Computer-Assisted Review” (CAR) is a prime example of Machine Learning Technology. CAR, also known as “predictive coding,” is an efficient e-discovery process that identifies relevant data and documents using AI tools. The English High Court in Pyrrho Investments Ltd v. MWB Property Ltd stated that the Court is satisfied that TAR provides adequate transparency and it discharges the discovery obligations of a party.[28] The U.S. District Court for the Southern District of New York, in Moore v. Publicis Groupe, validated the use of computer technology in performing the discovery process.[29]
In TAR, AI assesses a bulk of electronically stored information (ESI). AI understands the coding of sample input by human expert reviewers, then distinguishes the relevant data from ESI to generate the required document according to the document creation demand.[30] The Irish High Court in Irish Bank Resolution Corporation Ltd v. Quinn noted that “if one were to assume that TAR will only be equally as effective, but no more effective, than a manual review, the fact remains that using TAR will still allow for a more expeditious and economical discovery process.”[31]
In arbitration, AI-automated translation and transcription are practically welcomed as they are time-effective and provide instantaneous outcomes. For example, SeamlessM4T, a multimodal and multilingual META AI model, is an effective tool that provides numerous translations, including speech-to-speech, text-to-text, speech-to-text and text-to-speech. This tool supports various languages and facilitates cross-language communication.[32] Similarly, natural language processing (NLP) is another tool that is revolutionising the legal landscape of arbitration. In NLP, the computer processes, understands, interprets, and generates human language.[33]
If AI is utilised appropriately on data-driven work based on a specific pattern or template, the arbitration process can be optimised. Regarding consistency and precision, AI assists not only in detecting data errors, including missing values, redundant features, incorrect labels, duplicate records, or typos, but also helps correct them. AI is capable of employing logic and rules to protect data consistency collected from various sources. Due to these attributes, AI provides satisfactory assistance in areas such as data management, valuation, and calculation.[34] Specifically, in the context of arbitration, AI is the best tool for drafting and proofreading specific standardised segments of documents or awards, such as the chronology of events or a factual matrix of the arbitration case, subject to verification by a human. More importantly, AI can assist in calculating damages by utilising various methods, including the market value method, replacement cost method, diminution in value method, present value method, repair method, and lost profits method. However, under various assumptions fixed by arbitrators or parties, AI can use these methods to calculate damages at different dates, for instance, from the date of the breach till the date of the hearing or award.[35]
Cons
To assist in the arbitral process, AI is undoubtedly a beneficial tool; however, some cons and pitfalls exist regarding the use of AI in arbitration, which are discussed below.
Assistance from AI tools may raise data privacy concerns, particularly in cases where confidential, private, sensitive, and privileged data is input for AI automation without following proper measures and protocols. Various AI tools can retain the information or protected data that has been entered into the AI for numerous purposes.[36] For instance, the privacy policy of OpenAI states that it collects users’ personal information, including account, technical, and communication data, such as device and usage data, as well as log files.[37] There is also concern that in some situations, OpenAI may disseminate collected information to a third party without notifying the actual user.[38] Several examples illustrate the risks associated with the dissemination of information by AI tools. For example, in 2023, as confirmed by OpenAI, a software bug unfortunately exposed the payment details of ChatGPT users,[39] including their residential addresses, credit card information, and email addresses.[40]
The AI results of an input are based on complex probabilistic computations, despite employing logical reasoning. AI is often incapable of expounding on its algorithmic calculations; therefore, it lacks self-awareness at times.[41] As a result, AI may devise conclusions that lack adequate explanation of how it arrived at the given results, which is commonly referred to as a black box issue.[42]
One of the significant pitfalls is that AI may engender ethical concerns for arbitrators and arbitration counsel. In some instances, where they fail to make adequate disclosure of their use of AI, they may breach the rules and protocols of professional conduct, thereby undermining the integrity and fairness of arbitral proceedings. Arbitration counsel is bound by duty not to misrepresent fact or law. Hence, they may be held accountable or exposed to disciplinary sanctions if they entirely rely on AI for fact-finding and the conduct of legal research without substantiation.[43] Mata v. Avianca, Inc.[44] can be cited in this regard, whereby Mata brought a claim against Avianca Airlines, asserting personal injury caused to him when a metal cart smashed into his knee during a flight. To substantiate the claim, Mata’s counsel cited three cases. The opposing counsel and judge were unable to find judgments cited in the brief. This hitch occurred due to ChatGPT, as it had prepared the brief. The court considered such an act intolerable and imposed sanctions against the counsels by stating:
In researching and drafting court submissions, good lawyers appropriately obtain assistance from junior lawyers, law students, contract lawyers, legal encyclopedias and databases such as Westlaw and LexisNexis. Technological advances are commonplace and there is nothing inherently improper about using a reliable artificial intelligence tool for assistance but, existing rules impose a gatekeeping role on attorneys to ensure the accuracy of their filings.[45]
The arbitrators may also be held liable if they indecorously utilise AI. Arbitrators must not delegate their decision-making authority to AI, as this may raise ethical concerns. LaPaglia v. Valve Corporation is one of the most recent pending cases in the arbitration landscape pending before the U.S. District Court for the Southern District of California (U.S.D.C.).[46] In this case, LaPaglia lodged a petition before the U.S.D.C. and pleaded to vacate the arbitral award according to 9 U.S.C. §§ 10(a)(3), (a)(4), based on the fact that AI carried out the arbitrator’s adjudicative role. LaPaglia further asserted that the arbitrator stepped on the toes of AI in order to draft an arbitral award. LaPaglia substantiated his assertion with concrete facts, detailing how the arbitrator shared a story about being assigned to author an article and how he beat the clock and finalised this assignment by leveraging ChatGPT. LaPaglia stressed that if this story is read in conjunction with the arbitrator’s statement that he wanted to issue an award before leaving for a trip, it is evident that ChatGPT did the arbitrator’s adjudicative duty of writing an award.
Moreover, the impugned award was riddled with several telltale signs of AI. More importantly, the award cited facts without relevant citations, even though such facts were not only untrue but also were not part of the record. LaPaglia’s lawyer’s law clerk questioned ChatGPT as to whether AI or humans wrote various paragraphs in the award. ChatGPT responded that awkward paraphrasing, disorganisation, overgeneralization, confusion, redundancy, and inarticulacy suggest that paragraphs were written by AI rather than humans.[47] Despite the outcome of the case still being unknown, this case serves as an example and starting point for examining the limitations of AI in arbitration and the ethical and legal obligations that accompany it.
AI tools and technology are not equipped to replicate emotional and conscious intelligence, which is a unique characteristic of human beings. These characteristics are significant in the context of arbitration, as AI technology is not yet capable of rendering judgment in matters concerning fairness, justice, equity, public policy, and cultural aspects. These factors, therefore, are understandable by human arbitrators.[48] AI tools function independently. However, their competencies are influenced by the considerations and datasets that are input into them. For this reason, AI may exhibit biases that its programmers possess. Moreover, biases are generally perceived in arbitrators; nevertheless, it might be more problematic to rectify such biases from AI tools.[49]
Additionally, AI technology sometimes produces unreliable outputs due to hallucinations. For example, AI large language models generate fallacious results that are unreasoned and not substantiated by the given facts. AI hallucinations result from the utilisation of computational probabilities, but AI have no adequate knowledge to give an exact answer to a particular question; therefore, the tool fails to evaluate the correctness of the resulting output.[50] Mata v. Avianca is a prime example of AI hallucinations.[51]
Deployment of AI Arbitrators
Whether it is imaginable to glimpse into arbitration’s future in the context of the appointment of AI arbitrators, the employment of AI tools, or the use of generative AI in decision-making in arbitration raises an intriguing debate on the role of this technology in influencing outcomes.[52]
The effectiveness of arbitration hinges on the integrity and quality of arbitrators. The parties select arbitrators for various reasons, such as professional experience, cultural background, reputation, expertise, neutrality, prior judgments, quality of decision-making, skill, and sound past approaches.[53] While the arbitration legal landscape provides leeway for parties to decide on arbitral procedures,[54] employing AI to serve as a tribunal depends on various ethical and legal considerations, such as bias and the potential to evaluate the legal arguments advanced by the parties’ arbitration counsel. The competency, aptitude, and expertise of humans surpass those of existing AI technology. An arbitrator’s primary role is to adjudicate the dispute, which requires considerable intelligence and human input to render a fair and just judgment.[55]
Arbitrators give adequate consideration to intangible legal principles, which are subject to interpretation. For instance, reasonableness, proportionality, exceptional circumstances, and good faith. Some of these concepts lack clear definitions and are interpreted according to the specific circumstances and facts of the case in a particular context. By way of illustration, in a dispute that requires the interpretation of a contractual term, whereby, parties must agree on the price of a product or service in accordance with well-regarded commercial business ethics, that is, the principles of honest reputation and mutual benefit. In this situation, the arbitrators would hear the parties’ submissions and examine various elements, including contractual language, prior business dealings, the parties’ presumed intent, and the factual matrix of the case, existing precedent, and established practices.[56]
However, in the given context, arbitrators are required to imply a contractual term under the English legal regime. Before the implication of a term, the arbitrators must be satisfied that: first, the term must not be implied in the agreement on the sole ground that it appears to be fair. Second, it should be implied where the agreement fails to demonstrate practical and commercial consistency or to provide business and commercial effectiveness to such an agreement.[57] Third, the necessity and obviousness require the term to be implied. Fourth, implied terms must not be in contradiction with the explicit terms of the contract. Fifth, the implied term should be capable of clear expression.[58] Before considering implied terms, arbitrators must address various legal concepts, including business and commercial effectiveness, equity, obviousness, and reasonableness. Thus, it is suitable that human arbitrators should analyse, examine, and apply these concepts based on the specific facts of each case, given the limitations of AI tools. These concepts must be correctly applied, as their incorrect application could adversely impact the outcome of the case.[59]
When rendering an award, arbitrators not only consider applicable laws but also combine them with considerations of equity and fairness. Typically, the arbitral tribunal grants equitable relief and applies the equity principle as equity follows law. The application of equity requires the tribunal’s substantial discretion, which correspondingly entails an adequate understanding of the facts and law, along with an extensive social context. The application of equity also demands rendering a cogent decision based on parallel, competing interests by implementing ethical reasoning, consciousness, and perceptivity. The arbitral rules of various arbitration institutions confer power on arbitrators to determine disputes as ex aequo et bono or amiable compositeur, whereby, if the parties to the arbitration agree, the arbitrator is required to settle the dispute based on fairness and equity, rather than strictly applying general legal rules.[60] As a generative AI tool, ChatGPT is unable to perform a detailed and thorough analysis if parties confer ChatGPT with the power of amiable compositeur or request that it decide ex aequo et bono.[61] To that extent, human arbitrators can proficiently apply equity principles and grant equitable reliefs, whereas AI arbitrators lay down arms.
The legal requirement is to render a reasoned award; for instance, the UNCITRAL Model Law, under Article 31(2), provides that the arbitral award shall provide adequate reasons on which it is based, unless the parties agree that the award does not provide any reasons.[62] Similarly, the ICSID Convention, under Article 48(3), stipulates that an award must address and resolve every question before the tribunal and provide reasons for its decisions.[63] Practically, how an arbitrator makes a reasoned award requires a high level of humanness. The tribunal has to understand the facts of the case and assess the standing of both parties. Moreover, the tribunal must carefully consider the legal requirements governing the arbitration, including agreements, applicable rules, and relevant laws. Furthermore, the tribunal is required to hear arguments and evidence from both parties, including witness testimony, documentary evidence, expert witnesses, and all other relevant information. After that, the tribunal must weigh and assess such evidence and information. Then, the tribunal writes an award and issues it, delineating legal reasoning and findings behind every issue before the tribunal.[64]
Taking all this into account, the existing AI tools are capable of summarising a significant quantity of documents, data, information, and evidence related to a case. However, AI may lack in performing the sentiment analysis, and AI tools remain unsuccessful in weighing evidence as it can be done by a human arbitrator, who, after cross-examination, evaluates witness credibility and documents genuineness. However, simple arbitrations may be settled by reviewing documents alone. These simple cases may be suitable for AI arbitrators, which are less complex and involve minor commercial disputes, such as the dispute between a seller and purchaser that are generally settled through an automated online dispute resolution platform.[65]
In the current situation, AI tools can outperform arbitrators in summarising documents and conducting research; however, to verify the human element in arbitrators’ research, additional measures may be required.[66] Moreover, the responses or research generated by ChatGPT are formulated based on text data patterns. Such research provides an inadequate understanding of the intricacies of law and the nuances of disputes, which a human arbitrator more readily understands due to professional experience and years of legal training. ChatGPT significantly fails in drafting a reasoned award, as it remains unsuccessful in understanding and generating complex reasons, which is a requirement in arbitrations. Precisely, it cannot provide reasoning concerning complex arbitrations. The reasoning that it provides remains unsubstantiated and fundamentally flawed.[67] Therefore, it may be incompetent to make a satisfactory arbitral award for the parties.
Currently, AI may not be appointed as arbitrators because it cannot function in the same manner as a human brain. Hence, AI may perform the functions of a shadow arbitrator with no decision-making power. In the future, AI technology may be developed to the extent that it acquires the ability to reason and think, utilising deep neural advancements and being stimulated like the human brain. At that age, the AI arbitrator might be appointed to settle complex disputes, or the decision-making aspects may be delegated to AI tools.
* * *
Editor’s note: This article will conclude in an upcoming issue of Dispute Resolution Journal.
[1] Megan Beck & Barry Libert, “The Rise of AI Makes Emotional Intelligence More Important,” 15 Harvard Business Review 1 (2017).
[2] Kathleen Paisley & Edna Sussman, “Artificial Intelligence Challenges and Opportunities for International Arbitration,” 11 New York Dispute Resolution Lawyer 35 (2018).
[3] Han L.J. Van der Maas, Lukas Snoek & Claire E. Stevenson. “How Much Intelligence Is There in Artificial Intelligence? A 2020 Update,” Intelligence 87 (2021); A. Breakspear, “A New Definition of Intelligence,” 28 Intelligence and National Security 678 (2013).
[4] Robert Walters, “Robots Replacing Human Arbitrators: The Legal Dilemma,” 34 Information & Communications Technology Law 129 (2025).
[5] Gerd Gigerenzer, “Psychological AI: Designing Algorithms Informed by Human Psychology,” 19 Perspectives on Psychological Science 839 (2024).
[6] Anil K. Seth, “Conscious Artificial Intelligence and Biological Naturalism,” Behavioral and Brain Sciences 1 (2024).
[7] Pedro Zuidberg Dos Martires, Luc De Raedt & Angelika Kimmig, “Declarative Probabilistic Logic Programming in Discrete-Continuous Domains,” Artificial Intelligence 337 (2024).
[8] V.C. Manduva, “Current State and Future Directions for AI Research in the Corporate World,” 2 The Metascience 70 (2024).
[9] Crina Grosan et al., “Rule-Based Expert Systems,” Intelligent Systems: A Modern Approach 149 (2011).
[10] Lucila Carvalho et al., “How Can We Design for Learning in an AI World?,” Computers and Education: Artificial Intelligence 3 (2022).
[11] Cole Dorsey, “Hypothetical AI Arbitrators: A Deficiency in Empathy and Intuitive Decision Making,” 13 Arbitration Law Review (2021).
[12] John D. Kelleher, Brian Mac Namee & Aoife D’arcy, Fundamentals of Machine Learning for Predictive Data Analytics: Algorithms, Worked Examples, and Case Studies” (MIT Press, 2020) 13.
[13] Kan Yao & Yuebing Zheng, Fundamentals of Machine Learning (Springer International Publishing, 2023) 77.
[14] Ibid.
[15] Emam Hossain et al., “Machine Learning with Belief Rule-Based Expert Systems to Predict Stock Price Movements,” Expert Systems with Applications 206 (2022).
[16] See ROSS Intelligence, https://www.rossintelligence.com/about-us/.
[17] See Lexis+ AI, https://www.lexisnexis.com/en-us/products/lexis-
plus-ai.page.
[18] See Harvey AI, https://www.harvey.ai/.
[19] See vLex, https://vlex.com/.
[20] See LawGeex, https://www.lawgeex.com/.
[21] See Luminance, https://www.luminance.com/.
[22] Divyesh Kumar & Noor Afza, “Robotic Process Automation: The Emerging Technology,” Handbook of Artificial Intelligence Applications for Industrial Sustainability (CRC Press, 2024) 274; Tania Sourdin, Bin Li & Tony Burke, “Just, Quick and Cheap?: Civil Dispute Resolution and Technology,” 19 Macquarie Law Journal 17 (2019).
[23] Crenguta Leaua & Corina Tanase, “Artificial Intelligence and Arbitration: Some Considerations on the Eve of a Global Regulation,” 17 Romanian Arbitration Journal 31 (2023); Georgios I. Zekos, “AI in Arbitration and Courts,” Advanced Artificial Intelligence and Robo-Justice. 321 (Springer International Publishing, 2022); Cristina Ioana Florescu, “The Interaction Between AI (Artificial Intelligence) and IA (International Arbitration): Technology as the New Partner of Arbitration,” 18 Romanian Arbitration Journal 42 (2024).
[24] Emily Hopkins, “Machine Learning Tools, Algorithms, and Techniques,” 10 Journal of Self-Governance and Management Economics 43 (2022).
[25] Catherine Rogers, “‘Arbitrator Intelligence’ Is Here,” Kluwer Arbitration Blog (21 September 2014).
[26] K. Simpson, “Arbitrator ‘Intelligence’ and the Mysterious Brown M&M,” 52 The University of Toledo Law Review 27 (2021).
[27] Mohammad Ali Solhchi & Faraz Baghbanno, “Artificial Intelligence and Its Role in the Development of the Future of Arbitration,” 2 International Journal of Law in Changing World 56 (2023).
[28] Pyrrho Limited v. MWB Limited [2016] EWHC 256 (Ch).
[29] Da Silva Moore v. Publicis Groupe et al., No. 1:2011cv01279.
[30] G. Vannieuwenhuyse, “Arbitration and New Technologies: Mutual Benefits,” Journal of International Arbitration, 35(1) (2018); Mikko Luomala, Jyri Naarmala & Ville Tuomi, “Technology-Assisted Literature Reviews with Technology of Artificial Intelligence: Ethical and Credibility Challenges.” 256 Procedia Computer Science 378 (2025).
[31] Irish Bank v. Quinn [2015] IEHC 175.
[32] Loïc Barrault et al., “SeamlessM4T: Massively Multilingual & Multimodal Machine Translation,” Cornell University arXiv:2308.11596 (2023).
[33] Georgios I. Zekos, “AI in Arbitration and Courts,” Advanced Artificial Intelligence and Robo-Justice (Springer International Publishing, 2022) 321; Cristina Ioana Florescu, “The Interaction Between AI (Artificial Intelligence) and IA (International Arbitration): Technology as the New Partner of Arbitration,” 18 Romanian Arbitration Journal 42 (2024).
[34] Layan Al Fatayri, “AI in International Arbitration: What Is the Big Deal,” American Review of International Arbitration, https://aria.law
.columbia.edu/ai-in-international-arbitration-what-is-the-big-deal/ (2024); H. Surden, “Artificial Intelligence and Law: An Overview,” Georgia State University Law Review, 35, 1305 (2018).
[35] Kathleen Paisley & Edna Sussman, “Artificial Intelligence Challenges and Opportunities for International Arbitration,” 11 New York Dispute Resolution Lawyer 35 (2018).
[36] David Elliott & Eldon Soifer, “AI Technologies, Privacy, and Security,” 5 Frontiers in Artificial Intelligence 826737 (2022).
[37] OpenAI, “Privacy Policy,” https://openai.com/policies/row-privacy-
policy/; Camilla A. Hrdy, “Keeping ChatGPT a Trade Secret While Selling It Too,” 40 Berkeley Technology Law Journal 75 (2025).
[38] Albert Yu Sun et al., “Does Fine-Tuning GPT-3 with the OpenAI API Leak Personally-Identifiable Information?” arXiv preprint arXiv:2307.16382 (2023); Michael Douglas, “Responsibility of OpenAI for Defamation and Serious Invasions of Privacy by ChatGPT,” 42 Communications Law Bulletin 8 (2023).
[39] ChatGPT is a Generative AI tool that is a subset of AI tools that generates content based on users’ prompts and training data.
[40] Michael Kan, “OpenAI: Sorry, ChatGPT Bug Leaked Payment Info to Other Users,” PCMag (2023).
[41] Shizhe Liang et al., “Mathematics and Machine Creativity: A Survey on Bridging Mathematics with AI,” arXiv preprint arXiv:2412.16543 (2024).
[42] Warren J. Von Eschenbach, “Transparency and the Black Box Problem: Why We Do Not Trust AI,” 34 Philosophy & Technology 1607 (2021).
[43] Ana Fernández Araluce, “AI in International Arbitration: Unveiling the Layers of Promise and Peril,” 49 Iurgium 35 (2024).
[44] Mata v. Avianca, Inc., F. Supp. 3d, 22-cv-1461 (PKC), 2023 WL 4114965 (S.D.N.Y. June 22, 2023).
[45] Mata v. Avianca, Inc., F. Supp. 3d, 22-cv-1461 (PKC), 2023 WL 4114965 (S.D.N.Y. June 22, 2023).
[46] John Lapaglia v. Valve Corporation Case No. Pending 25CV0833 RBMDDL (S.D. Cal. Apr. 8, 2025).
[47] John Lapaglia v. Valve Corporation Case No. Pending 25CV0833 RBMDDL (S.D. Cal. Apr. 8, 2025).
[48] Megan Beck & Barry Libert, “The Rise of AI Makes Emotional Intelligence More Important,” 15 Harvard Business Review 1 (2017).
[49] Herbert B. Dixon, “Artificial Intelligence and Bias,” 63 The Judges’ Journal 37 (2024).
[50] Varun Magesh et al., “Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools,” 22 Journal of Empirical Legal Studies 216 (2025).
[51] Mata v. Avianca, Inc., F. Supp. 3d, 22-cv-1461 (PKC), 2023 WL 4114965 (S.D.N.Y. June 22, 2023).
[52] Kevin Chan, “A New Era of Maritime Arbitration: Ex Machina Determinations,” 40 Journal of International Arbitration 521 (2023).
[53] Datuk Sundra Rajoo, “Importance of Arbitrators’ Ethics and Integrity in Ensuring Quality Arbitrations,” 6 Contemporary Asia Arbitration Journal 329 (2013).
[54] Section 1(b) of the English Arbitration Act 1996, https://www.legislation.gov.uk/ukpga/1996/23/section/1; Article 19(1) of the UNCITRAL Model Law. See also Chand Bagh Foundation v. Messrs Rehman Brothers (Pvt.) Ltd. 2007 CLC 751; Haq Nawaz v. State 2005 YLR 1850; Travel Automation (Pvt.) Ltd. through Managing Director v. Abacus International (Pvt.) Ltd. 2006 CLD 497.
[55] Gizem Halis Kasap, “Can Artificial Intelligence (‘AI’) Replace Human Arbitrators? Technological Concerns and Legal Implications,” Journal of Dispute Resolution 209 (2021).
[56] OPI Gas Limited v. United Energy Pakistan Limited, Suit No.1683 of 2017, Judgment on 10 August 2017; Datuk Sundra Rajoo, “Importance of Arbitrators’ Ethics and Integrity in Ensuring Quality Arbitrations,” 6 Contemporary Asia Arbitration Journal 329 (2013).
[57] Barton and Others v. Morris [2023] UKSC 3; Attorney General of Belize v. Belize Telecom Ltd [2009] UKPC 10; Geys v. Société Générale, London Branch [2013] 1 AC 523; Lister v. Romford Ice and Cold Storage Co Ltd [1957] AC 555; Liverpool City Council v. Irwin [1977] AC 239.
[58] Barton and Others v. Morris [2023] UKSC 3; Marks and Spencer plc v. BNP Paribas Securities Services Trust Company (Jersey) Ltd [2015] UKSC 72; BP Refinery (Westernport) Pty Ltd v. Shire of Hastings (1977) 180 CLR 266; The Moorcock (1889) 14 PD 64; Reigate v. Union Manufacturing Co (Ramsbottom) Ltd [1918] 1 KB 592; Shirlaw v. Southern Foundries (1926) Ltd [1939] 2 KB 206; Luxor (Eastbourne) Ltd v. Cooper [1941] AC 108; Equitable Life Assurance Society v. Hyman [2002] 1 AC 408; The English Court in Equitable Life Assurance Society v. Hyman held that “a term will be implied if it is ‘essential to give effect to the reasonable expectations of the parties’ as diluting the test of necessity, . . . ‘[t]he legal test for the implication of . . . a term is . . . strict necessity. . . .’” [2002] 1 AC 408.
[59] Attorney General of Belize v. Belize Telecom [2009] UKPC 10; BP Refinery Ltd v. President, Councillors and Ratepayors of the Shire of Hastings (1977) 52 ALJR 20.
[60] Article 35 of the UNCITRAL Arbitration Rules 2010, https://uncitral
.un.org/sites/uncitral.un.org/files/media-documents/uncitral/en/arb-rules-revised-2010-e.pdf; Article 21(3) of the ICC Arbitration Rules 2021.
[61] Andrew Perlman, “The Implications of ChatGPT for Legal Services and Society,” 30 Michigan Technology Law Review 1 (2023).
[62] Article 31(2) of the UNCITRAL Model Law.
[63] Article 48(3) of the ICSID Convention.
[64] Jennifer Kirby, “What Is an Award, Anyway?,” 31 Journal of International Arbitration 475 (2014); Martin Platte, “An Arbitrator’s Duty to Render Enforceable Awards,” 20 Journal of International Arbitration 307 (2003).
[65] Phillip Landolt, “Arbitrators’ Initiatives to Obtain Factual and Legal Evidence,” 28 Arbitration International 173 (2012); Daniel Ben-Ari et al., “‘Danger, Will Robinson’? Artificial Intelligence in the Practice of Law: An Analysis and Proof of Concept Experiment,” 23 Richmond Journal of Law and Technology 1 (2016).
[66] Paul Bennett Marrow, Mansi Karol & Steven Kuyan, “Artificial Intelligence and Arbitration: The Computer as an Arbitrator—Are We There Yet?,” 74 Dispute Resolution Journal, 35 (2020).
[67] Andrew Perlman, “The Implications of ChatGPT for Legal Services and Society,” 30 Michigan Technology Law Review 1 (2023); see generally Paul Bennett Marrow, Mansi Karol & Steven Kuyan, “Artificial Intelligence and Arbitration: The Computer as an Arbitrator—Are We There Yet?,” 74 Dispute Resolution Journal 35 (2020).