Payel Chatterjee
Trilegal, Mumbai
payel.chatterjee@trilegal.com 

Aman Singhania
Trilegal, New Delhi
aman.singhania@trilegal.com 

Yuvraj Singh Sharma 
Trilegal, New Delhi
yuvraj.sharma@trilegal.com

Disputes are becoming increasingly international, complex and time sensitive. Both private and state parties alike across the globe have resorted to settlement of disputes through arbitrations, heavily relying on technology – due to its manifold benefits of privacy, fairness, cost control, and sectoral expertise. Introduction of simple technologies like e-mail, digital data maintenance, online case research and management, have transformed not only inter-personal relationships, but also our interaction with data in the legal context. Arbitration procedures devised under the framework of information technology are not just an impulsive creation, but rather a logical next step towards technology aiding the practice of law. 

It is a common cliché that lawyers are not entirely at home with technology and digitalisation. When the world was sprinting with the 21st century technological revolution, much of the legal industry was seen crawling or ducking underneath. Legal-tech companies and incubators have introduced innovative data repository, transfer and search engine tools to the market. However, to this day, technology in the legal context continues to be misunderstood, misused, and mistrusted, with the only limited acceptance in the field of cyberspace, tech security, intellectual property, and the like. 

Technology in arbitration: rise of machine learning 

The idea of integrating technology into adjudicatory processes is not as distant as it seemed a decade earlier. The cascading effect of Covid-19 has fortified the demand for mechanisms to deliver speedy yet cost effective dispute resolution. Legal professionals across the globe have been forced to rethink their relationship with technology and welcome it in their sometimes-obsolete chambers. Online dispute resolution (ODR) has emerged as the pivot point of the relationship between IT and law during the pandemic, and several jurisdictions have mandated court proceedings to be conducted through ODR mechanisms.

Empirical data from the international arbitration community hints towards improving arbitral efficiency through adoption of new technologies like cloud-storage, e-disclosure, video conferencing, and use of online platforms and technologies. The ICC Commission on Arbitration and ADR’s updated Report on Information Technology in International Arbitration (2022) reveals that 93 per cent of interviewees believe that IT has revolutionised arbitrations by streamlining processes and reducing burgeoning costs of arbitral processes.[1] The survey has also indicated that use of IT is increasing and leads to improvement in efficiency of arbitral proceedings.[2] In fact, a big shift is being witnessed among practitioners, arbitrators, and institutions such as the International Court of Arbitration[3] and London Court of International Arbitration, to adopt tech-tools to improve the speed at which proceedings are conducted.[4] 

These steps are somewhat consistent improvements over time, which are now well rooted and known amongst most of the arbitration profession across the globe. However, from the ashes of Covid-19, a path-breaking technology is now poised to yield a heavy influence in the coming decades and create genuine innovations in international arbitration in the form of artificial intelligence (AI). AI or machine language that can develop its own ‘thinking patterns’ is a crucial aspect in revolutionising effectiveness of arbitrations. At the same time, we have recently witnessed law firms being fined for reliance on fake opinions and precedents created by AI, doing away with their gatekeeping role. In one such instance the court acknowledged, however, that there was nothing 'inherently improper' about using AI as a research tool. 

Artificial intelligence – the new player in arbitration

Since 2017, the ICC has administered more than 500 cases under its Expedited Procedure Provisions without any significant challenges to date relating to a purported lack of ‘due process’.[5]  Given parties’ inclination in opting for such faster and predictable procedures in higher-value disputes, the stage is set for AI to take design control of arbitral frameworks. Arbitration specialists have begun contemplating potential uses of AI and are increasingly utilising it in the field of arbitration to streamline processes, enhance decision-making, and improve efficiency in the following manner:

• Document review and e-discovery: AI tools can review and categorise vast amounts of documents and data quickly, making it easier to identify relevant evidence, reducing the time and cost associated with document review and e-discovery.[6]
• Predictive analytics: AI algorithms can analyse historical arbitration cases and predict likely outcomes or settlements based on various factors, helping parties make more informed decisions about pursuing arbitration or negotiation.
• Case management: AI-powered case management systems help arbitration providers and practitioners manage their cases more efficiently, automating scheduling, tracking deadlines, and managing case-related documents.
• Online dispute resolution (ODR): AI-driven ODR platforms provide a digital environment for resolving disputes online. These platforms can automate communication, facilitate negotiations, and even provide preliminary decisions.
• Legal research and precedent analysis: AI-powered legal research tools can help arbitration practitioners find relevant case law and precedents quickly, saving time and improving the quality of legal arguments.
• Compliance and due diligence: AI can assist in assessing the compliance of arbitration proceedings with applicable laws and regulations. It can also aid in due diligence processes by identifying potential conflicts of interest.
• Data security and privacy: AI can enhance the security of sensitive information used in arbitration by detecting and preventing data breaches, unauthorized access, and ensuring compliance with data protection regulations.
• Decision support: While AI is not typically used to make final arbitration decisions, it can provide decision support by analysing evidence, identifying patterns, and presenting relevant information to arbitrators to aid in their decision-making.[7]
• Natural language processing (NLP): NLP technologies enable AI to assist in summarising legal documents, creating transcripts, and improving the quality of written communication.
• Virtual arbitrators and mediators: Some AI systems have been developed to act as virtual arbitrators or mediators, providing an automated, neutral party to facilitate the resolution process. While these systems are not intended to replace human arbitrators, they can assist in preliminary negotiations and fact-finding.

AI guidelines – the road to formalization 

Some jurisdictions have begun adopting AI techniques in different stages of criminal adjudication including in trial risk assessment, bail, sentencing and parole decisions. E-commerce players like PayPal and eBay have adopted AI to resolve millions of customer disputes through algorithm driven ODR.

Several arbitral institutions have already started contemplating incorporation of AI guidelines in their institutional rules governing arbitration. On 31 August 2023, the Silicon Valley Arbitration and Mediation Centre (SVAMC) led the way in publishing draft guidelines on the use of AI in arbitration (AI Guidelines).[8] These AI Guidelines are divided into three chapters, which are respectively addressed to: (i) participants in arbitration; (ii) parties and party representatives; and (iii) arbitrators. The AI Guidelines contain proposals on: (i) the uses, limitations, and potential risks of AI applications; (ii) consent and disclosure norms; (iii) safeguarding confidentiality; (iv) duty of maintaining diligence in use of AI; (v) non-delegation of decision-making responsibilities; (vi) respect for integrity of proceedings and evidence; and (vii) respect for due process and compliance with procedural issues.

The Ministry of Justice of New Zealand on 8 September 2023 has also come out with the draft Best Practice Guidelines for use of Generative Artificial Intelligence (AI) in Courts and Tribunals.[9] These guidelines are intended to assist participants to (i) navigate potential applications of AI; and (ii) perform informed and responsible usage of AI in arbitration. The guidelines are proof of concept of the increasing role of AI in arbitration and serve as a useful reference point for parties attempting to safely leverage AI capabilities.

Key challenges posed by artificial intelligence

Need for ethical guidelines and regulations and upholding confidentiality

Despite its many advantages, AI also poses several challenges. Experts have consistently raised concerns of AI’s inherent bias based on the historical arbitration data used to train the AI. The AI system may perpetuate or even worsen those biases, leading to unfair outcomes. Ensuring fairness and transparency in AI algorithms is a critical challenge along with ensuring compliance with privacy and data protection norms. 
The use of AI in arbitration also raises ethical questions, such as whether it is appropriate for certain decisions to be made by algorithms without human oversight. Parties need to ensure transparency is maintained from a procedural standpoint. Thus, as a pre-requisite, ethical guidelines for development and use of AI in arbitration need to be established. 

It is trite that AI will only have limited legal applicability and acceptability unless cohesive regulations to address AI-specific issues and usage in arbitration are adopted.

Need for transparency and human element

Secondly, many AI algorithms operate as 'black boxes', meaning that their decision-making processes are not easily understandable by humans. Lack of transparency in AI models used in arbitration can raise concerns about accountability, especially when parties need to understand the basis for decisions. AI would also struggle to understand the broader context of a dispute, including cultural nuances, emotions, and other factors that can significantly impact the resolution process. Arbitrators often bring a level of intuition and contextual understanding that AI may lack. Further, given the significant upfront costs associated with the implementation of AI systems, it also faces severe resistance from traditionalists who are accustomed to conventional dispute resolution methods. Overcoming this resistance and ensuring the acceptance of AI technologies is a significant challenge.

Future of AI in arbitration: where are we headed?

Addressing challenges posed by AI requires careful consideration, collaboration between technology developers and legal professionals, and ongoing efforts to update regulations and ethical guidelines. However, use of technology and AI need not be looked at with an often-heard cynicism of 'man v machines'. 

Developments in the area are not restricted to the western world. For example, the Indian Supreme Court has taken advantage of AI's potential by developing the SUVAS (Supreme Court Vidhik Anuvaad Software), which converts judicial papers from English into nine regional languages.

Striking the right balance between the advantages of AI and the protection of human-centric values is crucial for the successful integration of AI in arbitration. Ethical collaborations between AI and lawyers are only likely to revolutionise the legal sector, improve timelines, efficiency and drive down costs. With humans relying on technology for their most critical elements of life, there seems no plausible excuse as to why, simpler matters of adjudication of disputes cannot come to rely on technology and AI in a new paradigm shift. 

---

[1] https://iccwbo.org/news-publications/arbitration-adr-rules-and-tools/information-technology-international-arbitration-report-icc-commission-arbitration-adr/.
[2] 2018 International Arbitration Survey: The Evolution of International Arbitration, QUEEN’S MARY UNIVERSITY OF LONDON 32 (2018), http://www.arbitration.qmul.ac.uk/media/arbitration/docs/2018- International-Arbitration-Survey-report.pdf.
[3] ICC Arbitration Rules, 2021, ICC (Jan. 1, 2021), https://iccwbo.org/dispute-resolutionservices/arbitration/rules-of-arbitration/
[4] LCIA Arbitration Rules, 2020, LONDON COURT OF INTERNATIONAL ARBITRATION (Oct. 1, 2020), https://www.lcia.org/Dispute_Resolution_Services/lcia-arbitration-rules-2020.aspx.
[5] https://www.law.com/legaltechnews/2023/09/28/the-new-frontier-ai-in-arbitral-decision-making/?slreturn=20231009234802
[6] Gizem Halis Kasap, Can Artificial Intelligence (Al")" Replace Human Arbitrators? Technological Concerns and Legal Implications' (2021) 2021 SSRN 209, 213.
[7] Tania Sourdin, 'Judge V Robot? Artificial Intelligence and Judicial Decision-Making' (2018) 41 UNSW LJ 1114, 1121.
[8] https://thearbitration.org/ai-guidelines/ai-guidelines-1/ 
[9] https://www.courtsofnz.govt.nz/assets/4-About-the-judiciary/judicial-committees/judicial-committees/Artificial-Intelligence-AI-Advisory-Group/2023-09-08-Draft-Generative-AI-Guidelines-Non-Lawyers-Final-draft-for-consultation.pdf.