AI in International Disputes Practice

Breaking Ground or Old Hat?

Welcome again Elizabeth Chan and Eliza Jiang as co-authors for this newsletter issue! If you don’t already know them, you should. Both are innovators and thought leaders in the law and tech spaces.

This image was created for this newsletter by Elizabeth using a Generative AI tool

Over the past year, generative artificial intelligence (Generative AI) in the form of large language models (LLMs) like ChatGPT-4 has taken the world by storm. Legal practice is no exception. Many will recall the headline-making story of a New York lawyer who was sanctioned by a judge for relying upon non-existent case law precedent that he obtained from ChatGPT-4 and did not double-check, as well as the Texas federal court judge who has implemented a standing order that requires all litigants appearing in his courtroom to make a certification concerning the use of Generative AI in their submissions.  

Yet, international arbitration practitioners have relied upon tools powered by other forms of artificial intelligence (AI) for many years. Indeed, in an era marked by Big Data and an increasingly complex dispute resolution ecosystem encompassing broad document disclosure and evidentiary collection, the work of arbitration practitioners would be impossible to manage without AI. 

So, at a moment dominated by sexy headlines about the risks and opportunities presented by Generative AI in legal processes, this newsletter edition takes a step back to examine the “old hat” AI-powered tools and applications that have long supported international arbitration practitioners with document management, document review, document production, and arbitrator due diligence, among other tasks, and concludes by introducing how Generative AI-powered tools may supplement and enhance the lawyer’s toolkit.  

Managing and Producing Documents in the Age of Big Data

Document review and production are often necessary evils in international arbitration practice. These processes are perceived as time-intensive and low-value grunt work. Moreover, they are often a sore point in the outside counsel and client’s relationship because the client may be unwilling to pay the usual rates for time spent on such work. This pressure point creates an opportunity to implement AI-driven document review tools to reduce the time and costs associated with labour-intensive document review and analysis. E-discovery platforms such as Relativity, Luminance, EverLaw, and CS Disco employ machine learning algorithms to categorise, extract, and analyse information from vast quantities of documents. 

Each platform has AI-driven functionalities that enable users to swiftly identify pertinent documents through conceptual search (in addition to keyword search), data visualisation, and document clustering, which expedites the overall document review process. 

Conceptual search in the context of e-discovery is a powerful and innovative approach to information retrieval that goes beyond traditional keyword-based searches. Unlike keyword searches, which rely on exact word matches, conceptual search utilises artificial intelligence and natural language processing to understand the underlying concepts and context within documents. Conceptual search is, therefore, useful in e-discovery as it enables legal professionals to uncover relevant documents even when specific keywords or phrases may not have been used, thus reducing the risk of missing critical information. 

One of the biggest challenges of document review and production is ensuring that the choices made are consistently reflected across all similar and duplicate iterations of that document. Here, AI-driven functionalities can drive efficiencies and consistency, thereby assuaging lawyers’ concerns over the risks of inconsistent instructions or understandings, or plain old human error. 

Data visualisation of related documents presents document relationships, patterns, and key information graphically and offers a visual narrative of the document corpus. This allows legal teams to quickly grasp the structure of the data, identify important trends, and pinpoint critical documents. Moreover, it aids in developing effective case strategies by revealing patterns and connections that may not be immediately apparent through traditional text-based analysis. 

These e-discovery platforms also cluster related documents by content or themes, enabling the review of groups of documents relevant to an issue or fact in the arbitration. Clustering also helps identify patterns, trends, or commonalities within a document corpus, which can be crucial for building a coherent legal strategy or uncovering hidden insights. Another benefit of clustering is ensuring that strategic decisions are uniformly and consistently applied to the overall document set, including decisions on the treatment of privileged, confidential, or non-responsive information.

These tools enhance the speed, accuracy, and comprehensiveness of document review for production and disclosure, making them efficient at managing large volumes of electronic data in complex international arbitrations. As experienced counsel will know, these tools enable international arbitrations to unfold on faster and more efficient timetables than would be possible if their legal teams were stuck reviewing, categorising, and producing hard copy documents from within storage boxes or paper filing cabinets. 

End-to-end ODR Platforms

A key feature of end-to-end online dispute resolution (ODR) platforms like New Era ADR is the ability to automate and streamline dispute resolution. AI algorithms facilitate the intake of cases, intelligently categorise them, and allocate resources efficiently. Parties can initiate arbitration proceedings seamlessly, guided by AI-driven prompts and tools, simplifying the complex legal processes associated with dispute resolution. This level of automation reduces administrative burdens and ensures that cases progress smoothly, saving both time and resources. These features may also democratise the dispute resolution process, making it more accessible for self-represented parties who may not have in-house legal expertise. 

Parties can track the progress of their cases in real-time, access relevant documents, and receive notifications through user-friendly interfaces. AI technologies underpin these features, ensuring that parties are well-informed about the status of their arbitration proceedings. This transparency fosters trust and confidence in the process, ultimately contributing to more equitable and satisfactory outcomes.

Platforms like TrialView offer an AI-powered litigation workspace, with smart tools for legal teams to navigate a dispute's full lifecycle. TrialView provides a centralised platform for uploading, managing and interrogating case data, smart bundling tools that permit a user to create a bundle in seconds (with automatic pagination, tabbing, and indexing). Hyperlinking and cross-referencing tools also run in tandem with in-built court compliance checks, and late insertion features offer greater flexibility. Each of these features can save a lawyer dozens of hours and prevent some of the last-minute stress (and overtime paralegal costs) that crop up on the eve of a hearing or trial. 

Meanwhile, TrialView permits the same data to be examined and interrogated using AI intelligent search. A user can ask a question, and AI-powered tools will not only find the answer, but will direct the user to the exact excerpt, paragraph, and document. Entity search tools allow a user to find connections between key dates, actors, and events, with timeline building offering further insights. This capacity to really get to know the evidence is incredibly potent, removing the need to manually trawl through paper files for specific facts and data. Aside from the time and cost savings, these tools allow legal teams to spend more time doing ‘real’ legal work - including focus on strategy, corroboration, and persuasion, as well as allowing time to consider counter arguments and counter approaches. 

Another example of AI in action is TrialView’s witness statement preparation and deposition creation tool, which allows users to record the interview, generate an automatic transcript, and convert it to a hyperlinked witness statement. 

Finally, on the presentation side of things, TrialView offers smart evidence presentation tools that allow all parties to follow in real-time, with annotation and highlighting tools available to mark up documents and the hearing/trial transcript as the evidence unfolds. Parties may also follow remotely using an integrated video platform and team members in different locations can use built-in AI tools to discern potential material that can corroborate or undermine the propositions being advanced in the hearing/trial room itself with great speed.

AI-powered Legal Research Platforms

AI-powered legal research platforms have become indispensable tools in international arbitration. They provide arbitration lawyers with the technology to search for relevant precedents, jurisprudence, and legal sources across multiple jurisdictions. 

Platforms like Kluwer Arbitration and Jus Mundi harness machine learning to compile and enrich their extensive databases of international arbitration cases, treaties, conventions, and related legal documents. AI power offers great benefits to researchers as it underlies the technology that prepares case citations and helps create cross-links and references between various content sets. 

CaseText also employs an AI-driven approach, leveraging natural language processing to extract valuable insights from legal texts. Its platform also offers the capability to analyse and summarise legal research results and draft legal memoranda. 

LexisNexis and Westlaw, established names in the legal research landscape, also integrate AI into their platforms to enhance research capabilities by providing predictive analytics to suggest relevant cases, statutes, and secondary sources based on user queries. 

Interestingly, with the rise of Generative AI, many of these legal research platforms have unveiled new Generative AI-powered LLM chatbots to enable the legal researcher to engage in a question-and-answer exchange to facilitate the legal research journey. Such tools supercharge the legal researcher’s user journey and experience. They are especially valuable to both newer legal researchers (such as students) or those who aim to understand a new area of law very quickly. However, one must not forget that the data and data enrichment that legal researchers encounter have benefitted from AI-driven tools for many years already. 

AI-enabled Machine Translations

International arbitration is characterised by its inherently cross-border nature. Parties frequently come from different cultural and linguistic backgrounds, and evidence and testimony may be in multiple languages. Despite this added complexity, the usual goal of effective communication and understanding of legal content is paramount to achieving equitable dispute resolution. In this context, AI-powered machine translation has emerged as a game-changing technology, offering advanced linguistic capabilities that transcend traditional language barriers. 

DeepL, renowned for its neural machine translation technology, and Google Translate, a widely accessible and versatile translation service, represent two exemplary platforms that leverage AI and deep learning techniques to deliver precise and context-aware translations of legal documents and communications translations. 

DeepL, Google Translate, and similar AI-driven translation platforms are indispensable assets in international arbitration. They ensure that all participants can effectively engage in the process regardless of language differences, promoting fairness and impartiality in dispute resolution.

Again, such AI-powered tools are not new to international arbitration practice but have become commonplace in light of the frequency with which parties and their counsel must operate across languages and linguistic barriers. 

Conflict Management and Arbitrator Diligence

International arbitration often involves complex disputes with multinational parties and necessitates a rigorous approach to arbitrator selection and conflict management. Various services have emerged to offer arbitrator profile and conflict-checking tools, including Arbitrator Intelligence, Kluwer Arbitration’s Profile Navigator & Relationship Indicator, and Global Arbitration Review’s Arbitrator Research Tool (ART). 

Each tool uses a combination of AI, data analytics, and self-reported information to provide comprehensive insights into arbitrators’ performance and track records of arbitrators. This data helps arbitration practitioners identify suitable candidates for arbitrators based on empirical data rather than subjective assessments. Further, engaging in this process may add diversity to the prospective arbitrator pool by bringing additional candidates to the practitioner’s attention who meet the case criteria but who may otherwise not be known by the selecting counsel. Overall, these tools provide data-driven inputs to the arbitrator selection process and promote transparency and objectivity in the arbitrator selection process. 

On the other hand, these tools may also introduce subjective insights. In some instances, the data collected may include candid feedback from counsel who have appeared before those arbitrators in particular cases. While these insights would be of a different nature from objective data-driven insights, they may prove equally useful in helping practitioners identify arbitrators who meet their needs and who may otherwise have been unknown to them. 

AI-driven Data Analytics for Third-Party Disputes Funding

Predictive data analytics has become a powerful tool in litigation finance and third-party funding, transforming legal professionals’ strategic decisions. Notable platforms in this space include Lex Machina and Arbilex. 

By aggregating all data from documents filed on court dockets and leveraging AI along with human legal expert review to structure the data, Lex Machina provides insights like the quantum of damages, potential case resolutions, opposing counsel’s litigation history, and timing of the proceedings to enable predictions on various aspects of their cases.

Similarly, Arbilex employs machine learning algorithms to analyse historical case data, legal precedents, and financial metrics. This enables third-party funders to assess the potential risks and rewards of funding a particular case. Burford Capital, a third-party funder, enhances its legal finance modelling with AI. However, it acknowledges the limitations of AI due to the confidentiality of about 90% of commercial disputes, which are resolved through settlement. Despite this, Burford considers that integrating AI can improve assessment accuracy and speed by analysing factors like profitability and outcome likelihoods. However, the effectiveness of AI models depends on the quality of available data, highlighting the difficulty of relying solely on AI tools in legal finance.

These and similar solutions enhance decision-making within international arbitration, serving as a valuable resource for legal teams dealing with cross-border disputes.  These tools analyse historical case data to provide insights into settlement probabilities and potential third-party funding opportunities and enable arbitration practitioners to make informed decisions and negotiate settlements more effectively.

Burford Capital also uses AI to originate new business and identify potential cases by enhancing the process of case identification for investment. They leverage AI to combine public data with insights from past successful investments, employing heuristics and prompting techniques. This approach helps Burford to identify lawyers and cases that meet their investment criteria and discover instances where businesses have suffered harm but are unaware of their strong claims. Specifically, Burford has initiated projects to scrape the web for lawyers with specific profiles related to successful case types, thereby streamlining the process of finding new investment opportunities and assisting businesses in recognizing valuable claims.

This image was created for this newsletter by Elizabeth using a Generative AI tool

Is There Still Room to Break Ground?

Notwithstanding the number of AI-driven tools and providers that dispute resolution pracititoners are already familiar with and frequently use, there is still ample opportunity, within these same arenas, for innovation driven by Generative AI-powered technology. Eliza, a coauthor of this post, is also the Co-Founder of Lawdify, a solution that creates intelligent systems (AI agents) to run specialised, laborious, and high-stake tasks for legal professionals. She believes that the new generation of AI legal technology platforms like Lawdify will leverage the capability of the LLMs as the “superbrain” to process semantics and to contextualise and connect relevant concepts within legal documents so that a layer of true intelligence is added to the voluminous documentary records that disputes lawyers need to manage, navigate, and learn. These techniques will enable disputes lawyers to quickly generate work products like chronologies, dramatis personae, lists of issues, lists of relevant facts, and cross reference each to the underlying evidentiary documents in seconds. Lawyers will be able to promptly retrieve a document based on conceptual and semantic search (without the need to anticipate specific keywords) with Natural Language Processing. They could also “pivot” the underlying evidentiary record (just like a pivot table in Excel) according to their needs based on a variety of parameters, for example, display the list of documents supportive of a particular fact that bolters the arguments of the claimant on a specific legal issue and sorted chronologically. Indeed, LLMs are proficient at doing this.

The next frontier of AI-powered legal technology solutions could create AI agents that understand the objective of a task and are capable of autonomously running them end-to-end, such as tagging for privilege for responsive documents in a production exercise, creating a privilege log, retrieving relevant and material documents in response to production requests. 

With respect to reliability and accuracy, AI-first legal technology companies like Lawdify will adopt techniques borrowed from data science to provide high scores in answer faithfulness (answers based on a real and not made-up fact in the corpus), and answer relevance. They will set up guardrails like providing sources of underlying documents, creating a record of the reasoning behind each action taken by an AI agent and of the considerations leading to such action (like generating a list of documents that were not responsive to a specific request), implementing evaluation stacks to benchmark the answers against ones provided by human lawyers, and a rigorous user feedback loop to collect and monitor user comments and actions.

Concluding Thoughts 

AI-powered tools are far from new in international arbitration practice. Indeed, for many years, the robust practice and procedural approaches that have become commonplace have been driven by the efficiencies and opportunities that AI-powered tools have enabled. 

Yet, Generative AI remains poised to profoundly transform how international arbitration work is performed and offered. Its potential is vast, offering enhanced capabilities in legal research, document review automation, and even altering traditional billing models. The promise of AI-driven end-to-end online dispute resolution platforms could revolutionise how arbitration is approached, particularly for low-value commercial disputes. It has the potential to rebalance how arbitration is practised, opening doors to democratising access to the arbitral process for users who may be self-represented and without in-house legal expertise. 

In all events, just like the “old hat” AI-powered tools that have facilitated international arbitration as it exists today, lawyers must be among the early adopters of Generative AI-powered tools, finding ways to test and hone new skill sets to enhance their practices and offer greater value to their clients. 

Post Script

• If you are interested in more from the three of us, we collaborated on a prior edition of this newsletter and also recently published a long-form article in the Contemporary Asia Arbitration Journal, “Harnessing Artificial Intelligence in International Arbitration Practice.”

• TrialView commissioned the article featured in this newsletter, which was first published on their website in April 2024. Thank you, TrialView!

The opinions and insights presented in this newsletter solely represent the authors’ views. They are not endorsed by or reflective of the policies or positions of their affiliated firms or organisations.

Like What You See?

Subscribe for free to receive future newsletters & insights.

You can also subscribe to the parallel newsletter on LinkedIn.