Over the past year, the TransPerfect team has attended numerous arbitration events, including International Disputes Weeks in Dublin and London to Arbitration Weeks in Perth, Paris, Toronto, and Hamburg. Across countless panels covering strategic, legal, and academic topics, one subject has consistently sparked the most debate: Artificial Intelligence (AI).

As AI continues to transform industries, it has firmly planted itself at the forefront of discussions in the arbitration community. Having engaged with practitioners throughout the year, I’ve gathered valuable insights—particularly on generative AI (GenAI)—that are highly relevant in the arbitration space.

To explore this further, I had the opportunity to sit down with TransPerfect Legal Vice President Al-Karim Makhani, who brings a unique perspective after a decade as a disputes lawyer in England, the Middle East, and Hong Kong, and another ten years advising on technology. His insights highlight how AI is transforming the practice of law and arbitration in ways we couldn’t have imagined even a few years ago.

A recurring theme throughout these conversations has been the role of knowledge. As Kofi Annan famously said, "Knowledge is power. Information is liberating. Education is the premise of progress, in every society, in every family." In the context of AI, this means understanding its capabilities and limitations in order to harness its potential effectively in legal practice.

Change is the only constant, and AI is both a present reality and an inevitable part of our future. It’s a topic everyone discusses, but what exactly is AI? One of the key lessons I took from law school was the importance of simplifying complex concepts. True understanding means being able to explain something in a way even a six-year-old could grasp:

1. For a child – the ability for a computer to think and learn. With AI, computers can perform tasks typically done by people.
2. From IBM – AI is technology that enables computers and machines to simulate human intelligence and problem-solving capabilities. On its own or combined with other technologies (e.g., sensors, geolocation, robotics) AI can perform tasks that would otherwise require human intelligence or intervention.
3. From the recently implemented EU AI Act – “AI system” means a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments.

Lawyers have, in actual fact, been using AI for many years without batting an eyelid—let alone making disclosures every time it’s used.

1. Automating the Mundane
   • Email Rules: Think of how "If this then that" rules tidy up your inbox or sift out spam. It’s AI working quietly in the background.
   • Office Software: From organising data in Excel using complex formulas to the much-applauded CTRL+F for finding text, AI simplifies these tasks.
   • Case Searches: Remember the last time you searched for a case in an online database? AI was probably guiding your search parameters to the right results.
   • Voice Assistants: Tools like Siri, Alexa, and Google Assistant are so embedded in our routines that we often overlook the AI powering them. Lawyers frequently use voice commands to set reminders, schedule meetings, and search for information, all while benefiting from AI's ability to understand and respond to natural language.
2. Streamlining Document Review
   • Technology-Assisted Review (TAR): TAR is a form of AI based on machine learning or predicative coding. Essentially, TAR is an algorithm that can review and mark mass amounts of documentation for relevance in place of humans, as its been trained by an expert lawyer to replicate the decision making of that expert lawyer.
   • Time Saving: TAR has drastically reduced the amount of time a human spends on manually reviewing electronic data and documents. In average matters this tracks between 40% and 60%. However, the ceiling for larger, richer data sets can rise above 95%.
   • Sentiment Analysis: In addition to traditional document review, AI tools like sentiment analysis can detect tone or emotional nuances in arbitration-related communications, helping lawyers identify key documents that may sway a tribunal’s decision.

As AI continues to evolve, its potential within our field grows exponentially. But what exactly is GenAI, and how does it differ from what we're used to?

GenAI is a type of machine learning that can be used to create brand new content. This can include audio, text, code, images, and videos. Examples of platforms that utilise this include ChatGPT, Claude, Gemini, and Copilot.

This is made possible by large language models (LLMs), which are algorithms that have learnt how to predict the next word based on previous words in a sequence (the language pattern) and generate relevant and new conversational text in relation to a prompt or query. The potential for legal services is both exciting and terrifying in equal measure:

1. Drafting Documents: AI can draft basic legal clauses, modify them for balance or bias, and assist in due diligence. In arbitration, this could extend to drafting key documents such as pleadings, submissions, and procedural letters.
   • AI Assistants: Virtual assistants with natural language processing (NLP) capabilities can suggest amendments, sense-check human edits, and ensure consistency across legal documents, including arbitration agreements and terms of reference. Paired with playbooks and clause banks, these tools can help build an initial draft of pleadings, procedural orders, or correspondence, streamlining the document creation process in arbitration cases.
2. Document Handling: GenAI has vast applications and ability to independently carry out document-related tasks. 
   • Summarisation: Generating clear summaries of documents, allowing you to quickly understand the sentiment of your dataset.
   • Q&A: Being able to ask questions, and receive AI-generated responses alongside document references for human verification.
   • Chronologies: AI can help to contextualise how case documents fit together in a collaborative timeline, by identifying, categorising, and placing the evidence sequentially.
3. Legal Research: Using AI tools designed for cases, legal articles, research papers, textbooks, and other relevant materials can multiply the reach of a research analyst and be a useful starting point when searching for and structuring a legal query.
   • Legal Search Engine: NLPs can be used to retrieve relevant case law, statutes, and regulations. This could expand to answer simple legal questions but also as a legal search engine that can quickly search for relevant legal provisions and case law.
4. Natural Language Processing (NLP): Legal documents and communication are filled with complex language. NLP can analyse legal text, understand its meaning, and even generate summaries or translate legalese into plain English for clients. Using the below clause as an example:
   • Prompt: Translate this clause drafted in complex legalese into plain language for a non-legal expert
     
     "To the extent permitted by applicable law, [Website name] shall not be liable for any direct, indirect, incidental, special, or consequential damages arising out of or in connection with the use of the website or its content. [Website name]'s total liability for all claims arising out of or relating to the website or its content shall not exceed the amount paid by you, if any, for accessing the website."
   • AI Response: This clause says that the website isn't responsible for any harm or loss you might experience from using it or its content. If there are any problems, the most you could be compensated is the amount you paid to use the website, if you paid anything at all.
5. Predictive Analysis: AI can learn from vast amounts of legal data, identifying patterns and trends that humans might miss. This can lead to improved legal research, risk assessment, and even the prediction of future legal outcomes (though not replacing lawyer judgment!).
   • Arbitrator Analysis: AI algorithms can synthesise an international database of arbitrator information that provides details on gender, nationality, experience, and applicable laws, as well as a list of arbitrators who have competed against the arbitrator under research.
   • Case Strategy: AI can be used as a statistics-based prediction tool to determine which of its case prediction models is most suitable for predicting a given case based on past data.
6. Witness and Expert Evaluation: Through technologies like AI-powered polygraphs, AI is enhancing the reliability of arbitration by analysing truthfulness more accurately than traditional methods.
   • Polygraph Technology: AI-powered polygraph systems can analyse micro-expressions, speech patterns, and physiological signals to determine truthfulness and reliability.
   • Real-Time Transcript Analysis: AI can analyse live testimony against prior statements to identify inconsistencies in real time, helping counsel spot contradictions during cross-examinations.
   • Document Cross-Checking: AI can instantly compare witness testimony to relevant documents or emails, providing immediate feedback on any discrepancies between testimony and documentary evidence.
7. Advanced TAR: Building on traditional TAR workflows, GenAI offers the potential for enhanced document review capabilities in arbitration. Rather than merely classifying documents, AI will be able to provide explanations for its coding decisions, which is crucial for maintaining transparency in arbitration.
   • Simultaneous Issue Analysis: AI can run analyses across multiple issues at once, ensuring documents are accurately tagged for various aspects of a dispute—whether it's liability, damages, or procedural matters.
   • GenAI Review Models: Emerging GenAI applications will require far less manual input than earlier TAR models. These applications could be categorised into three potential workflows:
     • GenAI Assisted Review: GenAI is trained using a review protocol and a set of pre-reviewed documents by human lawyers to ensure it understands the parameters of the review.
     • GenAI Iterative Assisted Review: Starting with an assisted review, this method incorporates continuous feedback from human reviewers, who validate the AI's results and provide input to further train the system as the review progresses.
     • GenAI Autonomous Review: This model relies heavily on AI with minimal human intervention, conducting its own quality control based on the initial review protocol. It Careful oversight is necessary to avoid risks.

In conclusion, the realisation is clear: AI is not just an auxiliary tool but a fundamental component driving modern legal processes. The legal community, traditionally cautious and measured, now stands on the brink of a technological renaissance.

Returning to Kofi Annan's words, "Knowledge is power. Information is liberating. Education is the premise of progress," we must consider the knowledge AI brings not as a challenge to our expertise but as a catalyst for our growth and adaptation. Embracing AI means moving with the constant of change, preparing ourselves for a future where our analytical skills are complemented by the deep insights and rapidly processing capabilities of AI.

In arbitration, the journey with AI is just beginning. Let us proceed with curiosity, diligence, and an open mind, ready to harness the full potential of AI in our pursuit of justice and efficiency. By embracing these new tools—whether for drafting pleadings, reviewing evidence, or predicting outcomes—we equip ourselves for a future where technology enhances the speed and accuracy of our work.