Integrating advanced technologies into the practice of international arbitration

The potential of new technologies such as AI to transform the legal field is immense, promising greater efficiency, accessibility, and transparency. It is already having a dramatic impact on how we deliver dispute resolution and prevention services—and will continue to do so at a pace that is almost beyond imagination. Yet, alongside the opportunities that new technologies offer come significant risks. A single misstep in the age of social media and tech-driven legal practice—appearing online as a cat while presenting oral arguments to a judge, as one U.S. lawyer did, or forgetting to mute a toilet flushing during a U.S. Supreme Court videoconference—can turn legal professionals into a laughing stock and the subjects of viral “memes.”

Even more serious issues will arise as AI and other technologies become part of our daily work. Consider the U.S. lawyer in a personal injury case, who unwittingly made multiple submissions to a judge and the opposing counsel what later turned out to be entirely fabricated by ChatGPT. This mistake—due to an AI-induced error known as “hallucination”—was not only embarrassing but also resulted in serious sanctions against the lawyer.

The crucial lesson here is not to fear AI or create anxiety about such technologies taking over and making lawyers and arbitrators irrelevant. I believe AI will replace those lawyers who don’t understand and master the technology. Luckily, there are precautions to keep in mind that will help legal professionals harness the growing power of AI and reap its benefits without becoming the next meme, endangering client relationships, or violating ethical standards.

Data security and client confidentiality

First, it is imperative to safeguard confidential information. Confidentiality is one of the most attractive benefits of arbitration, but it can be easily compromised if information is entered in ChatGPT or on unstable or insecure technology that does not guarantee the protection of private information and data. While security is improving, platforms can still be problematic because AI models continuously “learn” from new data shared with them and also remember previously processed information. Lawyers who use AI will need to know how these systems work, determine whether they are proprietary, and if they can be trusted to maintain the confidentiality of client data. In the global practice of arbitration, lawyers must also be fully aware of data protection laws and constraints at the place of enforcement.

The human touch

Equally important is to not rely solely on AI to prepare a case. AI tools understand patterns and relationships and can generate responses, but conducting an independent analysis of the facts, the law, and the evidence is still essential. Because arbitration involves complex legal issues as well as the assessment of people's behavior and intentions, human expertise is invaluable to correctly interpreting the context, upholding ethical standards, and preventing unfair or adverse outcomes. The human touch is required to balance AI by examining results for hidden biases or errors, assessing subjective factors beyond quantitative data analysis, and taking into account cultural sensitivities when making decisions.

Lost in translation

AI-based machine translation tools are improving but far from perfect. For the most part, they provide a largely literal translation that could be incorrect, thereby affecting the evidence on which the dispute is decided, the integrity of the proceedings, and the reasoning of the arbitrator. As the number of international arbitration cases increases with cross-border trade and commerce, such translation tools are invaluable. But use them with caution and complement an AI translation with an academic or legal translator who has a deep knowledge of the legal field.

Full disclosure

Whether to disclose the use of AI in legal proceedings is currently one of the hottest topics facing arbitration practitioners. Draft guidelines are circulating, but we have yet to define what constitutes the “use” of AI and what ethical guidelines should be in place. After all, we don’t disclose that we use Lexis of Westlaw or Jus Mundi now for research, and the moral duty of confidentiality already exists. So why should a lawyer need to disclose the use of AI in their submissions—if it is simply another tool—and the lawyer understands its limitations? It remains to be seen how these guidelines will evolve and how influential they will be, along with the technology that is becoming integral to our work.

I present these ideas as both a cautionary tale to avoid becoming an embarrassing meme—and more critically, to provide a framework for applying AI and other technologies in the best possible ways. These are already transforming how parties and other stakeholders access our services to resolve and prevent disputes. It is not science fiction to imagine what AI and similar algorithmic models will eventually be capable of in the not-too-distant future, and how this will alter so much of what we did in the past.

As such, technology will have a broader impact on the legitimacy of decisions and the decision-making process. The legitimacy of decisions will always depend on the parties’ trust in the process, grounded in the quality of procedural justice. Otherwise, we will have, at best, powerful and complex technology-generating outcomes that parties will not trust as reliable.

The fundamental rules regarding technology should therefore remain the same. Integrate AI wisely into legal practice, balance the benefits of technology with an equal dose of caution, protect data to ensure confidentiality, and focus on transparency, accountability, and reliability. And finally, always maintain ethical standards to ensure trust and integrity.

Equally important to remember as AI evolves is to recognise that the invaluable human qualities we bring to the arbitral process— empathy, discretion, knowledge, judgment, and interpretive skills—can and must co-exist with advanced technology.

Today, technology is shaping the expectations that lawyers and clients have for assessing legal and arbitration services, and also creating opportunities to enhance those services to advance access to justice. But recognize that technology is not infallible. There is bias, bad data, and questions surrounding legitimacy and confidentiality. Is it possible for AI to reason, or consider novel legal questions that involve complex facts? Not yet. But that day will come, and we should be prepared.

Claudia Salomon is the President of the ICC International Court of Arbitration