The AI Advantage: Transforming International Arbitration

On 7 October 2025, Charles Russell Speechlys and Jus Mundi hosted a seminar in Paris exploring the development and use of artificial intelligence (“AI”) in international arbitration. The panellists comprised Simon Le Wita, Partner at Charles Russell Speechlys LLP in Paris, Monica Crespo Chanis, Head of Product at Jus Mundi, and Alexis Mourre, Arbitrator and Founder of MCL Arbitration. The seminar was introduced by Thomas R. Snider, Partner and Head of International Arbitration at Charles Russell Speechlys LLP in Dubai, and Jean-Rémi de Maistre, Co-Founder and CEO of Jus Mundi, and the panellist discussion was moderated by Jue Jun Lu, Partner at Charles Russell Speechlys LLP in London.

Introduction

The emergence of AI has taken us all by storm. AI has transformed the international arbitration field, leading to new complex legal and ethical challenges that must now be addressed. Its ability to process voluminous amounts of information, interpret legal principles at scale, and draft sophisticated documents promises big gains for parties, counsel, and arbitrators alike. However, these advantages cannot be looked at in a vacuum. There continue to be questions about due process, accountability, and regulation when AI is at play.

The field of international arbitration is experiencing a pivotal moment, one in which the potential of AI is undeniable, but the boundaries of its responsible use remain unclear.

What AI is doing in arbitration today

Monica Crespo Chanis explained that, in her experience, AI was now being used to perform tasks that map onto how lawyers think and work, including but not limited to summarising complex materials, administrative support, and workflow automation. The demand for research remains significant—roughly a third of legal professionals’ work is research-based—and general-purpose legal technology tools have to date struggled with the specificity of arbitration. This has driven the development of specialised AI systems, such as Jus Mundi. The capabilities such systems offer extend to even the most complex legal tasks in international arbitration, such as drafting, evaluating a set of documents, retrieving and processing information, and interrogating their contents. All of these processes, to an extent, mimic how we work as humans.

The practitioner’s perspective: cautious collaboration, a client-centric approach

For lawyers, the questions are practical: where does AI help, where does it hinder, and what do clients expect when their lawyers use it to supplement their work? Simon Le Wita explained that he sees the legal profession as one with a permanent duty to adapt to the “old world” of hard copy texts, having now been replaced with a “new world” of an abundance of electronic data and the ability to manipulate that data electronically.

Thus far, Mr Le Wita’s approach has been one of caution and trying to balance the obvious time and costs-saving advantages associated with AI against the need not to eradicate independent judgment and reflection. That approach is one which remains under constant review as lawyers work to test these new tools and explore how they can best help them. The real test however is whether clients are happy with the way in which their lawyers are using the tools at their disposal. If these tools are helpful and AI can collaborate effectively with lawyers, that can only be beneficial. Being transparent with clients is, in Mr Le Wita’s experience, crucial.

The arbitrator’s perspective: human judgment and discretion

Discussing his own perspective on the risks involved, Alexis Mourre explained that, as an arbitrator, he frequently sees excessive and unnecessarily lengthy written submissions being used by parties engaged in arbitral disputes. Rather than trying to address that issue head on, we are instead looking to legal technology to help solve it. In his view, AI can never properly replace the art of sophisticated legal drafting. That is not to say that AI will never to be able to provide effective written advocacy – no one knows what the future will hold and it is possible that AI will be able to formulate and apply logical reasoning in the future, which reflects the values, judgement, and background of the human who would otherwise be making the decision. For now, though, Mr Mourre is of the view that this remains a real challenge with the application of AI by an arbitral tribunal

Regulation and governance

The debate around the need for regulation of AI is moving quickly, and the central challenge that the legal industry faces at present is to define what precisely should be regulated and when. Alexis Mourre stated that, in his view, regulatory focus should be on the use of AI by arbitrators and institutions. He proposed the use of baseline principles such as a restriction over the ability to delegate adjudicative functions to AI, and the obligation to disclose the use of any AI tool with party consent being obtained at the outset. Without such safeguards in place, there is a risk that arbitral integrity will be compromised.

Monica Crespo Chanis agreed with Mr. Mourre, emphasizing that if AI tools are used, parties must be notified about their application and have the right to investigate which tools were used, how they were used, and to challenge any decisions made with the assistance of AI.

A tribunal’s scope to research

Questions from the floor probed the limits of a tribunal’s authority to use AI, and it was questioned whether arbitral tribunals could and should be permitted to use AI to conduct their own legal research without the need to first seek the consent of the parties. Mr Mourre took the view that arbitrators should not use AI without the consent of the parties, noting that an overriding principle of arbitration is that arbitrators may raise questions on their own motion and not use AI or the fruits of their own research to answer any such questions for them. As a matter of best practice, if a tribunal is unhappy with or has questions on submissions posed to them, the tribunal should invite the parties to clarify and/or raise specific enquiries addressing their concerns.

Access to justice

Another concern raised by a member of the audience was in connection with the gateway effect: starting with AI-assisted decision-making for lower value disputes could, over time, normalise delegation and push AI-assisted decision-making up the value chain in the name of claim management efficiency. Mr. Mourre emphasised the importance of legal values and the need for justice to be determined solely by humans. Ms. Crespo Chanis agreed; if AI is used in decision-making, it must be transparent and open to scrutiny and challenge at every stage. The right of a party to contest an AI-influenced decision is and should always be non-negotiable. For the moment, the focus of AI should be to assist with administrative measures. As our knowledge and understanding of risk management improves, AI could potentially play a different role in the future.

A balanced way forward

There is no doubt that, when used properly, AI provides tangible value and can enhance the foundations on which international arbitration is built. For practitioners, the focus needs to be on adding value for clients, coupled with transparency. For arbitrators, the focus needs to be on preserving fairness and exercising subjective human judgment. For institutions and legal technologists, accountability and human-centric design are essential. The current goal is to clearly grasp AI’s limitations and align incentives with core principles, ensuring that technology supports the system rather than the system supporting technology.

Originally published in the Daily Jus