
Fatsah Ouguergouz(Algeria)
Rector of the African Institute of International Law (Arusha, Tanzania)
Member of the Institut de Droit International
Chief Justice Zhang Jun,
Excellencies,
Moderator and Members of the Panel, Ladies and Gentlemen,
All protocol observed,
Warmest greetings from Arusha.
I am very pleased and honored to participate in this Fifth Seminar of the International Commercial Expert Committee, and would like to express my sincere thanks to its organizers for their kind invitation, in particular Ms. Long Fei, Director of the Coordination and Guidance Office for China International Commercial Court.
By way of introduction, I would like to clarify a number of conceptual points by defining what I mean by the expressions “artificial intelligence” and “international commercial justice” used in the title/topic of our panel.
Artificial intelligence may be defined as a set of computer systems capable of performing operations generally associated with human intelligence, such as analysis, reasoning, classification, translation, or content generation. As for international commercial justice, it should be understood broadly so as to encompass both disputes brought before national courts in transnational commercial matters and international commercial arbitration, mediation, and other alternative dispute resolution mechanisms.
I would also like to stress that a distinction must be drawn between, on the one hand, the use of artificial intelligence as such in commercial arbitration and, on the other hand, two other phenomena that have already been developing for several years, namely the digitalization of arbitration and online arbitration. The digitalization of arbitration refers to the replacement of physical media and procedures by electronic tools: electronic filing of written submissions, electronic transmission of exhibits, electronic signatures, or the use of case-management platforms. Online arbitration refers to proceedings conducted primarily or entirely in a digital environment, including hearings held by video conference, without necessarily involving the use of artificial intelligence. In June 2024, I had the privilege of leading a delegation of ten Chief Justices from the African continent to China, at the invitation of the Supreme People’s Court of China, and of visiting several judicial institutions at different levels of jurisdiction in Beijing, Shenzhen, and Guangdong. We were impressed by the degree of digitalization of judicial proceedings, particularly in terms of electronic filing and the very rapid processing of applications, without artificial intelligence as such necessarily intervening in those processes.
Artificial intelligence truly comes into play only when it is called upon to perform operations involving automated analysis or generation, such as semantic searches within arbitral awards, classification of evidence, reconstruction of chronologies, translation, predictive assessment of the likelihood of success, or preparation of a draft award. Thus, every use of AI presupposes a degree of digitalization, but not every digitalized or online proceeding constitutes arbitration using artificial intelligence.
In my view, the essential issue raised by the application and development of artificial intelligence in international commercial justice is not primarily technological or technical, or in other words to determine the extent to which artificial intelligence can contribute to faster and more accessible commercial justice. The answer to that question is relatively straightforward. In the field of international commercial justice, artificial intelligence can intervene at every stage of the dispute-resolution process: preliminary case analysis, legal research, evidence management, translation, organization of the proceedings, drafting of written submissions and, potentially, assistance in drafting the arbitral award, both with respect to the factual section and the legal reasoning itself.
The essential question raised by the application and development of artificial intelligence in international commercial justice is an ethical one. It is a question of determining the conditions under which artificial intelligence can contribute to faster and more accessible justice without compromising the independence of arbitrators, the confidentiality of arbitral proceedings, equality between the parties, and the right to a fair arbitral process. In other words, the question is how far one may go in the use of artificial intelligence in arbitration while preserving, in particular, the fundamentally human character of the adjudicative function.
This fundamental question has already been the subject of a number of studies conducted by institutions around the world. Thus, in 2024, the Silicon Valley Arbitration & Mediation Center (SVAMC) published Guidelines on the Use of AI in Arbitration. The Guidelines contain a number of recommendations addressed both to parties to a case and to arbitrators. In particular, all participants in international arbitration should be responsible for ensuring that their use of AI tools is consistent with their obligations to safeguard confidential information (including privileged, private, secret, or otherwise protected data). They should not submit confidential information to any AI tool without appropriate vetting and authorization. Parties, party representatives, and experts “shall not use any form of AI to falsify evidence, compromise the authenticity of evidence, or otherwise mislead the arbitral tribunal and/or opposing party(ies)”. Arbitrators, for their part, “shall not delegate any part of their personal mandate to any AI tool, particularly their decision-making process”; the use of AI tools “shall not replace their independent analysis of the facts, the law, and the evidence”. At the intergovernmental level, the Council of Europe has also issued recommendations aimed at ensuring that the use of artificial intelligence in the judicial field remains an instrument serving the public interest and has advocated an approach based on human oversight and critical review of artificial-intelligence systems, particularly in order to prevent the biases that may be embedded in such systems.
The leading arbitral institutions have also taken up the issue, including, for example, the Stockholm Chamber of Commerce Arbitration Institute (SCC), the London Court of International Arbitration (LCIA), the International Chamber of Commerce (ICC), the International Council for Commercial Arbitration, and the Chartered Institute of Arbitrators (CIArb) .
More recently, in February 2026, the United Nations Commission on International Trade Law (UNCITRAL) initiated a discussion on the desirability of developing international procedural guidance concerning, inter alia, transparency, confidentiality, equality of the parties, the reliability of AI-generated outputs, and the maintenance of human supervision. It notably recommended distinguishing between administrative assistance and substantive involvement in legal reasoning, clarifying the boundary between permissible artificial-intelligence assistance and impermissible delegation of adjudicative functions, as well as identifying low, medium, and high-risk uses of artificial-intelligence tools. It also recommended reflecting on “how to safeguard due process and equality of arms, including managing disparities in access to advanced artificial intelligence tools and ensuring fairness where one party relies heavily on artificial intelligence assistance”. It further recommended preserving ex aequo et bono decision-making as an exclusively human domain, explicitly recognizing that such awards — grounded in contextual human judgment and fairness beyond formal legal logic — should remain free from any artificial-intelligence intervention or assistance.
The principal African arbitral institutions have likewise progressively incorporated the issue of artificial intelligence into their research, training, and procedural-modernization activities. I will limit myself to institutions located in the main regions of the continent and in countries participating in China’s Belt and Road Initiative (BRI), namely Rwanda, Kenya, South Africa, Nigeria, and Egypt.
The Kigali International Arbitration Centre, the Nairobi Centre for International Arbitration, the Cairo Regional Centre for International Commercial Arbitration (CRCICA) and the Lagos Chamber of Commerce International Arbitration Centre have notably entered into partnerships with specialized platforms in order to publish and index non-confidential arbitral awards, access to which is facilitated by AI-powered search engines. These initiatives contribute to improving access to African arbitral jurisprudence, strengthening the predictability of outcomes, and increasing the international visibility of practices developed on the continent. In May 2025, the Association of Arbitrators (Southern Africa) NPC has adopted guidelines on the use of artificial intelligence, designed to promote its responsible integration into arbitration and dispute resolution proceedings; these guidelines aim to assist parties, arbitrators, and tribunals in using these technologies by defining best practices that preserve the integrity of the proceedings and ensure compliance with ethical requirements, while mitigating the risks that may arise from the use of AI. One may also mention that in Egypt, the Dispute Resolution Expert Manager (DRExM), which is AI driven, has been used to resolve construction disputes due to its ability to recommend the most appropriate dispute resolution technique, depending on the nature of the dispute, the evidence, and the relationship between the parties.
A striking example in Africa of the experimental use of artificial intelligence in the drafting of an arbitral award comes from South Africa. Indeed, during the first Johannesburg Arbitration Week, held in April 2024, Young AFSA, an affiliate of the Arbitration Foundation of Southern Africa (AFSA), which is an institutional partner of the China-Africa Joint Arbitration Center (CAJAC), organized a simulation entitled The Case of the Heavenly Twins: The Moot, the Man and the Machine. In this experiment, three teams of young practitioners presented their arguments before a tribunal composed of experienced arbitrators. At the same time, the University of Pretoria’s law school used ChatGPT 4.0 to generate an award based on the same case file; the participants were then given two anonymized awards: one drafted by human arbitrators and the other generated with the help of AI. They were tasked with determining which constituted the better decision without knowing who had authored it. A large majority of participants correctly identified the decision drafted by a human, demonstrating a solid understanding of the characteristics specific to human legal reasoning—such as the thoroughness with which case law is analyzed and the legal issues highlighted following an in-depth analysis of a case. However, the fact that opinions were split almost evenly regarding the “correct” decision shows that the one generated by AI was credible enough to rival human expertise; this experiment also provided a concrete assessment of a generative AI’s ability to organize the facts, examine the arguments, apply the relevant rules, and present reasoning in the form of an award.
Conclusion
The fundamental question raised by the use of artificial intelligence in arbitration is ethical in nature; it relates to the extent to which its use can reasonably be expanded while preserving the fundamentally human nature of the judicial function. There is therefore an urgent need to develop model clauses governing the use of artificial intelligence tools, with a view to preventing or minimising the potential risks arising from such use, notably by providing appropriate data protection, cybersecurity safeguards and clear provisions specifying the consequences of non-compliance.
Thank you for your kind attention.