The Use of Artificial Intelligence in Arbitration: Where We Are and Where We’re Headed
In recent years, integration of digital functionalities and electronic media has become increasingly frequent in arbitration proceedings. Video conferencing is now common for hearings, including sessions for the production of evidence. Pleadings and other documents are almost never submitted in physical form, giving way to cloud-based document management. New systems and applications emerge almost daily, promising to ensure that disputes are resolved with greater efficiency, saving time and resources for all parties involved.
The inherent flexibility of the arbitration process, with its guiding principle of party autonomy, means that artificial intelligence systems may well produce the next big technological leap in arbitration practice. After all, AI tools are not inherently incompatible with the Brazilian legal system. In fact, during the 2nd Symposium on Prevention and Non-Judicial Resolution of Disputes (2021), the Federal Courts Council (CJF – Conselho da Justiça Federal) affirmed that it is "admissible in arbitration to use artificial intelligence technologies to assist the parties and the arbitrator in the course of the proceeding" (Statement 106).
A number of applications of these new technologies are already available. Prompted by the legal community and the innovation market, parties have turned to software solutions to assist them in the crucial process of selecting arbitrators: Initiatives such as Arbitrator Intelligence and Global Arbitration Review Research Tool provide data collected from lawyer feedback on arbitrators' involvement in hearings, average time taken to issue awards, preferences, and procedural practices.
With this data in hand, parties can mitigate potential conflicts and widen the range of candidates, leading to greater racial and gender diversity in the field. The trend is for arbitrator appointments to become more rational and less subjective.
The primary use of artificial intelligence in arbitrations, however, continues to be data processing in complex disputes, helping lawyers sift through terabytes of digital documents to extract what is relevant for each case. The question that remains is whether arbitral tribunals could also make use of similar software to issue decisions – for instance, to comply with a document production order. English courts have already done so, and arbitral tribunals might follow the same path in the future, especially given the growth of e-discovery.
The future seems to hold even more challenging questions. Looking beyond AI as a tool available to flesh-and-blood arbitrators, what will society and lawmakers have to say about arbitrators made of wires and metal? AI arbitrators is highly controversial topic, mainly because automated arbitration faces barriers in various legal systems. Brazilian law, for example, requires the arbitrator to be "any legally capable person trusted by the parties" (Article 13, Law 9.307/96), characteristics which are obviously absent in a robot.
Still, provisions like these date back to a time when terms like "robots" and "artificial intelligence" were found in sci-fi novels, such as Isaac Asimov's I, Robot, which inspired a film of the same name many years later. A literal interpretation of the law might eventually pave the way for parties to consent to the resolution of specific disputes by means of AI programs, especially in cases requiring a swift result.
These developments suggest that "the time when humanity faced the universe alone", to use Isaac Asimov's words, may be coming to an end. As artificial intelligence reveals new frontiers and offers new alternatives in the legal arena, legal practitioners must be vigilant in ensuring that use of these tools respects the principles of due process, the right to a full answer and defense, and impartial and reasoned decision-making, especially in the realm of dispute resolution.
