Arbitration has become established in recent decades as one of the most relevant alternative methods of dispute resolution. Its flexibility, speed, and confidentiality make it especially attractive in complex business disputes, often involving international contracts, large sums of money, and high technical expertise. 

In parallel, we are experiencing a technological revolution driven by the rise of artificial intelligence (AI), capable of radically transforming entire sectors of the economy and society. 

The intersection of these two phenomena raises crucial questions: can arbitration be conducted by AI systems? To what extent can technology contribute to the efficiency of the procedure? What are the ethical and legal limits of its use?

The topic is not merely speculative. Concrete experiments are already underway, with software capable of drafting preliminary arbitration awards, analyzing precedents, estimating probabilities of success, and even simulating negotiation strategies. Given this scenario, the discussion about arbitration with artificial intelligence takes on urgent dimensions, as it involves redefining fundamental categories of private law, the autonomy of the parties, and the very idea of ​​arbitral jurisdiction.

I – Traditional arbitration and its foundations

Before analyzing the incorporation of AI, it is important to understand the pillars of arbitration. Unlike state jurisdiction, arbitration is the result of the will of the parties, who enter into an agreement to submit their disputes to arbitrators chosen by them. 

The procedure is characterized by private autonomy, procedural flexibility, and the principle of confidentiality.

Another key aspect is the specialization of the arbitrators. Often, the parties choose professionals with in-depth technical knowledge in a specific area, such as civil construction, energy, technology, and the financial market. This is precisely to ensure that the judgment is more closely aligned with the specificities of the dispute. 

At this point, arbitration already differs from the traditional judicial system, which is often slower and less specialized.

The question that arises is: if the selection of referees is based on their expertise, would it be possible to replace them with AI systems trained on vast datasets, capable of identifying decision patterns and applying complex reasoning quickly?

II – Artificial Intelligence: Potential in Arbitration

Artificial intelligence is already being applied at different stages of arbitration, albeit in a partial and auxiliary way. 

One of the methods used involves the selection of arbitrators. AI platforms can cross-reference information on past experiences, rulings, academic publications, and contact networks, helping parties choose arbitrators best suited to the profile of the dispute.

Furthermore, in arbitrations involving thousands of pages, such as those related to infrastructure, AI can accelerate the discovery phase by identifying relevant documents, detecting inconsistencies, and grouping evidence by similarity.

In legal proceedings, we can use AI for jurimetrics. In arbitration, algorithms trained on previous arbitral decisions can provide estimates of the probability of success, assisting the parties in procedural strategies and cost-benefit calculations for a potential settlement.

One of the most common uses, even outside of arbitration, is the use of AI tools to generate draft reports, which can be reviewed and adjusted by human arbitrators, reducing the time and costs of the procedure.

Furthermore, there are smart tools that allow you to automate deadlines, notifications, the organization of virtual hearings, and even the control of fees, bringing more efficiency to the arbitration process.

These applications are, to a large extent, already underway, indicating that arbitration is moving towards becoming increasingly technological.

III – The possibility of artificial referees

One of the most controversial debates concerns the use of artificial referees, that is, AI systems assuming the decision-making role. This hypothesis generates both enthusiasm and concern.

This idea was already present in discussions about the Brazilian judicial model, and now it's moving into the world of arbitration.

On one hand, there are proponents of the idea that algorithms could make decisions more quickly, objectively, and predictably than human arbitrators, eliminating personal biases and reducing costs. In less complex or mass disputes, this possibility would be even more attractive, allowing for fast and standardized decisions.

On the other hand, numerous objections arise. The principle of autonomy of will, even though it allows the parties to choose their arbitrators, can be strained by the idea of ​​entrusting the final decision to software. 

Even more serious is the issue of the motivation behind the decision: the opacity of machine learning algorithms can make it impossible to understand the basis of an arbitration award, violating the requirement for transparency and the right to a fair hearing.

Furthermore, there are practical questions: who would be responsible for any failures of an artificial arbitrator? The programmer? The arbitral institution? The parties who consented to its use? The civil liability regime in this scenario remains unclear.

IV – Legal and regulatory aspects

Brazilian law, governed by the Arbitration Law (Law No. 9.307/1996), does not expressly prohibit the use of AI in arbitrations. The legal text states that any capable person can be an arbitrator, which, in principle, would exclude machines. 

However, a broad interpretation of private autonomy could allow parties to entrust the decision to artificial systems, provided there is no violation of public order.

Internationally, some experiences are already emerging. In 2018, Estonia announced the creation of a pilot project for robot judges in small claims cases. Although aimed at the state judiciary, the experiment highlights the technological advancement in decision-making. 

In the field of arbitration, certain institutions already offer online dispute resolution (ODR) platforms, where algorithms assist in mediation or suggest agreements.

The regulatory challenge, therefore, lies in balancing innovation with fundamental safeguards. Arbitration cannot be reduced to a mechanical act of data processing; it is necessary to preserve the legitimacy, motivation, and fairness of the decision.

V – Ethics and algorithmic impartiality

Another crucial point is the issue of impartiality. If arbitration depends on the parties' trust in the arbitrators, how can we ensure that algorithms are not biased by training data that reflect historical prejudices or inequalities?

Research has shown that AI systems can reproduce biases related to gender, race, or social origin, compromising their neutrality. 

In the context of arbitration, this would be particularly serious, as it would undermine the credibility of the award. Hence the importance of clear ethical rules, requiring transparency in the data used, auditing of algorithms, and the possibility of challenge by the parties.

VI – Future Paths

For AI-powered arbitration to become legitimate and effective, certain steps seem indispensable. Initially, it is necessary to regulate the use of these new tools. This involves creating national and international norms that govern the use of AI in arbitration proceedings, establishing limits, responsibilities, and transparency standards.

Arbitration chambers must lead innovation by offering secure, auditable, and accessible platforms, as well as preparing arbitrators to use new technologies.

One of the main topics for deliberation should be in the field of algorithmic ethics, so the requirement for independent auditing, explainability of decisions, and mechanisms for correcting biases will be fundamental to preserving the legitimacy of the system.

AI literacy will be a substantial process in this initial phase, meaning that arbitrators and lawyers will need to master basic knowledge of AI, data analysis, and technology to adequately perform their roles in the new environment.

Therefore, the transition should be gradual, starting with auxiliary functions (process management, document analysis) and evolving towards a more prominent role for AI, always under human supervision.

Conclusion

Arbitration with artificial intelligence is no longer a distant future: it is a reality in gestation. Although still surrounded by legal, ethical, and practical challenges, the technology already demonstrates great potential to make procedures more efficient, accessible, and predictable.

However, it is crucial to recognize that arbitration is not merely a technical mechanism for dispute resolution, but also a space of trust, legitimacy, and justice. These values ​​cannot be sacrificed in the name of algorithmic efficiency. Artificial intelligence should be understood as a support tool, not as a complete replacement for human rationality.

Private law, through arbitration, has the opportunity to lead an innovative process of technological incorporation, capable of redefining paradigms and designing new ways of resolving conflicts. It is up to the legal, arbitral, and technological communities to build this future responsibly, balancing innovation with fundamental guarantees. 

Only in this way can AI-powered arbitration fulfill its promise: to unite the tradition of private autonomy with the revolution of machine intelligence, preparing for the challenges of an increasingly complex and interconnected world.

Read more: 5 technological trends that every public servant needs to know

BIBLIOGRAPHIC REFERENCES

BECKER, Daniel; CAMPOS, Ludmilla. Artificial Intelligence and Arbitration: A Brief Study on the Instrumentalization of Technology in Arbitration Proceedings. Privacy and Data Protection Magazine, no. 008.

BERGMAN, Deborah. Civil liability for the use of artificial intelligence in mediation and arbitration. 2022. Master's thesis. Autonomous University of Lisbon (Portugal).

HOECHNER, Lara de Melo Schneider Bier et al. Revolutionizing Law: the influence of Artificial Intelligence on mediation and smart contracts in Brazil. Avant Journal, v. 8, n. 2, p. 186-198, 2024.

MARTINS, André Almeida. Arbitration and technology: an overview. In: International conference estate, succession and autonomy – new assets and new trends and adfas 5th international congress. UCP Editora, 2024. p. 263-279.