Gisela Mation, Maria Dória, and Sávio Andrade

The potential of artificial intelligence is enormous and its use can bring about extraordinary developments. However, because it is not an easily applicable binary technology, there are a number of challenges to be faced in order to take full advantage of its potential.

In general, artificial intelligence operates with algorithms that perform inferences from data and accumulate knowledge and learning from previous experiences.

Over time, the system will improve itself and should always make it possible to trace the patterns detected so that we can understand its mechanisms and understand how the conclusions presented were reached.

However, this is not always the case. Machine learning is often not mappable, constituting a veritable black box that is neither accessible nor sufficiently intelligible by humans.

Moreover, the international community has already pointed out several cases of mistaken conclusions that occur when artificial intelligence is used for more sophisticated issues, different from repetitive operational tasks guided by humans.

Not surprisingly, global discussions are underway regarding problems related to the difficulty of tracking the conclusion of algorithms, the bias of the data used - leading to biased conclusions - and accountability arising from the use of artificial intelligence.

In Brazil, there is still no specific regulation on the subject, since the legal framework of artificial intelligence, approved by the House of Representatives in September 2021, is still pending in the Federal Senate.

However, it is imperative that, from the outset, the use of artificial intelligence observe ethical and legal principles and adhere to technical criteria of reliability, traceability, and accountability. It is essential to establish a minimum stringency to provide security for those involved.

Also in the second half of 2021, the Federal Judicial Board (CJF) held the II Conference for the Prevention and Out-of-Court Resolution of Disputes in order to seek efficient solutions to the large number of lawsuits in Brazil. Among the pronouncements approved, 106 stands out,[1], which encourages the use of artificial intelligence in arbitrations.

Today it is already possible to imagine, for example, the use of artificial intelligence in transcriptions and translations, document analysis and systematization, or even in supporting the appointment of arbitrators, for competence verification and conflict checking.

There are various possibilities of employing artificial intelligence techniques to contribute to the improvement of efficiency in procedural acts in arbitration proceedings. These acts include those performed by arbitrators as well as by lawyers and experts. They also cover activities performed within the proceeding itself and preparatory activities.

Necessary care

The intention to encourage the use of artificial intelligence in favor of dispute resolution is laudable and necessary. To achieve greater efficiency, the incentive must be accompanied by guidance on the limits for the use of this technology, the phases in which it can be applied, and the means of verifying how it has helped the result obtained. In other words, due process of law needs to be updated to accommodate the use of artificial intelligence.

In this context, the CJF was right to point out, in the justification of the aforementioned pronouncement, the need for "this phenomenon to observe the constitutional guarantees of due process of law, especially the right to participation and the right to an adversarial process.

The ponderation was pertinent not least because caution is even more recommended in arbitration proceedings. One should avoid giving the other party elements to support a possible action for vacatur, which could harm both speed and legal certainty, two of the main advantages of arbitration proceedings.

This is not to reject the use of artificial intelligence in dispute resolution. As the CJF rightly pointed out, its use can contribute to disputes being resolved more efficiently and with less delay. However, its employment deserves care and diligence.

In the absence of specific regulations, it is essential that lawyers be aware and work to ensure that the use of artificial intelligence tools respects a broad defense, the adversarial process, and the guarantee of the arbitrator's free persuasion.

The direct consequences of technological advances have already arrived and, more than ever, the parties need to be well advised by multidisciplinary professionals who not only master the law, but also have the knowledge necessary to understand, in real time, how the application of artificial intelligence can impact on resolution of the dispute. In this manner, they can act quickly on behalf of their client.


[1] Pronouncement 106: It is admissible in arbitration to make use of artificial intelligence technology tools to assist the parties and the arbitrator during the course of the proceedings.