Consideration on the Use of Artificial Intelligence as an Arbitrator in Arbitration

Since Alan TURING's article in Mind magazine, in which he asked the question “can machines think?”, [1]artificial intelligence has entered our lives. Since the day it entered our lives, it has come a long way by taking shape according to the needs of people; thus, it has progress so far that it has succeeded in replacing people today. This issue has both positive and negative effects. The fact that it has progressed enough to replace people shows that it not only has the effect of facilitating life, but also has enough danger to leave them unemployed. Artificial intelligence is now more prominent in the field of law, as it is actively present in every area.

Artificial Intelligence in the Law

Research shows that artificial intelligence, which is used effectively in many different fields, plays a more active role in legal fields compared to the past. It is possible to mention artificial intelligence supported case law search databases as an example.  Thus, reaching multiple judicial decisions on a particular subject through keywords has become so short that it now takes seconds.  However, AI is not only used for case law review and data collection, but also for drafting contracts. Thanks to artificial intelligence supported legal sub-base software such as Evisort[2], Leverton[3], etc., draw up contracts today has become much easier than in the past. Moreover, there are even software that perform legal analyses supported by artificial intelligence. As explained on its website, AI Lawyer[4] provides services such as document summarization, contract comparison, language support and legal advice.

Artificial Intelligence and Arbitration

It is evident that technology has made the implementation of international arbitration much easier than before.

First of all, arbitration is defined as “the situation where the parties who have a dispute over a right, by agreement between them, leave the resolution of the dispute to special persons and these persons resolve the dispute”[5]. The special persons who resolve the dispute are arbitrators.

When we consider this issue; globalization is occurring with the increase in the human population in the world and the widespread use of technology, and trade can now be carried out from one end of the world to the other. As a result of this, in case of disputes in international trade, the parties prefer to international arbitration, which is a more effective, preferable, and less costly dispute resolution method.

As a result of the parties being in different countries and the seat of arbitration being in different countries, technology comes into play to facilitate the preparatory stages of the arbitration proceedings. For example, technological communication between the parties and the ability to easily transcribe the statements in the hearing are simple examples.

The use of artificial intelligence is currently more limited; generally, these systems are systems that provide databases on arbitration and mediation. An example of this is the US initiative Dispute Resolution Data[6].

Dispute Resolution Data, a global database on international commercial arbitration, receives its data from ICC, ICDR, CEDR. The information it receives consists of data such as location, cost, results, adjudication of the case, etc. Currently, the use of artificial intelligence in arbitration is generally limited to providing information and draw up contracts.[7]

The Use of Artificial Intelligence as Arbitrators in Arbitration

Alan Turing says that “there is nothing that humans can do that machines cannot do”[8], which brings the following question to mind? Is there really nothing that humans can do that machines cannot do? Can artificial intelligence, which plays an active role in law, replace the decision-makers, which are the three pillars of the judiciary? Can we use artificial intelligence as arbitrators instead of arbitrators who are the decision makers in arbitration? It will be possible to answer these questions by first explaining the place of arbitrators in arbitration.

In arbitration proceedings, the arbitrators are the final decision-makers on the dispute. This is one of the reasons that distinguish arbitration proceedings from court proceedings. However, in order to reach a final judgement, arbitrators, like judges, listen to the parties, examine witness statements and, where necessary, refer to expert reports.[9] As a result of all these, it finalizes the dispute. Arbitral awards are binding decisions like court judgements [10].

In addition, arbitrators should have certain characteristics, the most important of which are that arbitrators should be independent and impartial [11].

When we analyze artificial intelligence in terms of arbitrators, it is possible to evaluate it from two different perspectives; the first one is the positive effect of artificial intelligence arbitrators on the resolution of the dispute, while the other one is the negative effect.

Since we emphasize the independence and impartiality of arbitrators, it may be fairer for artificial intelligence, which works as a machine and does not have any human emotions, to resolve the dispute according to the norms previously loaded into the system. Because a machine does not have human emotions, it cannot be the case that it is closer to a party than before or that it is tied to a place.

Data and norms are uploaded to the system and as a result, it resolves the dispute with the information it receives. It applies whatever the norm is. However, even here, it is unclear that the person who uploads the data to the system will perform software impartially. It is not yet clear how this will be supervised.

 At the same time, it is certain to resolve the dispute in a very short time compared to human arbitrators. In this direction, concrete disputes based on calculation can be much easier to resolve after being uploaded to the system algorithmically; property settlement as a result of divorce, calculation of inheritance shares…[12]

In the United States, not in arbitration proceedings, but in the dispute where they used artificial intelligence as a mediator; Smartsettle One[13], a dispute that could not be resolved by a human mediator for 3 months was resolved in less than 1 hour. [14] Since artificial intelligence has been used as a mediator and has been seen to resolve the dispute in such a short time, it also raises the question of why it should not be used as an arbitrator in an arbitration proceeding.

When we need to evaluate the negative aspects of artificial intelligence arbitrators to arbitration proceedings, the question is how to load the concepts such as good faith, equity, discretion of arbitrators, which are abstract concepts that can be grasped by human emotions, to artificial intelligence as an algorithm.

In courses such as basic concepts of law, introduction to law and throughout the faculty, every prospective lawyer is taught that law does not only consist of the application of normative law until graduation. In this direction, it is a serious matter of discussion how a machine, which is far from human emotions and conscientious feeling, will apply a rule such as good faith in a concrete dispute or how it will make a decision in accordance with equity.

Arbitrators do not simply apply norms. They make judgements in line with their observations, their professional experiences, their experiences in their social lives and also their conscience in a fair manner. While making these judgements, they make decisions by taking into account both the rule of good faith and the rule of good faith, which are both used in the contract and appropriate to the dispute; the rule of good faith, the use of discretion when equity requires it. It is not possible for artificial intelligence to use all of these abstract concepts at the same time. [15]

Using artificial intelligence arbitrators instead of human arbitrators only for the sake of speed and brevity of the proceedings is, in a sense, contrary to the “principle of fair settlement” , which we have explained as a positive effect of arbitration. This is because it is not always fair to apply the norms simply, but it is fair to apply the law in accordance with the principles of equity and good faith.

In addition, considering the margin of error of artificial intelligence, it is not very reliable to completely delegate a dispute to an artificial intelligence arbitrator.

However, not using a marvelous technology such as artificial intelligence in arbitration would also cause a great mistake. Because, as we have explained above, this technological product, which greatly facilitates human life, can be an excellent assistant to the arbitrators in terms of resolving the dispute at every stage of the proceedings, even if it is not used instead of the arbitrator.

For example, they can find precedent decisions related to the dispute and guide the arbitrator. It can make suggestions to the arbitrator about the current dispute, and facilitate the work of the arbitrator by translating. In this way, the smallest details that may be overlooked by the human arbitrator are reminded to the arbitrator by artificial intelligence, while the arbitrator can analyse the solution skills of artificial intelligence in the concrete dispute based on his conscience and abstract legal principles.

Another point that needs to be examined is the necessity for the parties to reach an agreement at the point where artificial intelligence can serve as an arbitrator. The most important point on which international arbitration is based is the freedom of will, and in line with the freedom of will, the parties can appoint anyone they want as an arbitrator or choose the arbitral appointment procedure as they wish[16].

At this point, it is also unclear how much the parties will trust and choose artificial intelligence arbitrators. In the period when artificial intelligence can be used as an arbitrator, it is certain that it will not have a widespread use unless the parties choose it. For this reason, the most reassuring issue is that artificial intelligence will continue to assist arbitrators at the point of dispute resolution.

If we enter a period in which artificial intelligence can perceive the abstract features that are unique to human beings, all these discussions can be reviewed again...

Conclusion:

As a result, the replacement of arbitrators by artificial intelligence in arbitration may have negative consequences. Since abstract concepts such as good faith, fairness, considering the circumstances of the current situation will be interpreted in line with the conscientious opinion of the arbitrator, it is not possible for artificial intelligence to understand and interpret them for the time being. Accordingly, it is highly likely that erroneous judgements will arise. Instead of being an arbitrator, artificial intelligence assisting the arbitrator will make the arbitration proceedings fairer, easier, less costly, and more accurate.

References

Akıncı, Z. (2016). Milletlerarası Tahkim. Vedat Kitapçılık

Akıncı, Z. (2021). Gelişen Teknolojiler ve Hukuk II: Yapay Zekâ. O. G. Güçlütürk & E. E. Aksoy Retornaz (Yay. haz.), Yapay Zekâ ve Tahkim. On İki Levha Yayıncılık

Doğan, V. (2019). Milletlerarası Özel Hukuk. Savaş Yayınevi

Dr., P. (2022). Hukuk Perspektifinden Yapay Zeka. E. Büyüksağiş (Yay. haz.), Yapay Zekâ ve Hukuk(s. 3-6). On İki Levha Yayıncılık

M. TURING, I.—COMPUTING MACHINERY AND INTELLIGENCE, Mind, Volume LIX, Issue 236, October 1950, Pages 433–460, https://doi.org/10.1093/mind/LIX.236.433)

WAQAR. (2022). The Use of AI in Arbitral Proceedings . OHIO STATE JOURNAL ON DISPUTE RESOLUTION, 37:3, 345–366.

YILMAZ, O. G., Using AI in Judicial Practice - Can AI Sit on the Bench in the Near Future?, Law and Justice Review, 10(19), 67-102