ARBITRATION IN THE AGE OF ARTIFICIAL INTELLIGENCE & DIGITISATION: EXPLORING THE ROAD LESS TRAVELLED
- By Tariq Khan & Meghna Nimbekar
Keywords: arbitration, artificial intelligence, digital data, legal reasoning, arbitrators, smart contracts, micro-data
Arbitration became a popular choice owing to the complexities and technicalities in litigation. The flexibility which parties enjoy is a key factor in choosing it as the preferred method of dispute resolution. The stakeholders in arbitration are always open to new technological advancements and innovations, so far as it makes arbitration cost- effective and time efficient. While it is true that there is no legislation which currently regulates the use of Artificial intelligence (“AI”) -enabled systems or software in arbitration, there has been a rapid growth in use of AI in the conduct of arbitral proceedings. Incidentally, one of the most significant, is the use of ‘speech recognition’ in international commercial arbitrations. Although, this transcription service increases the costs for the parties (as it involves logistical arrangements), it makes the use of court reporters or stenographers obsolete, as it records everything said in a hearing via microphones and thereafter, provides a real-time transcript. It is then checked by the stakeholders, subsequent to which, a final version is shared with all parties. Similarly, AI is also useful for interpreting the witnesses at the hearing as professional interpreters can be very costly. Also, when documents are in different languages, AI can translate thousands of them in a few minutes.
Clearly, the last two-decades have seen advancements in AI and how it has impacted the industries; however, the use of AI in arbitration remained limited to the aforementioned purposes. With an increase in disputes, especially during this unprecedented post-pandemic era when virtual hearings have become the order of the day, it has become necessary to inculcate AI in various stages of arbitration as there is an increasing pressure to deliver effective legal services.
The most unique feature of this method of dispute resolution/arbitration is that it is quite receptive to the vicissitudes of the business world. And although arbitration may not be instant to identify its opportunities, it does identify that the ball is in its court. Let us cite a clincher to this effect –
It is “[the] first online collective decision-making application based on the Influents Algorithm [which] is a proof-of-concept pilot, designed for a two-party conflict with an arbitrator/mediator.”
From the perusal of the above, it is gathered that an AI arbitration is already in existence. Let us now delve into the pragmatic style of this AI system.
AI is ‘no rocket-science’
The term was coined by John McCarthy in 1956 even though the path to understand ‘if machines can truly think’ had begun much earlier. Simply put, it is a process of combining chunks of data with extremely “powerful data processing systems and intelligent algorithms”. For instance, a smart contract becomes a self-executing contract provided the terms of execution (as put into the lines of code of the algorithm) are met.
This algorithm plainly follows what is called the “if – then” pattern so as to ascertain an outcome which is based on frequent (and featured) patterns of the data already saved. In other words, if ‘X’ is the information fed to the computer, then the result (whether it is ‘Y’ or ‘Z’) would depend upon the already saved data with the computer (software). What happens if the information/words provided are outside the scope of saved data/the memory slot of the AI, is however, beyond the scope of this article (but in a nutshell, “self-tutoring” is what the system resorts to, in such a case).
Fitting the traditional into the ‘not so novel concept’
It would not be out of place to mention (though not exhaustively) the credo of principles of arbitration at this juncture – party-autonomy, cost-efficiency, time-saving and privacy; an AI ought to be a barometer to measure the efficiency of these principles of arbitration in comparison to arbitration surviving without it.
Further, the merits of AI to arbitration may rightly be fit into a smorgasbord of factors, out of which the most important is “legal reasoning” – it forms the very basis of “the Award”. However, there is no strait-jacket formula for reaching the outcome, leaving the legal issue mostly to the ‘wide ranging and ever dynamic’ exegesis of the arbitrator. Resultantly, decisions are bound to be followed by errors and miscalculations.
Interestingly, an AI uses the past arbitral awards in order to analyse ‘actual legal reasoning’ and provides a well-reasoned advice as to the progress of companies and their arguments in the past. The approach of the arbitrators and the manner in which the arbitrator approached the ‘issue of damages’ previously, is also analysed. In other words, AI has an option of predicting results in advance. Thus, the ‘inherent uncertainty’ that dispute resolution is vested with would be ruled out.
Thus, the loose concepts of arbitration (the leeway of party autonomy coupled with the arbitrator’s mind), ought to push us towards an AI system. An upgradation, much like the one adopted by Cyril Amarchand as early as 31st January 2017 (in a Press Release), is the future of settling disputes if one has to walk with the pace of time –
The stakeholders possess bulk digital-data and find it difficult to sift through each document. Quite possibly, important texts are missed – but with the help of an AI, data is presented in an effective manner and only the most relevant part is surfaced. This digital data that we talk of, called the micro-data, is a means to efficiently conserve cost and time that is often created by bulk digital-data.
Another argument in favour of AI being time efficient in arbitration, is further supplied (with reference to) – research related issues, preparation of lengthy and expensive reports, an indefinite cross-examination and arbitrations with a foreign element:
(i) The human researchers may at times erroneously mention cases which are not updated, or their human eye could miss out on important and relevant points which may turn the whole case around.
(ii) Then come the legal experts who charge exorbitant fees for preparing reports for their clients. This report, howsoever short, ends up creating a hole in the pockets of even business giants (let alone the mediocre or a small-scale businessman).
(iii) Another area which needs much attention, is the course which is adopted by the arbitrator to manage the proceedings by issuing a ‘procedural order’ (which is a requirement in specific cases of arbitration). This detailed order issued by the arbitrator in such cases makes it difficult to put a full stop on cross-examinations (any time soon), among other things.
(iv) Arbitrations which have a foreign touch, often struggle with “seat” and “venue” issues. It is an arduous task for arbitrators to travel from one place to another; in addition to the expenses incurred.
Now, imagine an AI coming to our aid in all of the above-mentioned situations. The AI can easily do away with the shortcomings of a human researcher. It can save the time taken by the legal experts and the money that they charge can be considerably reduced. The process of cross-examination will eventually be completed in a short period (for example, by using transcription services). Further, foreign parties will not have to bear the brunt of an arbitration where the “seat” is outside their own country and the AI (whether assisting an arbitrator or handling the entire case by itself) will iron out the creases quite well In addition, the authors take the liberty of mentioning the nuance of using a “lie-detector” in case of cross-examinations in arbitrations. However, this is a far-sighted idea that may or may not be used in future.
Per contra, concerns may arise as to whether ‘human-touch’ would absolutely be done away with if AI is introduced as an aid in the dispute resolution mechanism. To put this to rest, we are of the view that a final check (at the stage of the “Award”) by a human arbitrator is what would further provide a safety-valve. Other threshold challenges like that of privacy, flexibility, loss of employment and a humongous investment; are bound to arise with the adoption of AI in arbitration. Perhaps the most lethal is the wrong use of technology, resulting in grave injustice as the arbitrators/the counsels are not tech-savvy. Nevertheless, to apprehend this at such a nascent stage, may be too prophetic, and only after a few trials, will we be able to come to a better conclusion to this effect.
However, no dispute resolution is (or can be) absolutely automated. It is thus concluded that “AIs most likely will not be superhuman but will be many hundreds of extra-human new species of thinking, most different from humans, and none that will be an instant God solving major problems in a flash.”
Tariq Khan is a Principal Associate at Advani & Co. and Meghna Nimbekar is a Law Graduate from Amity Law School Delhi (GGSIPU)
Preferred Method of Citation- Tariq Khan, Meghna Nimbekar, ‘Arbitration in the Age of Artificial Intelligence and Digitisation: Exploring the Road Less Travelled’ (ICAR, 4 August 2020) <https://www.investmentandcommercialarbitrationreview.com/post/arbitration-in-the-age-of-artificial-intelligence-digitisation-exploring-the-road-less-travelled>.
 Chris Smith & others, ‘The History of Artificial Intelligence’,
 Ibid.  Brett G. Scharffs, ‘The Character of Legal Reasoning’,
available: http://law2.wlu.edu/deptimages/Law%20Review/61-2Scharffs.pdf.  Ibid.
The views and opinions expressed in the article are those of the author(s) solely and do not reflect that of official position of the institution(s) with which the author(s) is /are affiliated. Further, the statements of the author(s) produced herein should not be construed as legal advice.