AI in Arbitration

Introduction:

Artificial intelligence, or “AI,” is the term coined to describe the general process whereby large amounts of data are combined with powerful iterative data processing systems and intelligent algorithms, thereby enabling the software to learn automatically from patterns or features in the data.

Arbitration is a document-intensive field of law that requires counsel and arbitrators to spend countless hours on legal research and document review. Due to an ever-growing demand for speed and efficiency, the present state of affairs cannot last.

The AI conclusions are reported to be remarkably accurate and often at a cost significantly lower than the countless hours a young lawyer would have to spend finding and attempting to analyze all the inputs. Thus, AI may serve to revolutionize the current disequilibrium in resources between parties who can afford the many lawyer hours such analysis may require and those who cannot.

Benefits of the use of AI in Arbitration

The use of Artificial Intelligence is sure to bring many advantages. They are few direct ones that are discussed below:

1. The main use of artificial intelligence in arbitration today is to review increasingly vast amounts of digital arbitral data held by parties and their counsel in order to determine what is relevant to the particular case and then to analyze that data and present it in a more effective manner. This use of AI to process arbitral data has, and will increasingly, help to correct the cost and time problem created by the digital data at issue in complex disputes today.
2. Expanding the use of AI to analyze arbitral awards to undertake actual legal reasoning and to provide reasoned advice about how companies and legal arguments have fared in the past, how arbitrators have decided issues, and how damages have been approached in similar cases.
3. AI offers the potential of predicting results in advance including, for example- Chances of success generally, and with a particular decision-maker;  Likely range of damages generally, and with a specific decision-maker;  Timing to a decision before a particular institution, and before a particular decision-maker; Likely costs to be incurred. It will help reduce uncertainty in any dispute resolution process.
4. AIs could also be used to assess evidence, which consists of arbitration of determining the relevance and materiality of documents. AIs could present a summary of the pieces of evidence produced by the parties and in the context of discovery or the analysis of the important quantity of documents, AIs could be more efficient than humans and less prone to mistakes.
5. AI may render the use of court reporters obsolete as the AI platform would be able to record the hearing via microphones and provide a real-time transcript with speaker identification for all concerned.
6. Arbitrators spend a lot of time on drafting standard sections of their arbitration awards, e.g., the parties, the procedural history, the arbitration clause, the governing law, the parties’ positions, and the arbitration costs. Arbitrators may save time and parties’ fees by delegating the drafting of such ‘boilerplate’ sections to AI machines.
7. Appointments of robots would be less vulnerable to challenge on grounds of conflict of interest or bias. Presumably, also, their decision-making process would be less likely to be tainted by the very human weaknesses of bias, illogicality or just having a bad day. 

Problems in the Implementation of AI

1. Policy Implication of the use of AI

Understandably, the prospect of machine arbitration raises a multitude of questions. The main question being whether it would be legally possible in the current legal framework?

The arbitration framework does not categorically rule out the use of technology in arbitral proceedings.

Article 19(1) of the UNCITRAL Model Law on International Commercial Arbitration states that “subject to the provisions of this Law, the parties are free to agree on the procedure to be followed by the arbitral tribunal in conducting proceedings”.[1] Article 19(2) states that  “failing such agreement, the arbitral tribunal may, subject to the provisions of this Law, conduct the arbitration in such a manner as it considers appropriate”, and also has “the power to determine the admissibility, relevance, materiality and weight of any evidence”.[2]

Article 19.1 of the Singapore International Arbitration Centre Rules provides that “the tribunal shall conduct the arbitration in such manner as it considers appropriate, after consulting with the parties, to ensure the fair, expeditious, economical and final resolution of the dispute”[3]Similar provisions can be found in International Chamber of Commerce Rules.[4]

However, though there is scope for development. There are several unanswered legal questions and uncertainties in the application of AI.

1. Who will have access to the necessary systems and data required to use the predictive capacity of AI to reduce litigation/arbitration risk? What will the cost of access be? For which purposes?
2. There is the question of the form and content of the decision itself. If a robot arbitrator renders its decision in the form of a code, can it be considered as an arbitral award? In France for instance, it would not be seen as a decision, because a decision needs to include legal reasons expressed in words to justify it. [5]
3. Decision by robot must also be inherently conservative, with the associated risk of perpetuating trends and stifling development. Because outputs are based on analysis of existing data, the ability to change or develop the law in response to changes in human thinking is stifled. Existing biases and assumptions are replicated and perpetuated. In short, it takes a human to think outside the box.

2. Data Privacy Concerns in the use of AI

Arbitration is a confidential process. In some special cases, such as in international commercial arbitration, arbitral awards in commercial cases are not published.[6]  Thus, when there is so much sharing of data happening in AI, it will be interesting to see how the intrinsic character of Arbitration will be preserved. Will access to awards and other arbitral data be open or closed? What impact will this have on data protection and privacy interests?

The current practices of international arbitral institutions with respect to the publication of selected awards in a redacted or summary form. The International Court of Arbitration of the ICC (ICC), International Centre for Dispute Resolution (ICDR), Singapore International Arbitration Centre (SIAC), the Stockholm Chamber of Commerce (“Stockholm Chamber”), and the Milan Chamber of Arbitration (“Milan Chamber”) publish redacted versions of selected awards usually with the party and possibly tribunal permission and typically excluding the names of the arbitrator, parties, and counsel, and the ICC publishes summaries of cases also excluding the names of parties and arbitrators and has started to separately publish the names of arbitrators sitting in their cases.[7]

To counter this hurdle, there are several start-ups that have started developing Arbitration databases. Few of them are- Global Arbitration Review Arbitrator Research Tool (GAR ART)[8], Arbitration Intelligence[9] and Dispute Resolution Data[10]

While all this data is helpful in gaining a deeper understanding of the commercial arbitral process, the current lack of access to the full reasoning of the award and the names of the arbitrators, experts, and counsel makes it insufficient for various aspects of AI analysis.

3. Ethical considerations in the use of AI

Arbitration practitioners could raise ethical reasons because of the absence of human qualities (e.g.: emotions) or due process defenses based on the so-called “black box”, which refers to the impossibility of directly explaining the results or predictions of the AI system.[11]

All the jurisdictions across the world have provisions for a fair trial. In the present interpretation of this clause, it is meant that the proceedings are conducted by a human because humans combine strict applications of the law with more subtle considerations of equity. It is tough for anyone would accept the legitimacy of robots as judges or arbitrators because they’re not human, they don’t have a heart, and they don’t apply equity.

4. Lack of explainability in AI tools

Firstly, the problem of lack of explainability in AI tools appears when the developers and implementers of these systems cannot explain how the program reaches a conclusion or a prediction. This is known as black-box systems and occurs especially within Deep Learning programs. This is problematic since as AI-enabled systems are becoming critical to life, death, and personal wellness, thee need to trust these AI-based systems is paramount.

There are several legislations that have made it compulsory for legal reasoning to be accompanied by the arbitral award. Not being able to answer legal reasoning questions can be especially problematic in judicial activity, where providing reasons and justifications is precisely one of its fundamental features. Indeed, in the International Arbitration, parties expect the arbitrators to reach duly explained and consistent decisions, which help legitimate the entire system.

For instance, regulations such as the Data Protection Directive and the General Regulation for Data Protection (GDPR) of the European Union establish a “right to explanation” to demand explanation and access to the logic behind the decisions made using AI.[12] Therefore, questions like: Why did the AI system make a specific prediction or decision? Why didn’t the AI system do something else? When did the AI system succeed and when did it fail? When do AI systems give enough confidence in the decision that you can trust it, and how can the AI system correct errors that arise? will not be answered. [13]

5. Lack of neutrality

AI may not be as “objective” and “neutral” as it is believed[14]. AI systems are trained, fed, and operated with data sets. The problem arises when the data used for training are themselves biased, which may result in AI systems becoming a vehicle to reproduce human biases and prejudices. [15]Similarly, AI systems used to assist the decision-making of arbitral tribunals could reproduce the trends that inhabit the cases entered as training data. Critical questions would be: Which cases will be used for this purpose? Who will choose them? How will diversity be guaranteed to take into account that many cases are kept confidential? In other words, a worldview or set of values that inadvertently inhabit previous cases and that are used to train the programs can influence how these systems behave and the decisions they make. For these reasons, the use of AI has the potential to affect the neutrality and equality between the parties, which are essential elements for maintaining the legitimacy of the International Arbitration regime.

Changes in Domestic Legislation of India for adopting AI

Speaking at an International Conference on ‘Arbitration in the Era of Globalisation’, organized by Indian Council of Arbitration and FICCI, the CJI said, “As we conceptualize international arbitration in a globalized era, we must also be cognisant of the synergistic opportunities available for international arbitration through the utilisation of disruptive technologies.”[16]

This shows that India is looking for measures to adopt AI in the Arbitration laws. With regard to this, there have been small proposed policy changes to accept Artificial intelligence in its early stages in India. As a result of the focus on developing institutional arbitration in India, the Bill proposed for amendment in the Arbitration and the Conciliation Act now makes a provision for the designation of AI by the Supreme Court (for international commercial arbitration) and the High Courts (for domestic arbitration). The Bill then provides that, in the absence of an agreed arbitration procedure between the parties, the AI (and not the appropriate court as per the current law) will be the appointing authority for arbitrator/s.[17]

However, the implementation of Artificial Intelligence in Arbitration will not be a simple path. There are a lot of developments required in other indirectly affected fields of law. Some are discussed below-

1. Data Privacy- The use of AI in arbitration involves a continual exchange and processing of data (often sensitive personal data or information) between the arbitral awards and the AI tool. There is no method to ensure the consent of the parties, arbitrators, and the liability when it is not taken. The existing laws of the Information Technology Act, 2000 though allow the use of Artificial Intelligence does not stipulate rules with regards the same. The government is planning to introduce the Personal Data Protection Bill; hopefully, it will be able to address the raising privacy concerns arising out of the use of AI.
2. The Intellectual Property Rights law will also have a significant impact. As of now, algorithms are not patentable as per section 3 (k) of India’s Patent Act. [18]Copyright of such algorithms has to be done as per the existing law. AI uses a large set of data collected from the public and analyses such collected data to give recommendations and suggestions as the result. The reusability of the same set of data by other service providers of AI will be considered as an infringement of copyright by the developer or not is unclear.

Conclusion

Regardless of how the above may sound, large international law firms already employ ‘data scientists’, ‘legal solutions architects’, and ‘heads of strategic client technology’, who focus on IT and AI solutions that would assist human counsel. New companies are being incorporated with a particular focus on AI solutions for legal research. An alternative solution, in order to limit the absolute objectivity of AIs, would be to subject the “award” rendered by the AI to a final check done by a human arbitrator. This could still be dematerialized as the smart contract inserted in the blockchain could already contain the name of the human arbitrator that will have to do the final check of the “award” given by the AI.[19]

Al has significant potential benefits for arbitration, but at what cost, and how this might impact arbitration that has to be analyzed.

At this point, we are faced with two possible paths: the creation of a legal framework for arbitration and AI; or the modification of existing international treaties (in addition to national legislation and arbitration rules). 

---

References:

[1] UNCITRAL Model Law on International Commercial Arbitration, 1985, With amendments as adopted in 2006, Article 19(1).

[2] Ibid.

[3] Arbitartion Rules of the Singapore International Arbitration Centre, 2016, Article 19.1.

[4] Rules of Arbitration of the International Chamber of Commerce Rules, 2012, Art. 25(1).

[5] https://www.hoganlovells.com/en/publications/the-future-of-arbitration-ai-robots-may-take-on-human-roles

[6] New York City Bar Association, International Commercial Disputes Committee, Publication of International Awards and Decisions, February 2014, available at https://www2.nycbar.org/pdf/report/uploads/20072645-Pu blicationofInternationalArbitrationAwardsandDecisions.pdf (“NYCBA Publication of Awards”).

[7] https://sussmanadr.com/wp-content/uploads/2018/12/artificial-intelligence-in-arbitration-NYSBA-spring-2018-Sussman.pdf

[8] See Global Arbitration Review Arbitrator Research Tool at https:// globalarbitrationreview.com/arbitrator-research-tool; and in this issue see David Samuels, The Unusual Suspects—Easier to Find with GAR’s ART, N.Y. Disp. Resol. Law. Vol. 11, Issue 1 (2018).

[9] See Arbitrator Intelligence at www.ArbitratorIntelligence.org and in this issue see, Catherine Rogers, Arbitrator Intelligence: From Intuition to Data in Arbitrator Appointments, N.Y. Disp. Resol. Law. Vol. 11, Issue 1 (2018).

[10] See Dispute Resolution Data at http://www. disputeresolutiondata.com/and in this issue see, Brian Canada, Debi Slate and Bill Slate, A Data-Driven Exploration of Arbitration as a Settlement Tool: Does Reality Match Perception? N. Y. Disp. Resol. Law. Vol. 11, Issue 1 (2018).

[11] http://arbitrationblog.kluwerarbitration.com/2020/02/06/could-an-arbitral-award-rendered-by-ai-systems-be-recognized-or-enforced-analysis-from-the-perspective-of-public-policy/?doing_wp_cron=1591080122.0984289646148681640625

[12] Watcher, S., Mittlestadt, B., Floridil, L. (2017). Transparent, Explainable, and Accountable AI for Robotics. Science Robotics, 2 (6). Recovered from: http://robotics.sciencemag.org/content/2/6/eaan6080.

[13] Schmelzer, R. (2019). Understanding Explainable AI. Forbes Magazine July 23, 2019. Recovered from:  https://www.forbes.com/sites/cognitiveworld/2019/07/23/understanding-explainable-ai/

[14] See, for example, the reference to AI as “neutral and bias-free” in Scherer, M. (2019). The Vienna Innovation Propositions, International Arbitration 3.0 – How Artificial Intelligence Will Change Dispute Resolution, Austrian Yearbook on International Arbitration, Volume 2019.

[15] See, for example, the controversy regarding Hewlett Packard computers with “racist” facial recognition systems at: https://www.wired.com/2009/12/hp-notebooks-racist/.

[16] http://www.ficci.in/ficci-in-news-page.asp?nid=20370

[17] http://arbitrationblog.kluwerarbitration.com/2018/11/24/proposed-2018-amendments-to-indian-arbitration-law-a-historic-moment-or-legislative-blunder-2/?doing_wp_cron=1591256804.7049129009246826171875

[18] http://www.ipindia.nic.in/writereaddata/Portal/IPOAct/1_31_1_patent-act-1970-11march2015.pdf

[19] https://www.ipg-online.org/data/cms_uploads/module_partner/publications/article%20%20arbitration%20and%20AI-blockchain.pdf