International Arbitration and Artificial Intelligence: Time to Tango?

Kluwer Arbitration Blog
February 23, 2018

Please refer tot his post as: , ‘International Arbitration and Artificial Intelligence: Time to Tango?’, Kluwer Arbitration Blog, February 23 2018,

Besides the inverted initialism, what does international arbitration (“IA”) and artificial intelligence (“AI”) have in common? Sure, both IA and AI are leading alternatives to the status quo: IA to traditional dispute resolution, AI to traditional methods of production. The former promotes freedom from the judiciary, the latter freedom from cognitive limitations. Beyond that, comparisons would appear, well, artificial.

Yet closer analysis reveals synergistic opportunities for both AI and IA, at a time when professions and their well-guarded domains are being threatened by disruptive technological forces. The use and impact of AI on the legal profession is slowly becoming a hot topic in legal, technology, and academic circles. Conferences in the international arbitration sphere have begun to address how the AI-IA alliance could play out in practice. Could the future of IA lie in AI? In this short post, I sketch possible ways in which AI-infused tools could help the international arbitration community provide greater value to stakeholders.

Enhanced Legal Representation

AI can augment human cognitive abilities and automate time-consuming labor. A number of AI-powered products and services already exist to help lawyers parse through submissions, identify better legal authorities, review documents and agreements (e.g. predictive coding), estimate costs, and predict outcomes. A number of start-ups are focusing on disrupting the legal industry, with some already offering case management and forecasting services to the international arbitration community.

Looking to the future, AI tools could play a significant role throughout the entire arbitration process. For example, it could recommend drafting suggestions for arbitration clauses, helping clients and lawyers identify blind spots and bulletproof their interests. Parties could agree to use AI for some aspects of the arbitration itself e.g. discovery, to lower costs. AI-infused products and services could help lawyers also better manage cases by, for example, diagnosing inefficiencies and automating management tasks. Clients could also pre-screen a legal team’s fit for a particular case (e.g. success rate, extent of prior experience, peer-reviewed evaluations), and obtain a second opinion on their legal team’s analysis. The potential for disruption is immense. While the technology underlying AI is still experiencing teething issues, its capacity to enhance the quality of legal services while lowering costs and inefficiencies cannot be ignored. International arbitration lawyers should seriously consider adding AI tools in their offerings to clients.

Enhanced Adjudication Services

AI could also help with the appointment of arbitrators, the preparation of the award, and the simulation of judicial review. Case management could be automated, or significantly streamlined with the aid of software, giving arbitrators more time to do what they do best: arbitrate. Synopses of longer awards (particularly of investor-state arbitrations) could be automatically generated to help readers navigate through decisions. Tribunal secretaries could be replaced by AI decision support systems. And perhaps one day an AI-powered “arbitrator” (or “AIA”, artificially intelligent arbitrator) could preside over a dispute. Ultimately, it would be up to the parties to appoint such “machine arbitrators”. As with any disruptive innovation, trust would be the overriding consideration (can I trust the AIA to make a fair and reasoned decision?). If the parties trust the AIA, then who is to stop them from using it, particularly in arbitration where freedom of choice is paramount?

Additional Institutional Services

Arbitral institutions could also offer additional services powered by artificial intelligence. As noted above, in institutional arbitrations, case management could be automated by software. AI could also be used to predict costs, duration, and, perhaps more ambitiously, the merits of an arbitration. In an effort to promote a speedy resolution of the dispute, arbitral institutions could, at the request of the parties, propose settlement ranges based on arbitrations of similar size and complexity. This could nudge the parties toward settlement.

These innovations will ultimately depend on the parties’ willingness to allow arbitral institutions to use their data to inform future predictions. A simple revision to the institution’s arbitration rules could do the trick. Confidentiality would be preserved as the parties’ data would be merely used as training data (size of arbitration, number of parties, duration, type of dispute etc.) and anonymized for algorithms to generate desired outputs (e.g. settlement ranges based on prior data points).

Parties may also benefit from AI-powered recommendations on how to resolve their dispute in more subtle ways, such as whether to employ an Online Dispute Resolution service to save costs.

And, who knows, to correct the diversity-deficit in arbitral appointments, a diversity algorithm could be employed to recommend arbitrators from a broader pool of candidates.

Better Insights For Scholars and Third-Party Funders

AI could also assist academics and third-party funders. Scholars would be able to benefit from more sophisticated data about cases and general trends in the law. Third-party funders would also be able to draw deeper insights to help decide which cases to fund.

Potential For Self-Regulation

If we are to think big, could AI and blockchain technology ultimately help parties resolve disputes themselves by using self-contained tools without the need for adjudicatory intermediaries?

Time To Sober Up?

While AI has the potential to transform international arbitration for the better, its use does not come without risks: bias, hacking, and the amplification of human mistakes. To guard against some these risks, AI decisions must be explainable and cannot operate as a “black box”. This is particularly important in the dispute resolution world where parties expect decisions to be reasoned.

The parties may also limit the use of AI technology in any aspect of their dispute. But if the business and legal case for AI proves to deliver a better and quicker service for less, then it will only be a matter of time until parties opt for AI-assisted dispute resolution (“AI-ADR”).

We now know that human beings do not always make rational decisions. But human irrationality can be predicted. Richard Thaler, winner of the 2017 Noble prize in economics, noted as much in his 2008 book Nudge (co-authored with Cass Sunstein), where he argued that if irrationality can be predicted, human decisions can be nudged to maximize better outcomes. AI should be welcomed by the arbitration community as a tool to improve efficiencies, reduce costs, and reach fairer—and more transparent—decisions.

Lucas Bento FCIArb is Chair of the AI & International Law Subcommittee at the New York City Bar Association. He is currently writing a book on AI and the Law (forthcoming 2019). The views expressed in this post are the author’s personal views, and do not reflect the opinions of Quinn Emanuel or of the New York City Bar Association.