The Future of International Arbitration May Not Be AI
As costs mount and cases drag on, some turn to AI, big data—but technology can only do so much.
Megan Turchi
International arbitration, just a few decades ago considered an obscure form of alternative dispute resolution, has risen to become the preferred method for handling cross-border business controversies.
In 2018, the International Centre for Settlement of Investment Disputes, a World Bank organization that administers dispute resolution proceedings between international investors, registered 56 cases, up from three in 1996. Before 1997 the ICSID had never registered more than four arbitrations in a single year; it has registered at least 38 every year since 2011.
Behind arbitration’s rapid proliferation lies the belief that private dispute resolution in front of a mutually agreeable arbiter not only produces more predictable, enforceable results than commercial litigation but also delivers, according to some experts, substantial cost advantages.
But in recent years, those advantages have eroded. The parties to complex, high-stakes arbitration matters have sought greater protections through due process and discovery, diminishing the prospect of high-speed, low-cost resolutions. And because every arbitration is unique, and its terms and details dictated by the parties and facts, there is no centralized mechanism for reining in the expanding costs.
Can Technology in Arbitration Save the Mediation Process?
To help restore arbitration’s efficiency, many in the arbitration community are turning to a familiar savior—the same solution that’s led wave upon wave of efficiency gains in nearly every industry and profession the world over: technology. A host of software solutions—some already in the market, others yet to launch—promise to streamline nearly every phase of the arbitration process, from selecting an arbitrator to calculating damages.
While the new technology may prove valuable, it appears unlikely to create the kind of revolutionary cost savings it has produced in industries like manufacturing and financial services. Seasoned arbitrators, while eager to see what technology can do for the process, express skepticism about its full potential.
“I’d love it if technology could help us and the tribunals,” said Santiago Dellepiane, a BRG managing director and co-chair of the firm’s Economics & Damages practice. “But I know that at the end of the day, human interpretation and judgment will drive the arbitration process.” That’s due in part to case specificity, Dellepiane said, and to the fact that, “even theoretically, there aren’t one-size-fits-all solutions that can be reliably implemented.”
For the business leaders who have come to prefer arbitration over litigation, it’s important to understand the range of technology in development. Some of it may help fulfill the original promise of arbitration for faster, cheaper and more satisfying resolutions. Some may fall short of fulfilling that promise but prove useful in limited circumstances.
Artificial Intelligence and Arbitration?
As in nearly every corner of the white-collar world, the technology with the greatest potential to disrupt international arbitration might be artificial intelligence. With its ability to replicate and augment human cognitive skills, automate time-consuming but simple tasks and process massive quantities of data, AI offers the potential to help manage cases and diagnose inefficiencies in the arbitration process.
For instance, AI could help the parties to a dispute choose their arbitrator by examining thousands of candidates’ track records in similar cases. It could recommend drafting suggestions for arbitration clauses, helping clients and lawyers eliminate errors, identify blind spots and ensure their interests are protected.
AI’s fundamental value proposition lies in its ability to streamline administrative tasks while freeing up arbitrators and lawyers to focus on the parts of the process that require the greatest amounts of human judgment: assessing the facts, constructing arguments and deliberating to determine outcomes.
“The potential for disruption is immense,” said Lucas Bento, a senior associate at Quinn Emanuel Urquhart & Sullivan who is writing a book on ethical frameworks for AI.
Bento points out that lawyers have been using AI-infused technologies like Westlaw, LexisNexis and Google for years. But he and other experts believe the time is ripe to seriously explore the technology’s use throughout the entire arbitration process in order to, among other things, enhance efficiency and reduce costs.
“This is the perfect time to experiment, particularly with technology,” he said. “One of the great advantages of arbitration is that it’s rooted on the parties’ consent, and so, barring anything contrary to applicable laws or policies, the parties are free to design the method and process to resolve their disputes.”
ICCA-ASIL Task Force on Damages Brings Technology and International Arbitration Together
Kathleen Paisley, an arbitrator based at Ambos Law in Brussels, London and New York, also sees strong potential for technology in international arbitration. She sits on the ICCA-ASIL Task Force on Damages, created in December 2016 to “to address an issue of particular importance, but too often overlooked, in the field of international arbitration: quantification of damages.”
In April 2018, the task force presented the prototype for a web-based damages application to the ICCA Congress in Sydney. Set to debut by the end of 2019, the tool was designed to provide arbitrators and practitioners a consistent approach to determining damages.
The ICCA explains that “[t]he app comprehensively and interactively maps out the various legal, financial, and policy concepts and issues that arise in damages. Harnessing the advantages of technology, the app allows users to navigate between these different damages concepts and visualize how they link and relate to one another.”
Once it launches, the tool will be available to anyone to use for free. Paisley points out that its launch coincides with a recent shift in how arbitrators look at damages.
“Arbitrators are taking a more sophisticated and hands-on approach to damages,” she said. “They are not always waiting until the end of the arbitration process to consider damages—they are being more proactive throughout.” The new application could help those arbitrators—and create a more level playing field by providing arbitrators and practitioners a consistent and robust approach to determining damages.
The application does not do calculations—the idea is to provide a map, to help make sure arbitrators don’t miss issues or find themselves lacking data they need to proceed.
“Given the complexities of the valuations at issue in many of the cases in which I sit as an arbitrator, no app can provide the answers,” Paisley said. “Instead our goal was to provide a resource that will allow everyone involved in the process to focus on the key damages questions and to provide the arbitrators with the information and expertise they need to answer them in a robust way.”
The task force stopped short of trying to provide answers with technology. That’s the part that will always, inherently, be dependent on the facts of each dispute, the parties’ businesses and the contexts they operate in. Normalizing constantly shifting factors like industry context, competitive dynamics and geopolitical factors in a way a computer could understand would be nearly impossible.
“There is a temptation to standardize aspects of arbitration,” said Dellepiane. “But nothing will replace understanding the business or understanding the concept of risk in individual, specific circumstances. In fact, tribunals recognize this, and what you see are more and more in-depth awards discussing the particularities leading them to reach their decisions.”
Risks to the Future of International Arbitration
As new tools continue to roll out, and international arbitration continues to grow, arbitrators, attorneys and business leaders should certainly be open to the potential for new efficiencies and improved outcomes. But they should also be aware of the potential risks. Here are a few to keep an eye out for:
Bias: Much conversation in the technology community centers around the risk that design bias (when creating AI) or confirmation bias (when using AI-based applications) could lead users to incorrect conclusions. If the humans that enter in data think the data is supposed to tell them one outcome, they may be predisposed to find that outcome, whether or not it’s supported in fact. That risk could be especially fraught for arbitration software that’s specifically designed to lead to fair outcomes built around conclusive facts. “A particular worldview or set of values may inadvertently inhabit the application’s ecosystem and influence how it ‘behaves,’ in given contexts,” Bento said.
Human error: Like all software, AI is created by humans. Humans make mistakes. If those mistakes find their way into an AI application’s code, it could lead the program to draw false conclusions or make other errors—ironically, similar to the errors AI is designed to eradicate. If the application is widely adopted, the errors could quickly spread across an entire industry. “AI’s potential is assistive in nature: to augment cognitive capabilities, improve organizational efficiencies, and help us correct biased and incorrect assumptions about the world we live in,” Bento said. But when used incorrectly, this assistive nature can exacerbate flawed data and flawed assumptions.
Barriers to entry: AI may become prohibitively costly if it is to truly consider the immense body of economic and financial literature, at least in its early days, for all but the largest law firms. That could tilt the playing field in a way that’s fundamentally unfair, enabling those that can afford the AI-infused services to achieve better results. The greatest hope for making AI more accessible seems to be that it becomes more open-source, which would make it available or at least affordable for most firms.
Access to data: Paisley also sees potential for technology more generally in international arbitration, both in managing the data in individual cases and eventually for predicting results, provided the necessary tools and data are available. “When we talk about AI or any form of machine learning to predict results, we still need access to both algorithms and an adequate data set, which is a challenge for a confidential decentralized process like arbitration,” Paisley said. “But the potential benefits in terms of encouraging settlement and limiting unmeritorious claims are extremely significant if we get there.”
People Still Power the Arbitration Process
At the end of the day, former BRG Managing Director David Saunders says there are three key components of arbitrations that will never be replaced: the arbiters, the counsel and the witnesses. “Technology will help experts,” he said. But in arbitration, “it’s unlikely that humans can ever be replaced by technology.”
This doesn’t mean Saunders thinks technology isn’t important. He expects to see more technology used for managing, analysing and presenting documents in arbitration proceedings. And he sees a place for AI, as well. “There will probably be advancements that lead to better speech recognition systems at hearings to make the provision of translations and transcripts more efficient,” he said.
As new applications and tools like that emerge, expect a lot of buzz, a lot of experimentation and at least a few breakthroughs.
Megan Turchi is staff writer for ThinkSet.