AI Technology and International Arbitration
Are Robots Coming for Your Job?
CIArb London Branch, January Meeting
Chair: Paul Rose. Moderator: Ben Giaretta.
Panel: Joanna Goodman, Dr Paresh Kathrani & Jonathan Leach.
Last year’s hot topic was ‘Will someone else fund my claim?’ Seminars and workshops on Third Party Funding attracted record registrations, packed attendance and often with a waiting list. The CIArb’s London Branch joint annual meeting with the LCIA, on the implications for third party funding for the arbitral process, held in September 2109 was no exception. The well-informed panel, including funders, lawyers and appointing bodies discussed how best to rub the magic lamp to conjure the funding genie.
Once again the London Branch has tapped into the zeitgeist. Its opening seminar for 2020 was a panel debate on the likely impact of artificial intelligence (AI) on the future of International Arbitration. Moderated by Ben Giaretta, the branch’s Honorary Secretary and Partner of Mishcon de Reya LLP, the panel comprised Joanna Goodman, a self-styled Legal Geek Woman and author of ‘Robots in Law: How Artificial Intelligence is Transforming Legal Services’, Dr Paresh Kathrani Director of Education and Training at the Chartered Institute of Arbitrators with a research interest in artificial intelligence, robotics, ethics and the law and Jonathan Leach, Head of Eversheds Sutherland’s International Arbitration practice in the UK, who generously hosted the event.
This time the panel was asked to look into its collective crystal ball. What impact will artificial intelligence have on international arbitration? Can robots replace the backroom lawyers, front row counsel or even the tribunal? The auditorium at Eversheds Sutherland’s London Offices was packed with those whose main concern may have been whether they would retire before they were replaced by a robot.
Joanna Goodman explained that as AI was here to stay it would be unwise to ignore it. Instead lawyers can use it as a tool to transform their services. Legal AI is done by converting natural language into machine learning managed by algorithm – the mathematicalisation of a task. We all assist with machine learning when using internet booking sites which require us to show that ‘I am not a robot’. Perhaps unknown to us, the captured traffic images are used to train autonomous vehicles.
Machine learning is particularly good for scheduled tasks, such as identifying terms and conditions, or checking properties against the land registry, which it can do fast and simultaneously. But it can also identify specific provisions, such as the contractual consequences of Brexit, even though the term ‘Brexit’ is not used. It has its flaws, such as confusing ‘ALL’ capitalised for emphasis, with the currency code for the Albanian Lek.
Nevertheless, Ms Goodman believed that AI was ‘perfect’ for dispute resolution as its prediction factors enabled parties to decide whether to settle. Ebay’s on-line processes, resolve 60m disputes a year, and China has No Court Room internet courts. Other sites, such as Legal Utopia help consumers identify legal problems and provide guidance for resolving them. It describes itself as ‘as social LawTech venture focused on utilising and leveraging artificial intelligence with a design-focused approach to deliver our mission to make law affordable and accessible to everyone’.
Conversely, DoNotPay.com, whose 19 year old creator claims it to be the world’s first robot lawyer, is a legal services chatbot, originally built to contest parking tickets, now available to ‘fight corporations, beat bureaucracy and sue anyone at the press of a button’. This Lawyer Bot successfully appealed $4m in parking fines in London and New York in 2016 with the consequential effect on local government revenue. It recognised that, as the process for appealing fines is formulaic, it was well suited to AI, which can drill down, identify and advise quickly and, therefore cheaply. Ms Goodman concluded her presentation with a reference to IBM’s Project Debater, an AI computational argumentation tool which assesses the persuasiveness of an argument as an aid to decision making.
Eversheds Sutherland’s Head of International Arbitration, Jonathan Leach helped put the burgeoning technology into a commercial context where Arbitration Tech could assist as a tool for case preparation and management. It can extract the meaning from documents, identify sentiment such as dishonesty and look for patterns in disputes. This would free up paralegals from churning documents and allow them to do ‘proper lawyering’. But for this to work, lawyers would need to become legal technologists. They will need to learn Tech.
Tribunals may be assisted by MRI imaging for detecting lying with a success rate of 85% compared with Judges who are only 60% accurate. Accurate lie detection might do away with cross-examination or identify it as irrelevant.
Accurately predictions of time and money, would result in more accurate pricing where law firms would charge less for work done by machines and more for the work requiring legal skills. Whether AI would replace counsel and arbitrators might depend on the type of AI and the parties’ respective investment in Tech. By 2050 there is a 50% chance of general AI being used. Mr Leach concluded however that AI is automation. It will not replace humans.
Dr Parash Kathrani took up the human theme and reminded the audience that there was a human element to the law which called for professionalism, confidentiality and trust, qualities which may be lacking when conversing with your Alexa. Dr Kathrani recognised that if lawyers are always looking back, the law would not develop. But he pointed out that technology, with its easily accessible, neat case summaries and nutshell learning, was undermining legal study, particularly of reasoning and the decision making process. Tech was not objective. It was code, coded by a human who was not immune from bias. The law may become code, and be developed as code, or even develop itself as code, but the quality of the resulting advice, argument or decision would depend upon the quality of the algorithm. There are currently many rival algorithms, promoted by different start-ups with no clear contender for the users’ trust. None can be said to be neutral. All provoke inherent suspicions of the unknown fuelled by science fiction. He referred to the on-line initiative ‘Advancing AI in Wisconsin’ set up to increase awareness of Tech and its impact on the community. Similar initiatives will emerge for other fields, including International Arbitration.
Contributions from the floor grounded the debate deeper into the here and now. As prediction could only be based on the past performance, an arbitration algorithm would need data feed from live arbitrations. An AI Arbitrator’s award may not be enforceable. The law in England and France require an award to be given by a human. The right to a fair trial is a human right. The film ‘Back to the Future 2’ was the most reliable vision of the future of its time. But the debate concluded where it had started. If you look into the crystal ball you will see that AI is coming. The available processing power dwarfs our understanding. Law firms will have to employ technologists just as Google do, to face the pace of change.
KIM FRANKLIN QC
Crown Office Chambers