A new generation of legal analysis tools is emerging. All of them rely on innovative use of data, and there is increasing use of machine learning to predict outcomes.

Disputes lawyers have long been pioneers in bringing technology into the practice of law, through the development of technology-assisted document review. Many firms, including ours, are now deploying machine learning and other tools which can assist with case preparation by automating tasks such as bundling, research, and drafting. But for disputes, the most significant innovation is in outcome prediction and claims analysis.

Predicting the outcome of a case is at the core of an experienced dispute lawyer’s skill. It helps parties decide whether to pursue a case, whether to settle and for what level, and how to approach case strategy. For centuries, lawyers have based their predictions on a combination of legal analysis, intuition and largely individual experience.

A New Generation of Tools

However, this analysis is becoming increasingly sophisticated, fuelled by a new generation of legal analysis tools. Lex Machina, a tech spin-out from Stanford University Law School, is one of the best-known examples. It mines litigation data to reveal insights about judges and opponents. It can show how likely a specific judge is to grant or deny a particular motion, the judge and opponent counsel’s experience of the type of case, and so on.

There are now many companies seeking to provide similar analysis, often targeting specific local markets. For lawyers and their clients, these tools can help to manage some of the uncertainty around disputes, not only around outcomes, but also budgeting and timescales.

Demand for Data

This relies, of course, on the body of available data. Not all jurisdictions have a public database of judgments, and the problem is more acute in arbitration where awards tend to be confidential. Here, one of arbitration’s greatest advantages over litigation may become a problem. There are many attempted solutions.

ArbiLex, a new entrant in this field, has based its system around investor-state arbitration, where awards are typically public. Arbitrator Intelligence, a non-profit initiative, aims to increase access to information about arbitrators and their decision making through post-award questionnaires sent to participants.

Effective deployment of machine learning will have many benefits. It will assist clients to better assess possible outcomes, which will result in more transparent settlements and fewer cases going to trial; better ROI. Attorneys will be subject to less confirmation bias and will be able to provide advice to clients supported by real data.

Ben Allgrove, Partner, Global R&D

Dispute Resolution Data works directly with arbitral institutions to collect anonymised data on cases, which subscribers can analyse.

Some tools seek to detail every aspect of an arbitrator’s profile and experience. GAR’s Arbitrator Research Tool contains over 250 profiles and continues to grow. Kluwer’s Arbitrator Tool is due to be expanded and re-launched in 2019. And it is not just private practice lawyers who will utilise new technology for disputes.

Technology for All

Some firms, such as LexPredict (recently acquired by Elevate), are helping parties to build their own tools to predict the outcome of litigation and arbitration cases.

Third party funding firms are reputed to be investing heavily in this area, for the obvious benefit: any improvement in their ability to pick winning cases translates straight to the bottom line.

And as any litigation funder will tell you, choice of law firm affects outcomes. In-house counsel no longer need to rely on experience and directories, but can turn to readily-accessible hard data. GIR’s FCPA Counsel Tracker lists the main law firms involved in every resolved case in the last decade. Premonition, a US-based technology company, sifts through court data to analyze the most successful lawyers.

New Court Innovations

Courts themselves are also beginning to explore this area. In 2013, the Chinese government required all Chinese courts to publish judgments on a platform run by the Supreme People’s Court. It is now the largest national database of court judgments in the world. This has allowed the Chinese courts to launch a package of technological developments.

In certain Chinese courts, litigants can consult an artificial intelligence system which evaluates possible litigation outcomes before the case is filed. A Shanghai court is piloting an AI system for judges which analyses and automatically collates similarly-decided cases for the judges’ reference. The system also can also conduct deviation analysis on draft judgments, to help maintain judicial consistency.

In April 2018, reforms were proposed to the French justice system which would allow AI to assist in the resolution of certain court cases. The new system will present parties with the predicted outcome of a dispute, followed by compulsory mediation, in an attempt to drive settlement and ease pressure on the courts.

The Chinese and French court reforms may come to demonstrate the effect of Goodhart’s Law – once you measure something, it changes. These systems highlight aberrant decision-making by judges and arbitrators, which may make such decisions less likely in the future. They also highlight when a party’s settlement proposals are out of step with the merits of its case, which may make settlements quicker and more frequent.

In the longer term we may even witness completely automated decision-making, where submissions are made to a proven algorithm rather than a court or tribunal. This has clear potential in areas such as high volume, low cost consumer adjudication.