Talking at the Sir Henry Brooke Annual Lecture last year, the Lord Chief Justice of England and Wales, Lord Burnett of Maldon, described the ability to predict the outcomes of cases as “one of the most exciting developments of the age”.
He explained that while in its infancy at present, AI, through machine learning technology, will soon have the ability to find patterns in highly complex decisions, and even perhaps to balance evidence with moral questions of right and wrong. As more data is inputted and the algorithms become more sophisticated, AI will continue to learn and improve.
As far back as 2004, Washington University compared the accuracy of their algorithm in predicting the outcomes of US Supreme Court decisions against the predictions made by a team of experts. The algorithm accurately predicted the outcome of 75% of 628 cases heard in 2002, compared with the experts’ 59%.
In 2017, Illinois Institute of Technology used the US Supreme Court’s database to expand the scope of the experiment to cover all the cases heard between 1816 and 2015. Their algorithm predicted the outcome of cases spanning nearly two centuries with 70.2% accuracy. Similarly, in 2012, researchers at University College London used machine learning to replicate judges’ reasoning in European Court of Human Rights judgments to predict the outcome of cases with 79% accuracy. The algorithm’s estimate is also helpful from a client or funder’s perspective, in that it is based on fact rather than emotion.
On a more serious note, In France the government has made a somewhat surprising intervention that seeks to limit the emerging litigation analytics and prediction sector. Article 33 of their Justice Reform Act prevents anyone from publicly revealing the pattern of judges’ behaviour in relation to court decisions and carries a maximum sentence of five years imprisonment for anyone who breaks the new law. A fierce debate raged about whether the names of judges should be removed from their decisions when they come to be published online.
Whatever the French government’s reasons for banning the statistical analysis of judges’ decisions, we envisage this to be a mere blip in the overall advancement of AI in the litigation sphere, there and elsewhere in the world.
For a firm which was one of the first to actively use litigation funding as part of its litigation strategy, we closely follow how AI may affect the decision-making process by third party funders.
As the litigation funding industry grows and evolves, AI will logically play an increasingly important role in assisting litigation funders with the process of assessing the merits of claims and, in turn, informing their decisions on whether to invest in certain cases.
By minimising the risk litigation funders take on, and simply by speeding up the decision-making process, litigation funding could become even more widespread, increasing access to justice – at least for those with predictably meritorious claims.