Whither the Arsenal?

By | 27th March 2013

No, this is not a post by a disgruntled North London fan fed up with the inability of Arsene Wenger’s team to win a trophy! We are on much more interesting territory!

It appears to me that there has been a lull recently in the flood of articles and comment on predictive coding or computer assisted review. I do not take that to mean that the subject has gone away. Indeed from our own evidence of cases involving predictive coding and lawyers wanting to know about it, the opposite is true.

It may be that the whole subject is becoming more mainstream or that the objections to the use of a computer to assist in the process of selection of relevant documents, after a seed set has been selected by a lawyer familiar with the issues in the case, have been overcome.

In fact, I am not sure that either of those is true.

What seems clear, is that there is an increasing acceptance on both sides of the Atlantic, and for all I know, in the other common law jurisdictions, that something has to be done about the spiralling cost of litigation. In England we are about to see reforms to the process of disclosure and the introduction of related changes to the rules relating to the need to prepare budgets for litigation coupled with both case and costs management by the judges. The changes are being introduced in a piecemeal fashion which I fear will lead to costly satellite litigation and uncertainty, but I also believe that the changes are overdue and that when lawyers become more familiar with the obligations to be placed upon them , they will start to welcome them. I am sure the clients will! After all, what clients have been saying to litigation lawyers for years is “What is it going to cost?” and in most cases they have been told that it is impossible to say and will depend on a myriad of factors so it will take as long as it takes at the relevant hourly rate.

You only have to think about taking your car to be serviced and being told that the garage has no idea how long it will take but that the labour charges are £x per hour, to see how unsustainable such an approach is becoming.

The purpose of this post about predictive coding is to draw your attention to a thoughtful article in the New York Law Journal by Thompson Hine partners, Barry Kazan and David Wilson.

Published on March 18th and entitled Technology-Assisted Review Is a Promising Tool for Document Production, the article examines the current state of play in the US and concludes that predictive coding is a useful tool in the litigators’ armoury, while acknowledging that it is not a panacea for all problems associated with large scale document review.

The article should be read in full but I want to mention two points about this case and highlight a recent case study published by Millnet:

Firstly, the authors cover the usual cases such as Da Silva Moore, Kleen Products and Global Aerospace on which I have commented before but they also discuss a case which is unfamiliar to me and which has not featured previously in my occasional musings on this subject. The case is In re: Actos (Pioglitazone) Products Liability Litigation in July 2012 where the court made an agreed case management order of the type which I suspect will become more common here. You can read the order by clicking on the link above. The interesting point about this case for me is that the order is extremely detailed. For example, it specifies the custodians whose documents shall be examined to create a “sample collection population” and provides for three experts to work with the vendor to train the software which was to be used to analyse the documents.

Secondly, the article makes the essential point that predictive coding is not all about a magic box of tricks. There is no substitute for “careful lawyering, sound judgment and good planning.” Amen to that!

Lastly, kCura has published a Relativity Assisted Review case study by Millnet. The case study features Associate Director Emma Kettleton and may be accessed here. Look out for a) the substantial costs savings, b) the fact that less than 30% of the documents were subject to human review and c) the fact that the senior lawyer involved reviewed “only”1519 documents in order to train the software as to what was relevant and what was not!

On these results, I think you can safely say that predictive coding is here to stay and, as Barry Kazan and David Wilson say, it is a promising tool in the lawyers’ arsenal.