Growing acceptance by court to embrace new technology: predictive coding in e-disclosure

arbtech2The High Court recently dealt with the issue of predictive coding used in e-disclosure.

Edward Spencer, associate at Taylor Wessing who acted in the case, explains the use of predictive coding in this case and considers how it may affect future litigation.

Pyrrho Investments Ltd v MWB Property Ltd & Ors [2016] EWHC 256 (Ch) concerns a claim brought by a former shareholder of an AIM listed company, under an assignment from the company itself, against a shareholder and number of former directors. The allegations include various breaches of fiduciary duty by the former directors in relation to loans and payments made by the company over a number of years.

What led to the use of predictive coding?

The parties to the proceedings are not in an equal position when it comes to documents. The vast majority of the documents are held by the company itself (the second claimant), which controls the backup tapes from the relevant time period.

Prior to full restoration of the backup tapes, it was already clear that the global data pool would run to many millions of documents. Accordingly, the parties entered into discussions about ways in which the data pool could be reduced to ensure a proportionate disclosure exercise. These discussions originally envisaged the application of keywords, followed by a manual review in the usual way. However, even after applying keywords the document universe requiring review was still vast. Taylor Wessing then suggested technology assisted review as a potential way to further reduce the cost of the disclosure exercise.

In the run up to a disclosure case management conference (CMC), the parties held a meeting, together with their IT service providers, to discuss the potential application of predictive coding. That led to the preparation of a witness statement in which the proposal by the fourth defendant, represented by Taylor Wessing, that predictive coding should be utilised was fully explained to the court.

By the date of the second disclosure CMC, the parties had reached further consensus as to the way in which the predictive coding exercise would be conducted, but still considered it prudent to seek judicial approval.

What were the challenges in pursuing this course of action?

The main challenge was reaching agreement as to the extent to which technology assisted review should be used by the second claimant when fulfilling its disclosure obligations.

The cooperation of the parties made this a much easier task than it could have been, but there remained issues as to the extent to which keywords should be applied to the data set before applying predictive coding.

Under what provisions did the court allow the use of e-disclosure?

The court has wide powers to give directions as to how disclosure should be conducted under the CPR in particular CPR 31 and CPR PD 31B. In conjunction with that wide discretion, the court is under an express obligation to actively manage cases to further the overriding objective which includes making use of technology (CPR 1.4(2)(k)). Furthermore, the overriding objective requires the court to deal with cases justly and at proportionate cost (CPR 1.1(1)).

What impact did it have on the e-disclosure process? Are there any drawbacks to the use of predictive coding?

The use of predictive coding in this case is expected to result in a cost saving of giving disclosure running into the high millions of pounds as against a traditional keyword search and linear manual review. It will also result in the disclosure process being conducted more quickly.

In terms of drawbacks, predictive coding, and technology assisted review in general, is a great tool to assist in e-disclosure, but it is not the holy grail. There is no doubt that employing a computer to undertake a large part of a review exercise can result in huge cost savings, but the technology is only as good as the training it receives. Another issue to consider is whether the type of case (be it the issues or the number of documents) is suitable for technology assisted review (not all cases are) and in particular which type of technology should be deployed.

In agreeing to use predictive coding, parties need to understand what they are agreeing to. This requires a level of cooperation above that usually found in litigation in order to ensure that the parties are clear as to the specific technology being deployed—including the detail of the process, confidence level and margin of error among many others. These are all things which should be discussed at the start, along with parties’ IT service providers, so there is less scope for misunderstanding later down the line.

How might this first judicial approval affect future litigation?

The use of predictive coding and technology assisted review is not new, but its uptake and use is not as widespread here as it could be. With the continuing court reforms, especially in respect of costs, judges and Masters are more and more willing to challenge parties as to why they are not making use of technology in order to simplify and reduce the cost of disclosure.

This judgment is significant because it signifies the growing acceptance by the court to embrace new technology. In future, parties are likely to be expected to cooperate more to ensure that disclosure is given in a proportionate manner—to what extent that means employing technology will depend on the facts of the case. There is however no doubt that the parties not employing technology or not cooperating to discuss its potential application may face questioning from the court.

Interviewed by Alex Heshmaty.

The views expressed by our Legal Analysis interviewees are not necessarily those of the proprietor.

Filed Under: eDisclosure

Relevant Articles
Area of Interest