skip to content

Predictive Coding, English Law, and the Disclosure Pilot Scheme

by Jennifer Roberts

desk-top-2-scaled

Linear review used to be the golden standard in edisclosure, particularly when the volume of an entire document collection tended to be smaller. It was industry practice to have human eyes review every document, and this practice carried over into electronic document review, often utilising large teams of document reviewers.

Whilst culling was commonplace, using objective metadata characteristics such as date and custodian, and also more subjective criteria such as keyword searches, linear review soon became impractical when modern data formats and communication technologies made the volume of information in a body of documents grow exponentially.

Use of techniques such as threaded review by inclusive emails and coding propagation do help with volume reduction. However, nowadays a document corpus could easily number in the hundreds of thousands, if not millions, of documents, and linear review is no longer practical for many cases involving electronically stored information.

English Law as it Pertains to Predictive Coding

The High Court has recently approved the use of predictive coding in disclosure in 2016, in Pyrrho Investments Ltd v MWB Property Ltd.  Citing the American DaSilva Moore  case (among other authorities), Master Matthews in Pyrrho approved the use of predictive coding due to the “enormous” expense of manually searching through the three million electronic documents associated with the case. In approving the use of predictive coding, Master Matthews provided these factors in favour of the decision (emphasis added):

(2) There is no evidence to show that the use of predictive coding software leads to less accurate disclosure being given than, say, manual review alone or keyword searches and manual review combined….

(5) The number of electronic documents which must be considered for relevance and possible disclosure in the present case is huge, over 3 million.

(6) The cost of manually searching these documents would be enormous, amounting to several million pounds at least, in my judgment, therefore, a full manual review of each document would be “unreasonable”… at least where a suitable automated alternative exists at lower cost.

(7) The costs of using predictive coding software would depend on various factors, including importantly whether the number of documents is reduced by keyword searches, but the estimates given in this case [are]… far less expensive than the full manual alternative…

(8) The ‘value’ of the claims made in this litigation is in the tens of millions of pounds. In my judgment the estimated costs of using the software are proportionate.”

It’s worth noting that while the court provides support for using predictive coding to satisfy disclosure obligations, several factors in Matthews’ opinion were specific to this case. The judgment notes that whether predictive coding is applicable to a case will depend on the particular case at hand. Not every case will involve millions of documents and correspondingly high costs of linear, manual review.

Justification for using predictive coding in the UK was further strengthened in the decision of Brown v BCA Trading Ltd & Ors, the first case to validate the use of predictive coding where one party contested it. Additionally, the High Court’s recent decision in Triumph Controls UK Ltd & Ors v Primus International Holding Co & Ors indicates some reluctance to trust simple TAR 1.0 systems that rely on humans to educate the algorithms via a single initial seed set of documents. Predictive coding is at its most effective when using continuous active learning from multiple lawyer and reviewer decisions.

Predictive Coding and the Disclosure Pilot Scheme

Following the lack of specific guidance in Pyrrho, Brown, and Triumph, the Disclosure Pilot Scheme came into effect in the Business and Property Courts as of 1 January 2019. The new disclosure scheme aims to bring about a wholesale culture change in how disclosure is done by setting forth specific disclosure models that provide guidelines for how much searching and investigation is necessary prior to the disclosure stage.

Although the Disclosure Pilot Scheme does not specifically mandate that lawyers use predictive coding, it does mandate that lawyers of opposing parties “promote the reliable, efficient and cost-effective conduct of disclosure, including through the use of technology,” and that they use ”software or analytical tools, including technology assisted review software and techniques… coding strategies, including to reduce duplication…prioritisation and workflows” where broader searches are mandated by the court.

Predictive coding has the potential to enable parties to more quickly comply with the new disclosure pilot models, whilst also keeping costs under control, and in some cases may actually be required in order to help with conformity to the Disclosure Pilot Scheme.

For more information on how predictive coding works on the Everlaw platform, get our white paper “Predictive Coding in eDisclosure” and tune in to our webinar with ACEDS, “Improve Review Outcomes with Predictive Coding on Everlaw.”