Insights

In Brief: A recent decision in the Supreme Court of Victoria is paving the way for the use of technology assisted review (TAR) in civil litigation within Australia. 

The decision in the case of McConnell Dowell Constructors (Aust) Pty Ltd v Santam Ltd & Ors is the first in Australia to order the use of predictive coding to assist in the efficient and practical management of large-scale litigation. 

It follows the adoption of the technique in jurisdictions including the UK in Pyrrho Investments Ltd v MWB Property Ltd and Ireland in Irish Bank Resolution Corporation Ltd & Ors v Quinn & Ors

What is predictive coding?

Predictive coding is essentially an algorithmic tool that utilises keyword searching to ‘train’ a computer to review a set of documents and identify those documents that are relevant to a matter. It involves the human review of an initial set of documents with the system applying what was learnt from the initial review to predict the relevance of documents in alternative sets. 

The key purpose of the predictive coding method is to identify and remove documents that are not relevant to a matter and to reduce the time it takes to conduct a manual review of documents. This translates into significant time and cost savings for clients, particularly in large-scale discovery. 

How predictive coding was used to review documents in the discovery process

The contract the subject of the dispute in McConnell Dowell Constructors (Aust) Pty Ltd v Santam Ltd & Ors generated in excess of 4 million documents, which was reduced to 1.4 million through the utilisation of de-duplication software. The Court was concerned that the manual review of documents in this quantum would have been exorbitant, with His Honour Vickery J suggesting that it would have taken a junior solicitor in excess of 10 years to review the documents, if one minute was spent on each. 

At the suggestion of a court-appointed special referee, the parties adopted a document management protocol and agreed to utilise predictive coding technology to assist in the review of documents. His Honour endorsed the report of the special referee and the agreement of the parties to use predictive coding in the discovery process.

The decision in McConnell Dowell Constructors (Aust) Pty Ltd v Santam Ltd & Ors also came ahead of a series of Practice Notes in the Victorian Supreme Court, designed to promote the use of electronic technology to facilitate the efficient conduct of litigation. 

The Federal Court has also since published Practice Notes which explicitly refer to the use of predictive coding. 

Risks of using predictive coding to review documents

Whilst the use of predictive coding is being praised for its ability to facilitate the efficient and cost effective review of large sets of documents, its effectiveness is dependent on the parameters set by the individual user. There are potential risks, including disclosing documents which are privileged or confidential in nature. Parties may also be subjected to adverse costs orders when large quantities of irrelevant documents are disclosed. It is therefore important to ensure that the review algorithm has been tailored to a reliable data set and that the initial review is undertaken by an experienced practitioner. 

Predictive coding could soon become common practice in Australian litigation 

Although the parties in McConnell Dowell Constructors (Aust) Pty Ltd v Santam Ltd & Ors were in agreement on the use of predictive coding, the Court’s endorsement of this type of technology to manage large-scale litigation with excessive quantities of data is clear. It is expected that other Australian Courts will follow the approach taken by Vickery J and that predictive coding will soon become common practice in Australian litigation. 

 

This article has been published by Colin Biggers & Paisley for information and education purposes only and is a general summary of the topic(s) presented. This article is not specific legal advice. Please seek your own legal advice for any questions you may have. All information contained in this article is subject to change. Colin Biggers & Paisley cannot be held responsible for any liability whatsoever, or for any loss howsoever arising from any reliance upon the contents of this article.​