Document Categorization In Legal Electronic: Benefits And Effectiveness Of Technology Assisted Review

TAR in the management of commercial litigation

Discuss about the Document Categorization in Legal Electronic.

An interplay of humans and computers, Technology Assisted Review (hereafter, TAR) in the management of commercial litigation, is viewed as an empowering, modern, cost-effective and efficient alternative to the exhaustive manual and time-consuming manual review of legal documents[1].    

This paper is contextualised with an understanding of what TAR is in order to critically examine the benefits and effectiveness and weighs them against the limitations that the technology presents.                 

The decision judgment handed down by Vickery J in the case of McConnell Dowell Constructors (Aust) Pty Ltd v. Santam Ltd & Ors (No 1) [2016] VSC 734 (McConnell) is considered amongst a number of cases and incidents that TAR was utilised. It will be critically discussed and argued that the benefits that flow from the use of TAR in civil litigation do outweigh the problems that will be identified in this regard and superior results are accomplished through the use of the technology.

TAR extensively relies on computers for purposes of carrying out an electronic review of documents.  It relies extensively on predictive coding. This is a type of coding technique that automatically makes predictions based on the needs of a researcher.[2] TAR has been recognised as a process of document discovery, whereby it is possible for a human to look to interact with a computer program with the intention of then locating documents that are relevant to a particular case.[3]  

TAR takes the input from humans regarding documents that are labelled relevant or non-relevant and draws deductions vis-a-vis other similar documents. The computer software will finally order the number of documents chosen for review by relevance so as to then be able to guide the process of review. 

The Courts opinion in McConnell Dowell Constructors (Aust) Pty Ltd v. Santam Ltd & Ors (No 1) [2016] VSC 734

Justice Vickery's decision in the case of McConnell could be said to reflect a growing trend by international jurisdictions to use TAR to simplify the document discovery process in cases of civil litigation. His Honour endorsed the idea of using predictive coding in the discovery process.  In the view of Justice Vickery, arbitration normally involves issues pertaining to the implementation of a contract. Therefore, use of technology to de-duplicate will help in the reduction of about 4 million documents to be read, to about 1.4 million. This is an aspect that was identified in the Irish Bank case[4]. In the view of the High Court, TAR is an accurate method of identifying documents. Consequently, Justice Vickery recognised that the time and expense involved with using traditional methods to review this kind of material would be prohibitive. As a result, there was a need to look to use TAR to reduce the volume of material to be reviewed to a far more workable number of documents in the thousands.

The Supreme Court's decision in this case is significant because it was the first in Australia to order TAR's use to reduce the time and effort otherwise typically involved with discovering documents in cases of civil litigation (see Vickery, 2015 and Warren, 2015 – see also Money Max Int Pty Limited (Trustee) v. QBE Insurance Group Limited [2016] FCAFC 148). In so doing, the decision of the Court in this case in Australia followed several other major western jurisdictions, including the UK[5], the Republic of Ireland[6] and the United States of America[7].

The decision in this case, arose after a number of Practice Notes which endorsed the use of electronic technology for purposes of ensuring that civil litigation is effective. A good example is Practice Note SC Gen 5, which recognised the use of technology in matters of civil litigation. From this Practice Note, it was observed that technology plays an important role in modern communication, and a large number of communication processes are now carried out through the use of technological tools, and this includes the storage of communication documents (Supreme Court of Victoria, 2017).    

Consequently, this Practice Note looked to expressly encourage the use of electronic documentation in civil litigation so that the use of ‘hard copy' will be the exception that needs to be justified (Supreme Court of Victoria, 2017). This is because, in the large cases in particular, it is believed that TAR will usually be an accepted method for completing document discovery in civil litigation that a court may order regardless of whether the parties involved actually consented to it (Supreme Court of Victoria, 2017 – see also Federal Court of Australia, 2016).

Due to the extensive development of Australia as a commercial hub, the courts and practitioners have a number of opportunities,[8] so the legal system must operate in an efficient way to cater for the needs of individuals and businesses. The benefits considered in regards to TAR's use in cases of civil litigation include significant time and cost savings.[9] It has been found that the amount of data to be reviewed in litigation is significantly increasing on an annual basis meaning that there is a need to look beyond the traditional review process to reduce both the time and costs involved for the end client.    

It was also discovered that cases that have used analytics went through four times as many documents as those cases not utilising the technology.[10] It is understood that TAR's use may serve to significantly reduce how many documents need to be reviewed in a particular case. This means that it is easier to efficiently and quickly review legal documents that are useful to a particular case[11]. This in turn helps the courts to implement the provisions contained in section 56(1) of the Civil Procedure Act. This section requires the courts to facilitate a quick and fast resolution of disputes.

Note that, predictive coding helps to identify the most relevant documents to use in litigation, and avoids an overlap of documentations, a process that is also known as Jaccard index. This is an index where independent assessors are able to identify a number of documents as relevant.[12] Note that, when there is an overlap between the independent assessors, chances of a delay occurring are high, and this may be expensive to the litigants, thus frustrating the intentions of using TAR. Moreover, this can affect the accuracy of getting the most needed documents for review.    

Additionally, concept clustering, where technology clusters similar types of documents together and disregards irrelevant clusters of documents can aid in achieving the overarching purpose. Outstandingly, despite the fact that the use of TAR can increase the speed with which the document discovery process in completed, nothing is lost in terms of accuracy when it comes to those documents that are relevant in a specific instance (as long as the process has been appropriately set up).

Benefits of using TAR in civil litigation

This is given to the fact that unlike a human assessor that could potentially suffer review fatigue, TAR does not experience such weaknesses and chances of it being consistent and reliable are high. Additionally, lawyers who use the TAR system can benefit from the conceptual searches that TAR allows. It will enable them to get relevant documents that can enable them understand the concepts of the case they are handling.  TAR can create greater uniformity between courts as well as ensure public confidence in the legal system, as it will seek to reduce waste of public resources arising from adjournments and settlements in situations where parties simply cannot afford trial.

Technology assisted review has the capability of helping marginalized litigants to ensure that they are able to have an access to justice. There are a number of ways that TAR assists marginalized litigants. One such was is the capability of TAR to save legal costs. Because TAR uses technology, it is easier for litigants to have an access to a number of legal information that are relevant to litigants, and this is because they are generated electronically.

This saves time and costs of accessing legal aide, since, it avoids the use of human reviews for purposes of having an access to documents that can help in a litigant’s case. Human reviewers normally charge a high price for their services; thus, TAR assists in eliminating them.  

Note that, there are a variety of ways that marginalized litigants can use TAR to assist them in their legal cases. Nonetheless, TAR uses an algorithm that searches through a large amount of data, for purposes of getting relevant information that can be useful in building a case or defending oneself. It is very fast and more accurate than the use of a manual review. On this basis, TAR helps in saving time, which can be used to focus on other areas that are of interest to a marginalized litigant. 

TAR is very useful in reducing the legal costs of business organizations. For instance, the application of TAR begins with the review of an expert, who has extensive knowledge of the case that the business organization is facing. The expert will then code the documents and come up with an algorithm which will enable the business organization to have an access to thousands of data that can help in their case. This process is very efficient, and has the capability of saving up to 80% of the legal costs associated with business litigation.

Therefore, the capability of a business organization to have an access to millions of data, through the use of technology is the main reason that TAR helps in reduction of business legal costs. Moreover, the reduction of business legal costs cannot create an increase in litigation cases, because it is cheaper. This is because it will help to reduce conflicts that may lead to litigation, because of the ease in which people will have an access to justice.      

Problems associated with TAR's use in cases of civil litigation include the fact that its use needs to fulfil the standard of reasonableness that is associated with discovery's production of documents to produce viable results. [13] The use of TAR is also somewhat flawed because it is only as reliable and reasonable as the human element involved so that the costs and time taken could be significantly increased.[14] In addition, although the TAR process may be quicker than more traditional review processes, estimating the number of training rounds TAR requires producing viable results and/or lower the document count in a particular case.[15] TAR also has limitations regarding the type of documents it can review and analyse.     

It cannot be used for some old hard copy documents or data sets, including drawings or spreadsheets.[16] This may appear due to the fact that the copy is too poor or theoretically void for the system to be trained reliably. Finally, there is a need for senior lawyers to review the results that are generated. This is because, unlike a linear review that may be able to rely solely upon junior lawyers, TAR's use needs documents to be reviewed by an expert in the field with the particular law firm in question (i.e., the senior lawyer).[17] This final step in using TAR may appear problematic for senior lawyers since it may be necessary for them to allocate further time to evaluate the results generated in a given case.   

Another problem that may arise out of the use of TAR is the notion that a large percentage of documents may be left unreviewed. This is because TAR prioritizes a review, based on a certain algorithm, and this may result to ignoring a large percentage of documents, that may be useful to the litigant.

Conclusion

Whilst it is clear that there are a number of particular benefits that are associated with the use of TAR in cases of civil litigation involving significant numbers of documents, there are also a number of notable disadvantages that are associated with the use of TAR that could impact upon the effectiveness of the process. However, it must be noted that a number of Australian and International judges have made positive extra-judicial comments about the technology and a number of organisations have publicly expressed enthusiasm for predictive coding in regulatory matters. Vickery is also noted in a leading case of VICBAR CPD as denoting that an important factor that a modern day judge should consider is reducing the time of hearing a case, and eliminating delays[18].

It has also been proven through facts and statistics that TAR is efficient and does not prolong the process of examining documents. With the proper training and legal knowledge, human judgment, strong judicial management, technological literacy and open mindedness, TAR can help in eliminating delays and save money during the process of legal litigation.      

Given that TAR's use has been recognised as being efficient, cost effective, and able to produce just results and be fair to litigants through its consistency, responsiveness and timeliness, the fact that it is still reliant on a human element should not rule out TAR's use to aid the achievement of justice in a given case. Moreover, future research and work on the use of TAR should address issues such as how technology can help in improving the manual review of legal documents[19].

References

Civil Procedure Act 2005 (NSW)  Uniform Civil Procedure Rules 2005 (NSW)

Federal Court of Australia (2016) Practice Note – Technology and the Court Practice Note, 25 October.

Da Silva Moore v. Publicis Groupe et al (2012) 287 F.R.D 182 Irish Bank Resolution Corporation Limited & Ors v. Sean Quinn & Ors [2015] IEHC 175.

Butterfield, W. P., Crowley, C. R. and Kenney, J. ‘Reality Bites: Why TAR's Promises Have Yet to be Fulfilled’ (DESI V: Workshop on Standards for Using Predictive Coding, Machine Learning and Other Advance Search and Review Methods)

Cormack, Gordon and Maura Grossman, Engineering Quality and Reliability in Technology Assisted Review (2016).

Douglas, Oard et al., Evaluation of information retrieval for E-discovery (2010), 347 (365), ARTIFICIAL INTELLIGENCE & LAW 347

Herbert L. Roitblat et al., Document Categorization in Legal Electronic Discovery: Computer Classification vs. Manual Review, 61 J. AM. SOC’Y. FOR INFO. SCI. AND TECH. 70, 70 (2010).

Paskach, Chrstopher, Eli Nelson and Mathew Schwab, The Case for Technology Assisted Review and Statistical Sampling in Discovery Position Paper for DESI VI Workshop (2015).

Technology and the Law Committee (2017) ‘Technology Assisted Review’ (Law Tech Essentials, Law Institute of Victoria, June)

John Tredennick, TAR for Smart People: How Technology Assisted Review Works and Why It Matters for Legal Professionals’ Catalyst (2015).

Marilyn, Warren, A vital commercial hub in the Asia Pacific region: Victoria – a commercial hub (2015).