ANALYTICS AND MACHINE LEARNING FOR E-DISCOVERY AND INVESTIGATIONS
CD: To what extent are analytics and machine learning (ML) tools being used to assist with e-discovery and corporate investigations? Are you seeing rising awareness of, and demand for, these technologies?
Chu: In Asia, lawyers have preferred using a traditional linear review approach for two key reasons: linear review is an easy-to-understand concept and teams have predictability in terms of the number of documents required for review. However, case data volumes are growing – we are regularly seeing matters with more than one terabyte of data – and traditional linear review is no longer efficient when dealing with large volumes. This is driving some progress toward the use of analytics and machine learning (ML) among e-discovery teams in this region. The most common use case that has been accepted here is email threading, as legal teams can understand how it works and recognise the efficiencies it brings to a review. There is also an increasing openness to predictive coding to help prioritise the review, from most likely to be relevant down to least likely. In any case, expertise is still a critical factor in helping the lawyers feel comfortable with how underlying algorithms work.
Kippin: There is a growing interest in analytics and ML tools, and they can deliver significant benefits in terms of time and cost savings, and strategically prioritising documents for review. However, there is still a level of expertise required to use these tools properly, so that results are accurate and defensible. The practitioners and teams who are seeing significant benefits in reducing data sets and the time needed to find facts are typically trained in the technology and understand how the algorithms and statistics work. As demand increases, legal teams will need to grow their technology proficiency and involve experts who know how to implement and optimise analytics to conduct e-discovery faster and more cost-effectively.
Oct-Dec 2021 issue
FTI Consulting