Disputes in Focus: Quick Q&A on Developments in Forensic IT & AI

Quick Q&A with Jason Alvares and Chris Wheatley

Jason Alvares and Chris Wheatley, members of BCLP’s in-house Forensic Technology team, have over 20 years combined experience working in the eDiscovery and digital forensics industry. They have worked on a wide range of cases from asset tracing to multi-party and multi-jurisdictional fraud investigations to follow-on damages claims and commercial disputes, arbitration and litigation.

Jason and Chris work alongside the legal team to manage the processing, review, and hosting of data and documentation using our in-house eDisclosure platform and other tools. They help us ensure that the eDisclosure exercise is undertaken as efficiently and effectively as possible. They do this by using advanced analytics and artificial intelligence.

What has been your recent experience of emerging trends and challenges in eDiscovery?

In our view, one of the main emerging trends is simply the increasing amounts of different types of data/documents that need to be collected for potential disclosure in litigation and investigations.

This became overwhelmingly clear over the course of the pandemic, with the increased use of instant messaging platforms such as WhatsApp, Slack, Microsoft Teams and Skype. Each of these platforms require different techniques to collect the data, and the data types produced are different and often incompatible with each other. This means we spend more time working with our client to ensure that this data is collected in an acceptable format so it can be searched and reviewed alongside other documents within a case. As more and more platforms and applications hit the market and gain popularity, this issue is only going to become more prevalent when scoping and estimating timelines for disclosure processes.

Technology assisted review – what recent developments are making it easier to handle huge datasets?

Technology assisted review, or TAR, is something that has come on leaps and bounds over the last few years, and can result in significant cost savings compared to a traditional manual review. It ranges from the use of keywords and concept clustering, through to the use of artificial intelligence and predictive coding.

In terms of recent developments in TAR, we would highlight the widespread and consistent use of continuous active learning (or CAL for short).

We were part of the BCLP team that secured a winning judgment defending our clients last year in a significant noise nuisance and breach of contract claim. In this case, we collaborated closely with the legal team, clients and third party custodians to understand key issues to aid with document collation. We then utilised a variety of technologies, including CAL, to work on large data sets to narrow down and focus in on documents for disclosure.

CAL is a form of predictive coding where a model learns the themes and topics of documents that are likely to be relevant or not based on decisions by a human reviewer. The model can then be applied to identify potentially relevant documents and (i) may remove the need to review an entire set (most useful in the largest of cases) or (ii) prioritise and re-order documents for review within a set, which may lead to a shorter or speedier review.

CAL is also great for quality checks over a review. CAL can be run over an already completed review set, and it can be used to return what we would term as “Conflict” documents. These are documents that are coded as relevant, that the model would deem irrelevant, or vice versa. This reduces the risk that a document may have been overlooked or miscoded, and is a great way to ensure the quality of the review. 

What are your practical tips for those in client teams managing a document retention and collection process?

Our tips to help the process run smoothly and hopefully make your/the responsible person’s life much easier:

1. Have a good understanding of your company’s IT architecture and consider the data sources that are always going to be required, if there is an investigation or litigation launched. The obvious ones are emails and documents from shared drives. But consider other forms of communication and data types such as databases, mobile phones and instant messaging.
   • Can your IT team prepare (and maintain) a summary for you, which answers questions like: What service centrally manages your emails? Are instant messages backed up in the same way as emails? What is the policy for retention of data after an individual leaves the company and how long has that policy been in place?
2. With the growing data volumes, and the ever expanding range of data that has to be collected, it is worth ensuring that you or your IT team has a mechanism for exporting these volumes of data to your eDiscovery experts. Examples of this may include a secure file transfer system with a large capacity, or it could be as simple as an encryptable hard drive that the data can be loaded onto and delivered.
3. Without delay following commencement of an investigation or litigation, introduce the eDisclosure experts to your IT teams. This will allow for better collaboration and understanding with regards to managing the scale of the volume of data as well as finding a suitable procedure for obtaining and transferring the data.