October 5, 2017
North East law firm for businesses, Muckle LLP, is one of the first firms in the region to adopt artificial intelligence (AI) technology to help speed up the disclosure process on a multi-million pound dispute claim.
The dispute resolution team at Muckle, headed by partner Susan Howe, is using predictive coding – also known as technology assisted review (TAR), to analyse hundreds of thousands of disclosure documents linked to an ongoing multi-million pound claim.
The first case initially involved one million disclosure documents, which was cut to 660,000 after a standard keyword search, and then the use of this innovative technology enabled Muckle to reduce that number down by more than 90 per cent to 35,000.
Traditionally, the documentation would have to be manually analysed and assessed for relevance by members of the team, but TAR removes these first few relevancy stages, in turn saving time and the associated costs.
Susan says: “Using this type of technology is relatively new and is now being recognised by the court. Firms in London are investing heavily in AI systems but, regionally, it’s not widely used yet.
“The dispute resolution team at Muckle increasingly works on large claim cases where disclosure documents can be in the hundreds of thousands, so having access to groundbreaking software such as TAR is game-changing.
“A process that would traditionally involve a team of trainees or paralegals turning the pages in a room for a year is now significantly condensed thanks to this cutting-edge computer programme.”
Predictive coding is the use of keyword search, filtering and sampling to automate portions of an e-discovery document review. The aim is to reduce the number of irrelevant and non-responsive documents that need to be reviewed manually.
There is an initial level of investment required from the legal team to teach the program keywords and test the algorithm on a set of example documents. The algorithm analyses the characteristics of the documents and learns from the lawyers’ decision-making to enable it to then identify similar documents and rank them in relevance.
Susan adds: “There is a lot of work involved in training the computer, but once this is done, it is incredible. And while the cost of using the technology is significant, in the context of substantial claims like our current one, it is worth the effort and the expense to be able to drop 75 per cent, or more, of your documents in a single swoop.
“As a litigator, it’s understandable to feel uneasy having to rely on technology, but once you use the system, and understand you’re the one training it, then it becomes very easy to trust. The court is now encouraging the use of this type of technology so we will definitely use it again on future large claims,” she concludes.
The use of predictive coding in the disclosure process was given judicial approval for the first time in the UK last year in the case of Pyrrho Investments Ltd and another v MWB Property Ltd and others .