Today, a special interest group (SIG) of the International Legal Technology Association (ILTA) released a draft best practices guide for the use of active learning (AI-based machine learning techniques) in eDiscovery and eDisclosure management. The group is seeking feedback from professionals across all industries.
The proposed guide was compiled over a six month period by litigation support experts from Norton Rose Fulbright, Eversheds Sutherland, Squire Patton Boggs, A&O Shearman, Linklaters, Ashurst, Fieldfisher, DLA Piper, Eltemate (a Hogan Lovells Technology Company), Travers Smith, RPC and CMS, with the following objectives:Provide a technology-agnostic blueprint for implementing active learning projects most efficiently and equitably'The group said.
By way of background, the SIG explained: “The practice of electronic discovery (or e-discovery) in England and Wales is primarily regulated by court rules and practice guidelines, which encourage parties to consider the use of e-discovery software tools to search for documents using keywords, email threads and other traditional techniques. Technology-assisted review has been used effectively in recent years. However, The use of AI in “active learning” is not subject to specific guidance., It could have a negative impact on the judiciary Where technology is inconsistent across parties or across industries.
James MacGregorThe MD of Ethical E-Discovery, who has played a key role in the project, told Artificial Lawyer that, at least in the UK, “there are no specific rules on the use of AI in e-discovery, or so-called active learning. ” “You're supposed to use technology for review when you have more than 50,000 documents, but the rules don't say how to use machine learning.” [i.e. Active Learning systems].
However, he added that the consultation will not focus specifically on new generative AI tools, but rather on AI approaches with NLP machine learning at their core, which have been around for several years.
Meanwhile, a retired High Court judge Victoria MacLeodCommenting at a recent ILTA conference, “It is essential that e-discovery technology is efficient, consistent, open and defensible. In the face of the developing field of AI in law and court case backlogs, If adopted, this guide will reduce conflict between opponents over the technicalities of active learning.This will relieve some of the burden on courts to resolve disputes regarding e-discovery mechanisms.”
She adds: “Active learning in particular distinguishes itself from traditional search techniques because of its self-regulating capabilities. The power of such processes must be balanced with clarity. This guide is an important step in that direction.”
The SIG hopes that the final version of the document will be approved by the Master of Courts of England and Wales and the Civil Procedure Rules Board. However, McGregor added that many countries lack clear rules on the use of active learning/AI, so if approved, the guide could be useful at a global level, including in the United States, the world's largest eDiscovery market.
The current version of the document is available for peer group review and feedback. The group has advised that to access this draft document, please contact: email address:
The consultation will close on Friday 20 September 2024, after which the document with redactions resolved will be submitted to the UK's Archivist.
—
UPDATE: Speaking of ILTA, Artificial Lawyer founder Richard Tromans will be speaking on a panel about law firms' use of AI at ILTACon in Nashville this August.
