Active Learning Best Practice Guide

This consensus document has been prepared by a working group (setup by our ILTA Europe - Litigation Support Special Interest Group (SIG)), comprising of many of the Heads of eDiscovery from London's largest law firms, with further feedback incorporated from more than 50 sources across the eDiscovery industry heralding from other legal jurisdictions such as the US and Australia and with input from retired Judge, former Master Victoria McCloud.

The impetus for creating this guide stems from the recognition that while Practice Direction 57AD encourages the consideration of "software or analytical tools, including technology assisted review software and techniques" Section 9.6(3)(a), it does not provide specific guidance on how these tools and techniques should be applied in practice, and the developing field of Active Learning raises concerns as to the need for transparent guidance lest collateral significant disputes arise where parties adopt different or bespoke processes.

The purpose of this document is to establish a standardised approach that can be followed by all parties in a dispute in the courts of England and Wales. By doing so, we aim to encourage greater adoption of machine learning technology in the disclosure process, ultimately reducing both the cost and time associated with conducting disclosure.

Key aspects of the guide include:

• Explanations of Active Learning methodologies
• Guidance on when and how to end an Active Learning review
• Approaches to validating the results of an Active Learning process so that the disclosure review can be brought to an agreed end

We believe that this guide will serve as a valuable resource for legal practitioners litigating in English Courts, promoting consistency and efficiency in the use of technology-assisted review. By providing clear, practical guidance, we hope to address the current uncertainties that have hindered the widespread adoption of these powerful tools.