Key Takeaways
• ILTA has introduced a Generative AI Guide to regulate how AI is used in legal disclosure processes within England and Wales.
• The guide addresses an urgent legal gap in Practice Direction 57AD, which currently lacks guidance on GenAI usage in disclosure.
• Experts warn that undisclosed and unstandardized AI tools could lead to biased outcomes and costly legal disputes.
• The guidance will undergo public consultation and is expected to be finalized by September 2025.
The International Legal Technology Association (ILTA) has released a new Generative AI Guide designed to provide a governance framework for using generative artificial intelligence (GenAI) in legal disclosure.
The initiative responds to a pressing legal challenge: the rapid adoption of AI in litigation without clear standards or oversight.
The guide specifically targets litigators operating within England and Wales’ Business and Property Courts, addressing the lack of direction in Practice Direction 57AD, which governs disclosure procedures but does not currently include guidance on AI-assisted reviews.
Background: A Legal System Facing AI-Driven Complexity
The integration of GenAI into legal workflows has accelerated in recent years, offering both efficiency and complexity. In traditional disclosure, parties are required to share relevant evidence.
However, as firms increasingly use AI to review and categorize massive volumes of documents, the need for standards on validation, transparency, and oversight has grown critical.
— Dr. Victoria McCloud, former High Court Judge and litigation consultant
Guide Development and Expert Collaboration
The Generative AI Guide was developed collaboratively by a panel of industry leaders:
• Co-chaired by Fiona Campbell (Fieldfisher) and Tom Whitaker (Burges Salmon)
• Contributions from David Wilkins (Norton Rose Fulbright), James MacGregor (Ethical eDiscovery)
• Additional expertise from DAC Beachcroft’s Jonathan Howell, Jamie Tomlinson, and Imogen Jones
The document builds on the 2024 Active Learning Guide—a document submitted to the Master of the Rolls, which is still under review—and is intended as a companion rather than a replacement. Its goal is to bring consistency to how GenAI is applied in the legal discovery process.
Core Areas Covered in the GenAI Addendum
The guide outlines best practices for using GenAI in disclosure, emphasizing structured workflows, validation, and accountability. It offers clear recommendations in the following areas:
• Appropriate use cases for generative AI in the disclosure process
• Techniques for prompt design and human oversight of AI outputs
• Validation strategies including elusion testing, and precision/recall metrics
• Integration with existing Active Learning workflows
• Transparency and documentation standards for auditability
— Fiona Campbell, Fieldfisher
Legal Implications: Toward Standardization and Reduced Costs
The use of proprietary AI models without regulation may lead to unfair advantages, especially when each party in litigation uses different tools with unknown biases or operational logic.
By setting industry-accepted rules, the guide aims to limit disputes around technical procedures and cut down on the time and cost associated with challenging the validity of AI-reviewed disclosures.
— James MacGregor, Chair, ILTA Litigation Special Interest Group
The official launch of the Generative AI Guide is scheduled for May 6, 2025, during an ILTA-hosted event at Fieldfisher’s London offices.
A panel including the guide’s contributors will present the framework and discuss its potential legal and operational impact.
The document will then enter a public consultation phase, during which the broader legal and tech communities can submit feedback. The final version is expected to be published in September 2025.
Spendesk & Dust Partner on Scalable AI Agent Rollout Poppins Raises €5M for Dyslexia Therapy App
For more news and insights, visit AI News on our website.
