Detail of a dark still life with legal treatises, a quill, and manuscripts

Legal Professional Privilege in the Age of Legal AI

How legal professional privilege applies to enterprise legal AI, what recent commentary in England & Wales tells us so far, and how Harvey is built to support the confidentiality discipline that privilege requires.

Working on a US matter? Read our paper on attorney-client privilege and AI.

Download White Paper

Summary

As AI becomes part of everyday legal work, lawyers in England & Wales face a familiar question in a new context: does using AI risk waiving legal professional privilege? The English courts have not yet ruled directly on the point. However, recent commentary does offer early insights, with close attention to what kind of platform is used, whether confidentiality is preserved, and who can access the material.

This paper examines how legal advice privilege and litigation privilege apply to enterprise legal AI. It explores recent commentary from the judiciary, the Law Society and the Bar Standards Board, how Harvey’s design supports confidentiality, and the governance questions that arise as AI moves from analysis to action.

Written by Laetitia Nappert-Rosales, Associate Legal Counsel, EMEA, Harvey

Key Takeaways

  1. In England & Wales, confidentiality is in many ways the linchpin of privilege and privilege has adapted to cloud storage, email and e-discovery platforms by asking a key question: has confidentiality been maintained?. Large language models raise this question again.
  2. Recent judicial commentary distinguishes open, public-facing AI tools from closed, secure enterprise platforms. In Munir, the Upper Tribunal observed that putting client material into an "open tool" like ChatGPT risks waiving privilege. Whereas the Chancellor of the High Court has said it is “hard to see” how a lawyer’s use of a secure AI system could have an impact on privilege.
  3. The recurring questions are what kind of platform is involved, whether confidentiality is preserved, and who has access to the privileged information. Whether a tool is labelled “open” or “closed” matters less than how it is deployed. Training, human review, retention and contractual terms all count.
  4. Harvey is built for professionals, not general consumers. Beyond enterprise agreements with stringent contractual confidentiality protections, we offer authenticated workspaces, no training on customer data, no retention by AI model providers beyond ephemeral processing, no human review, customer-configurable retention settings and regional data processing in customer-selected regions.
  5. Privilege is highly context-specific and not a guarantee any vendor can provide in the abstract. Well-designed legal technology can make the discipline of preserving it easier to maintain.
  6. As AI enters the agentic era, governance must exist at the point of action. That means questions regarding which actions the system may take autonomously, which require human review, and which systems the tool can reach, will be key as this debate evolves.
Download White Paper