Disputes 101: Legal privilege, old principles, new challenges

The established principles of English legal professional privilege are deceptively simple to state, but harder in practice to apply. Never mind the sounding of a📣new-legal-development-klaxon📣 for Aabar v Glencore, and the vexed question of how to protect privilege when using AI.

This is the third blog post accompanying the 2026 Disputes 101 seminar series. 

Old principles

The test for English legal professional privilege, as we know, has two heads:

• Legal advice privilege which protects confidential communications between a lawyer and a client made for the dominant purpose of giving or receiving legal advice in a relevant legal context, and evidence of such communications.
• Litigation privilege which protects confidential communications made for the dominant purpose of obtaining information/evidence or advice in connection with the conduct of adversarial litigation that is in progress or reasonable contemplation.

For over 20 years now, legal advice privilege under English law has been artificially hemmed in by the Court of Appeal's decision in Three Rivers No 5 and its requirement that the “client group”, for privilege purposes, where a corporate is concerned, is only those individuals authorised to seek and receive legal advice (and not any/all employees).

For at least 10 years (if not since Three Rivers No 5) it has been clear that where a lawyer interviews an employee of a corporate client, for example to gather facts, where there is no litigation in reasonable contemplation, and that employee is not part of the Three Rivers No 5 “client group”, it is going to be very difficult to protect those notes as privileged as lawyers' working papers: see RBS Rights Issue.

In relation to litigation privilege, there are a regular supply of cases looking at whether a report was in fact prepared for the dominant purpose of litigation or for some other purpose. In a case last year, Krishna v Gowrie, the High Court re-emphasised that “the court must take a realistic and, indeed, commercial view of the facts” when discerning the purpose. This meant that a valuation report was found to be “part of the ongoing and developing set of hostilities” between the parties and was therefore privileged. Ultimately courts will look at substance over form, which cuts both ways.

New law: Aabar v Glencore and intra-client communications

There had already been one privilege development from the Glencore securities litigation: the ending of the shareholder rule, a longstanding, but somewhat obscure, principle which said that a company could not assert privilege against its own shareholders, save in relation to documents that came into existence for the purpose of hostile litigation against that shareholder. This was confirmed by the Privy Council in other litigation. 

Now there has been another. This one is likely to have a bigger impact (if upheld on any appeal). The court held that “intra-client” communications (i.e. communications between individuals who form part of the narrowly defined “client group”) can be privileged if they are for the dominant purpose of seeking legal advice. The court thought of them as “client working papers”. And they are a wider category than documents that disclose the substance of privileged communications, or so-called inchoate communications (e.g. a draft unsent email to a lawyer).

To quote from the judge “It follows, for reasons that I have sought to explain, that Glencore is entitled to assert legal advice privilege in respect of intra-client documents provided that those documents were created with the dominant purpose of seeking legal advice,… [and not only] insofar as the documents disclose the substance of a communication between members of the client group and a lawyer made for the dominant purpose of giving or receiving legal advice or the document was intended to be a communication to a lawyer but was not sent.”

The examples given by the court were:

• One member of the “client group”, the day before they are due to meet their lawyer, writing themselves notes for the meeting.    
• One member of the “client group”, who will not be attending a meeting with the lawyer, emailing another member of the “client group” with information or thoughts in preparation for the meeting. 

New challenges: AI and privilege

The rapid adoption of AI tools raises novel questions for privilege and disclosure. They are covered in detail on our sister blog on investigations: Protecting legal privilege when using AI. 

In essence the danger areas are:

• The fact that AI inputs and outputs are potentially disclosable in a dispute.
• The issue of whether you can maintain confidentiality when using an AI tool, confidentiality being a core requirement of privilege. Although there are arguments to the contrary, it is safest to assume that use of public consumer-facing tools will not meet this requirement. This means assuming that the inputs and outputs will not be privileged and that the use of such a tool will result in the loss of protection in any previously privileged material.
• The use of AI tools by non-lawyers for legal questions which it is safest to assume (although there are arguments to the contrary in certain circumstances) will not be protected by privilege.