Claire Levene
Solicitor
Lorem ipsum..
Director & Solicitor | Personal Injury & Medical Negligence
When AI Breaks the Seal: How ChatGPT Could Accidentally Destroy Legal Professional Privilege
In the legal world, few protections are as fundamental, or as fiercely guarded, as legal professional privilege. It is the invisible shield that allows clients to speak openly with their lawyers, confident that what is said in private will remain private. Without it, the justice system would struggle to function effectively.
But a recent court decision has highlighted a modern and unexpected threat to that shield: generative artificial intelligence.
As tools like ChatGPT become embedded in everyday professional life, many people are turning to them to summarise documents, sense-check advice, or draft follow-up communications. The convenience is undeniable. Yet for clients handling sensitive legal advice, that convenience may come at a steep cost.
The Foundation of Legal Professional Privilege
Legal professional privilege protects confidential communications between a lawyer and their client made for the purpose of giving or receiving legal advice. It exists to encourage honesty and full disclosure. A client who fears their communications could later be exposed might hesitate to provide the complete facts their lawyer needs.
The protection is powerful, but it is also fragile.
Privilege only survives as long as confidentiality is maintained. Once privileged material is voluntarily disclosed to a third party, the law generally treats that disclosure as a waiver of privilege. In simple terms, the protective seal is broken.
How AI Can Put Legal Professional Privilege at Risk
A recent court decision in UK v Secretary of State for the Home Department (AI hallucinations; supervision; Hamid) [2026] UKUT 81 (IAC) has brought this principle sharply into focus in the context of generative AI. The court held that sharing privileged client letters and decision letters from the Home Office into an open-source AI tool, such as ChatGPT, is to place this information on the internet in the public domain, and thus to breach client confidentiality and waive legal privilege. However, they did go on to say that closed-source AI tools which do not place information in the public domain, such as Microsoft Copilot, are available for tasks such as summarising without these risks.
Meanwhile, in New York, a Federal court adopted a similar analysis.
The implications are significant.
Many users view generative AI as a neutral productivity tool, no different from a digital notebook or drafting assistant. But legally speaking, entering text into a third-party AI system may be treated as disclosing it to an external party.
For privileged legal advice, that distinction matters enormously.
A Common Scenario
Consider a typical situation. A client receives detailed legal advice from their solicitor. The document is dense and technical. Wanting a clearer overview, the client copies the advice into ChatGPT and asks for a summary or a simplified explanation.
From the client’s perspective, this may seem harmless.
From a legal perspective, however, the client has just shared privileged material with a third party. If the court treats that disclosure as voluntary, privilege may be considered waived. Once lost, privilege cannot easily be restored, and the advice may potentially become disclosable in litigation.
The very protection designed to safeguard the client could be undone by a few keystrokes.
Convenience Versus Confidentiality
The rise of generative AI has created a new tension between efficiency and legal confidentiality. These tools excel at quickly digesting information and assisting with drafting tasks. But their use raises complex questions about data handling, third-party access, and the legal consequences of disclosure.
For clients and lawyers alike, the key issue is simple: privilege depends on secrecy. If privileged material leaves the protected lawyer-client relationship, even for convenience, that secrecy may be compromised.
Practical Lessons for Clients
The safest approach is straightforward: never paste legal advice into generative AI tools.
If clarification or simplification is needed, the better option is to return to the source — the lawyer who provided the advice. Lawyers can summarise, explain, or prepare follow-up documents without risking the loss of privilege.
Organisations should also consider updating their internal policies on AI use. Employees who routinely use AI for drafting or analysis may not realise that uploading legal advice could have legal consequences for the entire organisation.
A New Risk in a Digital Age
Generative AI is transforming how professionals work, including within the legal sector itself. Yet the fundamental principles of privilege have not changed. Confidentiality remains the cornerstone.
What has changed is how easily that confidentiality can be broken — often unintentionally.
In the past, waiving privilege might have required deliberately sharing advice with a third party. Today, it may happen with something as routine as asking an AI assistant to “summarise this document”.
Recent Case Law on AI and Legal Professional Privilege
On 26 March 2026 in the Court of Appeal of Ireland, Ms Justice Costello provided guidance in [2026] IECA in respect of the use of AI (see paras 72 – 76):
“(I) Parties are entitled to use AI to assist in carrying out research in respect of their case provided that they do so responsibly and do not, even inadvertently, mislead the court, by advancing propositions or relying upon supposed authorities which in fact have no foundation at all and are simply hallucinations.
(II) In all cases where they do so, they should expressly inform both the other parties and the court of their use of AI in this regard.
(III) A self-represented party is responsible for the ultimate written or oral work in their case just as much as the lawyers representing the parties are.
(IV) it is important therefore that any party who uses AI as part of their research independently verifies the accuracy of their submissions and the authorities cited as supposedly establishing the propositions advanced.
(V) No authority should be cited by a party who has not actually verified that it is a genuine Judgment of the court and that it is – or at least arguably is – authority for the proposition contended for”.
Whilst not binding in the UK, it is useful guidance which may well be adopted by courts in the UK.
The lesson for clients is clear: legal advice should stay between you and your lawyer. In the age of artificial intelligence, protecting privilege may require more caution than ever.
If you require legal advice, contact one of our team today for a free 30-minute consultation.
0151 306 3694 Online enquiry