The UK’s immigration tribunal has commented on the risk that documents covered by legal advice privilege could lose their privilege if uploaded into a generative AI tool. This decision raises several questions for employment tribunal litigation, with AI increasingly used by both sides in employment claims.
Legal professional privilege
Legal professional privilege protects two types of communications:
- Confidential communication between a lawyer and their client, for the dominant purpose of giving or receiving legal advice (“legal advice privilege”); and
- Communications between lawyer and client, or either of them and a third party, made for the dominant purpose of pending, reasonably contemplated or existing litigation (“litigation privilege”).
Privilege cannot apply to documents which are no longer confidential. For example, if a document has been posted online, its privilege has been waived, because the document is publicly available and has lost its confidentiality. Privilege can be waived intentionally or inadvertently.
Loss of privilege due to AI use
The recent Immigration and Asylum Chamber decision in UK v Secretary of State for the Home Department [2026] UKUT 81 (IAC) considered the consequences of a solicitor’s use of AI tools to help draft letters or summarise Home Office decision letters.
In its judgment, the Tribunal stated (our emphasis): “We also observe that to put 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… 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.”
The Tribunal did not expand on its observation. Its use of the terms “open” and “closed” source are potentially confusing, as these typically refer to how the software’s code has been licenced, not whether the actual tool itself stores information in a private manner.
It is more likely that the Tribunal was highlighting the difference between the public versions of generative AI tools and private enterprise versions. The former will often collect and reuse uploaded information to train their algorithm and respond to other users. Enterprise versions usually provide a closed platform with higher data security, where uploaded information is not shared outside the business or reused for training.
The Tribunal’s decision, at least when applied to public AI tools, is consistent with the guidance issued to UK judges on AI use, which advises that “Any information that you input into a public AI chatbot should be seen as being published to all the world.”
There is still some risk that privilege could be lost when using enterprise versions, depending on the specific terms and conditions governing how uploaded information will be used. However, in a recent speech to the City of London Law Society on this issue, the Chancellor of the High Court reassured that “assuming the lawyer has used a secure system, I would suggest that it is hard to see how that could have an impact on privilege.”
Implications for employment disputes
This recent immigration case would not bind an employment tribunal in Northern Ireland or elsewhere in the UK. The judgment considers the requirements for maintaining a client’s legal advice privilege, which is likely to be a larger concern for lawyers than their clients.
In employment disputes, there is a more immediate question on whether legal advice privilege is lost if the circumstances are reversed, with a client (whether employee or employer) inputting legal advice they have received into a public AI tool.
It is difficult to see how legal advice privilege could be maintained over the original advice in this scenario. There is the further potential that the resulting material generated by the public AI tool will not be privileged either.
There are many foreseeable scenarios where these issues could occur. For example, an employee may receive legal advice about a potential constructive unfair dismissal claim. They may then decide to represent themselves and rely on a public AI tool to build on that advice, by asking it to summarise the advice, assess whether they should resign and prepare a resignation letter.
On the other hand, a HR team may have received legal advice on a dismissal, which they have shared with the manager conducting the disciplinary process. The manager may then upload the advice note onto a public AI tool (potentially in breach of the organisation’s policies) to create a meeting script or to consider whether they should dismiss, or how they could justify a decision to dismiss.
It is likely that parties in employment disputes will regularly seek disclosure of prompts or documents uploaded to AI tools, to test whether there are any discoverable documents. Disclosure is, of course, subject to general requirements around relevance, and there are other forms of privilege which can apply, including litigation privilege. Parties will need to keep the potential issues in mind when using AI tools before and during employment tribunal claims.
For legal guidance and advice regarding AI Policies or Employment Law queries, please contact Jack Balmer or another member of our Employment Team for more information.
While great care has been taken in the preparation of the content of this article, it does not purport to be a comprehensive statement of the relevant law and full professional advice should be taken before any action is taken in reliance on any item covered.
