Can solicitors in England use ChatGPT?
What the SRA, the Law Society and the courts expect.
Confidentiality, checking authorities and supervising AI work.
Yes: nothing in the SRA’s rules forbids it, and many firms use it daily.
Three duties come with it.
Keep client information out of unsafe tools, check every authority the tool produces, and supervise anyone who uses it.
Most AI advice online answers this question under American rules, which do not apply here.
Confidentiality comes first
In August 2026 the SRA issued a warning notice on the misuse of AI.
It said that entering confidential client information into public AI tools without safeguards may breach confidentiality and data protection duties.
The Law Society’s guide, Generative AI: the essentials, makes the same point.
It advises against feeding confidential information into tools you do not control.
Privilege adds a second reason for care, because a disclosure you did not intend can put it at risk.
In practice that means an approved tool on firm accounts, with a signed processing agreement and training switched off.
Our guide to putting client information into ChatGPT sets out the six checks.
Check every authority
In R (Ayinde) v Haringey [2025] EWHC 1383 (Admin), a pleading cited five cases that did not exist.
The Divisional Court called it wholly improper to put fake cases in a pleading.
It also said every lawyer has a duty to check the accuracy of research before relying on it.
AI tools invent authorities with complete confidence, and the invented ones look right.
Treat every case, citation and quotation an AI tool gives you as unverified until you have read it in a trusted source.
Supervise it like any other work
The SRA’s notice also named weak supervision of juniors and support staff using AI.
The standard does not drop because a chatbot drafted the first version.
A supervisor who signs off AI-assisted work owns it exactly as if a trainee had written it.
Where it helps without the risk
- First drafts from your own precedents, reviewed by the fee earner.
- Client intake and conflict-check paperwork, where a person still decides whether to act.
- Chasing clients and other parties for documents, approved before anything goes out.
- Explaining a process in plain English for a client letter, checked before it is sent.
Three rules for the firm
No client information in a tool the firm has not approved. No authority in any document until someone has read it in a trusted source. Supervisors sign off AI-assisted work as their own.
Write those rules down.
Our one-page AI policy template is a starting point, with a clause for regulated firms.
This guide explains the position; it is not legal or regulatory advice.
Frequently asked questions
Can a solicitor put client information into ChatGPT?
Not into a public tool without safeguards.
The SRA says that may breach confidentiality and data protection duties.
What happens if AI invents a case citation?
The lawyer answers for it.
The High Court called fake citations in a pleading wholly improper in Ayinde.
Does the SRA ban AI?
No.
It expects the same standards for AI-assisted work as for any other work.
Keep reading
Get started
See it on your own business
Opafex is live in beta.
Tell us your first agent’s job and we’ll set you up.