Skip to content
SafeLegalAI

72 on the record · +1 this week

guidance diffSRA & UK RegulationAI Governance

UK legal-profession AI guidance: version histories for solicitors and barristers

The version record of the SRA, Bar Council, and Law Society guidance on generative AI — what each says now, and when it changed.

Daman Kaur

Three bodies guide how lawyers in England and Wales may use generative AI: the Solicitors Regulation Authority, the Bar Council, and the Law Society. Their documents are separate from the judiciary’s guidance for judges, and each has its own revision history. This is that record.

None of the three prohibits generative AI. All three land in the same place: the lawyer is accountable for the output, must verify it, and must keep confidential client information out of public tools.

Solicitors Regulation Authority

The SRA’s “Compliance tips for solicitors regarding the use of AI and technology” carries a formal “Updated” date of 9 February 2026. An earlier version carried an “Updated 15 February 2022” date and was still live in the 18 October 2025 archived capture.

Version Date Status
Earlier version Updated 15 February 2022 Superseded
Current 9 February 2026 Current

The current version restates the SRA Standards and Regulations rather than adding AI-specific rules. Its key points: using AI does not dilute a solicitor’s personal responsibility for the work; AI output must be verified, given the risk of hallucinated or fake citations; no identifiable client data should go into an AI tool without informed consent; it should always be made clear to clients where they are interfacing with AI; and the COLP and firm governance framework must remain fit for purpose. The SRA separately published a Risk Outlook report on AI in the legal market.

Bar Council

The Bar Council’s “Considerations when using ChatGPT and generative AI software based on large language models” is ethics advice, not formal guidance under the BSB Handbook. It has two substantive versions.

Version Date Status
v1 30 January 2024 Superseded
v2 25 November 2025 Current

Both versions require barristers to verify all LLM output, to keep privileged and confidential client information out of the tools, and to retain full personal accountability. The November 2025 update added references to newer tools, refreshed the document for recent High Court case law on AI misuse, and expanded the list of risks (including training-data bias and confidential data entering training sets).

Law Society

The Law Society’s “Generative AI – the essentials” is a living guide, revised repeatedly rather than re-issued as clean numbered versions. Its page does not render a stable revision byline, so the intermediate dates below are revision windows bounded by archived captures, not exact edit dates.

Version Date Status
Original c. November 2023 Superseded
Revision 2024 Superseded
Revision c. March 2025 [approx.] Superseded
Current June 2026 Current

The through-line is unchanged: solicitors remain accountable for AI output and must verify its accuracy. The June 2026 revision is notable for what it now cites — the referral of two solicitors to the SRA over incorrect AI-generated information submitted to court in Rodney v Gee’z Micro Bar & Pitstop. The profession’s own guidance now points back at the enforcement record.

Why the version record matters

The consistent core — accountability, verification, confidentiality — has not moved since 2023. What has changed is context: each update folds in newer tools and, increasingly, the actual disciplinary cases. Guidance that starts citing named referrals is guidance responding to a live enforcement trend, not a hypothetical one.

Sources