SRA Guidance on AI for Solicitors: What It Requires

Does the SRA allow solicitors to use AI? Yes, the SRA has never banned it, and on 9 February 2026 it updated its compliance tips confirming that its existing Standards and Regulations, not new AI-specific rules, govern how you use it. The guidance is not a green light for any particular tool and it is not a ban. It is a reminder that the responsibility never leaves the solicitor.

Firms keep asking us for "the SRA’s AI rules" as if there were a separate rulebook to comply with. There isn’t. What exists is guidance that maps AI onto duties you already hold. Once you see it that way, the compliance job becomes concrete and finishable.

Where the SRA’s AI guidance actually lives

Two documents carry it. The first is the SRA’s Risk Outlook report, "The use of artificial intelligence in the legal market," published on 20 November 2023. It groups the risks under seven headings, bias, errors, scale, confidentiality and privacy, accountability, regulatory divergence, and crime, and notes that "errors" includes hallucination, where a model produces plausible but false information. The same report recorded that three-quarters of the largest solicitors’ firms were already using AI, roughly double the figure three years earlier.

The second is the SRA’s "Compliance tips for solicitors regarding the use of AI and technology," updated 9 February 2026. It does not add new obligations. It restates the ones that apply: keep the client’s best interests central, protect confidentiality, do due diligence on third-party platforms, run risk and impact assessments before you deploy, and be clear to clients "where they are interfacing with AI." The report is explicit that the SRA regulates outcomes, "not necessarily the specific systems" firms use to achieve them.

Why this stopped being theoretical

On 6 June 2025 the High Court handed down a judgment in Ayinde v London Borough of Haringey and Al-Haroun v Qatar National Bank. In the first, five of the cases cited in the grounds for judicial review did not exist, they had been generated by AI. In the second, 18 of 45 authorities cited were fabricated, and others were misquoted or irrelevant.

Dame Victoria Sharp, President of the King’s Bench Division, held that generative tools such as ChatGPT "are not capable of conducting reliable legal research" and that lawyers have "a continuing professional duty to check accuracy against authoritative sources." The court imposed a £2,000 wasted-costs order in the Ayinde case, referred the solicitors to the SRA and the barrister to the Bar Standards Board, and set out a ladder of possible sanctions running from wasted costs through contempt of court to prosecution for perverting the course of justice. That is the enforcement reality behind the polite word "accountability."

What the SRA actually requires of you

Strip the two documents and the case down and four duties do the work:

• Confidentiality. Client information must stay confidential and securely stored. The specific trap is a fee-earner pasting client material into a free, public tool that may retain or train on it. That is a confidentiality breach whether or not anything goes wrong afterwards.
• Competence. You must understand a tool well enough to review its output critically, including recognising when it is wrong. "The AI said so" is not a defence, and unchecked citations are exactly what the High Court sanctioned.
• Accountability. A solicitor owns the work regardless of the tool. The SRA expects the Compliance Officer for Legal Practice (COLP) to be responsible for regulatory compliance when new technology is introduced, and firms should record material AI use, including to their insurer.
• Client transparency. It should be clear to clients where they are dealing with AI, with an alternative for those who are unable or unwilling to.

Turning the guidance into something a firm can sign off

This is governance, not a technology project. In practice it is three moves: name the approved tools and ban the rest for client work, write a one-page AI policy fee-earners will actually follow, and make sure supervision and engagement terms reflect it. The most common gap we find first is shadow AI, staff already using unapproved consumer tools on live matters, which is precisely the confidentiality risk the SRA flags.

For the practical question of which tools are safe and how fee-earners can use them day to day, our companion guide on SRA-compliant AI for law firms covers what is allowed and what is not.

We work directly in SRA contexts and build the workflows around how your fee-earners actually operate. Our 90-minute AI readiness audit tells you, in plain English, which uses are worth it and how to keep the regulator comfortable. Book a call to talk it through.

General information for UK firms, not legal or compliance advice. Verify your obligations against current SRA guidance and, where needed, take professional advice. Last updated: 16 July 2026.