Aipedia: Solicitors and law firms
The SRA's warning notice of 17 August 2026 tells regulated firms and individuals that using AI does not reduce or transfer their professional responsibilities. It matters because the SRA has said it will have regard to the notice and can take disciplinary action against those who ignore it.
Part of: AI in solicitors and law firms
In short
The SRA published its "Misuse of AI" warning notice on 17 August 2026, after receiving 42 reports of potential AI misuse between July 2025 and July 2026. It targets two problems: inaccurate court and other documents, including hallucinated citations, and client confidentiality when data goes into AI tools. Solicitors remain accountable however work is produced, supervisors answer for juniors' AI use, and COLPs must take reasonable steps on compliance. The SRA's June 2026 supervision guidance adds that AI outputs need human review, scrutiny and professional judgement.
Two SRA publications now define the regulator's position on AI in legal practice.
| Document | Date | Status | Main message |
|---|---|---|---|
| Misuse of AI warning notice | 17 August 2026 | In force; the SRA will have regard to it | AI has no separate legal personality; responsibility stays with the solicitor and firm |
| Effective supervision guidance, new AI section | Updated 12 June 2026 (first published 23 November 2022) | Guidance | Firms should consider what effective supervision of AI-assisted work looks like |
| Risk Outlook: AI in the legal market | 20 November 2023 | Research and risk report | Hallucination, confidentiality, bias, accountability and emerging insurance approach |
The notice applies to all SRA-regulated firms and individuals, including in-house solicitors. Its accompanying news release gave the figure of 42 reports of potential AI misuse between July 2025 and July 2026 and said investigations covering citations, supervision and confidentiality are ongoing, without saying how many.
Inaccurate documents. The SRA says every authority cited to a court must be genuine, relevant and verifiable, and must actually support the point made. Submissions must be properly arguable. A hallucinated case or a real case quoted for a proposition it does not support both fall short.
Confidentiality. Client information should only go into an AI tool where there are contractual, technical and organisational safeguards. The data should stay in a secure environment, be inaccessible to unauthorised third parties, not be used for training unless explicitly authorised and not be kept longer than necessary. Putting client material into a public tool may breach confidentiality and permanently waive privilege.
In-house solicitors get a specific warning: take care with tools built by the business, and watch for conflicts between the employer's interest in AI and your own professional duties.
The notice does not create new rules. It maps AI misuse onto existing standards:
The notice also cites recent cases, including Ayinde and Cork v Smith [2026] EWHC 1199 (Ch) on authorised persons' responsibility for unchecked hallucinations, and Brett v SRA, which called misleading the court "one of the most serious offences".
Supervisors stay accountable for AI-assisted work by trainees, paralegals and unauthorised staff. The COLP must take reasonable steps to ensure compliance. The June 2026 supervision guidance adds that an authorised individual must keep ultimate responsibility for legal services delivered with AI assistance. In practice, firms should expect to show:
The Upper Tribunal's view in Munir that a supervisor who misses a junior's fake citations may be more culpable shows why these controls matter. Our page on AI legal research and fake citations covers the court cases in detail.
The SRA says it "can and will act" where firms fail to consider the notice. Possible routes include investigation, internal sanctions and prosecution before the Solicitors Disciplinary Tribunal. The courts are already referring solicitors: Ayinde and Al-Haroun both produced SRA referrals, as did Munir for a COLP. The SRA has not published outcomes of its AI investigations yet, so the level of sanction for solicitors remains to be seen.
The notice is a conduct document, not an insurance one, but the two overlap. PI under the SRA minimum terms responds to civil liability claims, and the closed list of permitted exclusions contains no AI exclusion. Regulatory fines and penalties are a permitted exclusion, so an SRA fine for AI misuse is usually not recoverable. A client's negligence claim arising from the same error usually would be, subject to the policy terms. Insurers may ask about AI controls, and a firm that can show it followed the notice is in a stronger position at renewal and in defending a claim. See Does solicitors' PI cover AI mistakes? and AI, confidentiality and privilege.
Apex arranges PI for solicitors and law firms. A broker can review how your primary and top-up layers treat AI-assisted work and how to present your controls. See also the AI in law firms hub.
The SRA published the "Misuse of AI" warning notice on 17 August 2026. It is in force and the SRA has said it will have regard to it when considering conduct. It followed an AI section added to the effective supervision guidance on 12 June 2026.
The SRA said it received 42 reports of potential AI misuse between July 2025 and July 2026. Investigations covering citations, supervision and confidentiality are ongoing, but the SRA has not said how many there are or published outcomes.
Not expressly. It says client information should only go into AI tools with contractual, technical and organisational safeguards, and that entering client material into public tools may breach confidentiality and permanently waive privilege. Most firms therefore restrict client data to approved enterprise tools.
Usually not. Fines and penalties are a permitted exclusion under the SRA minimum terms. A civil claim by a client who suffered loss from the same AI error is a different matter and would usually fall within cover, subject to the policy terms.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.