On 17 August 2026, the SRA published a warning notice on the misuse of AI in legal services. Whilst it leaves more questions than answers, it makes one thing clear: blaming AI is not an acceptable excuse.
Two key concerns
Hallucinations. AI tools can confidently generate fake cases, citations and “facts” that look entirely plausible. Given the potential negative impacts on clients, courts and trust and confidence in the legal profession, those accused of misuse may face regulatory action, wasted costs orders and even contempt proceedings.
Confidentiality. Entering client information into public AI tools risks breaching confidentiality and can permanently waive legal professional privilege. Both free and paid-for tools carry risk.
The case law
The courts are taking this seriously. In Ayinde v Haringey LBC, fabricated citations led to a wasted costs application and regulatory referrals, and made clear that reliance on AI was no defence. BCP v A Mother and Cork v Smith reinforce that professionals must verify AI-assisted work and that lack of intent to mislead is irrelevant. In UK v Secretary of State for the Home Department, the Tribunal confirmed that putting client letters into an open-source tool such as ChatGPT places that information in the public domain.
Accountability
AI has no legal personality. Users remain fully accountable for accuracy, and supervisors can breach their own duties if a supervisee submits unreviewed AI-generated content.
What firms should do
- Verify everything before it reaches a client or court.
- Know where your data goes before confidential information touches an AI tool.
- Supervise properly, in line with the SRA’s updated Effective Supervision
- Build real governance, not just a policy nobody reads.
The takeaway
Used correctly, AI is valuable, but it does not dilute professional responsibility. Judgement, verification and accountability remain with the human lawyer. See our full blog here.