What it means for a UK law firm
AI hallucinations have become a professional conduct issue for lawyers in England and Wales. In Ayinde v London Borough of Haringey [2025] EWHC 1383 (Admin), the Divisional Court considered cases in which authorities that did not exist had been put before the court, and referred lawyers to their regulators. The court said that freely available generative AI tools “are not capable of conducting reliable legal research”.
AI hallucinations are treated by the courts as a supervision problem as much as a technology problem. In UK v Secretary of State for the Home Department (AI hallucinations; supervision; Hamid) [2026] UKUT 81 (IAC), the Upper Tribunal said it does not matter whether false citations are inserted by a trainee or by ChatGPT: the qualified lawyer with conduct of the matter must make sure documents are checked before they reach the tribunal.
An AI hallucination comes from how large language models generate text, so no single software update removes the risk. Grounding answers in source documents and checking every citation reduce the risk; review by a qualified person is still needed.
Example
Illustrative scenario
A paralegal uses a chatbot to find authority on a costs point. The chatbot returns several case names with neutral citations, and one of them does not exist. The fee earner checks each citation on Find Case Law before the skeleton argument is filed, and removes the invented case.
How Chancery handles it
Chancery checks every quote in a workflow output word for word against the stored document and labels it with the result, such as Passage matched or Passage not verified. When the documents do not answer a question, Chancery says so. A solicitor still reviews each section before an output can be approved; see Chancery Verify and our article on AI hallucinations and supervision.