A new LexisNexis report titled _Scaling up: The AI-dependent lawyer_ finds that 83% of UK legal professionals are concerned about lawyers relying on inaccurate or fabricated information generated by artificial intelligence (AI), a figure that has climbed sharply from 57% at the beginning of 2024. AI hallucinations, where an AI system confidently produces information that is factually wrong or entirely invented, now top the list of lawyer AI concerns, ahead of confidential data leakage (53%) and excessive reliance on the technology (52%). The survey of 543 UK legal professionals found that AI use is near-universal: 94% now use AI for legal work, with 74% using it at least once a week and 34% using it every day. Legal research is the most common application, cited by 69% of respondents, followed by document summarisation (62%), drafting (53%), and document review (53%). One in ten respondents now describe themselves as dependent on AI to do their jobs, rising to 15% at large law firms. The findings follow a series of high-profile AI errors in legal practice. Pinsent Masons was criticised by a judge after a junior lawyer's AI-hallucinated research cited an insolvency rule that did not exist. Sullivan and Cromwell apologised to a New York judge after fabricated citations appeared in a court filing. More than three-quarters (77%) of those surveyed believe AI is causing clients or members of the public to misinterpret the law, while four in five (81%) said they feel more comfortable using AI grounded in specific legal sources.
Why this matters
The gap between near-universal AI adoption and persistent, deepening concern about hallucinations is the defining tension in legal AI right now. Lawyers are using the technology at unprecedented rates but trust in its outputs has not kept pace with deployment, and the high-profile judicial criticism of AI errors at named firms has made risk management around AI outputs a reputational as well as a professional conduct issue. The SRA's earlier warnings about AI misuse, combined with the LSB scrutiny of supervisory standards, suggest that regulatory expectations around AI governance in law firms are hardening even as the technology embeds itself into daily practice.
On the Ground
Law firm risk and innovation teams, alongside legal technology practices at larger firms, face immediate pressure to build or refresh AI governance frameworks that satisfy both the SRA's professional conduct standards and client expectations. The core work involves reviewing technology licence agreements for AI tools in use, marking up data processing agreements to ensure client data is not used to train third-party models, drafting AI governance policies that set firm-level use standards, and preparing regulatory impact assessment memos for new tool adoption. A trainee on an AI governance matter would assist with vendor due diligence questionnaires for AI tool suppliers, review technology licence terms against the firm's data security requirements, and help draft the AI acceptable use policy sections dealing with verification obligations for AI-generated research.
Interview prep
Question you might get
“Given that AI hallucinations have led to judicial criticism of named firms, what steps should a law firm take to govern AI use responsibly in legal research and drafting?”
Sign up free to see the full answer
A model answer you can lift into an interview — how to frame this story for a partner.
Sign up freeMy notes
saved