Using artificial intelligence in legal practice
16 Nov 2026
Lawyers face significant new responsibilities if they are to make safe and effective use of AI. Jon McLeod looks at the role of communications at the heart of new guidance for the profession.
Many of us enjoyed a peaceful morning’s reading and concentration during the recent Microsoft outage.
It is clear that technology in professional practice can prove a distraction. And for those who are banking on AI becoming a source of quick and cheap time-saving hacks in legal practice are being told to think again.
A new practice note from the Law Society provides an excellent framework for those law firms – and that means all of them – increasingly bringing AI into the heart of their practices.
What is clear is that there a significant responsibilities – some new, some old – for law firm leaders and practitioners looking to harness the AI revolution.
Many boil down to getting communications right – especially in dealings with clients and internally.
The guidance builds on the warning notice issued by the Solicitors Regulation Authority in August (see Oliver Simpson’s commentary on this), though this is actually quite limited, covering the issues of AI hallucinations and the risks of inputting client confidential information into AI systems, which emerged as early concerns regarding the use of the technology by regulated lawyers.
It is apparent that all types of AI use do not alter in any way the responsibility of the lawyer their firm for its output. “You should apply an appropriate level of human involvement and safeguards to mitigate risk, for both [the] business of law and [the] practice of law cases, based of the level of risk posed by the particular AI use… You remain responsible for verifying final outputs and ensuring compliance with your professional, regulatory and legal obligations.”
Interestingly, given recent events, the paper sounds a warning about the concentration risk that might arise from relying on a single provider’s technology: “A disruption affecting that provider could affect multiple systems, workflows and legal matters simultaneously.”
It is clear that technology in professional practice can prove a distraction. And for those who are banking on AI becoming a source of quick and cheap time-saving hacks in legal practice are being told to think again.
Liability for AI usage in the legal profession is addressed in the UK Jurisdiction Taskforce’s legal statement on liability harms under the private law of England and Wales (see Gina Watson’s commentary on this). For lawyers, in essence, existing professional, regulatory and legal obligations continue to apply; liability cannot be transferred to an AI system as it does not have a separate legal personality.
Client communications are a key area of focus withing the Law Society’s new note; the interaction with the client needs careful handling, and many lawyers are being confronted by clients who have pre-consulted AI’s view of their case or who sift the facts of the matter using AI tools. Both these steps are clearly problematic, and there is work for the lawyer to do to set out to clients the ground rules which will limit unbridled AI use:
- AI generated material may be inaccurate
- Inputting privileged advice into AI systems is likely to result in the loss of that privilege
- Clients who generate large quantities of AI-spawned material may face additional costs as lawyers will be obliged to check that material
No one benefits from an AI-generated workload which muddies the waters when it comes to the true facts and the law.
A further point of interest raised by the Law Society is the risks inherent in using AI in the preparation of witness statements or of witnesses themselves. The Court of Appeal has already found that such use could constitute (or be part of) impermissible “coaching”. Under Momodou, coaching a witness risks their evidence no longer being classified as their own. In an extreme case, the evidence could end up being excluded by the court.
More widely, the court can use its own jurisdiction to govern anything the lawyer does in using AI which might be considered to fail the obligation to
- Act in a way that upholds the rule of law and the proper administration of justice
- Uphold public trust and confidence in the legal profession, and
- Act with honesty and integrity
Lawyers bringing cases with hallucinated authorities have already faced tough sanctions including findings of contempt. Coming unstuck in court could be the end of the long a difficult road for the lawyer who has failed properly to engage with AI from the get-go. This alone makes the Law Society’s guidance a worthwhile bedtime read – that is, if nightmares are to be avoided.