AI-generated legal advice can give clients false confidence, overlook crucial facts and even put confidential information at risk. Kevin Rogers, partner and chair of Northamptonshire law firm Wilson Browne Solicitors, examines the growing use of public AI tools by people seeking legal help and explains why they should never replace advice from a qualified solicitor.
For some time now many (not all) new enquiries, particularly those with no existing relationship, arrive armed with a detailed synopsis of their “very strong case”. Frequently these run to ten pages or more and are packed with references to damages, consequential losses, punitive awards, human rights breaches etc.
These summaries are usually less helpful than their authors intend. It is not just WB seeing this trend. A recent Legal Futures article reported that 77% of lawyers believe AI is causing clients and the public to misinterpret the law. Clients are misleading themselves.
So what is wrong?
The obvious issue is that AI can raise expectations and give clients an inaccurate understanding of their rights and remedies. Lawyers have always dealt with the equivalent of “my mate down the pub said…”, and managing expectations is part of the job.
My greater concern is that important facts can be missed. The more outlandish an AI summary is, the easier it is to dismiss. The more plausible and focused it appears, the more tempting it becomes to accept parts of it without sufficient scrutiny. Cases built on shaky foundations eventually collapse when tested by a good opponent.
Solicitors cannot outsource their obligation to investigate thoroughly. If a client says, “This is everything you need to know”, proper due diligence still has to be done. Sometimes a handful of questions can completely change the outcome.
Major risk
My biggest concern is legal privilege and confidentiality. Litigation clients may unknowingly waive privilege if they feed confidential information, legal advice or litigation strategy into public AI systems, particularly where information may be retained or used to improve the service.
Imagine a chat history identifying weaknesses in your case, discussing settlement strategy, or revealing information you would rather keep private. The consequences could be significant. We routinely warn litigation clients not to input confidential details or legal advice into public AI platforms.
Costs and protecting clients
Solicitors are paid to tell clients what they need to hear, not what they want to hear. I do not want clients spending money debating AI-generated conclusions. They are paying for legal advice, not validation of online research.
The answer is education. Increasingly, we explain to clients that we must start from scratch and establish the facts independently. Clear pre-enquiry guidance on our website – warning of the AI Pitfalls – helps here.
Does AI have a place?
Despite my criticism of public AI as an advice tool, we use secure, specialist legal AI within the firm. Used properly, it is a research tool that improves efficiency and access to resources. It is not a substitute for legal analysis.
The profession must embrace AI responsibly while ensuring it never replaces proper investigation, professional scepticism or independent thought.
My view, as a Legal 500 Leading Partner with over 27 years in law remains simple: Whilst there are undoubtedly some simple legal areas that AI can help with, if you need a solicitor, use a solicitor, not AI. The risks and delays are not worth it.
