A Shropshire law firm is warning of the dangers of using Artificial Intelligence to generate legal advice.
Lawyers at Lanyon Bowdler, which has offices across Shropshire, Herefordshire and North Wales, say there are growing concerns that people think they can obtain adequate information about legal matters simply by asking AI.
The reality, however, is different and searches often result in a lack of accuracy with AI-sourced information which can sometimes include the fabrication of case law.
Other problems highlighted with AI responses include detailing outdated legislation and amalgamating UK law with that of other countries.
There are also real fears within the legal profession that using unreliable AI information could lead to further litigation problems while threatening personal data security, resulting in a loss of legal privilege.
Brian Evans, managing partner of Lanyon Bowdler, said: “The use of AI is increasing and will continue to do so, and there is no doubt that AI is going to prove extremely beneficial in many areas. But we need to issue a warning and urge people to exercise caution when it comes to legal advice and data security.
“AI is easy to use when looking for answers to problems but the thought that anything generated by AI is accurate and safe to use is misplaced, and potentially very dangerous. I would strongly advise anyone looking to access legal advice through this medium not to do so.
“There was a recent article in the Law Society Gazette concerning the use of AI generated documents in court proceedings, which acts as a very good example. The article reported on an employment tribunal case involving claims of various forms of discrimination.
“Following the remedy judgment, the claimant applied for reconsideration and submitted a number of AI generated documents, including a 47-page overview document.
“The employment judge noted that the documents appeared to have been generated using AI, were lengthy, highly repetitive, and written in language that was difficult to understand.
“The judge commented that, while it may be tempting for parties to rely on AI when preparing submissions, it is generally more clear and effective for the court or tribunal to understand arguments when litigants set them out in their own words.
“In addition to the clarity issue and the need to address a court or tribunal in a fit, proper and relevant way, there are a number of potential risks in using AI as a basis for legal advice.
“These include relying on outdated legislation, omitting up-to-date case law or amendments – or even placing faith in what could easily be the invention of legal precedents, laws and cases that simply don’t exist.
“Another common mistake people make when using AI is failing to understand that AI draws on information from across the world and this can result in a mix-and-match of laws and procedures from different countries.
“Something that applies in one country is unlikely to apply here in the UK, so the information you are being given by AI may have no relevance to your particular situation.
“There are also serious concerns about parties involved in litigation inputting data about the case into a public AI system to assist them in dealing with the case.
“This could very well amount to putting the documents in the public domain, which means that they may no longer be protected from disclosure to the other side in the litigation process.
“Parties don’t realise this and would be horrified to learn that the advice letter they had from the solicitor, or the detailed barrister’s opinion setting out the strengths and weaknesses of their case, which they put into a public AI system to obtain a summary, might then have to be handed over to the other side as part of their disclosure obligations.
“Advice should be taken before putting any information relating to a legal case into an open access AI model.”
