Solicitors are confronting a new ethical puzzle as clients increasingly lean on artificial intelligence to draft legal instructions, often pushing back against professional advice with AI-generated rebuttals. The trend is testing the boundaries of the Law Society of Scotland's conduct standards, which make clear that solicitors can only act with client permission.
Thomas Mitchell, a partner at RTA LAW LLP, said he has observed a marked rise in client emails that bear the hallmarks of AI composition-formal language, bold headings, and bullet-point lists of queries, some entirely unrelated to the matter at hand. "The email comes in from a client striking a more formal tone than usual, often accompanied by bold, titled paragraphed headers," he said. "Often there is a list of questions affixed to the bottom, some relevant to the issue at hand, some not."
When clients trust the machine over the lawyer
The friction intensifies when clients challenge legal advice because an AI program disagrees. Mitchell described it as taking "the classic, 'my mate down the pub says you should have done this' argument to a whole new level." Such conflicts force solicitors to debate the soundness of their own counsel with a machine that has no duty of care.
Yet the professional obligations remain unchanged. The Law Society's standards require that practitioners work only on their client's instructions. The difficulty, Mitchell said, is determining whether a list of AI-spun questions represents the client's true wishes or the machine's invented priorities. This ambiguity poses a direct challenge to the traditional model of client instruction.
As the use of generative tools in legal work grows, many solicitors are seeking AI for Legal guidance to better understand the technology and its implications for professional conduct.
Ghost authorities and wasted costs
The risks of unchecked AI use in legal settings have already produced cautionary tales. In Sabrena Rodney v Gee'z Micro Bar and Pitstop and others and R (on the application of Ayinde) v Haringey LBC, courts encountered fake, AI-generated legal authorities. These so-called "AI hallucination" cases serve as stark reminders that generative AI can fabricate plausible-sounding nonsense with confidence.
Mitchell warned that "examples of generative artificial intelligence being used carelessly is becoming more common and practitioners should be careful to not let AI do their thinking for them." The silver lining, he said, is that existing conduct principles-duties to clients, the court, and the profession-remain intact, even if a few have momentarily forgotten them.
The regulatory vacuum for clients' AI use
While solicitors must comply with strict conduct rules, there is no equivalent framework governing how clients use AI to formulate or challenge legal advice. Mitchell posed the core question: "how will clients' use of AI tools be regulated when it comes to the provision of instructions?" The answer remains unclear, and in the meantime, practitioners must navigate each AI-infused client interaction on a case-by-case basis.
He expressed concern that lawyers are being asked to act on instructions generated not by the person paying the bill, but by the algorithm that person consulted-an algorithm that may have been fed incomplete or inaccurate information.
Why this matters for legal professionals
For solicitors, the immediate takeaway is practical: AI does not alter the fundamental duties owed to clients, but it does create new risks around verifying the authenticity and relevance of instructions. Mitchell invoked the computer science maxim "garbage in, garbage out," suggesting that as long as clients feed poor information into AI tools, the output will be equally flawed. This means lawyers will need to spend more time unpicking AI-generated queries, educating clients on the limits of the technology, and documenting instructions carefully to guard against future disputes.
Your membership also unlocks:
