AI Won't Make You a Lawyer
On 27 February 2013, the Faculty of Law at the University of Cambridge hosted a debate that attracted considerable attention within legal circles in the United Kingdom and beyond. What captured that attention was not merely the stature of the speakers, but the motion itself: "Those Who Wish to Practise Law Should Not Study Law at University."
Speaking in favour of the motion, Lord Jonathan Sumption, one of Britain's most distinguished jurists, argued that studying law is not necessarily the best preparation for legal practice. In his view, legal practice depends primarily on analysing facts and evaluating evidence rather than on identifying and interpreting legal rules. He further argued that these practical skills could be developed through the study of other academic disciplines. Opposing the motion, Professor Graham Virgo rejected the view that legal practice is simply about facts and evidence. He argued that law is a learned profession, grounded in scholarship and intellectual engagement, and that legal education develops the legal mind while drawing on a wide range of other disciplines. He concluded that, although studying law is not the only path to legal practice, it remains the best preparation. The motion was ultimately defeated by a narrow margin, yet the question it raised has never lost its significance.
Today, more than a decade later, the essence of that same question has returned in a new form, prompted by artificial intelligence: Is the ability to access legal texts instantly, retrieve case law, and draft legal submissions enough to make someone a lawyer?
In my view, the answer remains no.
Artificial intelligence is an extraordinarily capable tool when used properly. It saves time, assists with legal research, organises information, compares legal texts, and helps prepare initial drafts. When employed appropriately and under qualified human supervision, it represents a genuine enhancement to legal practice.
Its impressive capabilities, however, should not blind us to its limitations. Experience has already shown that AI can misstate legal rules, cite cases or authorities that do not exist, or generate fabricated information presented as established fact. If it misunderstands either the law or the facts, it will undoubtedly reach the wrong legal conclusion. That is where the real danger lies.
The practice of law is not simply about finding statutes and case law. It is about understanding facts, identifying their legal significance, selecting the applicable legal rule, and accepting professional responsibility for the advice given to a client or a court. That responsibility cannot be assumed by a machine, regardless of how sophisticated its data-processing capabilities may become.
A well-known quotation commonly attributed to Samuel Colt, the inventor of the revolver, illustrates the point: "God created men, but Samuel Colt made them equal." Whether or not the quotation is historically authentic, it offers an instructive analogy. Just as a firearm may give a weak person the appearance of strength, artificial intelligence may give a non-lawyer the appearance of expertise. Yet the appearance of strength does not create courage, just as the appearance of knowledge does not create experience.
The analogy goes further still. A firearm in the hands of an unqualified person may pose a danger to others. Likewise, artificial intelligence in the hands of someone who lacks adequate legal training may become a source of serious harm, particularly where people's rights, freedoms, and property are concerned. However capable the tool may be, it does not relieve its user of the duty to possess knowledge or the responsibility to verify its output.
For that reason, I believe the proper place of artificial intelligence is to remain an assistant to researchers, an assistant to legal drafters, and a tool for improving the efficiency of legal work—not a source of legal knowledge, nor a substitute for legal reasoning or professional judgment.
Artificial intelligence has transformed the tools of legal practice, but it has not transformed its essence. Human beings will remain the source of legal knowledge—through legislation, legal scholarship, interpretation, and application. They remain the custodians of professional conscience and those who ultimately bear responsibility for their opinions before clients, courts, and society. Artificial intelligence, however advanced it may become, will remain a valuable tool. It may help perform some of a lawyer's tasks, but it cannot confer a lawyer's knowledge or experience. Those are not produced by algorithms, but by education, practical experience, professional judgment, and the willingness to accept responsibility.
