Across the world, artificial intelligence is being welcomed as a transformative force in law, promising faster research, smoother drafting and quicker problem-solving. Yet a group of Ibadan-based lawyers has stepped forward with a clear warning: AI must never be allowed to replace the human conscience that sits at the heart of justice. In separate conversations with the News Agency of Nigeria, these legal practitioners said that while AI can assist lawyers, it can also become a dangerous weapon if misused. Their main concern is the growing temptation to rely on fabricated or “hallucinated” legal authorities generated by AI tools. When a lawyer submits a court process containing a fake case or a citation that does not actually exist, that is not simply an error of research. In the eyes of the legal profession, it is a breach of ethics, an act of professional misconduct and something that should attract sanctions. The lawyers stressed that AI has many capabilities, but it cannot exercise judicial conscience. It cannot weigh competing moral claims, feel the weight of a precedent, or understand the deeper principles behind a legal argument. As the Chief Justice of Nigeria, Justice Kudirat Kekere-Ekun, has also reminded the legal community, technology is only an instrument of justice. It can help judges do their work, but it can never replace the constitutional duty of judges to determine legal rights and obligations. In other words, no matter how sophisticated algorithms become, the final say in any matter must belong to a human being who can be held accountable, who has taken an oath, and who understands that law is not just data but a living system of values.
Mr Oluseun Abimbola, a former Attorney-General of Oyo State and a Senior Advocate of Nigeria, spoke particularly strongly on this point. He said that citing suspicious or bogus authorities is a clear breach of the Rules of Professional Conduct, and it is a practice aimed at confusing the court. In his many years at the bar, he said he had occasionally come across counsel who tried to rely on questionable citations, but he noted that such behaviour was rare. Most judges he had appeared before did not simply accept what was thrown at them. Instead, they took the time to fact-check citations before relying on them in their rulings and orders. When a judge discovered that counsel had presented a fabricated authority, the judge would reprimand that lawyer and could go further by imposing sanctions. Abimbola was asked at what point an unverified AI citation becomes a punishable offence. His answer was clear: the moment it is presented to the court. Whether the citation is made orally or written into a brief, if it is later found to be AI-generated and wrong, the counsel citing it becomes liable. The judge could impose an order of costs against the lawyer or refer the matter to the Legal Practitioners Disciplinary Committee for a full investigation. On the question of personal costs, Abimbola pointed to the famous United States case involving the Avianca airline, where lawyers used ChatGPT to generate legal research and ended up citing fake cases. He said that kind of sanction was appropriate and could be adapted in Nigeria, though the amount might vary. He described personal cost as a minimal penalty for a lawyer’s failure to verify AI-generated authorities. If it happens repeatedly, Abimbola warned, it can no longer be excused as negligence. At that point, it begins to look like intentional deception. And where simple costs are not enough, the full disciplinary machinery of the LPDC should step in. While he acknowledged that such misconduct was still rare in Nigeria, he added that the increasing use of AI might escalate the problem, which is why every counsel must verify, verify and verify again before placing anything before a court.
The chairman of the Nigerian Bar Association’s Ibadan branch, Mr Adekola Kareem, offered a grounded perspective on how much AI is actually being used by lawyers in the city. According to him, the use of AI for legal research is not yet widespread among his members. In fact, many lawyers in Ibadan remain largely indifferent to AI, and only a small number use it for legal research or litigation. Most still depend on physical books, law reports, the Nigerian Weekly Law Reports, trial books and other traditional research materials. Kareem said the practice was simply not prevalent in his jurisdiction, at least not for now. But he did not dismiss the issue. He revealed that the NBA Ibadan branch has concluded plans to organise a seminar on the ethical use of AI, and that a research fellow and consultant on AI has been engaged to make a presentation at the next meeting of the branch. That suggests the legal community is not sleeping on the challenge; it is preparing to educate its members before problems become widespread. When asked whether the Rules of Professional Conduct should be amended to specifically deal with AI-related misconduct, Kareem said the existing rules are adequate. His reasoning was straightforward: misleading the court is already a violation, no matter what tool was used to create the misleading content. If a lawyer presents fabricated authorities to the court, that lawyer has committed professional misconduct and is liable to be punished by the Legal Practitioners Disciplinary Committee. There is no need for a new rule to say that lying with the help of AI is still lying. The old rule already covers it. What needs to change, perhaps, is the awareness and discipline of lawyers, not the fundamental ethical framework. For Kareem, the answer lies in training, orientation and peer accountability rather than a frantic rewriting of the rulebook.
Another legal practitioner and political analyst, Mr Oluwatobi Fatoki, gave a more confident assessment of the ability of Nigerian courts to protect themselves from fake AI authorities. He argued that the courts already possess enough institutional tools and judicial knowledge to detect fabricated cases. He reminded people of the age-long principle that a lawyer may not know all the law, but must know where to find it. That principle remains mandatory. Fatoki said courts do not simply accept authorities hook, line and sinker. After lawyers finish their presentations in the courtroom, judges retreat to their chambers to study the arguments, marshal the points together and form their decisions. They have research assistants and court researchers whose job is to verify citations. If a case is fabricated, a careful judge will know. Judges routinely direct counsel to produce certified true copies of the authorities they have cited, and the front-loading system under the High Court Rules provides another layer of protection. Superior courts also subscribe to electronic law reports such as LawPavilion, Legalpedia and the Nigerian Weekly Law Reports, which make verification easier. Fatoki acknowledged, however, that many lower courts and customary courts do not have the same access to these electronic resources, and that makes detection harder in those forums. Still, he aligned himself with the position of the Chief Justice when he said that AI cannot exercise judicial conscience. It is a guiding tool, not a tool that dispenses the exact mind of the court. In his memorable phrase, the mind of a judge, like the intestine of a man, cannot be known by AI. That is a vivid reminder that legal judgement involves an inner humanity, an ability to read between the lines and understand context, that no algorithm can replicate. Fatoki also called for compulsory AI literacy in the curriculum of law faculties, so that young lawyers learn how to use these tools responsibly from the very beginning of their careers. He warned against the practice of feeding clients’ briefs into AI programmes, blindly generating court processes and then submitting them without independent verification. That, he said, is an abuse that will inevitably harm clients and erode public trust in the legal profession.
Younger voices in the legal community are also thinking carefully about these issues. Mrs Funmilola Onalaja, a Legal Associate at Mondial Solicitors, said she personally does not use AI to draft court processes. She was careful not to make assumptions about how her colleagues do their work, but she said she believed in verifying any information that comes from an AI tool. Onalaja pointed out that the cost of accessing legal databases is a serious challenge for many lawyers. Not every lawyer can afford subscriptions to the major electronic law reports, and this financial pressure can make AI seem like an attractive shortcut. The convenience of AI is real. It can produce answers instantly, in a way that saves time and energy. But that very convenience carries a danger: it makes it tempting to accept information without checking whether it is true, accurate or even real. Onalaja said practical training on the correct use of AI could help young lawyers avoid the problem of hallucination, which is the term used when an AI model confidently invents a fact, a citation or a case that does not exist. Her point was not that AI should be banned or feared, but that lawyers must learn how to work with it safely. That means verifying every citation, checking every source and treating AI as a research assistant whose work must always be reviewed by a professional human mind. It also means being honest about the limits of one’s own knowledge and not allowing the seductive speed of technology to weaken the discipline of careful legal work. The legal profession has always depended on precision, patience and truthfulness. Those values do not change simply because the tool changes. If anything, the arrival of AI makes those values more important than ever.
Taken together, the voices from Ibadan tell a balanced story about artificial intelligence in Nigerian legal practice. They do not reject AI as a useless or evil invention. They see it as a useful assistant, a powerful research tool and a sign of the future. But they are also determined to keep it in its place. AI can find cases, summarise arguments and even suggest language for a brief. What it cannot do is stand in the shoes of a lawyer and take responsibility for what is placed before a court. It cannot apologise when it makes a mistake, it cannot feel the weight of a client’s life or liberty, and it cannot exercise the conscience of a judge. The legal profession is built on trust. Judges trust lawyers to bring them genuine authorities. Lawyers trust judges to rule fairly and honestly. Clients trust lawyers to protect their interests. If AI is allowed to break that chain of trust through fabricated cases and careless citations, the damage will be severe. The solution, according to these Ibadan lawyers, is not to bury our heads in the sand. It is to educate lawyers, strengthen professional discipline, provide better access to legal databases, and teach every law student how to use AI without being deceived by it. The courts already have safeguards, but those safeguards need to be supported and extended, especially in lower courts where resources are scarce. The disciplinary bodies must be ready to punish those who deliberately use AI to mislead the court. And every individual lawyer must embrace the ancient duty of verification. In the end, the message is simple: let AI do the heavy lifting, but let human beings do the thinking. Let machines help us find the roads, but let human wisdom decide which road to take. Only then can technology serve justice instead of undermining it.