Practical Perspectives On The Use Of Ai By The Nigerian Judiciary

1. Introduction

1.1 I am delighted and deeply honoured to have been asked to share some thoughts with Your Lordships and Ladyships of the Lagos State Judiciary on the occasion of the Legal Summit being held as part of the New Legal Year ceremonies for the 2025/2026 legal year.

1.2 The theme of the summit is topical and apt given current developments in our profession and in the world at large. Artificial Intelligence (AI) has become so ubiquitous and is such a buzzword that the Lagos State Judiciary cannot afford not to be in the vanguard of the judiciary's understanding and adoption of AI in Nigeria.

1.3 I have read with great interest the lead paper presented by my good friend and colleague, Mr. Olumide Babalola. He is a subject matter expert, and this is evident from the detailed and comprehensive exposition of the issues contained in his paper.

1.4 I will try not to repeat or rehash anything he has said, save for the purpose of further elucidation. The concept of AI, and of technology in all its ramifications can be intimidating and esoteric to the uninitiated and it is useful to do all we can to make it less so, so potential end users are not discouraged from maximising its benefits.

1.5 To this end, and in consonance with the theme of the summit, my intervention will focus on identifying some of the practical use cases for AI that can be of immediate benefit to our judiciary as well as some of the practical dangers that we must be wary of.

2. What is AI

2.1 Mr. Babalola has provided us with a textbook definition of AI. It is evident from reading his paper though that there are different levels of advancement and complexity in AI. Indeed, AI has been with us for much longer than the present buzz would make us believe, and there are many low hanging fruits that a lot of us have not yet utilised to anything like their full capacity.

2.2 Thus, for example, we have had speech to text technology with us for a long time, which could have relieved our judiciary of the tedium of writing in long hand and could also have relieved our court stenographers, for those who use them from having to type every word uttered.

2.3 We have had websites with us for a long time and the capacity to have certified copies of the judgments of our courts published in real time as is done in other jurisdictions is not dependent on AI, but on our willingness to utilise these technologies.

2.4 Also, as Mr. Babalola addresses in detail in his paper, the technology that underpins e-filing has been with us for a long time, but we have not deployed it in the manner that enables us to reap its true benefits. Rather, it has been deployed in a manner that retains and reinforces human intervention, often further complicating the process, rather than easing the burden of filing processes in our courts.

2.5 Whilst these technologies may be underpinned by AI in some form or the other, the point I seek to make is that they are not new and that what has been missing is our willingness to harness and use them maximally. From my interaction with technology nerds, it has become evident that most of us do not utilise more than a minimal percentage of the basic computing capacity of our laptop computers or even our cell phones. Consequently, I urge that we take a back-to-basics approach to the use of AI and technology generally. Indeed, I am in total agreement with and cannot overemphasise one of the concluding recommendations of Mr. Babalola's paper, which is the emphasis on the need for continuous training and education programmes for both the Bar and the Bench on the use of technology for enhanced productivity and efficiency.

2.6 I will now focus on one of the most important use cases that I think we can employ AI for in the immediate.

3. Facilitating research and eliminating conflicting judgments

3.1 The greatest use case for AI in my view is the way in which it can assist our courts to establish what the position of the law is on any issue of law that comes before them in respect of which there are different contending positions. I was engaged recently to give expert evidence on Nigerian law in a matter that is pending before the Royal Courts of Justice in England, Alame & Ors v Shell Plc & Anor,¹ and I was embarrassed by the fact that on some of the issues of Nigerian law that I and the opposing expert on Nigerian law were asked to express our opinions on, we were both able to find authority supporting diametrically opposing positions.

3.2 One of the issues was regarding a question whether the law on limitation of actions is a federal matter or a state matter under Nigerian law. As we may be aware, each state of the federation has its own limitation law, but there is no general federal limitation law as such, except one places reliance on the application of the Limitation Act of 1623, as a Statute of General Application. The question that had then arisen was as to whether the limitation law of a state could be applied to restrict the application of rights conferred by a federal statute, namely the Oil Pipelines Act.²

3.3 Our research established that between the decision in SPDC v Farah³ in 1994 and the decision in Nigerian Agip Oil Co. Ltd v. Erekeosima & Ors⁴ in 2023, there had been at least 13 decisions of the Court of Appeal arriving at conflicting and divergent views on the question. More worrisome was the fact that several of these decisions that decided the issue in one way made no mention of preceding decisions of the same court that had gone the other way.

3.4 Another related issue was regarding the question whether an objection challenging the competence of an action based on the statute of limitation is a matter of procedure or a matter that goes to the jurisdiction of the court to adjudicate. Once again, we, the experts on Nigerian law were able to provide diametrically opposing decisions of the Nigerian courts on the issue. In Chigbu v. Tonimas Nig. Ltd. & Anor,⁵ the Supreme Court no less had stated categorically that limitation is a matter of procedural law. However, counsel was able to come up with contrary decisions of the same Supreme Court in which issues relating to limitation were equally categorically stated to be matters that go to the jurisdiction of the court and which could deprive it of the power to adjudicate.⁶

3.5 My purpose in this paper is not to debate the correctness of either of the positions canvassed on these issues, but to highlight the challenge we face in Nigeria from the lack of clarity as to the position of the law on several issues and the role that AI can play in resolving the problem.

3.6 When these issues arose in the Alame Case, the Nigerian law experts were challenged by the presiding judge, Hon. Mrs. Justice May DBE, that this situation must be a reflection of poor advocacy on the part of the Nigerian Bar. Justice May concluded that the courts in reaching these decisions would have placed reliance on the cases that were cited to them by counsel. Justice May's position was that where counsel only choose to cite authorities that support their position to the courts and do not highlight the existence of contrary authority or seek to distinguish such authorities, they contribute to this problem.⁷

3.7 Whilst I have my reservations about whether this is the sole cause of the problem, I think AI can assist in addressing the problem to some extent. As Mr. Babalola notes in his paper, citing Unity Bank Plc v Igala Construction Co. Ltd.,⁸ "The court can rely on any authority whether cited by parties or discovered in the course of its own research in arriving at a decision." This being the case, I am of the view that where there are conflicting arguments or conflicting authorities on any issue of law debated before our courts, the courts should utilise the opportunity created by AI to identify and review "all" the previous authorities on the point with a view to establishing the position of the law with as much certainty as practicable.

3.8 I am not oblivious of the various challenges that attend upon this suggestion. Not the least of which is the enormous workload that our judges carry and the attendant strain they operate under. Nonetheless, it should be our aspiration for our judiciary to aim for this level of excellence. I recollect many years ago, having a conversation with the late Justice Emmanuel Olayinka Ayoola, JSC of blessed memory, who was one of my mentors. He told me that as a first instance judge of the High Court of Western Nigeria, he would often identify an issue arising from arguments presented in a case before him and decide to go way over and beyond the arguments presented by counsel, including sojourns into the jurisprudence of other countries to analyse the issue and arrive at a decision. It is no wonder that he attained the heights he did, including representing the Nigerian judiciary with honour in various postings abroad.

3.9 I believe that with the advent of AI, those judges who wish to do so can address this issue of conflicting judgments with more ease. The research capabilities that AI platforms provide is miles removed from the manual research that the likes of Hon. Justice Ayoola would have had to contend with. As Mr. Babalola puts it, "GenAI can assist by rapidly scanning vast legal databases, summarising relevant case law, ... tools that can save hours of manual research time. ... Modern systems can sift through massive legal databases to retrieve relevant cases far beyond what manual research permits."

3.10 I would couple this with a recommendation that our judges should each have research assistants who are technically savvy and legally sound, and who can make full use of the advantages of AI. Indeed, we should be working towards the position where the best brains being produced by our law faculties compete for the opportunity to serve as research assistants to our judges, and where such research assistants, in turn, become prized candidates for recruitment by law firms.

3.11 There is no gainsaying the fact that the younger generation are more attuned to technology than the older ones. Engaging research assistants in the manner I have recommended will serve the dual purpose of providing our judges with ongoing assistance in their journeys to familiarise themselves with technology in general and AI in particular, whilst providing invaluable training for the young lawyers coming into the profession.

3.12 I will now speak briefly about two of the challenges with AI, which Mr. Babalola has also dealt with extensively in his paper.

4. Deepfakes and the need for greater vigilance

4.1 The capacity of AI and related forms of advanced technology to generate evidence that is false is mind boggling. One only needs to go to the movies to realise that the line between reality and imagination has become more and more blurred. In the context of the adversarial litigation in our courts, this has multiple repercussions. In contested cases, it places a greater onus on counsel and the courts not to accept any evidence other than "real" evidence at face value. As Mr. Babalola notes in his paper, the ease with which photoshop technology can place an individual in a photo whereas he/she was miles away from the scene is a cause for concern. This is even more so in uncontested cases. We are in an age now where the courts could be presented with evidence of service of processes on an individual, inclusive of photographic evidence, in circumstances where the individual was never served, or as is trending at the moment with the case of the mysterious Mr. or Ms. Tali Shani, where the individual never even existed in the first place!

5. Hallucinations

5.1 The other area is the tendency for AI to hallucinate. I have experienced this personally and the quality of the hallucination is alarming. Hallucination is not a problem if AI would provide you with an answer that is obviously outlandish or wrong. The problem with AI hallucination is that it presents you with answers that appear watertight in their logic and language. This increases the risk that you will be lulled into a false sense of security. I believe this is what has happened in all those instances in which counsel have cited AI generated authorities to the court confidently assuming they were on firm ground.

5.2 I had a similar experience when I asked ChatGPT to provide me with authorities on a point of procedure relating to arbitration-related litigation. ChatGPT provided me with a lucid two-page response that was exactly what I wanted to hear, inclusive of citations from rules of court, with the specific order and rule number. Alas, when I cross checked the provisions of the rules cited by ChatGPT, it was not even on a related topic! Had I relied on the output from ChatGPT without crosschecking, I would have been added to the growing list of lawyers who have had their reputation dented by uncritical reliance on AI.

6. In conclusion

6.1 There is great potential for AI to revolutionise the efficiency and productivity of our judiciary, if we deploy all the available tools in a methodical manner. It will require an increased focus and investment on training the Bar and the Bench on the optimum use of these It will also require increased but selective and well-advised investment in the hardware and software that supports the use of these technologies.

6.2 Experience has shown that the purveyors of these technologies are not altruistic and will be too happy to sell technology that does not move the dial forward in any significant manner. To avoid this, the judiciary is well advised to seek solutions that include hands-on and ongoing training and support from the technology supplier through the entire life cycle of the use of the technology with payment tied to deliverables and milestones in terms of adoption and usage.

6.3 Over and above technology, however, we must address the unreasonable workload that our judges operate under. I read a post recently on the WhatsApp platform of the Justice Reform Project (JRP), in which there were statistics of the judge to population ratio in several countries. I was shocked to see how far-removed Nigeria is from the average.

6.4 The truth is that no amount of AI or technology will address this shortfall in the manpower and infrastructure required to make our justice sector work efficiently. In the Alame Case I referred to earlier, in which I gave expert evidence in England in February, the case in question was the only case the judge was handling for the entire period of more than one week that I was there. I understand that this is standard practice in the commercial courts in England. There was no long cause list or court room packed full of litigants waiting for their cases to be called (or not called) or adjourned as the case may be. Unless we get our governments and our judiciaries to aspire to the provision of this level of manpower, an efficient justice sector will remain a mirage and our judges will continue to operate sub-optimally.

Footnotes

1 [2025] EWHC 1539 (KB) < https://www.judiciary.uk/judgments/alame-and-others-v-shell-and-others/ > - accessed 21 September 2025.

2 OIL PIPELINES ACT, CAP 07 L.F.N. 2004.

3 [1995] 3 NWLR (Pt 382) 148.

4 [2023] LPELR-60874(CA) 26.

5 [2006] LPELR-846(SC).

6 Nasir v C. S. C. Kano State [2010] 6 NWLR, Pt. 1190, 253.

7 There is clearly some merit in this accusation, and I am aware that many of our colleagues at the Bar are of this persuasion. However, I am also aware of instances in which our courts have had their attention drawn to conflicting authorities and have decided the matter in favour of one line of argument as against the other without seeking to analyse or resolve the conflict. Our firm has been involved in two of such matters recently. In Anibaba v Dana Airlines Limited & Anor (2025) 9 NWLR (Pt. 1994) 377 and in Hempel Paints v OIS (2025) LPELR- 81602(CA). In Anibaba we urged the Supreme Court to depart from its previous decision in Jenyo v Akinreti (1990) 2 NWLR (Pt. 135) 663 and though the Supreme Court decided the appeal in our favour on the point and impliedly overruled Jenyo v Akinreti, it did not do so expressly, thus leaving room for argument as to the correct position of the law on the topic. Similarly, in Hempel, we drew the attention of the Court of Appeal to a contrary decision of the Court in Limak v Sahelian Energy & Integrated Services Ltd (2021) LPELR-56408(CA) and although the Court agreed with us and decided the issue in our favour, it was silent as to the previous decision in the Limak case, again leaving cause for uncertainty as to the position of the law.

8 (2021) 10 NWLR (Pt. 1785) 407.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.