Chief Justice Sundaresh Menon: Keynote Lecture at the Judicial Executive Programme

I. Introduction

1. I am especially delighted to address this combined audience of participants from the Judicial Executive Programme (“JEP”) and the Judicial Perspectives Programme.

2. To the international participants of the JEP, a particularly warm welcome to Singapore. The JEP is the Singapore Judicial College’s (“College”) flagship executive leadership programme that we open to the international community of judges as part of our commitment to promoting judicial dialogue and the bonds of judicial families. We curated this out of the need for a programme that brings together current and aspiring judicial leaders from around the world, and which serves as a platform for mutual exchange. Over the years, we have been improving and upgrading the programme, and it has been a very helpful forum for these purposes.  The other great draw of the JEP is the opportunity to learn from and build networks with like-minded individuals from around the world. I was therefore delighted to hear that we have, this year, a terrific cohort of 29 judges and court leaders from 11 jurisdictions across Africa, Asia and the Americas.⁽¹⁾ Each of you brings unique and much-valued perspectives on what it means to be a judicial leader, and I very much look forward to hearing from you during the discussion to follow.

3. We are also joined by colleagues from the Judicial Perspectives Programme, which is part of the College’s suite of foundational ‘101’ programmes. Most of those foundational programmes are directed at the ‘how’ questions of judging; for instance, how one might write accessible and well-reasoned judgments, or how one might critically and forensically analyse evidence. The Judicial Perspectives Programme complements those more technical subjects, by addressing the equally significant questions of the ‘what’ and ‘why’: for instance, what does the role of a judge entail? And why is it important that judges do what they do?

4. These questions are relevant to judges and court leaders at all levels, hence this particular series of combined plenary sessions, which bring together participants of various seniorities and roles. 

5. Over the past day and a half, you have unpacked four aspects of the judicial role: as Adjudicator, System Reformer, Learner and Leader. My remarks this afternoon are focused on the last of those components – judicial leadership. 

6. We live in a time of tremendous transformative change, much of which has been driven by advances in new technologies, especially the advent of Generative Artificial Intelligence (or “Gen AI”). 

(a) The scale and speed with which Gen AI has entered the mainstream is nothing short of remarkable. It is hard to believe, but ChatGPT was launched less than three years ago. It reached 100 million monthly active users just two months after its launch, a feat which led some to term it the fastest-growing consumer application in history.⁽²⁾ By mid-2025, two and a half years after it was launched, ChatGPT had reached 700 million weekly active users.⁽³⁾ Gen AI is now everywhere: Google’s world-leading search engine is now augmented by Gemini AI;⁽⁴⁾ Microsoft has integrated its Copilot AI into its widely used Microsoft Office suite of productivity tools.⁽⁵⁾ I think it can fairly be said that Gen AI is not just in the mainstream; it is now ubiquitous. 

(b) In addition to their growing presence, AI tools are also becoming increasingly capable. Several AI models have now passed the most challenging third level of the Chartered Financial Analyst exam, when as recently as two years ago they could clear only the first two levels.⁽⁶⁾ In a similar vein, Gen AI models developed by Google and OpenAI achieved gold medals at the International Math Olympiad in July this year, bettering the silver medal performance in 2024.⁽⁷⁾ The same is true when one considers the length – and, by extension, the complexity – of the tasks that Gen AI tools can complete. One recent study found that the length of software engineering tasks which Gen AI tools were able to complete had roughly doubled every seven months for the past six years.⁽⁸⁾ In short, Gen AI is getting exponentially better at completing increasingly complex tasks.

7. Given that trajectory of Gen AI’s growth – both in terms of user adoption and technological capability – we are very likely still, as my friend, Professor Richard Susskind, puts it, only at the “foothills” of what Gen AI can do.⁽⁹⁾ At a time when there is perhaps an excess of information for any human intelligence to manage without aid, AI presents us with options to augment and enhance human intelligence. But this raises a number of issues. And how we navigate those options requires thoughtful responses that call not so much for virtues of knowledge or skill, as they do temperament and discernment – which might be better referred to as wisdom. It is that capacity – wisdom – that will enable judicial leaders not just to lead in times of great change, but through times of great change, as stewards of a justice system that has and must continue to navigate seismic changes in a way that ensures that the pursuit of ‘Justice’ remains a concrete reality and not just a remote ideal.

8. It is that interaction – between artificial intelligence and human wisdom – that I want to speak on today and, in particular, how we, as judicial leaders, should think about that interaction. My address is structured in three parts, all seen through the lens of that interaction between human and artificial intelligence:

(a) First, I set the stage by making the case for why judiciaries have to take the lead in court reform. 

(b) Second, I turn to discuss what a culture of judicial leadership might look like, or what values a judicial leader should embody in an age defined by change and uncertainty.

(c) Finally, I touch on how we might seed that culture of judicial leadership, and how we might develop judicial leaders in that mould.

II. Why do we need judicial leadership?

9. First: why is it important that judiciaries take the lead in court reform? I make three points.

10. First, I would suggest that judicial leadership of court reform is not merely an option; judges are in truth duty-bound to lead in this process.⁽¹⁰⁾ One might argue, and for a long time many did hold the view, that the role of judges is just that – to judge – in other words, to adjudicate disputes according to law. On that view, the judge’s role is somewhat more passive – it is to operate within the parameters of the prevailing rules and processes of the justice system, and not necessarily also to think of changing them or of innovating and finding new systems that might be better fit for purpose. I suggest that such a viewpoint is predicated on an unduly narrow conception of the mission of judicial office. I accept that adjudication is an important, indeed even a critical part of what judges must do, but it is essentially one facet among many of the judicial role. That role, fully expanded and fully expressed, is to uphold the rule of law and to ensure the fair and efficient administration of justice.⁽¹¹⁾ And if we accept that our true mission is directed more broadly to the administration of justice, then judges must be concerned not only with the exercise of adjudication, but also with ensuring that the systems and processes through which justice is administered remain fit for purpose and accessible to our users. 

11. Even the best judge cannot properly administer justice if the processes through which she must do so are inaccessible to those who need them most.⁽¹²⁾ In one of my first speeches as Chief Justice, I said that the best system in the world is nothing more than a curio on the shelf if the people who need justice cannot reach it – it may be nice to look at from afar, but it does not serve anyone.⁽¹³⁾ And that drives my strongly held view that the role of the court goes beyond the determination of particular cases, to looking at the justice system, continuously exploring it, and ensuring that it remains fit for purpose. 

12. The upshot is that our duty to the administration of justice demands that we serve not only as the operators of the justice system, but also as its stewards, responsible for its reform and its upkeep, and in particular for ensuring its accessibility. That last point on access is a real passion of mine, because accessibility is foundationally critical to ensuring that our judiciaries remain relevant to our societies. I would go so far as to say that our concern extends to anything that fundamentally affects or impacts the rule of law, or has a real prospect of doing so. This includes the rise of Gen AI – which is going to transform societies, and the lives of people in our societies – and extends to matters like the fitness of legal education and the formation and development of our lawyers. This view of the judicial role – as entailing being concerned with anything that affects, or has the potential to seriously affect the rule of law – explains why I speak and comment on such a broad range of issues that, some of which might not traditionally be seen as part of the judicial remit. But I think that they are, for as long as they have the potential to impact the rule of law, they occur on our watch, and we need to think about their implications for how justice is accessed and administered.

13. Gen AI has the potential to transform not just the ways in which judges adjudicate disputes, but also the place of adjudication in a justice system. It is probably too early to speak of ‘robot judges’ ⁽¹⁴⁾ (even though that was very fashionable when ChatGPT first came out), but Gen AI has already demonstrated potential to fill roles we previously thought only humans could fill. In some jurisdictions, judges sitting as neutral evaluators play an important role in offering prospective litigants a preliminary assessment of the likely outcome of their dispute in order to facilitate settlement negotiations.⁽¹⁵⁾ It is easy to imagine AI tools being harnessed within the justice system to offer that kind of preliminary, non-binding evaluations, at a scale that would otherwise be unachievable if human assessors were the key and only legitimate resource. 

14. I spoke at an event at New Delhi last year,⁽¹⁶⁾ where I sketched the scale of the access to justice deficit. According to statistics from the World Justice Project, it is estimated that some five to six billion people have no meaningful access to law or to justice. This is the product of a combination of factors: distance and geography, lack of resources, lack of literacy, or a lack of awareness amongst those people that they have a legal issue. As I said to that audience, every Chief Justice and every person with responsibility for a justice system should be deeply concerned that that is the state of our planet,⁽¹⁷⁾ because if people are unable to access justice, then a fuse has been lit for potentially dire outcomes. That is the scale of the problem; there simply are not enough judges or lawyers to ever get all those people the access that they need to justice. 

15. In this connection, technology is the essential force multiplier, and we shall have to think about how to harness it responsibly and effectively. Technology is particularly suited to addressing this problem when one considers the type of justice needs that most of the people who live outside the protection of the law have. Their problems are not the complex, difficult issues that are the subject of deliberations in apex courts; the vast majority of them are relatively straightforward matters like issues relating to land, neighbourly disputes, community difficulties or problems with registration. 

16. The challenge this poses is primarily one of scale, and this brings me back to my main point, which is the potential that technology – serving as a force multiplier – holds for addressing this aspect of the justice gap. To illustrate this point, let me make a brief reference to my visit to the Shenzhen Intermediate People’s Court just a week ago. My court has an annual meeting with the Supreme People’s Court of the People’s Republic of China, and I had asked for this year’s meeting to be held in Shenzhen, as the Shenzhen Intermediate People’s Court has been doing pioneering work on incorporating AI into the process of adjudication and I wanted to see what was being done and it was most impressive. The Intermediate People’s Court has been working on a pilot to incorporate AI, which has been operational in Shenzhen for some months. A team of technologists, working with a group of about a hundred judges, have disaggregated the judicial task into various discrete, smaller components. They constructed an application software system, within which AI executes defined tasks, drawing on a ringfenced bank of materials. In this way, the AI tool extracts certain ‘basic information’ – from information relating to the parties, through to the background facts, issues, submissions, the prevailing reasoning and possible approaches for resolving the case – and presents this to the judge in a way that makes the material more manageable. The system has been tried and tested, and I was told that it is expected to be integrated into the national judicial system in China by the end of November, where it will make a profound difference to the ability of China’s judges to manage their staggering caseload.

17. This also illustrates the importance of context when thinking about how AI might be harnessed to address the justice gap. Jurisdictions which do not have a large caseload may find the case for using AI at scale far more limited (though AI can certainly still be used for discrete tasks, such as legal research or summarising documents). But where the problem presents in the form of extremely large caseloads – such as in China, which has a caseload of 40 million, or the backlog in India of some 30 million cases – one cannot realistically hope to meet that sort of challenge simply by running a model that works elsewhere or by insisting upon what may be regarded as ideal processes. Instead, we must think of how we might use technology to address the problem before it becomes overwhelming. Context is key, and the pilot in Shenzhen demonstrates that it is possible to harness this hugely powerful tool – AI – in a way that allows us to address seemingly insurmountable challenges of scale and magnitude.

18. My second point on why judges must lead in matters of court reform is that we are particularly well-placed to lead.⁽¹⁸⁾ I think it would be a grave mistake for judiciaries to abdicate responsibility for systems reform to a bureaucracy of non-judges who may not have that essential blend of experience and perspective that judges and administrators working daily at the coalface of our justice systems would have. It is that perspective that makes us best placed to identify the pain points and evaluate the best solutions. 

19. I was just speaking with Justice Michelle Gordon of the High Court of Australia, who told me about the reform to introduce electronic filing in her court. Owing to various constraints, she ended up leading the effort, sitting with her staff and getting input from technologists to design the system from the ground up. As I guessed, and as she confirmed, that system works uniquely well because it was shaped and developed with the involvement of judges, who work the system every day and therefore knew what information was needed and when, and who were therefore able to design a system that met their needs and circumstances. 

20. Now, this does not mean that judiciaries go at this alone. Systems reform is necessarily and deeply a multidisciplinary endeavour. It entails working in teams with other subject-matter experts, such as – in the context of technology-based reform – technologists, data analysts, computer scientists and software engineers. This was reflected in the make-up of the team in Shenzhen – as mentioned, the team comprised about 10 to 12 engineers and technologists, alongside about a hundred judges. At each stage, the judges disaggregated and defined the judicial task to be tackled, and the technologists developed the solutions. The technologists then went back to the judges with their proposed solutions, which the judges either validated or varied. This developmental loop of iteration and reiteration was repeated several times, until the team was satisfied with how it was working.

21. In saying this, I do not mean to suggest that any of us must therefore become expert computer scientists. But we should all at least acquire a working knowledge of the subject. We have to be conversant with the key technologies– what they can do, and more importantly, what they cannot do. And we need to keep abreast of those developments, so that we will be able to interface between and coordinate efforts across these contributors from various fields and disciplines.⁽¹⁹⁾

22. My third and final point is that judges and judiciaries should lead court reform because these reforms ultimately affect us and the way we work. In short – leading is in our interest. We see this both at the level of the individual judge as well as of the system.⁽²⁰⁾ From the individual’s perspective, there is surely a compelling interest in being an active participant rather than a mere passenger in matters that will affect one’s work and career. To remain passive in the face of change is to surrender our ability to influence that change, and to shape the future that we must inhabit. And, from the institution’s perspective, there is huge value in getting judges and administrators at all levels involved in reform, because the success of these reforms will depend, at least in part, on securing the buy-in of those who must operationalise it. Stakeholders who have participated in and contributed to the process of change are far more likely to work hard to make it work. 

23. To sum up, I suggest that we must take the lead in court reform because: (a) it is our duty to do so; (b) because we are best placed to do so; and (c) because it is in our interests to do so.

III. What values should a judicial leader embody?

24. I turn next to the question of what a culture of judicial leadership might look like. I had the privilege of giving an address to the Rwandan judiciary in 2022. In that address, I suggested that judicial leaders should aspire to embody these three qualities: (a) vision – meaning the ability to identify and set a course; (b) initiative – because leadership is about taking action to better the lives of the people around you and the systems that you own or operate; and (c) courage – to stay the course in the face of obstacles, and also to acknowledge when you have made a wrong call and have to start again.⁽²¹⁾

25. A lot has changed since 2022. While those qualities remain relevant, I suggest three other factors to keep in mind, because of the very different operating environment that we function in. You might think of these as three “Be’s” or for that matter, 3 “C’s”: be curious, be critical and be courageous. 

(a) First, be curious – because the vision we chart must be filled and shaped by a sense of openness and a willingness to try the new, even if that means upsetting the status quo. 

(b) Second, be critical – because not everything will work out as predicted, and we will need to develop a healthy sense of scepticism, so that we remain discerning and disciplined even in this time of seemingly unbounded opportunity. 

(c) Third, be courageous – because sometimes things will pan out, and sometimes they might not, and we shall need courage both ways. 

A. Be curious

26. Let me begin with the value of curiosity, which has at least three facets.

27. The first is a sense of creative dissatisfaction. This is the conviction that there must be a better way to do things. It means refusing to be a slave to the received wisdom of the time. It manifests as a willingness to question the status quo, and to reconsider practices and workflows – practices that might have been developed at a time when the problem statement was very different, and that might have worked well for decades, but which may no longer be the best option.

28. Let me give you an example to illustrate this point. Many of us grew up in justice systems where the courts’ primary interface was with lawyers. When I was a young lawyer, self-represented persons (“SRPs”) were very much an exception; most of the time, courts dealt with lawyers. And because courts dealt primarily with lawyers, they could write their rules and construct their processes in a way that relied heavily on the special jargon that judges and lawyers use. That legal jargon was useful because it condensed otherwise complex legal ideas into compact, shorthand phrases. While that might have been fine in the 1970s, the 1980s and perhaps even into the 1990s, when lawyers remained the primary interface with the courts, that reality has since changed. Our courts and many courts around the world are now facing growing numbers of SRPs, and that requires that we revisit our old operating paradigm – under which justice was dispensed primarily by a priesthood of judges and lawyers – and examine its fitness for purpose in light of current realities.

29.   For SRPs, navigating the court system (and even the court building) can be a most distressing and frustrating experience. To help me get a better understanding of the impediments that SRPs face in accessing justice in our court building, my team invited me to walk through the user journey of an SRP seeking to file a bankruptcy petition. To say it was challenging would be a grave understatement. A prospective applicant had to go to eight different locations within the courthouse. You would go to one place to get a document, and you would then be instructed to go to another place to have it processed, and so on. And this felt like a hassle even though I had a team of people leading me and showing me where to go. The SRP does not have the benefit of that guidance and probably is already under the stress of facing bankruptcy. The danger is that an SRP who encounters the justice system in that way sees what seems a broken system that does not work; or worse, works only for the privileged few who can afford the services of lawyers to assist them. 

30. Today, we have addressed this gap, and such filings can be made in one place, at a single stop. The point is this: just because we have always done things in a certain way does not necessarily mean that that is how things have to be done. We have to be realistic, and constantly ask ourselves if there is a better way of doing things, and of responding to the realities of current circumstances. The willingness and ability to reimagine the art of the possible is particularly important given that the true value of Gen AI lies in its potential to completely transform our processes and create novel solutions that were previously unimaginable.⁽²²⁾

31. That aspect of creative dissatisfaction must be accompanied by another facet of curiosity – that is, the drive to explore. This is the willingness to consider uncharted territory; to be willing to learn about and to try new things and new ways of doing things. It is that type of curiosity that drives us to seek information on emerging technologies and their potential applications; and to try new tools, recognising that experimentation often requires tolerance for uncertainty, imperfection, iteration and reiteration. 

32. Undergirding all of that is humility. The curious mind must be an open mind. Robertson Davies, the Canadian novelist, said: “The eye sees only what the mind is prepared to comprehend.”⁽²³⁾ Openness is key, because we cannot contemplate possibilities that we have already (and perhaps subconsciously) closed our minds to. The curious leader is therefore one who recognises that her own worldview is limited and therefore actively seeks and values the perspectives of others. That culture of openness is essential at the institutional level as well. And it was this desire to broaden our perspectives that caused us to establish a dedicated Access to Justice Programme Office (“A2J Programme Office”), which became so successful and so central that we upgraded it to a fully-fledged division this year. The A2J Programme Office has embarked on countless projects aimed at improving the experience of court users (including the project to improve the user journey for filing bankruptcy petitions, which I spoke about earlier), and has been a critical player and an important driver of our transformation into a more user-centric and outward-looking institution.⁽²⁴⁾

33. Tying those distinct but related threads of curiosity together, I suggest that the curious leader is one who is willing to reimagine the familiar, embrace the novel, and one who is open to diverse perspectives.

B. Be critical

34. I turn next to the importance of thinking critically. 

35. We live in a time of seemingly limitless opportunity, and because of that, it can be tempting to want to try everything. My younger colleagues refer to this as ‘FOMO’ – the fear of missing out. The options are endless, and we have to be critical in our evaluation of those options, especially in deciding what we want to embrace or pursue. This entails discipline – so that we adopt AI tools for the right reasons, and not just because they look impressive; and (b) second, a strong sense of discernment – of being able to tell what the AI tool’s strengths are, what its weaknesses are, and how well it will fit or work in the context of our system, so that we make good decisions about what tools to use and where to deploy them. 

36. Let me pause here to make an important point. Our goal is to make good decisions – decisions that are well-grounded, well-informed, and that are arrived at through a proper process. These may not always be the ‘right’ decisions, or ‘successful’ decisions. There is an important distinction to draw here: we must be prepared that even a good decision may not always be, with hindsight, the ’right’ decision. After all, the inescapable reality is that we may sometimes get it wrong. And so, what is key is that we focus on making good decisions, and on setting up systems that support and facilitate the making of good decisions.

i. Critical acumen and discipline

37. What do I mean by ‘discipline’? Discipline entails that ability to choose among the many options that our curious minds identify. 

38. As leaders, we need to be clear and precise in thinking about how the tools meet our ends, rather than just pursue the acquisition of those tools for their own sake. My colleague, Justice Aidan Xu, who is our Judge-in-charge of Transformation and Innovation, said in a recent interview that “we must approach our exploration and testing in a principled way, so that we do not go chasing after every single new shiny thing”.⁽²⁵⁾

39. We do this is by adopting an outcome-centred – rather than tool-centric – approach to our exploration of Gen AI.⁽²⁶⁾  In other words, we begin by articulating the problem statement to be addressed, as well as our desired outcomes, before considering whether and if so, how, those ends might be achieved with the development of an appropriate tool. In contrast, the danger of adopting the tool as the starting point is that we risk unduly fixating upon it, even after it becomes clear that the tool’s features and capabilities, whilst perhaps impressive, are not actually needed to achieve the outcomes we desire.

40. Let me explain this by reference to a project we embarked on some years ago. Motor accident cases comprise a sizeable proportion of the State Courts’ caseload. In Singapore, we have a ‘Road Accident Guide’, which contained numerous drawings depicting various typical scenarios of traffic accidents. For each scenario, the Road Accident Guide would state what the likely outcome was, in terms of the apportionment of liability. Based on this, we developed an outcome simulator, the Motor Accident Claims Online (or “MACO”), to help parties involved in motor accidents make decisions about their motor accident claims by giving them an assessment of the likely result – in terms of liability and quantum.⁽²⁷⁾  The predictive elements in the tool are not powered by predictive AI models; they are driven by deterministic algorithms (ie, “If A, then B”) which provide an answer based on input variables like the types of vehicles or the circumstances under which the accident took place.

41. The short point is this: if we cultivate the discipline of remaining focused on the desired outcomes instead of fixating on the tool and how shiny it looks, then we will realise that in many contexts and for many applications, a Gen AI tool may not be necessary, or even helpful. We may find – as was our experience with the development of MACO – that our desired outcomes could be better achieved with far simpler and cheaper tools.

42. Taking this point a little further, an outcome-centred approach will also help us realise that the most meaningful change can sometimes come from low-tech or no-tech solutions and reforms.⁽²⁸⁾ Let me give you an easy example. When the COVID-19 pandemic hit, courts needed to keep running and lawyers needed to keep working. The switch from sitting in a courtroom to having hearings over Zoom was heralded as a great leap forward for the use of technology in the courts. But videoconferencing technology is by no means new or high-tech – the technology had been around for years. But the pandemic supplied the push that was needed, and when people tried it, they had their eyes opened to its potential. This illustrates the importance of starting with one’s desired outcomes, and then working backwards to identifying the change needed to achieve those outcomes. Oftentimes, you will find that you will not need the shiniest or most advanced technology to get you there.

ii. Critical acumen and discernment

43. The second aspect of being critical is discernment. This means looking beyond the hype that often surrounds technological tools to understand what they are actually capable (and incapable) of, and where they may be most appropriately deployed. Having an understanding of what is actually possible is an important part of discernment.

44. The other critical component of discernment entails moving beyond the rather simplistic view of AI as a monolithic ‘smart machine’, towards disaggregating and understanding it in terms of its distinct capabilities and functions.⁽²⁹⁾ Our collaboration with Harvey AI to develop tools for our Small Claims Tribunals illustrates this point. It is tempting to think of that project, at the finish line, as a fully integrated AI-powered solution capable of assisting SRPs with anything and everything. But that was not in fact how it was developed. The product you now see has been and will continue to be developed in stages. In our linguistically diverse society, the first issue that court users face often is a problem of language. AI had proven quite capable at translation, and so we started with that, and incorporated that into the system. AI is also very good at summarising information and data, and that was the second function we built into the system – a tool which summarises case documents and thereby supports SRPs in preparing for hearings by helping them understand the facts, evidence and arguments raised by the other party, so that they are able to articulate their arguments in a more succinct and effective manner.⁽³⁰⁾

45. And so, it is important that we disaggregate AI into its various and distinct functions and capabilities. Those capabilities are by no means uniform, and lie instead on what has been referred to as AI’s ‘jagged frontier’⁽³¹⁾  – highly sophisticated in certain domains, yet surprisingly limited in others. The contours of that jagged frontier are likely to vary depending on the type of AI and the specific field of knowledge in question.⁽³²⁾ It is imperative that we carefully consider where the particular application of AI in question sits on that jagged frontier, so that it is appropriately and safely deployed. A failure to appreciate this could be deeply problematic. For instance, attempts at deploying predictive AI in the criminal justice system have revealed several now well-documented shortcomings of such AI systems.  

46. On the other hand, generative AI tools – tools which generate text, images and code – have shown far more promise. We know, for instance, that Gen AI is excellent at identifying patterns and extracting relevant insights according to specified criteria. We have sought to harness these strengths by building Gen AI tools to assist court users with preparing their cases.⁽³³⁾

47. One example is Pair Search,⁽³⁴⁾ a search facility which we developed together with the Government Technology Agency of Singapore (or “GovTech”). Pair Search enables any court user to research court judgments and legislation. Besides finding relevant case law, Pair Search can summarise or extract key principles from cases, categorise the search results thematically and even draft summaries or outlines based on the user’s prompts. Another example of our use of Gen AI is our project with Harvey AI, which I have already mentioned.

48. These projects are reflective of how we in Singapore have sought to approach the adoption of AI. That is, by first assessing the range of AI’s capabilities, and then leveraging its strengths in areas where AI can most meaningfully improve the experience of court users and the productivity and work quality of our staff.⁽³⁵⁾

C. Courage

49. That brings me to the third value of judicial leadership – courage. These include decisions to act, decisions to stop, and decisions to say ‘no’. Each requires courage of a different sort.

i. Courage to act

50. Let me start with the courage to act. 

51. The challenge we face is not simply one of venturing into uncharted waters, but rather one of staying the course in the face of resistance, criticism and the weight of conventional wisdom pulling in the opposite direction.

52. Take, for example, our decision to shift from paper filing to electronic filing some twenty-five years ago, which met with considerable resistance.⁽³⁶⁾ Lawyers complained that it had increased the complexity of filing and thereby the costs of litigation; while judges baulked at having to refer to electronic copies of documents on a system they felt was unstable and slow. The then-Chief Justice, Chief Justice Yong Pung How, firmly believed that it would work, and saw it through. This, coupled with the high-paced effort to clear the substantial backlog which then plagued our system, was hard on the Bar. But today, looking back, we are grateful that our leaders then had the courage to persevere, because those painful but necessary steps laid the foundation upon which everything else in our justice system – which is, today, very kindly regarded in many circles – now rests.

53. Today, electronic filing has become so integral to our operations that the idea of not having had it or doing away with it would be unthinkable. But at the time, the conventional wisdom was that it was not a good idea, and that it was not going to work. And that makes this an apt demonstration of the qualities I have spoken about – discernment, discipline, curiosity, openness, and having made the decision to go ahead, having the courage to see it through. 

ii. Courage to stop

54. On the other hand, you will sometimes find after you start on something that it is not quite working out. That sort of situation requires a different type of courage – courage to acknowledge that the experiment has not panned out, and that it is time to call it.  

55. This can be hard to do – at both the personal and institutional levels – particularly if substantial goodwill has been poured into the venture. Sometimes, for instance, there might be a tendency to persist because of how much you have sold an idea. But true courage requires a willingness to step back and say ‘enough’. 

56. That is probably far easier said than done. But that is another dimension where humility becomes very important, because persisting can lead to even greater humbling or humiliation later on. I think it is a mark of good stewardship that frees productive resources that could be channelled elsewhere. 

iii. Courage to say ‘no’

57. Yet another facet of courage is the courage of restraint or, put simply, the courage to say ‘no’. This could mean saying ‘no’ to AI applications that may seem impressive at first blush, but which do not serve any clear institutional purpose. It can be particularly important where taking a particular step seems to result in immediate benefits whereas in the longer term there may be problems that may or may not surface.

58. Let me give you a quick example. Many courts around the world have issued AI governance frameworks – for instance, guidelines on the use of AI by court users. We have issued an AI governance framework, but our approach has been quite different. Whereas other frameworks tend to be rather prescriptive – as to when AI can and cannot be used, or as to what must be disclosed, etc – our framework, in essence, reminds and cautions court users that they ultimately bear responsibility for any document they file, regardless of whether any AI tool was used or how it was used.

59. Why did we do that? There was some pressure to err on the side of prescriptiveness, because several other jurisdictions were publishing frameworks of that ilk. But Justice Xu and I discussed it, and we had three concerns with taking an unduly prescriptive approach.

(a) First, the sheer pace at which the technology was advancing and transforming meant that there was a real risk that detailed and compendious prescriptions would quickly run out of date.

(b) Second, it was unclear if those prescriptions could in fact be effectively policed. For instance, if we imposed a requirement to disclose the use of AI in the preparation of any document, would it have been feasible to check every single document? And even if we did, were there reliable means of ascertaining whether AI was or was not used?

(c) Third, the fundamental point was that court users should be responsible for the documents they file. This would be the case regardless of whether the court user enlisted any assistance (human or AI) in doing so.

60. For these reasons, we chose to adopt a relatively ‘light touch’ governance framework which, in our view, rightly focuses on the basic and essential point – that anyone who files a document should bear full responsibility for what they file, regardless of whether AI (or indeed any other form of assistance) was used in the preparation of that document.

61. Those three qualities – curiosity, critical acumen and courage – complement and reinforce each other and I suggest that they ought to shape our thinking on judicial leadership. Standing alone, each paints a picture which is incomplete. Curiosity is, on its own, directionless. Courage alone is indulgent. And critical acumen, without more, is unproductive and can lead to paralysis. But combine the three, and we have, I think, the beginnings of wisdom, and a solid foundation for the sort of leadership we will need to navigate the most pressing challenges of our time.

IV. Conclusion: How might we develop a culture of judicial leadership?

62. Let me conclude by sharing some thoughts on the third and last of the questions I had opened with: How might we begin to cultivate a culture of leadership that embodies curiosity, critical acumen and courage? While this remains a work in progress, let me offer three suggestions for how we might approach this.

63. The first relates to creating the conditions necessary for leadership to flourish at all levels – especially at the grassroots. As I have said on previous occasions, leadership is not an appointment, title or office. Leadership at its core, is about a willingness to act and thereby initiate change driven by the desire to better things for those around us. That is an inherent human quality that we should all strive to cultivate. Anyone – regardless of seniority – can be a leader. And if anyone can be a leader, we should empower everyone to lead.⁽³⁷⁾

64. Each court will have its own way of incubating innovation and canvassing ideas for change in their system. In Singapore, the junior judiciary – comprising the State Courts and the Family Justice Courts – identifies their own strategic objectives and designs their own initiatives to achieve those goals by developing annual workplans.⁽³⁸⁾ To do this, they encourage judicial officers at every level to think of how they can make their court, their workplace, their part of the justice system, function better or more optimally.⁽³⁹⁾

65. My second suggestion concerns the role of judicial education. A few years ago, we made a strategic decision to significantly upgrade our investment in judicial education, and that has been one of the best decisions we made. The College has been instrumental in leading our efforts, and these programmes are among its key pieces in the judicial education space. But as important as the College is in driving and setting the agenda for judicial education, its work represents only half of the story.⁽⁴⁰⁾ The project of judicial education is, at bottom, a partnership between the College and its learners. The College exists to equip and resource learners and help them learn how to learn, but ultimately every learner has a responsibility to take ownership of their professional development.⁽⁴¹⁾ To that end, the College has developed a number of tools and resources. We have a Judicial Competency Framework with 14 competencies that act as a scaffold for the development of Singapore judges. This Framework not only informs the College’s programming, but also facilitates self-directed and self-paced learning. It equips judges to identify the competencies they need to hone, and to pursue opportunities for developing them.⁽⁴²⁾ 

66. My final suggestion pertains to the value of building networks and communities of leaders and learners, both within and across jurisdictions. This is something I hold close to my heart, because I firmly believe that we, as judges, are all part of the same judicial family. Judges have a unique familial bond with one another because we stand unique in this shared endeavour to administer justice, and there is an enormous amount we can learn from each other, regardless of our particular laws, our systems of justice, or our legal traditions. We are all in the business of dispensing a fundamental human value – justice – and that is why we have so much to learn from each other. 

67. At the institution-to-institution level, the Singapore Judiciary is a champion of the value of international engagement and exchange. We regularly convene international engagements with our counterparts from abroad on a range of issues, from transnational commercial law to technology to judicial education and pedagogy. We have benefited immensely from those platforms, as we hope our counterparts have from our participation in them. 

68. That sentiment is true also at the people-to-people level, and is perhaps best exemplified by programmes like the one you are now participating in, where you get to sit with, interact with and learn together with colleagues and friends from other jurisdictions. As we come alongside each other we are also learning from each other and in the process we cross-pollinate and cross-fertilise our perspectives so that we come to a better understanding of what judicial leadership might entail and how we can do it better.

69. I hope that these remarks have given you some food for thought on the why, what and how of judicial leadership. We now have some time for questions, and I very much look forward to hearing from you and discussing any questions you might have. Thank you very much.