Justice Goh Yihan: Speech at the Inaugural Singapore-India Conference On Technology

INAUGURAL SINGAPORE-INDIA CONFERENCE ON TECHNOLOGY

“Artificial Intelligence in Judicial Training and Education: Potential Use of Artificial
Intelligence in Training and Education of Judges”

Saturday, 13 April 2024

Response by the Honourable Justice Goh Yihan
Supreme Court of Singapore

Introduction

1                 Good evening. I am delighted to have the opportunity to respond to the excellent speeches made earlier.

2                 It is heartening that there is broad agreement of the importance in the training and education about AI. It is clear that advances in AI will create opportunities for those willing to work with AI. However, it is vital that legal and judicial education provide lawyers and judges with what is needed to take advantage of these opportunities.⁽¹⁾

3                 With these points in mind, and without intending to summarise or repeat what has already been said, I make three points in response: (a) what needs to be taught about AI, (b) what should be taught apart from AI, and (c) possible new teaching methods arising from AI.

What needs to be taught about AI?

4                 I begin with what needs to be taught about AI. It is clear from the speeches that AI tools will revolutionise the practise of law. As such, there needs to be instruction on how to use AI within the legal field, much like how LexisNexis and Westlaw are already integrated into legal education. In saying this, I leave aside the quite obvious proposition that students will need to know about the substantive law generated by AI, such as whether AI should be ascribed legal rights or be exposed to legal liability.

5                 It is not hard to imagine a future where students will routinely use AI tools in their legal education. Because there will be a financial impact associated with these programs, education providers need to decide on which tools to purchase.⁽²⁾ This may in turn depend on which AI tool ends up being the front-runner much like how LexisNexis and Westlaw have emerged ahead of their competitors. But the wait for a front-runner should not delay matters. There are other ways to expose students to AI tools. Indeed, in the United States, it is common to have law students work with lawyers to assist startup and other new technological companies.⁽³⁾ This exposes students to tools being used in the real world. More broadly, at this stage, it may not be the mastery of a particular AI tool that matters, but the general familiarity of the use and relevance of such tools. It is not inconceivable that something similar can be worked out for judicial education, even as one needs to be careful about the conflict of interests.

6                 Equally importantly, judges need to be taught about the limitations of AI and how these may affect the very tools they use. For example, in the Guidelines for Use of Generative AI in Courts and Tribunals issued by the Courts of New Zealand to their judges, it is stressed that judges need to understand not only the capabilities of Generative AI chatbots, but also of their limitations. The Guidelines stress, for instance, that such chatbots are not search engines and certainly do not provide answers from an authoritative database. It is also important to be mindful of the data used to train the chatbot, and if such data is representative of the local context. Perhaps most importantly, judges need to be taught about biases that may be inherent in any output from a Generative AI chatbot. Ultimately, judges need to be taught how to discern inaccurate output even as they rely on AI tools to enhance their work. In a sense, this is similar to how we always check the reliability of the primary source but in a different context altogether.

7                 Finally, judges, especially those who will set judicial policy going forward, should be taught about how the use of AI can potentially affect the judicial and legal system in a fundamental way. For example, Zuckerman suggests that lawyers’ use of AI tools to augment their skills may blunt their professional aptitude.⁽⁴⁾ The more AI is used for legal research, the less may be needed from lawyers. Thus, as more of their traditional functions are handed over to machines, lawyers’ forensic experience will diminish. Lawyers may acquire narrower skillsets, which can have a profound impact on how the legal system functions and, in so far as judges are drawn from lawyers, how judges will become.⁽⁵⁾ Zuckerman makes the analogy with airline pilots. As more flying is done by onboard computers, pilots accumulate less experience of flying and may find it more difficult to react to unexpected emergencies. This is not an argument against AI flying planes. But the point is that AI may well affect legal expertise.⁽⁶⁾ Further, Zuckerman posits that if AI were to become commonplace in the future, and lawyers and judges all have AI, would there be the need for an appeal against AI? Can there be a continued utility in an appeal, whose institution rests on, among others, error correction?⁽⁷⁾ Judges who drive judicial policy will need to be exposed to the possibility of these foundational ideas, even if they sound far-fetched today. Indeed, these scenarios may turn out to be very ones to guard against materialising.

What needs to be taught apart from AI?

8                 It seems obvious that much needs to be taught about AI. But we must not neglect what needs to be taught apart from it. To begin with, judicial educators might want to refocus on areas that AI tools have not been able to replicate.⁽⁸⁾ It has been suggested that law schools must continue to train students in three areas: (a) engage in high-level critical analysis; (b) provide creative solutions to complicated problems; and (c) provide emotive client-focused representation.⁽⁹⁾ These three areas require students to integrate knowledge from different disciplines to provide creative solutions to important contemporary problems.⁽¹⁰⁾ As creativity and interdisciplinary skills will continue to be human domains for a bit longer, they should form the basis for the training of future lawyers and judges. There are other skills that AI cannot yet replicate, and it would be beneficial for legal educators to build on those skills so that future lawyers/judges and AI complement each other.⁽¹¹⁾

9                 Further, there is a humanity about the law that we cannot neglect. As a personal example, recently, I wanted to return a mouse I purchased on Shopee. I mailed it out at a drop-off point but when the seller received it, he reported that there was no mouse. I was summarily denied a refund. I turned to the Shopee system, which asked me to submit any “evidence” I had. The only “evidence” I had was a close-up photo of the address label on the envelope with a bulge that, so I argued, contained the mouse. I received a response later saying that Shopee was not giving me a refund. No further explanation and no grounds of decision. I felt dissatisfied but there was no right of appeal or an avenue to seek permission to do so from a human. I eventually reached out to the seller to say that I really did not mail out an empty envelope. He replied to say he understands. In the end, I found some closure. Leaving considerations of proportionate justice aside, the point is that that justice may be seen to be done through the machine and not the judge. And that is something we need to guard against and educate our judges about, even as we harness the benefits of AI.

10                 As Justice Allsop so eloquently put it, logic, definition, and precise taxonomy may solve many problems, but human judgment in relation to others is central. As he put it, to mistake the machine for the Master will lead to the end of the spirit of the law as human and free. The danger is not the machine becoming human; it is the human becoming the machine. There is a humanity about the law, especially in criminal and family cases, that needs to be cultivated.⁽¹²⁾

11                 Ultimately, judicial education must undergo a process of transition from traditional knowledge acquisition to a systemically different one, built on the basis of technologies such as AI.⁽¹³⁾ For a future lawyer or judge, it is important to obtain not only a certain amount of theoretical knowledge, but also practical skills in the operation of AI tools. These skills include those that complement AI. But at the same time, judicial education in the age of AI cannot forsake the very humanity that lies at judging another human being.

New methods of teaching

12                 Finally, perhaps the most exciting aspect of AI in education is not only about content, but how it enhances the very teaching process. In a briefing post, the UK Parliament noted that AI tools have the potential to provide different ways of learning and to help educators with lesson planning, marking, and other tasks.⁽¹⁴⁾ There is no reason why AI cannot achieve the same in judicial education.

13                 We need only to look at how AI has revolutionised training in chess to understand the profound impact that AI can have in education. I still remember the debate from years ago about whether Gary Kasparov can defeat the IBM chess computer Deep Blue. Today, the debate is no longer about whether human chess players can defeat chess engines but how to work with chess engines to improve their game. Certainly, this may be why India has emerged as a top chess country, with five Indian players at the Candidates now being held in Toronto. Indeed, AI can help decision-makers learn; specifically, it can help them learn strategic interactions by serving as artificial training partners and thus help them to overcome a bottleneck of scarce human training partners.⁽¹⁵⁾ Leveraging the staggered diffusion of chess computers, researchers find that they did help chess players improve by serving as a substitute for scarce human training partners.

14                 It is not hard to imagine a future where judges hold moot sessions with very real AI avatars in place of reluctant classmates who awkwardly play the role of witness or opposing counsel, often with comedic but at times ineffective effect. Given how humans learn better by repeated immersion as opposed to rote learning, judges who have the opportunity of training with otherwise scarce human training partners should benefit immensely. They can also do so without concerns about causing real-world consequences until they are assessed to be ready. But in using these tools, students should also be aware of their limitations, especially being desensitised to what would otherwise be human behaviour. Indeed, research has shown that chess computers were not a perfect substitute, as players training with them were not exposed to and thus did not learn to exploit idiosyncratic human mistakes,⁽¹⁶⁾ much like mine.

15                 Apart from being a teaching tool, AI can also revolutionise the assessment method. For example, AI could lead to collection algorithms to provide detailed and personalised feedback from students. AI systems can show student abilities, repeat lessons, and design personalised learning plans for each student.⁽¹⁷⁾ Assuming confidentiality and privacy issues can be addressed, might it be possible for AI to “listen in” on a judge conducting a trial and provide feedback and personalised training programmes? This is not too far-fetched. Indeed, many of us have Amazon Alexa or Google Assistant which detects user patterns and gives user-tailored recommendations on what to buy and what to listen to. AI would thus be the powerful if persistently present virtual teaching assistant that provides truly personalised learning for individual judges.⁽¹⁸⁾ And what’s more, building on the collective wisdom of other learner judges.

16                 Ultimately, AI in the classroom has the potential of revolutionising judicial training. Stakeholders will need to learn how AI can be effective at delivering educational outcomes and provide training and guidance for judicial educators using these tools.

Conclusion

17                 To conclude, I hope I have left some food for thought through these three points.

18                 Thank you very much for what has been a really thought-provoking conference on the impact of technology on the law. It may perhaps be in education that the greatest future for progress lies. And it is rightly where one of our focuses needs to be.

19                 Thank you.

_{(1)       William Connell and Megan Hamlin Black, “Artificial Intelligence and Legal Education” (2019) 36 The Computer & Internet Lawyer 14 at 15.
(2)       William Connell and Megan Hamlin Black, “Artificial Intelligence and Legal Education” (2019) 36 The Computer & Internet Lawyer 14 at 16.
(3)       William Connell and Megan Hamlin Black, “Artificial Intelligence and Legal Education” (2019) 36 The Computer & Internet Lawyer 14 at 17.
(4)       Adrian Zuckerman, “Artificial intelligence – implications for the legal profession, adversarial process and rule of law” (2020) 136 LQR 427 at 441.
(5)       Adrian Zuckerman, “Artificial intelligence – implications for the legal profession, adversarial process and rule of law” (2020) 136 LQR 427 at 441.
(6)       Adrian Zuckerman, “Artificial intelligence – implications for the legal profession, adversarial process and rule of law” (2020) 136 LQR 427 at 441.
(7)       Adrian Zuckerman, “Artificial intelligence – implications for the legal profession, adversarial process and rule of law” (2020) 136 LQR 427 at 445.
(8)       Posting of Christian B. Sundquist to A Place to Discuss Best Practices for Legal Education Artificial Intelligence, Algorithmic Knowledge and the Future of Law Schools, https://bestpracticeslegaled.albanylawblogs.org/2018/04/09/artificial-intelligence-algorithmic-knowledge-and-the-future-of-law-schools/ (2018, April 9).
(9)       Posting of Christian B. Sundquist to A Place to Discuss Best Practices for Legal Education Artificial Intelligence, Algorithmic Knowledge and the Future of Law Schools, https://bestpracticeslegaled.albanylawblogs.org/2018/04/09/artificial-intelligence-algorithmic-knowledge-and-the-future-of-law-schools/ (2018, April 9).
(10)       Mateus de Oliveira Fornasier, “Legal Education in the 21st Century and the Artificial Intelligence” (2021) 19 Revista Opiniao Juridica 1 at Part 2.2.
(11)       William Connell and Megan Hamlin Black, “Artificial Intelligence and Legal Education” (2019) 36 The Computer & Internet Lawyer 14 at 16.
(12)       James Allsop, “The Legal System and the Administration of Justice in a Time of Technological Change: Machines Becoming Humans, or Humans Becoming Machines?”, a lecture delivered at the 2023 Sir Francis Burt Oration on 21 November 2023.
(13)       M A Yavorskiy, I E Milova and V V Bolgova, “Legal Education in Conditions of Digital Economy Development: Modern Challenges, in Svetlana Ashmarina and Anabela Mesquita, The Sundarbans: A Disaster-Prone Eco-Region (2020) at p 461.
(14)       UK Parliament POSTnote 712 (23 January 2024).
(15)       Fabian Gaessler and Henning Piezunka, “Training with AI: Evidence from chess computers” (2023) 44 Strategic Management Journal 2724.
(16)       Fabian Gaessler and Henning Piezunka, “Training with AI: Evidence from chess computers” (2023) 44 Strategic Management Journal 2724.
(17)       Mateus de Oliveira Fornasier, “Legal Education in the 21st Century and the Artificial Intelligence” (2021) 19 Revista Opiniao Juridica 1 at Part 3.
(18)       Mateus de Oliveira Fornasier, “Legal Education in the 21st Century and the Artificial Intelligence” (2021) 19 Revista Opiniao Juridica 1 at Part 3.}