In November 2024, the Singapore Courts hosted the International Association for Court Administration (IACA) Conference in Singapore for the first time. The event, themed “Building Trust in The Judiciary,” gathered 250 judges and senior court administrators from over 40 countries to discuss key topics like AI in the justice system and access to justice. Thomson Reuters is proud to be a sponsor at the event.

Representing Singapore, Mr. Tan Ken Hwee, Chief Transformation and Innovation Officer of the Singapore Courts, delivered a keynote on AI in the judicial system. The keynote highlighted the various aspects of AI applications to legal texts. Mr. Tan addressed both the transformative potential and the ethical challenges of AI in legal contexts. Read on for the highlights of his address to the legal community. 

Mr. Tan Ken Hwee, Chief Transformation and Innovation Officer, Singapore Courts, delivered a keynote at the IACA conference

Courts can use AI to produce case summaries 

AI tools can provide reasonably reliable summaries of cases to the judiciary, shared Mr. Tan. “You can now feed an AI or large language models (LLMs) the material in a case and ask it to generate a summary and explain it to you, to give you a sense of what the case is all about,” he said. 

Whilst “you cannot expect it to be 100% comprehensive”, warned Mr. Tan, it is “highly usable”. 

AI can help give you a quick flavour of what a particular case is about. But when it comes to eliminating the risk of inaccuracies altogether, this is not the way to look at it. Mr. Tan reminded the audience that “…hallucination is, in a sense, a feature and not a bug of this technology.” 

AI for gathering evidence 

Another significant area where AI can make a difference is gathering evidence. Mr. Tan addressed the rising potential for increased efficiencies.

“…The ability of LLMs (large language models) to be able to help us sift through evidence and synthesise it and give us a composite document summarising the evidence is potentially a huge game changer,” he said. 

Scheduling cases with AI 

Should courts be using AI to schedule cases? Mr. Tan noted that the courts could use algorithms to optimise time. Specifically, to minimise the amount of unscheduled time and assigned time. 

“You can optimise court usage. You can optimise the time of judges much better if you have such technology. But the key thing here is this is not really a particularly good use of generative AI. It is instead something that can be done with a lot of algorithms and constraint programming techniques that is available, but it is nevertheless something that should still deserve a certain degree of attention, but it’s often talked about in the rubric of artificial intelligence, and so it is something that I think deserves attention.” 

Helping underrepresented litigants prepare for cases 

Another area of significant interest to the Singapore courts would be the ability to help underrepresented litigants prepare for their cases. As Mr. Tan mused, the use case would be to organise thoughts and materials to prepare for court hearings with AI. “This is something that we are actively exploring,” he said.   

“We think that this is especially valuable for unrepresented individuals, to help them to organise and to structure their material, which in turn then helps the court decide that case,” he said. 

Possible “AI” use cases in justice system

Traffic light system for use of AI in legal tech 

In summary, Mr. Tan proposed a set of traffic lights to navigate the use of AI in legal tech. 

“I think the prediction of outcomes is something that we might want to be extremely careful about,” he warned. 

“The use of AI to decide cases is also something we need to be extremely careful and possibly consider to be a red-light area. There is a huge difference in LLMs when it comes to language versus knowledge. 

“What if the AI tells you that there’s an 80% chance that the plaintiff is the one that should win this case and not the defendant? What do you do with that number? Right? So, I propose that those should go into the red zone.” 

Conversely, access-to-justice use cases, and the possibility of using AI to help draft documents, both in court and out of court would suit comfortably within the yellow zone, according to Mr. Tan. 

“I will also say, well within the green-zone, and this is perhaps the areas that are really receiving a lot of attention, would be the use of AI to do evidence review, to generate summaries, and to do speech-to-text transcription,” he added. 

Traffic light system for use of AI in legal tech, proposed by Mr. Tan.

Thomson Reuters Singapore would like to thank Mr. Tan Ken Hwee for enabling Legal Insight Southeast Asia to share his commentary with our readers. 

Thomson Reuters Legal Solutions for Judiciary and Courts

Explore court management software solutions that enable court clerks, staff, and lawyers to administer their courtroom duties efficiently and securely. View more here.

Event photo