The Minister of Justice of Lower Saxony in Germany Visits TRAIL

TRAIL had the privilege of hosting Minister of Justice Dr Kathrin Wahlmann and her delegation from the Ministry of Justice of Lower Saxony in Germany on 21 February 2025. The delegation comprised Judge Dr Daniel Stolz, Judge Dr Michael Henjes, Ms Verena Brinkmann, Ms Julia Zwake and Mr Patrick Dahm, a German lawyer who has been working as a foreign lawyer in Singapore. They were particularly interested in TRAIL’s research on the legal and technical ramifications arising from the use of AI in the legal industry, with a focus on the courts in particular.

Hosting the delegation was A/Prof Daniel Seng, Co-Director of TRAIL, along with Masters of Law students from his AI, IS and the Law class. After a brief introduction of TRAIL, Prof Seng explained the workings of Large Language Models (LLMs), which represents the most interesting and topical development that has spearheaded the use of AI in the legal space. Prof Seng focused on the hallucination problem with LLMs, illustrating it with the infamous “How Many Rs are there in Strawberry” problem, and also referred to the Mata v Avianca, Inc case, where two New York lawyers were sanctioned for using ChatGPT to generate briefs submitted to the courts in the Southern District of New York that had actually contained three fake cases with fake quotes and citations.

Prof Seng then referred to the Singapore Supreme Court Registrar’s Circular No 1 of 2024: Guide on the Use of Generative AI Tools by Court Users as well as the UK’s AI: Guidance for Judicial Office Holders, noting that both flagged the issues of hallucinations and emphasized that responsibility remained on users of these tools to check the accuracy of any documents generated by Generative AI. Prof Seng also shared some examples to illustrate the difficulty that users face in verifying that the information generated by such tools is accurate. For a typical use case such as  summarization, even Apple was forced to withdraw its AI feature after receiving complaints that the tools were making repeated mistakes in its summaries of news headlines.

The discussion included how the risks and problems associated with the use of such tools in the courts could be managed, the limitations that an LLM tool trained only on a particular judge’s judgments would pose, and how reliance on LLM tools would short-circuit legal training, especially for junior lawyers and members of the bench.

The Masters’ students also shared their perspectives on how they managed the problems associated with the use of LLM tools.

Professor Andrew Simester, Dean of NUS Law, highlighted that with technology changing the legal landscape, it is imperative for a law school to introduce students to these rapid advancements to prepare them to work with technology in the legal profession.

A/Prof Daniel Seng giving an overview of the problems with LLMs

(left to right) Judge Dr Daniel Stolz, Judge Dr Michael Henjes, Minister Dr Kathrin Wahlmann

Minister Dr Kathrin Wahlmann talks about the need for courts to adopt AI to improve efficiency and speed up court processes

Mr Ulices (Master’s student) describes how he would use AI tools in practice

Mr Patrick Dahm (right) translating the discussion for Minister Wahlmann

Joining the session were Master’s students Ms Tsui and Mr Kaushik

Professor Andrew Simester giving his insights into legal education

(First row from left to right) Judge Dr Daniel Stolz, Judge Dr Michael Henjes, Minister Wahlmann, A/Prof Daniel Seng and Dean Andrew Simester. (Second row from left to right) Mr Patrick Dahm,  (Master’s students) Ms Gail Wong, Ms Jung Chih Lin, Ms Tsui Chloe Sin Wai, Mr Shivam Kaushik, Mr Fadel Muhammad, Mr Aguilar Romeo Arthur Ulices and Ms Watcharavasunthara Chalida.