GPT-Legal Tool to Provide Efficient Summaries for Over 15,000 Court Judgments in LawNet Database
Starting from September, Singapore’s legal resource database, LawNet, will introduce a new generative AI tool called GPT-Legal, aimed at enhancing research efficiency for lawyers. Announced by Chief Justice Sundaresh Menon at the annual mass-call ceremony for new lawyers, GPT-Legal will be able to generate summaries for over 15,000 court judgments stored in LawNet.
The large language model tool, developed by the Singapore Academy of Law (SAL) and the Infocomm Media Development Authority, was trained on legal data from LawNet’s comprehensive repository, which includes court judgments, academic articles, journals, and textbooks. The tool will feature built-in trust and safety features to ensure accuracy, such as an accuracy score for each summary and the ability to pinpoint corresponding paragraphs, allowing users to verify the information quickly.
In addition to GPT-Legal, SAL is launching two initiatives with Microsoft to help lawyers navigate AI in legal practices. The first is an online course on generative AI, providing an understanding of the technology and its applications, while the second is a guide for lawyers on how to craft effective AI prompts.
AI’s role in the legal sector continues to expand, with applications such as document drafting, due diligence, and even case outcome predictions. Legal analytics tools like Lex Machina and Solomonic are already in use in other jurisdictions, and Singapore is exploring AI tools for the Small Claims Tribunal as well.
The legal profession in Singapore is increasingly recognizing the importance of embracing AI technology to improve service delivery, and the introduction of GPT-Legal marks a significant step in this evolution.
