Collection
Applications and Evaluation of Large Language Models in the Legal Domain
Large Language Models (LLMs), including both proprietary and open-source, are presently being applied in many diverse domains, including finance, healthcare, the soft sciences, engineering, and education. In the legal domain, LLMs are increasingly being used by researchers in many diverse tasks and, despite cautions issued by the U.S. Federal judiciary and guidelines established by the European Bar Association, legal practitioners are starting to rely upon them as well. For instance, GPT-3 has been applied for statutory reasoning on U.S. statutes. GPT-4 has been shown to be able to pass the Uniform Bar Examination (UBE), at times outperforming humans. LLMs have also been leveraged for tasks such as sentence-level annotation of legal text, investigating digital evidence, contract drafting, as well as others.
In contrast to these positive developments, there are several potential risks that can arise from the application of LLMs in the legal domain. For instance, generative AI models including LLMs have been found to hallucinate and generate inconsistent text while summarizing legal documents. Moreover, LLMs are known to exhibit various types of social biases that can be particularly harmful in the legal domain.
The purpose of this collection is to showcase various applications of LLMs in the legal domain, including analysis and generation of various types of legal text, as well as attempts to explore and mitigate the potential risks and ethical concerns associated with such applications of LLMs in this domain.
Editors
-
Jack G. Conrad
Jack G. Conrad is Director of Applied Research Emeritus with Thomson Reuters, Minneapolis, USA. He focuses on a broad range of technical application areas involving AI, ML and textual data processing. For over two and a half decades, he has delivered critical artifacts and infrastructure for research and business directed projects across a diverse spectrum of domains. He has published more than 50 peer-reviewed research papers and has 9 patents. He is past president of the International Association for Artificial Intelligence and Law and has served on the IAAIL Executive Committee for 8 years. Email: jackgconrad@gmail.com
-
Kripabandhu Ghosh
Kripabandhu Ghosh is an Assistant Professor at the Department of Computational and Data Sciences, IISER Kolkata, India. He has been working in Legal Analytics since his PhD. His papers have received awards at the two most recognized international conferences in AI and Law – Best Paper award at JURIX 2019 and Best Student Paper award at ICAIL 2021. He has organized several AI-Law events, including a workshop with an international conference, shared tasks such as AILA and SAIL. He is currently serving as an Editor Board member of AI and Law journal. Email: kripaghosh@iiserkol.ac.in
-
Debasis Ganguly
Debasis Ganguly is presently an Assistant Professor in the University of Glasgow, UK. Prior to that he was a research scientist in IBM Research, Ireland. His general research interest is directed towards mitigating cognitive biases from AI systems, developing privacy preserving AI models, and working towards developing more interpretable AI and information retrieval systems. Over the years he has published over a hundred research papers in top-tier conferences and journals, such as SIGIR, CIKM, ECIR, TOIS, IPM etc. Email: Debasis.Ganguly@glasgow.ac.uk
-
Saptarshi Ghosh
Saptarshi Ghosh is an Associate Professor of Computer Science at Indian Institute of Technology, Kharagpur, India. His research interests include Legal analytics, Social media analytics, and Algorithmic bias and fairness. He has published more than hundred research papers, and presently leads a Max Planck Partner Group focusing on algorithmic bias and fairness. His works on AI & Law have been awarded the Best Paper award at JURIX2019 and the Best Student Paper Award at ICAIL2021 conferences. He is presently the Section Editor on Legal Information Retrieval for the AI and Law journal. Email: saptarshi@cse.iitkgp.ac.in
