Abstract: We have been developing the PROLEG (PROlog-based LEGal reasoning support) system to simulate judgement reasoning after all the relevant facts are determined. In this work, we modify the PROLEG system to support arranging issues in civil litigation. Given a desired effect requested by one party, our interactive system (which we call int-PROLEG) automatically calculates possible justifications for the desired effect using PROLEG rules related with the Japanese civil law and case rules established by the Japanese supreme court.

Bio: Ken Satoh is a professor from National Institute of Informatics in Japan. He has been studying logical foundations of artificial intelligence for 30 years and published more than 150 papers (See http://research.nii.ac.jp/~ksatoh/). Currently he is very much interested in application of logical framework to law such as a support system for judges and a compliance mechanism of AI. In preparation for this research topic, he entered the law school of the University of Tokyo in 2009 and passed the Japanese bar exam in 2017.

Abstract: Legal documents often tell stories, and NLP tools can help us identify and quantify the elements of those stories. In this talk, I will present research on using information extraction technologies -- syntactic parsing, semantic role labeling, entity resolution -- to the task of automated understanding and interpretation of legal documents. Further, I will explore machine reading comprehension using pre-trained language models as an alternative to the more structured approaches. The main application is a large corpus of 30,000 collective bargaining agreements from Canada.

Bio: As Assistant Professor of Law, Economics, and Data Science at ETH Zurich's Center for Law & Economics, Elliott Ash works on empirical analysis of the law and legal system using techniques from econometrics, natural language processing, and machine learning. Elliott's research has been published in the Journal of Law and Economics, Economic Journal, Journal of Public Economics, Cornell Law Review, Georgetown Law Journal, Journal of Politics, Political Analysis, and the proceedings of leading conferences such as the Association of Computational Linguistics. That work has been funded by an ERC Starting Grant as well as additional grants from the Swiss National Science Foundation, U.S. National Science Foundation, the Turing Institute, the Washington Center for Equitable Growth, and the Hasler Foundation. Prior to joining ETH, Elliott was Assistant Professor of Economics at University of Warwick, and before that a Postdoctoral Research Associate at Princeton University’s Center for the study of Democratic Politics. He received a Ph.D. in economics and J.D. from Columbia University, a B.A. in economics, government, and philosophy from University of Texas at Austin, and an LL.M. in international criminal law from University of Amsterdam.

Abstract: Concepts are one of the most important and the most investigated types of mental representations in cognitive science. They are considered to be the crucial building blocks of human knowledge and to enable different types of conscious and non-conscious inference. In particular, they play essential role in the process of language acquisition and use. In computational modeling of legal reasoning there exists a tension between the line of research stressing the role of concepts and, on the other hand, the stream which avoids conceptual modeling. I am going to explore the strengths and weaknesses of either of the approaches and outline a proposal of a middle ground theory. The three main facets of this theory are (1) linking conceptual modeling to linguistic indicators; (2) indicating a catalogue of different conceptual structures and (3) employing the notion of conceptual net and conceptual variations.

Bio: Dr Michał Araszkiewicz is an assistant professor (adiunkt) in the Department of Legal Theory at the Jagiellonian University in Kraków and holds a PhD in legal theory. Michał Araszkiewicz has published extensively in the field of legal theory and in the area of AI and Law. He specializes in theories of legal reasoning and argumentation, legal interpretation, case-based reasoning as well as in normative aspects of Artificial Intelligence, including the right to explanation. He is currently a member (Secretary-Treasurer) of the Executive Committee of the International Association for Artificial Intelligence and Law (IAAIL) and of the Steering Committee of JURIX. He served as the President of the ARGDIAP association (argdiap.pl) – a NGO focused on the problems of argumentation, dialogue and persuasion. He has co-organized numerous scientific events including the JURIX 2014 (Conference Chair) and four consecutive editions of XAILA – The EXplainable and Responsible AI in Law workshops at JURIXes 2018-2020 and ICAIL2021. He is also a legal advisor, partner in Araszkiewicz Cichoń Araszkiewicz Law Firm (acrlegal.pl). In legal practice he specializes in the field of legal regulation of AI as well as in Intellectual Property, Data Protection and broadly understood Protection of Information.

Abstract: Online Dispute Resolution (ODR) is an alternative to traditional litigation that helps litigants without an attorney and improves court efficiency. Most ODR systems require a neutral facilitator, but there is a shortage of facilitators that limits the adoption of ODR systems. This technology uses computational linguistics and machine learning to (1) monitor cases to detect situations requiring immediate attention and (2) automate selection of standard text messages appropriate to the current state of the negotiations. This technology can compensate for shortages of facilitators by improving the efficiency of experienced facilitators, assisting novice facilitators, and providing autonomous facilitation.

Bio: L. Karl Branting is a Chief Scientist in Human Language Technology department of The MITRE Corporation. Dr. Branting was the 2004-2005 president of the International Association for Artificial Intelligence and Law, a 2000-2001 United States Supreme Court Fellow, and recipient of NSF Career and Fulbright Senior Scholar grants while a professor at the University of Wyoming. Dr. Branting has a Ph.D. in computer science from the University of Texas at Austin, a J.D. from Georgetown University, and a B.A., magna cum laude in philosophy, from the University of Colorado.

Abstract: This presentation will consider the question of how we ensure that AI-enabled technologies, when adopted by legal practitioners, are applied in an effective and trustworthy manner. Using fact-finding in civil and criminal proceedings as a case study, a task to which AI-enabled technologies have been applied for over 15 years (often called “e-discovery” in the US), we will consider challenges to the trustworthy adoption of advanced technologies in the real-world practice of law. We will then consider the means to meeting those challenges, focusing chiefly on the development of practically viable protocols for gathering the evidence needed to ground the adoption and use of AI-enabled technologies in a sound understanding of the technologies’ effectiveness. Having reviewed the specific circumstances of legal fact-finding, we will conclude by drawing generally applicable lessons for ensuring that, to the extent that AI is adopted in the service of the law, it is done so on the basis of an informed trust.

Bio: Dr. Bruce Hedin is a leading expert in the assessment of the effectiveness of advanced search and analytics technologies at performing legal tasks. As a consultant, he supports clients in the design and oversight of sampling and measurement protocols to validate the results of AI-enabled review technologies and provides guidance to counsel engaged in meet-and-confer discussions regarding the use of AI-enabled review and retrieval processes. Dr. Hedin has contributed to a number of initiatives that seek to ensure that, to the extent AI is adopted in the service of the law, it is done so on the basis of an informed trust. Example initiatives include the US National Institute of Standards and Technology’s Text Retrieval Conference (US NIST TREC) Legal Track (where he served as coordinator from 2008 through 2011) and the Law Chapter of IEEE’s Ethically Aligned Design, First Edition (where he served as a member of the multi-disciplinary drafting team), which establishes a framework for the trustworthy adoption of AI in legal applications and the justice system at large. Dr. Hedin has contributed articles to publications such as Perspectives on Predictive Coding and Other Advanced Search and Review Technologies for the Legal Practitioner, The Journal of Artificial Intelligence and Law, Current Challenges in Patent Information Retrieval, and New York Law Journal. Dr. Hedin earned his Ph.D. from Stanford University and his B.A. from Cornell University.

Abstract: Modelling the arguments relevant to the consideration of legal cases has been a central topic of AI and Law since its very beginnings, and a number of approaches have been developed. I will describe how these various approaches can complement one another by being seen as addressing different stages of the reasoning process. Each stage can be modelled using argumentation schemes. I will look at one stage in particular, the move from facts to reasons to decide for one or other of the parties, and present some novel argumentation schemes to model this stage.

Bio: Trevor Bench-Capon has researched into Artificial Intelligence and Law since the early 1980s, initially at Imperial College London, but since 1987 at the University of Liverpool, and has written some 300 journal and conference papers. Although formally retiring in 2012, he has remained research active and continues to hold an Honorary Professorship at Liverpool. He has been Programme Chair of both ICAIL and JURIX and was President of the International Association for AI and Law from 2001-3. He is Editor in Chief of AI and Law Journal. He has also been active in the Computational Argumentation community and was Organising Chair of the inaugural COMMA conference in 2006.

Abstract: Open government policies provide citizens the right to access government documents through Freedom of Information laws. However, in the UK and many other countries, exemptions to Freedom of Information laws prohibit the disclosure of specific types of information that might cause harm to individuals, organizations, or nations. Therefore, it is vital that government documents are manually reviewed to identify and protect any sensitive information before the documents can be released to the public. The transition to digital documents (including emails) has brought new challenges in reviewing documents for public release, and there is a need for information retrieval technologies to be able to assist the digital sensitivity review process. However, automatic sensitivity identification is a challenging task, since sensitivity is typically context-dependent and results from a combination of factors, such as who said what about whom and when. I will discuss some of the challenges in automatically identifying sensitive information and present some of our work on developing approaches for assisting human reviewers to sensitively review large collections of digital documents.

Bio: Graham McDonald is a lecturer in Information Retrieval (IR) in the School of Computing Science, University of Glasgow. His main research interests focus on developing novel approaches for automatically classifying sensitive information and developing active learning strategies that can adapt to and learn from human users in decision support systems. Indeed, his PhD thesis was titled A Framework for Technology-Assisted Sensitivity Review: Using Sensitivity Classification to Prioritize Documents for Review, and he has published research on sensitivity classification, active learning strategies and assisting human reviews in top Information Retrieval venues such as SIGIR, ECIR, CHIIR, TOIS and IRJ. His research on sensitive information has been at the forefront of shaping solutions within the archival and governmental sectors to enable the safe release of digital information from public bodies. Moreover, he is part of a consortium that is developing the first large-scale technology-assisted sensitivity review platform for the UK Government. This has resulted in the first UK Government department successfully sensitively reviewing and transferring a large collection of digital documents to The National Archives in October 2021.

Abstract: Westlaw Quick Check is an established brief analysis product, competing with products from CaseText, Lexis, and Bloomberg Law. This talk will tell how the Quick Check team learned what customers want and how to provide solutions despite major unsolved basic research questions in AI and Law.

Bio: Tom Vacek is a senior Applied Research Scientist at Thomson Reuters. Since 2012, he has had significant contributions to the Research Recommendations, Litigation Analytics, and Quick Check products and has publications at venues for legal AI and basic machine learning research. He has a Master's degree in Computer Science from the University of Minnesota.