Special Issue of the journal Artificial Intelligence on “AI & Law”

The access to justice crisis is one that cannot be effectively solved without the automation of legal services. The automation of legal services cannot be efficiently done without efficiently automating legal reasoning. Legal case-based reasoning (CBR) provides a method of obtaining explainable and strong predictions for legal issues that lawyers would typically predict on the basis of analogy to prior decided cases. Automating explainable predictions with regard to these sorts of legal issues is difficult without resort to CBR. Wider adoption of CBR in the legal realm therefore has the potential to increase the scope of legal services that can be automated. Despite this potential, as of early 2018 there were no open-source or commercially-available tools for building legal case-based reasoning systems. This paper describes an open-source tool named docassemble-openlcbr[5] designed for ease of use by legal professionals in implementing CBR in the development of automated legal services.

Computational models of relevance in case-based legal reasoning have traditionallybeen based on algorithms for comparing the facts and substantive legal issues of aprior case to those of a new case. In this paper we argue that robust models ofcase-based legal reasoning must also consider the broader social and jurisprudentialcontext in which legal precedents are decided. We analyze three aspects of legalcontext: the teleological relations that connect legal precedents to the socialvalues and policies they serve, the temporal relations between prior andsubsequent cases in a legal domain, and the procedural posture of legal cases,which defines the scope of their precedential relevance. Using real examples drawnfrom appellate courts of New York and Massachusetts, we show with the courts' ownarguments that the doctrine of stare decisis (i.e., similar facts should lead to similar results) is subject to contextual constraints and influences. For each of the three aspects of legal context, we outline an expanded computational framework for case-based legal reasoning that encompasses the reasoning of the examples, and provides a foundation for generating a more robust set of legal arguments.

Legal Reasoning Models

Using background knowledge in case-based legal reasoning: A computational model and an intelligent learning environment

Resumo: In this paper we propose an agent-based hybrid Intelligent Tutoring System (ITS) for Legal domain by combining Case-based Reasoning (CBR) and Rule-based System (RBS). This system has been developed taking into consid- eration a Problem-based Learning as a pedagogical approach. The idea is to engage law students in interactions with the ITS based on the resolution of legal problems. The start point of these interactions occurs when ITS submits a pe- nal situation to law students. Then, the students are expected to learn through two fundamental but different skills to solve legal problems. First, should have to know how to retrieve relevant cases and legal concepts about the cases, and second, how to use them effectively as examples for justifying positions in a legal argument. Abstract: In this paper we propose an agent-based hybrid Intelligent Tutoring System (ITS) for Legal domain by combining Case-based Reasoning (CBR) and Rule-based System (RBS). This system has been developed taking into consid- eration a Problem-based Learning as a pedagogical approach. The idea is to engage law students in interactions with the ITS based on the resolution of legal problems. The start point of these interactions occurs when ITS submits a pe- nal situation to law students. Then, the students are expected to learn through two fundamental but different skills to solve legal problems. First, should have to know how to retrieve relevant cases and legal concepts about the cases, and second, how to use them effectively as examples for justifying positions in a legal argument.

AI and Law: A fruitful synergy

A fuzzy legal case-based reasoning system has been developed for CISG (United Nation, Convention on Contract for the International Sale of foods). Since law is adversarial, there are usually at least two opposing viewpoints. Although either side may be quite reasonable, only one side wins in the end. An argument between two agents, such as the plaintiff and defendant, in a legal domain is very useful for classify a vague concept, because it can provide an explanation from the two sides. On the basis of the proposed structural similarity measure approach, a fuzzy legal argument is developed. This approach simulates the cognitive process of human beings in the legal argument. It can be used for students to learn the skill of the legal argument, or provides an advice for the attorneys in practice.

The paper gives ontologies in the Web Ontology Language (OWL) for Legal Case-based Reasoning (LCBR) systems, giving explicit, formal, and general specifications of a conceptualisation LCBR. Ontologies for different systems allows comparison and contrast between them. OWL Ontologies are standardised, machine-readable formats that support automated processing with Semantic Web applications. Intermediate concepts, concepts between base-level concepts and higher level concepts, are central in LCBR. The main issues and their relevance to ontological reasoning and to LCBR are discussed. Two LCBR systems (AS-CATO, which is based on CATO, and IBP) are analysed in terms of basic and intermediate concepts. Central components of the OWL ontologies for these systems are presented, pointing out differences and similarities. The main novelty of the paper is the ontological analysis and representation in OWL of LCBR systems. The paper also emphasises the important issues concerning the representation and reasoning of intermediate concepts.

LEGAL ARGUMENTS AND REASONING IN THE CONSTITUTIONAL LAW-GOVERNED STATE: THE COMMENTARY The interdisciplinary research on legal argumentation presented in this volume, entitled Legal Arguments and Reasoning in the Constitutional Law-governed State: The Commentary (edited by Monika Florczak-Wątor and Andrzej Grabowski), is primarily inspired by the theory of constitutional law-governed state developed in Italy, Spain, and Latin American countries, by scholars proposing doctrines of positivist or postpositivist constitutionalism and neoconstitutionalism. As explained by Andrzej Grabowski in the “Introduction” [pp. 23–29], the theory is focused first and foremost on legal reasoning as it is conducted in the process of judicial law application and with particular stress on how it is affected by constitutional norms and values. Legal theory on its own does not seem to possess sufficient means to examine legal reasoning in constitutional law-governed states adequately—such an endeavour might be done far better with the help of dogmatics of constitutional law. Hence, this commentary on 91 arguments, topoi, and legal reasoning schemata result from the research team’s joint efforts composed of 18 legal theorists and constitutionalists.

LEGAL ARGUMENTS AND REASONING IN THE CONSTITUTIONAL LAW-GOVERNED STATE: THE COMMENTARY The interdisciplinary research on legal argumentation presented in this volume, entitled Legal Arguments and Reasoning in the Constitutional Law-governed State: The Commentary (edited by Monika Florczak-Wątor and Andrzej Grabowski), is primarily inspired by the theory of constitutional law-governed state developed in Italy, Spain, and Latin American countries, by scholars proposing doctrines of positivist or postpositivist constitutionalism and neoconstitutionalism. As explained by Andrzej Grabowski in the “Introduction” [pp. 23–29], the theory is focused first and foremost on legal reasoning as it is conducted in the process of judicial law application and with particular stress on how it is affected by constitutional norms and values. Legal theory on its own does not seem to possess sufficient means to examine legal reasoning in constitutional law-governed states adequately—such an endeavour might be done far better with the help of dogmatics of constitutional law. Hence, this commentary on 91 arguments, topoi, and legal reasoning schemata result from the research team’s joint efforts composed of 18 legal theorists and constitutionalists.

INTRODUCTION The prototype proposed in Section 11.5 would transform legal information retrieval into AR. If, as argued there, some of the legal knowledge representation frameworks of Part I's computational models of legal reasoning, argument, and prediction can be annotated automatically in case texts, then a legal app could accomplish more than conceptual legal information retrieval. It could support cognitive computing. This chapter describes a cognitive computing environment tailored to the legal domain in terms of tasks, interface, inputs, and outputs and explains how type systems and annotations based on the computational models will help humans frame hypotheses about legal arguments, make predictions, and test them against the documents in a corpus. A hypothesis predicts how a legal issue should be decided, such as: – The plaintiff should win the issue of trade secret misappropriation where defendant deceived it into disclosing its confidential product data even though the information could have easily been reverse engineered. – The plaintiff can still show causation even though more than six months elapsed between the vaccination and the onset of the injury. – Plaintiff's claim for conversion should fail where she had not actually caught the baseball, even though the defendant intentionally interfered with her attempt. Posing and testing legal hypotheses like these is a paradigmatic cognitive computing activity in which humans and computers can collaborate, each performing the intelligent activities that they do best. Humans know the hypotheses that matter legally; the computer helps them to frame and test these hypotheses based on arguments citing cases and counterexamples. The type system annotations will enable a conceptual legal information system to retrieve case examples relevant to the hypotheses, generate summaries tailored to the users’ needs, construct arguments, and explain predictions. The chapter discusses challenges that still need to be addressed in order to construct these new CCLAs. How can the computational models of case-based, rule-based, and value-based legal reasoning and argumentation be integrated with conceptual legal information retrieval? What roles do the type system and pipelined text annotators play in this integration? What kind of manual conceptual annotation of training sets of documents will be required? What will CCLAs look like? How will they help humans to frame and test legal hypotheses?

Lawyers use a reasoning process known as legal reasoning to solve legal problems. Legal expert systems could potentially help lawyers solve legal problems more quick and adequately, enable students to study law at school or at home more easily, and help legal scholars and professionals analyze the law and legal systems more clearly and precisely.In 1992, Hajime Yoshino of Meiji Gakuin University started a “Legal Expert Systems” project. This “Legal Expert” project is funded by the Japanese Ministry of Education, Science and Culture and is scheduled to run from May 1992 to March 1998. Yoshino organized over 30 lawyers and computer scientists to clarify legal knowledge and develop legal expert systems.This project covers a wide range of technologies such as the analysis of legal knowledge, the analysis of legal rules on international trade (United Nations Convention on Contracts for International Sale of Goods (CISG)), legal knowledge representation, legal inference models, utility programs to develop legal knowledge bases, and user interfaces. This project, which ends in March 1998, will focus on developing comprehensive legal expert systems as the final product. In this issue, we present 12 papers written by “Legal Expert” project members.In this number, Hajime Yoshino gives are overview of the legal expert systems project, explaining its aims, objectives, and organization. Six papers that follow his introduction include three on case-based reasoning. Legal rules are given by ambiguous predicates, making it difficult sometimes to determine whether conditions for rules are satisfied by the facts given of an event. In such cases, lawyers often refer to old cases and generate hypotheses through analogical reasoning.Kaoru Hirota, Hajime Yoshino and Ming Qiang Xu apply fuzzy theory to case-based reasoning. A number of related systems have been developed, but most focus on qualitative similarities between old cases and the current case, and cannot measure quantitative similarities. Hirota et al. treat quantitative similarity by applying fuzzy theory, explaining their method using CISG examples.Ken Satoh developed a way to compute an interpretation of undefined propositions in a legal rule using adversarial case-based reasoning. He translated old cases giving possible interpretations for a proposition into clauses in abductive logic programming and introduced abducibles to reason dynamically about important factors in an old case to the interpretation suiting the user’s purpose.Yoshiaki Okubo and Makoto Haraguchi formalized a way of attacking legal argument. Assume that an opponent has constructed a legal argument by applying a statute with an analogical interpretation. From the viewpoint of legal stability, the same statue for similar cases should be applied with the same interpretation. We thereby create a hypothetical case similar to the case in question and examine whether the statue can be interpreted analogically. Such a hypothetically similar case is created with the help of a goal-dependent abstraction framework. If a precedent in which a statue has been applied to a case with a different interpretation – particularly complete interpretation – can be found, the opponent’s argument is attacked by pointing out the incoherence of its interpretation of the statue.Takashi Kanai and Susumu Kunifuji proposed a legal reasoning system using abductive logic programming that deals with ambiguities in described facts and exceptions not described in articles. They examined the problems to be solved to develop legal knowledge bases through abductive logic programming, e.g., how to select ambiguities to be treated in abductive reasoning, how to describe time relationships, and how to describe an exception in terms of the application of abductive logic programming to legal reasoning.Toshiko Wakaki, Ken Satoh, and Katsumi Nitta presented an approach of reasoning about dynamic preferences in the framework of circumscription based on logic programming. To treat dynamic preferences correctly is required in legal reasoning to handle metarules such as lex posterior. This has become a hotly discussed topic in legal reasoning and more general nonmonotic reasoning. Comparisons of their method, Brewka’s approach, and Prakken and Sartor’s approach are discussed.Hiroyuki Matsumoto proposed a general legal reasoning model and a way of describing legal knowledge systematically. He applied his method to Japanese Maritime Traffic Law.Six more papers are to be presented in the next number

A hybrid information retrieval (IR) system to drive case-based reasoning (CBR) in legal domains is presented. The paper points out the assumptions made and processes involved in a typical mental conceptualization an intuitive model of how human beings represent, recognize and retrieve concepts. The world of general discourse is simplified into a narrow domain, limiting the set of concepts, terms, and structures. The limitation allows the application of stereotypical scenarios. Legislators' professional work creates, based on the logic of law, a natural basis for an indexing and retrieving structure which can be used as a front-end processor for a legal CBR system. The practical application of the combined IR-CBR system is demonstrated by experiments in Bankruptcy Case Law. The natural language description of a novel situation (case text) is the input, which is automatically classified for the production of similar cases' lists.

Modelling argumentation based on legal cases has been a central topic of AI and Law since its very beginnings. The current established view is that facts must be determined on the basis of evidence. Next, these facts must be used to ascribe legally significant predicates (factors and issues) to the case, on the basis of which the outcome can be established. This thesis aims to provide a method to encapsulate the knowledge of bodies of case law from various legal domains using a recent development in AI knowledge representation, Abstract Dialectical Frameworks (ADFs), as the central feature of the design method. Three legal domains in the US Courts are used throughout the thesis: The domain of the Automobile Exception to the Fourth Amendment, which has been freshly analysed in terms of factors in this thesis; the US Trade Secrets domain analysed from well-known legal case-based reasoning systems (CATO and IBP); and the Wild Animals domain analysed extensively in AI and Law. In this work, ADFs play a role akin to that of Entity-Relationship models in the design of database systems to design and implement programs intended to decide cases, described as sets of factors, according to a theory of a particular domain based on a set of precedent cases relating to that domain. The ADFs in this thesis are instantiated from different starting points: factor-based representation of oral dialogues and factor-based analysis of legal opinions. A legal dialogue representation model is defined for the US Supreme Court Oral Hearing dialogues. The role of these hearings is to identify the components that can form the basis of an argument that will resolve the case. Dialogue moves used by participants have been identified as the dialogue proceeds to assert and modify argument components in term of issues, factors and facts, and to produce what are called Argument Component Trees (ACTs) for each participant in the dialogue, showing how these components relate to one another. The resulting trees can be then merged and used as input to decide the accepted components using an ADF. The model is illustrated using two landmark case studies in the Automobile Exception domain: Carney v. California and US v. Chadwick. A legal justification model is defined to capture knowledge in a legal domain and to provide justification and transparency of legal decisions. First, a legal domain ADF is instantiated from the factor hierarchy of CATO and IBP, then the method is applied to the other two legal domains. In each domain, the cases are expressed in terms of factors organised into an ADF, from which an executable program can be implemented in a straightforward way by taking advantage of the closeness of the acceptance conditions of the ADF to components of an executable program. The proposed method is evaluated to test the ease of implementation, the efficacy of the resulting program, the ease of refinement, transparency of the reasoning and transferability across legal domains. This evaluation suggests ways of improving the decision by incorporating the case facts, and considering justification and reasoning using portions of precedents. The final result is ANGELIC (ADF for kNowledGe Encapsulation of Legal Information from Cases), a method for producing programs that decide the cases with a high degree of accuracy in multiple domains.

We present a computational model of dialectical argumentation that could serve as a basis for studying elements of legal reasoning. Argumentation is well-suited to decisionmaking in the legal domain, where knowledge is incomplete, uncertain, and inconsistent, We model an argument both as information structure, i.e., argument units connecting claims with supporting data, and as dialectical process, i.e., an alternating series of moves made by opposing sides. Inspired by the legal domain, our model includes burden of proof as a key element, indicating the level of support that must be achieved by a particular side to an argument. Burden of proof acts as a move filter and termination criterion during argumentation and determines the eventual winner. We demonstrate our model by considering two examples that have been discussed previously in the artificial intelligence and legal reasoning literature. INTRODUCTION As the artificial intelligence (AI) and legal reasoning communities are well aware, most decisions are reached against a background of incomplete, uncertain, and inconsistent knowledge (i.e., weak theory domains; Porter, et. al., 1990). The most widely used AI methods for reasoning under uncertainty either rely on an absence of outright contradictions (e. g., probabilistic reasoning; Pearl, 1987) or are unable to support motivated decision-making in the face of inconsistent information (e.g., default reasoning; Ginsberg, 1987). Both solutions put the problem of deciding what to believe outside their respective domains of discourse. Choosing the proposition with highest Permission to copywithoutfee all or part of this material is granted provided that the copies are not made or distributed for direct commercial advantage, the ACM copyright notice and the title of the publication and its date appea, and notice is given that copying is by permission of the Association for Computing Machinery. To copy otherwise, or to republisb, requires a fee andlor specific permission.