Artificial Intelligence and Legal Theory at Law Schools

Currently, Artificial Intelligence and Law is an interdisciplinary field somewhere on the border between Computer Science and Law. For many reasons, this interdisciplinarity hampers its further development and the transfer of its results into both legal theory and legal practice. In this position statement, we argue that AI and Law should become part of legal theory, or rather a “school” of legal theory, to enable the field to eventually become an integral part of legal education and academic legal research. The paper discusses relationships between this revised conception of legal theory and other subjects of legal education, including jurisprudence, legal methods, legal informatics (law and computers) and legal research and writing, as well as consequences for the collaboration of academic lawyers and computer scientists in the field. 1. PROBLEMS WITH THE INTERDISCIPLINARITY OF AI AND LAW The growth of the field of AI and Law requires sustained institutional support. It must be possible to obtain a Ph.D in AI and Law and have a reasonable opportunity to obtain a position as a professor at a university. Once in this position, it must be possible to offer courses and seminars in AI and Law, at least at the graduate level, to obtain research grants, to build up a team, and to offer a Ph.D. program, so as to nurture the next generation of scholars in the field. This next generation should have the same opportunities, to assure the sustainability of research in the field. Moreover, if the field is to have any practical relevance, results from AI and Law research need to be systematically integrated into legal education and become part of the normal curriculum at law schools. Continuous interaction with and feedback from legal practice is necessary, if the field is not to devolve into an entirely “academic” exercise, in the pejorative sense. At the moment, it is doubtful these conditions are met, except perhaps in The Netherlands. In the US, all AI and Law professors are now at computer science departments. A few have, or have had, a joint position at a law school. Most AI and Law PhDs have been in computer science. It has proven very difficult for these PhDs to continue their work in the field. If they do not also have a law degree, it is quite unlikely, if not impossible, for them to obtain a position at a law school. Surely any requirement for dual degrees, perhaps even multiple PhDs, would be a severe impediment to the continued growth of our field. A PhD in computer science may enable one to obtain a position at a computer science faculty, but this does not imply an opportunity to continue work in AI and Law. In Germany the situation is even more desperate. The field of AI and Law is essentially dead here. It is part of the field of Legal Informatics (Rechtsinformatik). Interestingly, Legal Informatics is a legal discipline. All professors working in the area of Legal Informatics are at law schools. But AI and Law no longer plays a substantial role. Legal Informatics has become, for the most part, synonymous with Information Technology Law, i.e. it addresses substantive legal issues related to information technology, such as software patents, data privacy, data protection, or the law of electronic commerce. There are a small number of law professors who continue to consider legal informatics to be about legal applications of information technology, in particular Herberger and Rüßmann at the University of the Saarland. Their institute has worked on a range of topics, including uses of online databases and the World Wide Web for legal research, eLearning methods for legal education and the use of business process modeling tools to model cross-border legal transactions within the EU. The proposal we want to present in this position paper for addressing these issues is to move towards integrating AI and Law into the field of legal theory. With time, AI and Law should become “just” a methodology for conducting research on legal theory. The goal should be to establish the field of AI and Law, if not the name, as a regular part of legal education and academic legal research. In the next section, the current status of legal theory in legal education is assessed. We will find that legal theory appears itself to be in anything but a healthy state, and thus may be open for reform and renovation. After that, the proposal for integrating AI and Law into legal theory is presented in somewhat more detail. This is followed by sections on the consequences of this proposal for legal education and further collaboration with computer science. We conclude with a discussion of some risks and challenges. 2. THE CURRENT STATUS OF LEGAL THEORY Legal theory as a field of study is not well defined. Its relationship to legal philosophy and jurisprudence is not clear. In his recent book, “The Renaissance of Legal Theory from 1965 to 1985” [4], Hilgendorf remarked that some consider legal theory to be nothing more than innovative legal philosophy, especially when this innovation makes use of results from other disciplines. An indication of this is Hilgendorf’s list of the most important topics in the German discourse on legal theory during the 20 year period which is the focus of his book: