Technology can take on many forms. Technology can range from TI-83 calculator I used when I was a high school student to the phones many of us take for granted that do far more than just make a simple phone call.
In the legal world, whether we attorneys want to recognize it or not, we already use quite a bit of legal tech. The computers we use to research, the programs we use to draft documents and emails, the phones we use to negotiate are all pieces of legal technology. One of the more fascinating and disruptive technologies that has come to prominence in recent years is legal analytics. Legal analytics, to put it simply, is deriving actionable conclusions from a set of legal data. Typically, this is done via some form of statistical analysis.
I had the privilege of interviewing one of the leaders in the field, Professor Kevin D. Ashley, author of Artificial Intelligence and Legal Analytics: New Tools for Law Practice in the Digital Age. Professor Ashley is a Professor of Law and Intelligent Systems at the University of Pittsburgh. (CSL = me, KDA = Professor Ashley.)
CSL: What led you to developing your expertise into computer modeling of legal reasoning and cyberspace legal issues?
KDA: After practicing for five years at White & Case on Wall Street, I married a woman I had met when I was studying at Harvard Law School. She was happily teaching at Westfield State College in Massachusetts, so I moved there. Computer law was a developing field at the time; I thought that I would get a Master’s degree in computer science to have some technical background before practicing in that area. I enrolled at the University of Massachusetts at Amherst, and soon met Prof. Edwina Rissland, who was interested in applying AI to modeling reasoning with examples in mathematics and law. This was the early 80’s; I had never even heard of AI, but I was quickly hooked. For my PhD, I developed HYPO, a program that reasoned with cases in trade secret law. Given a trade secrets scenario, it made arguments for the plaintiff’s side drawing analogies to past cases. Then it responded to the arguments by distinguishing the plaintiff’s cases on behalf of the defendant and citing cases as counterexamples.
[As you can see, Professor Ashley was well ahead of his time considering he was working on applying artificial intelligence to modeling legal reasoning in the 80’s when computers were hardly a household item!]
CSL: What is the biggest misconception you think lawyers have about legal technology and computer modeling’s role in the practice of law?
KDA: I think that many lawyers may mistakenly assume that computer programs can read like humans do. Given the claims made on behalf of commercial programs, one cannot blame them if they believe that programs can read. Today, as far as I know, programs can extract certain kinds of useful semantic information from legal texts, but they are still far from being able to read these texts like humans do.
CSL: What do you see as the biggest benefits of legal analytics to the practice of law? That being said, why do you think many lawyers remain resistant to the use of legal technology?
KDA: Legal analytics make possible new legal applications that will change the practice of law. By connecting computational models of legal reasoning (like HYPO and its successors) directly with legal texts, the new legal apps will support conceptual legal information retrieval that can better take into account the reasons why attorneys are seeking particular information and the uses to which they intend to put the information they retrieve. The new legal apps will support collaboration between humans and computers in which each performs the kinds of intelligent activities that they can do best. This collaboration will help attorneys formulate legal hypotheses about the strength of arguments or the likelihood of outcomes and test them against corpuses of legal texts.
If attorneys are resistant, I think it is understandable. A lot of claims are made for the capabilities of the new technologies, but the techniques are still the subject of on-going research. Even the methods for objectively evaluating the new capabilities are under development. Why should attorneys be any less skeptical than the researchers themselves? As with other areas of research, the more we know the more interesting questions we discover.
CSL: Do you think lawyers should develop a familiarity with building AI programs for the legal industry?
KDA: Law firms should develop a culture of legal process engineering. Firms should take stock of the data resources they have available including text corpora such as their court filings, briefs, and memoranda. They should identify the kinds of tasks and processes that could be performed if certain kinds of information could systematically be extracted from that data. Today, text analytic techniques, including natural language processing, machine learning, argument mining, and network diagrams, can help to extract and apply useful semantic information from the data. A firm could set up an AI committee of interested partners, associates, and technical staff to identify and assess the opportunities. Associates need to learn how technicians evaluate the limitations of the AI technology, as well as the potential, so that they can ask the right questions of commercial providers.
CSL: For those wanting to learn more what resources would you recommend someone look into?
KDA: Firms should connect with university researchers in computer science. Firms can hire technical graduate students in paid internships. The students can help firms understand how to pre-process textual data, apply machine learning, and evaluate the resulting ML models. Two conferences, the International Conference on Artificial Intelligence and Law and the Jurix conferences, regularly serve an active academic research community in the field of AI & Law. Their proceedings are available from ACM Press and from IOS Press. Artificial Intelligence and Law, published by Springer, is the journal of record for the field. If I may, my new book, Artificial intelligence and Legal Analytics, New Tools for Law Practice in the Digital Age, from Cambridge University Press, provides a good foundation.