AI and the legal librarian: revolution or evolution?

Technology is undoubtedly changing the way the legal libraries operate and the open-endedness of possibilities is thrilling…

Whether or not we are on the cusp of a revolution of strong AI, it seems clear that software increasingly will perform legal tasks that were previously reserved for lawyers, paralegals, and information professionals. AI may not replace lawyers, but AI will absolutely replace legal work. A great deal of the money paid by clients for those legal services will therefore go to the architects of that software.

Will this be an information oligopoly, with the benefits of AI accruing to a small number of publishers and software companies, at the expense of law firms, law schools, and libraries? Or will the benefits be widespread, like the creation of electric power?

The answer, in part, lies with how broadly the providers of legal services and information participate in artificial intelligence. If information professionals cede the “heavy lifting” work of AI to software companies, they will be unlikely to participate in the gains of the new economy. However, if information professionals themselves use AI tools, conduct their own experiments in supervised learning, and learn how to gain new insights from legal data, there is great potential for a broad-based renaissance in legal services, in which everyone has the opportunity to participate.

Understanding the AI workbench

What kinds of insights could law libraries draw from AI tools? At the simplest end of the spectrum, AI tools can help libraries draw important insights from in-house data. For example, law firms can use tools such as IBM’s Watson Developer Cloud to convert documents from images to structured text, build chatbots that understand natural language to triage patron requests, or build search engines that run semantic searches over the firm’s KM system.

Libraries can use AI tools to help change the way that law school clinics deliver legal services. For example, the Georgetown University Law Center conducts the Iron Tech Lawyer competition, in which nontechnical law student teams work with clients to create apps for justice using Neota Logic. Expert systems such as Neota Logic or A2J Author may not solve every legal problem, but they can help address simple, recurring legal problems at scale. Indeed, one-to-many software tools may be one of the only scalable ways to address the access-to-justice problem. Information professionals in law schools could be the lynchpin to this kind of development.

Artificial intelligence tools also can offer key insights about the cost of different matters in a law firm, helping firms that want to offer fixed-fee engagements compete more effectively for business. One problem that law firms face is that their client-matter billing data isn’t structured enough to understand the distribution of costs for different types of matters.

Sophisticated clients are increasingly demanding that their law firms provide alternative fee arrangements, such as fixed-fee billing. However, without structured data, law firms have faced the impossible task of combing through historical billings to add metadata such as the matter type, stage, lawyer, rate, time, and cost – all to understand how to price future matters. AI tools offer the prospect of extracting insights even from unstructured data. This avoids tedious re-keying or data entry for past matters.

Data mash-ups offer even more interesting possibilities. When firms can compare their own internal billing information with public information such as dockets or public law collections, they can provide quantitative marketing materials to potential clients. Instead of touting the firm’s expertise generally, firms can demonstrate that they achieve better results faster, and at lower costs.

Perhaps further in the future law libraries will be combining docket sheet data, judicial opinions and AI tools, such as IBM Watson’s Personality Insights, to analyze the actions a particular attorney took to generate a positive or negative outcome from a judge. Extrapolating upon that idea, and with a large enough dataset, a law library may be able to create a tool to anticipate reactions from judges or even opposing counsel, to allow their attorneys to best tailor their approach in court.

The open-endedness of the possibilities is pretty thrilling. Law libraries collect so many different kinds of information that it’s hard to say ahead of time what cool insights they can derive with AI tools. Data scientists will tell you that it’s often hard to know what insights are even possible until they get their hands dirty with the data sets.

This extract is from ‘The Evolution of the Law Firm Library Function’, by Ed Walters, CEO of Fastcase and Sean Tate, AI sandbox product manager at Fastcase, published by ARK Group. The book is available to purchase here.

The impact of technology on the future role of the legal librarian will be discussed at the Legal Libraries conference on 10 October, in London. To find out more and to book your place, click here