I, Robot
At 15, I became self-aware. I started working in the public library. Maybe it was earlier. If my memory banks are not malfunctioning and have not been tampered with, I started out shelving books. I do know that I read all the books in the young adult section of the library. Encyclopedia Brown. Ramona. Jo. The Witch of Black Bird Pond. And a book about young Quakers in love. I ended up reading the Large Print books in the adult section. Barbara Cartland. Mary Stewart. Elizabeth Cadell. Westerns. Gothic romances. Mystery suspense. Before I got to anything too risqué, I went to college.
I took all the required courses, but discovered acting and the wonders of the arts, and the joys of serendipity, of finding out things you can’t explain or understand because you were wandering the library stacks and found Radclyffe Hall’s The Well of Loneliness. Reading books, writing poetry, playing instruments, acting, watching TV, sports took me places I never imagined. Made me risk. But I had to knuckle down, finish my studies, and graduate, so I did.
During my entire time in college, I worked at the university library and the law library. After college, I continued to work at the law library while I decided what my next steps would be. One of my various jobs involved being an evening circulation desk supervisor. I found I liked helping students and faculty at the law library circulation desk. And I wanted to get better at it. I asked the law library director at the time about how I can answer some of the questions I got and he gave me a copy of Legal Research in a Nutshell, and I was off and running. It may have also been then that I started trying to compile lists of answers to frequently asked questions. Or maybe later. But I now had my career path. I was going to be a legal reference librarian. Not just any legal reference librarian. The world’s greatest legal reference librarian, because, if you’re going to do something, why not excel at it? Why be mediocre? Where’s the fun in that?
So I embarked on my career. I attended library school, then law school, and became a law librarian. And thus began my professional quest to determine what makes a great legal reference librarian, and be one. I created legal research guides. I compiled FAQ lists to expedite answering questions at the reference desk. I taught legal research classes. I did one-off in-class lectures. I attended conferences and made sure to attend all the relevant educational programs. I wrote articles on how to research topics. I compiled bibliographies. I created crossword puzzles related to legal research topics to make them more interesting. And I presented at conferences on all types of topics.
But the question of how to assess greatness in legal reference librarianship dogged me. How do I become what I aspired to be? There were several possible models of greatness. The authors of the legal research textbooks and guides, and columns on legal reference work in law library publications. Colleagues who combined in-depth knowledge of legal bibliography and substantive law with commitment to service. Colleagues whose assistance patrons clearly valued. User satisfaction surveys, maybe? But, in the back of my mind, I kept on thinking that another general index of greatness had to be speed and accuracy – how quickly you can respond correctly to a patron request? How wonderful it would be if I answered every question at the reference desk correctly, quickly, and efficiently? I would not waste the user’s time. They would come away with answers they liked each time. I, Robot.
But, is great legal reference really about answering the question asked, or is it about the user being satisfied that they have been helped? A user could have the wrong answer, but still value the assistance they received because the librarian helping was friendly, empathetic, and cared? Remembered their names? Asked them about themselves, their day? That does not compute. If wrong answers are okay, then how do you assess greatness? What makes a great legal reference librarian? Could the greatest legal reference librarian be, a robot?
With the advent of ROSS, the “robot lawyer”, it’s becoming more real that there could be a robot law librarian.
And, at a recent conference, someone called me the “Google” of international law librarians. Is legal reference librarianship just like being an AI law librarian to users – just a source for answers, or is there more?
Would users be satisfied if, instead of a human being, they were served by a robot at the law library reference desk? Would interacting with a machine be as valuable to a library patron? Would a robot law librarian care about the user as a person? Is the legal reference transaction just about the user asking a question and quickly getting the correct answer? Is there satisfaction for the legal reference librarian in just being a source for answers? Being a Westlaw Answers, an Answers.com, an Ask Jeeves, a Quora, a Medium Daily Digest, a Google?
A colleague thinks that a robot law librarian could answer about 90% of the reference questions asked, but for the other 10%, you need someone with deep expertise who is able to answer complicated questions with thought. A user with a complex question does not want a long-winded answer or a short, quick answer. A user wants a legal reference librarian who is aware of the context and can give answers that draw on their extensive knowledge of the relevant resources suitable to the question and the user. That is the measure of greatness – extensive knowledge of resources and thoughtfulness in responding.
So, to become the great legal reference librarian I aspire to be, do I deepen my knowledge, the artificial intelligence? Do I improve my emotional intelligence? Do I let the arts back in again, and risk being wrong, but having a satisfied user because I have been creative, caring, and thoughtful of their needs, and personalized, humanized the reference transaction? Do I “wander the library stacks” in a reference interview to discover new, interesting things I might not understand, but perhaps new ways to help a user?
There are no awards given for great legal reference librarianship, just for greatness in law librarianship generally. Maybe if there were one, there would be publicly acclaimed general models for what it means to be a great legal reference librarian, and I can stop wondering if the greatest one is a knowledgeable, caring human being or, a robot. Should I, to become the great legal reference librarian I seek to be, become, finally I, Robot?
Thanks for your helpful column Lyonette. I wonder also about a third possibility: that the question posed is not the right question to get the “right” answer? Sometimes (always?) the framing of the question is just as important as the answer. While a “robot” could give an answer to a question posed by a human, is it also able to say “I think I understand what you are trying to discern, and here is a better way to ask the question…”?
Kari
Thanks, Kari. Great point!
Lyo, 1. I think you are a GREAT legal reference librarian. 2. there should be an award. 3. I disagree with your colleague about the 90% to 10% ratio of asked and answered.
A human doesn’t go to the reference desk for just an answer to a question. The reference desk visit may be for the non verbal feedback that their question isn’t ridiculous or the validation that their question is novel, significant and interesting or the verification that they are asking the correct question.
I think that the interaction of asking can be as important as the answer.
I think Kari & Shaunna hit the nail on the head–your output is only as good as your input (garbage in, garbage out). It is often only through a dialogue or an iterative process that both the client and the researcher are able to hone in on (and express using the correct lanaguage) the precise information need of the client. Many of my research questions come with background material, sometimes a saga’s worth, that is required to understand the crux of the issue or the requester is relying on the assumption that I have the requisite institutional knowledge in order to properly place the information need in context. AI is iterative and will continue to improve, but it can’t quite yet navigate these kinds of contextual factors. I would suggest that for AI to be useful for complex research (not just reference) in most legal organizations, it will require significant inputs of institutional knowledge. We struggle enough with doing KM properly, or even at all–how do we document our learned organizational history in a way that a robot can properly interpret and apply it? Input isn’t only the query someone enters into a machine; it’s all of the information that machine relies on to deliver a response.
I’m not threatened by AI or robots–I’m excited about the possibilities but aware of its current limitations.
Hello from France,
I was asking myself the same question. I don’t feel threatened either.
Here’s why :
1. I’ve read a bit about ROSS : see (in French). Like its legal AI counterparts in France, it is, fundamentaly :
a. a new generation search engine, which uses natural language processing powered by assisted machine learning and semantic analysis. Worse, ROSS is restricted to one legal domain : bankruptcy
b. an aid to decision making. ROSS does not anwer
c. a weak AI. Not the (theoretical) artificial general intelligence (also called strong AI)
2. with a huge investment, a so-called legal librarian AI could be developed, able to answer straight questions and dialog with the patrons to narrow down their needs. But it took three years to develop ROSS from Watson (quite costly, no ?) and what about the quality of the questions and, moreover, the suggestions ? And, more important, doesn’t Google already do that ?
3. good law librarian researchers do much more. They foresee the implications and future needs of their patrons, they train lawyers, have communication and IT skills, guess what to buy, know how to search out of the databases and the Web
4. law librarians have worked with computers, electronic documents management systems, the Internet, CMS, blogs, social networks … We’ll work with “AI”.
Here is the blog post I forgot to mention after “see …” :
Intelligence artificielle en droit : les véritables termes du débat (in French)
http://www.precisement.org/blog/Intelligence-artificielle-en-droit-les-veritables-termes-du-debat.html
And here is another post about the evolution of the librarian researcher’s (documentaliste in French) future :
Documentaliste : un nouveau métier, de nouveaux noms (on French)
http://www.precisement.org/blog/Documentaliste-un-nouveau-metier-de-nouveaux-noms.html
Bot posts may be long but have an executive summary (or TL;DR).