Intersection of Law and AI with Paul Callister
This week Host Alan Olsen interviews Paul Callister, a dynamic individual with a diverse background, takes us on a journey through his educational path from chemical engineering to philosophy, followed by his transition from practicing law to becoming a law librarian and professor.
As a former attorney in a family law firm, Paul sought to find fulfillment and a platform where his talents could shine. He shares insights into his academic pursuits, including obtaining a master’s degree in library and information science, which granted him a unique perspective as both a lawyer and librarian. This hybrid role allowed him to delve into the intricate relationship between law and information access.
The discussion takes an intriguing turn as Paul delves into the historical significance of language, literacy, and law in societies such as ancient Greece, Egypt, and the modern Anglo-American tradition. Drawing parallels between language structures and societal structures, Paul highlights how the accessibility of legal information has influenced governance and citizen participation throughout history.
Paul also provides valuable insights into the emergence of generative AI, particularly in the realm of legal research. He explains the distinction between AI models like ChatGPT, designed for conversation, and the emerging generative AI tools used for advanced legal research and retrieval. Paul discusses the potential for AI to reshape legal research methodologies, enhance decision-making, and create a divide between firms with access to these advanced tools and those without.
Wrapping up the episode, Paul and Alan explore the ethical considerations surrounding the use of AI, particularly in academic settings. Paul shares his perspective on integrating AI tools like ChatGPT into the classroom, emphasizing the importance of guiding students and helping them develop critical thinking skills while leveraging AI for research and analysis.
In this intellectually stimulating episode, listeners gain insights into the evolving role of law librarians, the transformative power of AI in legal research, and the ethical nuances of AI integration in academia. Join Alan Olsen and Paul Callister for an engaging dialogue at the intersection of law, information science, and technology.
Hi, this is Alan Olsen and welcome to American Dreams. My guest today is Paul Callister. Paul, welcome to today’s show.
Thank you for having me.
So Paul, you got a unique background for the listeners, can you take us through, you know your journey. So let’s start at college.
Sure, I went to Brigham Young University in Provo was a started in chemical engineering and ended in philosophy, which I graduated in and then got married about that time and then went to Cornell Law School, where I had an interesting experience practice law was my family’s law firm in Glendale, California, near Los Angeles, doing mostly tax related and transactional work. And was not particularly fulfilled or happy and not anxious to take over a family business, which may seem odd.
In my eighth year, there or after my eighth year, I entered a program at the University of Illinois, graduate schools for a library and information science, it was getting a master’s degree, and I did it online, so I could keep working. So most of the program was online, and got a degree, a master’s degree from there, which effectively made me a librarian as well, as an attorney.
I got to ask this question. Okay. So you went from a family business, right? into an academia, position Law Library? What is it that in that transition that you were seeking to gain,
um, a lot of it had to do with, there’s not one thing, but one of the driving things was finding a place where my talents could allow me to excel in a way that I felt satisfied and that would eventually bring in a decent income for my family. And while law librarianship and librarianship generally is not well paid, I am now a library director and, and make sufficient for our family’s needs. So well, when
you when you look at law and its information environment, why does that become so important to us, as a society
in which it really makes a difference. For instance, the Greeks used a alphabet that was democratic, it’s called the demotic script, and they wrote everything on stone, steal it, and they believe that the worst thing you could do was hired the law. And they were, you know, they had slavery, but they were, they had a democracy for a period. And they had assemblies and diecast. And people would participate in that. You compare that to, for instance, the Egyptians, and the Egyptians form of writer they wrote on papyrus, and they had some very interesting features of their legal system.
But one of which is you had to have a scribal class of priests, because their language is not purely phonetic. It has silent determinants. And for law, every concept I could find that was a legal concept has a silent glyph determined it is a glyph, a higher glyph. That is a scroll, which means it’s connected to writing.
And you would know that as the priest, but you would not know that if you were someone who was illiterate, which this priestly class is really the only literate class. And so their society to dependent much more on having a priestly class mediating, it didn’t matter who took over Egypt, they had to have the same priestly class.
So things like that can make a huge difference in how laws and laws function where you have Greece where everyone’s participate in democracy and hearings, court hearings, they have a assembly of like 500 that listen to things you know, we live famous one that that sentence, Socrates to death, unfortunately, and Greece where you have depend on this scrobble class to run the country.
And that has caused me to wonder how you test a hypothesis is China because its language is fundamentally different or was So there’s simplified Chinese now is because it’s much more much more difficult to learn, you have to have a scrobble class at an enlightened class.
And that brings up a different kind of society than say we have here in the West. They are Anglo American European tradition, where we very literacy became extremely important with the Bible. And after the Bible came political tracks and legal tracks and legal books, and became much more accessible to a much broader class of people.
Yeah, the the world there. You walked in through the Greeks, the Romans, and Egyptians, here are through society today. And obviously, the world has become a lot smaller with the Internet. Sure. And the question, I have the now we have generative AI. Right, and I’d say you did some work on. On generative AI, there’s been some funny things happen when people try to do their court cases on using AI. Yeah, well, let walk us through the impact that AI you see AI having in this industry?
Well, we need to distinguish two types. The first is what the public is going to have, which is chat, GPT and Bard and some of the others that are available. And they are not specifically trained on legal texts, then there are a huge movement that’s really starting to peak in the law vendors, the ones that hold lots of legal information, including legal information that’s edited, and, and a commentary on our primary law, our statutes or cases, IX regulations, etc, which are really the valuable stuff. And those vendors are launching what’s called generative AI, with retrieval.
Augmented generation, meaning it has a search function, it will go out and pull information in instead of just having a training base that is designed to converse with you. That’s what chat GPT has, it’s this is an oversimplification, but you have something designed to converse with you that has some background on the topic. And because it’s baked into the language, you get things that are relevant, some of the time now are a lot of the time. But with the legal profession, we’re going to see a very different kind of, of AI, as they implement these search device.
And they’re talking about, it’s not just a keyword searching and more, we’re talking about a vectorized search. And what I mean is they’ve taken language and broken it down words, into dimensions. 300 dimensions is probably what was used in chat, GPT sentences, they’ve turned into vectors, all the things are related to all the words and sentences that are related to that co occur within a certain period or that have certain properties and characteristics.
And so they’re able to manipulate that and when you’re searching, you’re searching by vectors and not keywords themselves.
And so you can get very different kinds of results and find relationships that you wouldn’t other wise, fine, I think on what’s going to happen on the vendor side is that is very expensive, what they’re doing. And so it will continue to create a class of lawyers, small firm lawyers who do not have the majority of law practice in the country who have access to these tools, and maybe another class of lawyers who do not. And already that has happened even without generative AI. Does that answer your question?
It does. It doesn’t. It seems like right now we’re at the very beginning of something big. Yeah. So if I if I can then moves through. We talked about you as a law librarian, but also your law professor. Well, it’s a big signee teaching.
Right now copyright law. And I’ve just proposed Of course, AI and the law, which will be fun. Yeah, copyright law was big. I tried to avoid that as a new library. And I was always been come to as a librarian to help kind of guide people on these copyright issues that they were in use in class, have time to call you or seek counsel. And usually I would tell them where there’s danger and not tell them yes, it’s okay. But there’s danger of these issues. And so I started participating on a national level with our Library Association. And I started getting into it.
And eventually I was asked to teach it, I also taught cyber law for a long time, and that had a heavy copyright component. And so it’s just been a natural progression. And I’ve written some about copyright issues, from my own growth, and published on the issues. So that I enjoy it, I enjoy the intellectual stimulation of it, I enjoy teaching it to students. My class and finding out is known as a hard class, but I still get people taking the course. And I brings me a lot of sense of fulfillment, and some utility, I can actually help out.
I’m not I don’t practice law, I’m not University counsel, but I think sometimes can say, No, don’t protect University, from some liability.
So is is using chat GPT to write it paper in your class, ethical, I would copyright law,
which here the university default is that you don’t use it unless the professor allows it. So, I have a policy, that word I specify that they can use it. And they explain why they use it. I will allow it. So, it’s going to be probably in my advanced I also teach advanced legal research different courses, and American legal research to foreign students. That’s research course. And I will specify when we will use chat, G, PT or other devices.
It’s a huge topic right now, some of my colleagues, I was on vacation, unfortunately, still, my college just came back excited from a conference with legal research and writing instructors who focus on first year students teaching them how to write briefs and things.
And they’re excited about well, maybe this will be helpful. Maybe this’ll help them brainstorm, maybe this will help them organize their thoughts. And others you say Not over my dead body, they have to learn that themselves how to organize. And there’s similar arguments on the research side, which is really what I do that they need to struggle on their own, or will this help them organize a research plan?
What do you see is the future of libraries with this technology environment progressing the way that it is,
um, the future is really in librarians. We’re going to need pa they may be called information sciences that was scientists, because the debate my first day of library school at Illinois, but I’m like, our role will be to help. Exactly. Teach the questions. And a generative AI is meant to be used as a conversation, not just a one-time question like on Google. But repeated, you get the best results by interacting it and following up with questions.
And I have spent a lot of time learning about generative AI through chat GPT. And I think a bard and other systems could and I’ve, I’ve learned that you have to keep following up question after question. And you get answers that you didn’t expect sometimes by questions that you with questions. You didn’t really ask about it, but you get inside knowledge. And that’s an area where I kind of trust where I’m learning a lot less so. So
Well Paul its’ been a pleasure having you with us today on American Dreams.