Transcript
Speaker: This episode is sponsored by LexisNexis, supporting information professionals as leaders in AI innovation. Explore resources and tools designed specifically for your role at LexisNexis.com, Backsplash Community, Backsplash Info Pro, and connect with them at AALL this summer in Cleveland.
Speaker: Well, hello everyone. Welcome to the very first episode of the American Association of Law Libraries All In podcast. Now, since this is our pilot episode, we wanted to try a couple of things out.
Speaker: See what sticks, see what resonates, and most importantly, bring you into the experience with us. First, I want to give a special thank you to the members who submitted the name All In.
Speaker: It perfectly captures what we're about, being all in on information, innovation, and inspiration. Now, since today's episode is about AI, I figured I'd go straight to the source.
Speaker: So I asked ChatGPT how to pronounce all in as it's spelled, A-A-L-L-N. And it told me, hey, y'all. So to everyone listening, man that's going to be my thing. I'm going to start the episode with, hey, y'all.
Speaker: If you like that greeting, go ahead and drop a note in the comments and let us know if it sticks. Now, the other option was to go full key and peel Mr. Garvey on it. You know, a all in, similar to AA run.
Speaker: But since I'm from Texas, I think I'm going to stick with A, y'all. Now, today's episode is really a bit of a social experiment. We're bringing together an an art kid, a theater buff, and two musicians who also happen to be law librarians and legal techies.
Speaker: and we're going to have a real conversation about generative AI. Our guests today are true OGs in the space, the host of The Geek in Review, Marlene Gebauer and Greg Lambert.
Speaker: We're really excited to have you both here for episode one of the All In podcast. And there's a little bit of irony here. I had the chance to be on episode 51 of Geek in Review, and Katie was on episode 171.
Speaker: So it's only right that we return the favor and welcome you to both Episode one, Greg, Marlene, welcome to All In. Thank you it's nice to be here. All right, Katie, Andre, it's good to be on A-Y'all. know, you're from Texas, so I feel like that word works. yeah not Now to kick things off, we want to frame this conversation in the right way.
Speaker: This episode is but all about the good, the bad, the ugly of generative AI. But before we get there, we want to start with the why and the perspective behind it. So I'm going to turn it over to my co-host, Katie Brown, to get us started.
Speaker: Thanks. I'm really excited for this conversation. We all know with generative AI right now, we either love it or we hate it. So let's start with our first question, Greg and Marlene. What inspired your Beyond the Model series? And how was your perspective on generative AI and the legal space evolved from the early days to where we are right now?
Speaker: All right. So that I'll take this one Marlene, since I've and it's been been writing this one. So we started, we've done the podcast, we've done the blog and we've experimented with a few things, TikToks, Instagram, and and just, I think I may be too mature. We'll say that too long in the tooth for some of these. saying so You weren't on I'm still on Facebook. um so
Speaker: we we started this Substack channel, and one of the things for doing the Substack channel was to try out some different things. And I really decided that what I wanted to do was create...
Speaker: creates create some content that, the you know, maybe a little bit more experimental. And Marlene and I had a guest on a few months ago, Anusha Gillespie, who we've known for for a number of years. She's been on on the show a couple of times at least. And she had written a novel, kind of a science fiction kind of fantasy novel. But really, once you scratch past the surface of it, it was really about why life sucks at a big law firm, for especially for a a woman lawyer in the first few years. And in so i was after talking with her, i had an idea for
Speaker: a basically just a blog. Um, it was like, why did AI legal research suck for so long? And then all of a sudden it, you know last summer I started noticing that I was getting actually pretty decent results coming back all of a sudden. Um, and I had talked with ah Ed Walters, uh, who's now at Clio. Um, and, uh, he, he and I were at a conference and he he explained that to me. And I went back and I i like did some research and I ended up with like 65 pages of this dense, you know, it's, it's because of the, you know, the way they're indexing it and they're doing a gentic brag and, and you know, all these, all these things that they were doing now in the backend.
Speaker: um And I was like, Oh my God, I can't, I can't write this. So I took, I borrowed from new show and I thought, well, you know, it'd be interesting to kind of have a story. So, you know,
Speaker: Ed became and one character and I became another character and I invented a few more. And it started off part one was just explaining why it was that all of a sudden legal research was better, but it was in a fictional law firm setting. And I thought it was just much better way of telling that story. And then every time I would finish, I'd be like, okay, I'm done. I'm not going to do anything. And then I would get another idea and I was like, well, what about Harvey versus Legora? Oh yeah. Let's do that. What about a gentick AI? And, you know, it's like agents. and And then the last one was, uh, uh,
Speaker: And I think this is the last one. We'll see. We'll see. Yeah, I was going to say, if you're saying every time, it's the last one. I'll believe it when I see it. Famous last words. The last one is three years in the future, which is kind of makes it hard to like fill in the gap.
Speaker: you know some So maybe maybe part six will be you know six years in the future. We'll we'll see how it goes. Yeah. It talks about how there are things like dark factories, which is an economics term for when processes in a factory are so automated, then there's no humans involved in the automation. You don't need to turn the lights on. Everything can be running in the dark. And I was thinking, well, you know, there's going to be processes that we do due diligence review, things like that, where there's going to be major portions of it that will probably be so automated
Speaker: that you, you know, that you, you won't need the lights on in the office for people to do it because you may not have people doing that part. So anyways, that wraps up this week. So we'll, we'll see ah if we continue or not.
Speaker: Yeah, I'll say that I definitely have been reading along because it is it is episodic, as you said. And it it does kind of leave you on some of these tech cliffhangers to know, you know, what's the next topic potentially that's going to be discussed. So for anybody out there who hasn't been following along, definitely go and check it out because I think that's, you know,
Speaker: Everybody wants to know what the what the next piece next piece of the the story is going to be. So i I wonder if this really is going to be the last one. We'll see. No, it's funny you said that, Katie. For those who wondering, Katie is the theater buff, Marlene is the art kid, and Greg and I are both the musicians. But we also have- Audrey's much better musician. He's toured internationally. I've just been a local. Yeah. But, uh, we also have interesting perspectives, uh, with Greg and Marlene both operating
Speaker: in law firms, in the legal tech space, Katie in the academic space, and myself previously in law firms, but now in government. And I had the opportunity to speak at the HBA Bench Bar Conference on last week about e-discovery and AI. And I got a really interesting question about access to justice and how AI works.
Speaker: can assist with that. And I think there are a lot of of different people doing their part for that. We have, i think, the LawDroid, Tom Martin and the LawDroid conference is coming up, and I think, next week or the week after that. But I want to direct this question to Marlene. And ah so we're going to talk about the good. We're going to be really positive right now. Where are you seeing generative AI truly deliver value in practice today? what's working better than expected, and how are law librarians helping to drive that impact?
Speaker: Well, I mean, it's certainly working a lot better than it did when we first when it first came out. you know I think that everyone was was a little dazzled, but also you know quickly skeptical.
Speaker: And you know I think that that's a healthy thing you know with with this tool because it does generate content. So you know in terms of of what we're seeing it do really well, we're seeing it do...
Speaker: Things like you know summarizing and analysis between documents, comparing documents. Some of them do redlining now, and that's incredibly helpful, I think, for for attorneys to be able to compare and contrast the differences. They offer workflows at this point, and I think people are starting to dip into that a little bit more. Attorneys are starting to look into that more.
Speaker: Just, you know, having something pre-programmed as opposed to having to create a really extensive prompt. And, you know, some of them work better than others. you know, I just heard this morning that one of the tools that that we have did an excellent job in in that and and saved tons and tons of time. So definitely a time saver.
Speaker: But, you know, I've been been writing and talking about how it's sort of beyond the time savings in terms of what it can do. And you have to, we have to look at value in a different way, because we can't really measure time anymore, because the time is much shorter, but the value of of the output is still there. So, you know, are we looking at in terms of output? Are we looking at in terms of the description model, you know, flat fees, like, how do we want to,
Speaker: look at this and And I think it opens up the ability for us to have conversations with clients to um to to find out what is important to them and and sort of shape what we're doing with AI around that. In terms of what librarians are doing with this, I mean, they're they're doing a ton and and many are are in the forefront of this. Some of the the main products that we hear in the news are are things that the that the library handles. So, you know things like a co-counsel or things like a VLEX. You know, there is a research component to them.
Speaker: And, you know, librarians are definitely getting involved in that. And, you know, also, I think as as some of the other tools are becoming more, kind of library related, they are also starting to get involved in those. there's There's a lot of blending and crossover that's happening right now, which I think is actually healthy. I think it's healthy for the competition. I think it's healthy to keep the different, the vendors on their toes and make sure that they're delivering the best possible resource or tool rather.
Speaker: And you know our you know librarians are actually you know Leading the trainings on that, comparing tools, that is that's one of the main things we we hear a lot is, well, what is this one good for and what is that one good for?
Speaker: And while there is certainly a need for experimentation on on all of those things, the like the library staff is is one of the key groups that do that type of comparison and do that type of testing and you know are able to convey that to the rest of the firm.
Speaker: So I want to follow up kind of in this. And this is a, is this a good or is this potential a bad question on that, on that last bit that you were talking about? So we are seeing a number of, and I'm kind of thinking about, right, we had the Fastcase into V-Lex and then we recently had the Clio purchase. We're also seeing a number of,
Speaker: resources that would have been in right people's tech stacks that now they are either purchasing or we've got these Gen AI pieces on top. So is this a good thing or a bad thing as we think about our tech stacks in some of these firms? What do you all think about that?
Speaker: I think it's a good thing. And, and you know, there's there's still room. you know, I'm still seeing new new new vendors come up with new tools. So it's it's not like suddenly...
Speaker: Innovation is dying because of some of these you know consolidations. I mean, you do have to pay attention to that, of course, but there's still a need for point solutions.
Speaker: There's still you know innovation that's happening out there that you know that that some of the other larger vendors aren't able to do as quickly. And yeah i don't i mean I don't think it's really any different than than sort of what we've done in the past where it's like, okay, you know we need to to make decisions about our tech stacks. It's just, I think...
Speaker: We need to make decisions about them more quickly. And you know we need to put you know smart people in the room to figure out, OK, where are the overlaps? you know Where do we need something that don't need something? and I think that is something that the library has done.
Speaker: for years. I mean, I know i did it. Like there are a million products out there where there is a significant amount of overlap, but there's this one special thing that this one does that is is absolutely critical. And so, yeah, you know, you still have to kind of weigh, is that still necessary or can you use something else?
Speaker: Yeah. And I think one of the things that librarians bring to the table that is actually underappreciated Is that librarians are doers. They're actually in the products. They're using the products. They understand the good and bad and ugly of the products does it stick with our theme here. and that i almost handle And you know, they also understand how it connects across the tech stack as well. You know, and and the thing with like, you know, our it t groups are great at making sure that, you know, the network is running, that everything is secure, the databases are there, but they're not experts in the external information that needs to be brought into the firm. And I think that's one of the values that we get, you know, and you team that up with with knowledge management style folks that are really good with augmenting the internal information that's developed by the firm.
Speaker: You know, it's a good one, two, three kind of punch. And so, you know, it One of the things that I've been encouraging all my librarian friends is, you know, kind of keep up on top of of what's going on, be able to speak the language of the KM and IT folks and the innovation folks. And, you know, and stay engaged in that because there's, they do not have the experience with actually using the tools and understanding what it is that the tool is accomplishing. for the firm. And that is so valuable. And people need to take advantage of that.
Speaker: ah Yeah. And I think I want to add from my perspective, I think law librarians are key in translating that information because here at the Harris County Robert W. Hainsworth Law Library, we will not let any of our patrons near any generative AI products. um And that's one, of course, on the vendor side, they've been very cautious about providing that type of access to, you know we mostly deal with pro se patrons, but
Speaker: Just from an A2J perspective, we know we'll never have enough lawyers to scale up to kind of solve this issue. But I've seen some really cool and interesting things with, I think it's the North Carolina, they have a A.I. Leah. and I think that, I think, and on your show as well. And being able to do that. And I think the law librarians are in a position because we're, not only you know dealing with the technology and the research perspective, but we're dealing with the people, especially if you're public facing. So we're you know hearing the type of issues they come in with and we can be the people who like translate that into you know these people, these legal aid organizations who are really building some cool tools, I think.
Speaker: The people, these legal organizations, we went to, I went to the Legal Services Corporation Innovation and Technology Conference in January. And just some of the things that some of these legal aid organizations are doing are really amazing. And they're doing that to to be able to help people. So I think we we play an important role in having these conversations. with the tech companies, the legal technology companies, as well as the legal aid organizations and the people themselves. So.
Speaker: Andre, I love that you mentioned Leah. I had the opportunity to see that for people who don't know, it's a really cool tool that's out there that they have in North Carolina. And it ended up launching just timing wise, right before the floods happened in Asheville. And I know it made a huge difference from an access to justice point of view. And so I think that is, you know, I know we're focusing on generative ai That is a chat bot tool that they had created um that really made a huge difference. And so we do have these instances, right? These are falling to some of those, the good, right? we can We can embellish our tech stack. We can have these opportunities with access to justice that happens. But we also know that there are some bad things out there as well. So I just I just want to drop that down there. yeah i just I just wanted to to follow up on like both of of you. It's like that this is a big area, this this A to J area. And and not just because, you know, people are saying, OK, you know, we need this. But I mean, it's a big investment area, too. You know, I sit on the advisory board for TLTF and, you know, we have people that are in that space that are on that board. And, you know, we have vendors that come in and present in that space. So, I mean, there's there's definitely interest there.
Speaker: And, you know, I think it's going to be a big it's going to be a big area. I mean, that's definitely one of the good areas where it can offer help where help is not available now.
Speaker: That's so good to hear, too, because I think what we read in the news now for, you know, when people are online, all we hear is the negative with pro se with the bad filings. Right. That's everything we read. So it's so nice to hear that there's positive governance issue. Yeah. this is mayi Education and governance issues. Let me share a story that may offset that a little bit.
Speaker: I know that I've i've talked with the with litigators, both at my firm and at other firms, who are saying that there are now pro se filings that typically they were able to quickly go in and get dismissed. But now they're actually creating such good arguments, um, on the issues that they're no longer just, you know show up to the court or, you know, submit a filing and have it dismissed that they're actually having to spend a lot more time on this because the generative AI tools are actually helping craft better arguments for the pro says. And so I, you know, it's, ah you know, my, my old Oklahoma, uh, County law library and self comes out and says, oh, good. yeah that mean It needs to be a challenge because you know it's just because someone doesn't have the resources to afford a big big time lawyer doesn't mean they don't have a legitimate legal issue.
Speaker: so it's you know i'm so We hear a lot about the you know hallucinations and the bad filings, but there's also some really good filings out there. And i I imagine there's been great successes where pro se litigants have used the AI tools to have success.
Speaker: Yeah, it's just so much easier. I mean, back when I was working in in academia for half a click, you know, I remember, I mean, it was a state school and people came in pro se and you were allowed to only help them so much.
Speaker: And, you know, as as a lawyer, i mean, I'm sitting there and it's like, I can direct you to the digest. I can direct you to the books, but I can't. They have a question. I'm like, I know the answer to that question, but I cannot tell you. And it just killed me. And so this I mean, the fact that they can do this so much more easily is just such a huge benefit.
Speaker: Yeah, that's really interesting that you mentioned that because of the whole unauthorized practice of law aspect of in academia. It also happens, I know, according to counties. Yeah. But it's it brings up a really interesting issue that I know is happening right now within the generative AI space. And it's this whole idea of, is that chat bot potentially with unauthorized practice of law? I don't know if you all have seen a that that filing that happened with the insurance company where the it was an individual who
Speaker: ChatGPT actually told her to fire her attorney. And so the insurance company is now going after OpenAI i and just wondering what your thoughts are on that. Good good luck with that. i mean, I think the case was closed and then she wanted to reopen it again fired her lawyer. But it was a lot of, at least what I'm i'm hearing are fictitious type of arguments. Yeah.
Speaker: Yeah, that's probably an example of the bad. Well, we needed a transition to the bad at some point. Well, well and I mean, that's the thing is, so you know, some of these are holding themselves out there. You know, Elon Musk this week, you know, pointed to his new Grok 420, which, you know, I still roll my eyes every every time he he comes out with a new model and it's some...
Speaker: you know reference to drugs or sex. and so But you know he was saying that that he uses it for his taxes and he uses it for his you know legal advice. and And so, and then, you know, Anthropic is coming out with the, and saying, you know, hey, we've got a skill here that does does legal work and open AI is claiming to be both you know, healthcare care and legal. So, you know, I think it this is what happens when you get the wild, wild west. There's no regulations really that have any legs on them at this point. And everyone is just kind of going, well, let's, let's just see how far we can push it. It's going to be interesting to see how and if You know, a well-regulated industry can come in and tamp some of this down or or if they do it, do it at all.
Speaker: I think you know it's a prime example of like ai like a need for AI education and and a need for for governance and a need for guardrails because it is like this is all very experimental and at this this point and it's changing very rapidly. And you know people need to understand you know what the tools can do and what they can't do you know and what their their strengths and weaknesses are.
Speaker: And, you know, then, you know, you have to have, you know, a real governance plan around that. So what you can use it for, what you can't use it for, what type of information you can put in, not put in you know, and then, and then basically the, the, you know just sort of a safety net for, for people to understand sort of what's safe, what tools are safe to use for certain types of things so that they can feel comfortable using these things. So, so that you don't end up with, uh, you know, all of these, uh, you know, hallucination cases that, that, uh,
Speaker: Damien Charlton is tracking, yeah which I think because like last time I checked was like up to 800 or something. I think it's like 1800 now. Oh gosh. So it's it's just, you know, people clearly don't know. And, you know, i'm thinking back to your pro se example. I'm, I don't know it was that case or a different case, but like they basically held.
Speaker: the pro se person to the same standard. Like you need to know what tools you're using and what they do. So it was it was no different than than what an attorney would be held to.
Speaker: So maybe the courts have to do that. mean, if you're talking pro se people. Yeah. Like who's going to, who's going to do the education for the pro se people, you know, maybe it's Andre, you know, it's like, you know, maybe it's the state schools. don't know, but they, know, it's one thing with a firm. I mean, most firms I would hope have that type of education and, and, and guardrails and governance, but for a pro se person,
Speaker: you know i'm I'm not sure where you would where you would find that right consistently. Yeah, I think i think you've you've made up a really valuable point. with And we we have a a a class that we go out and teach to people all over Harris County called AI and U. And just to educate them, I think from my perspective, partnering with the courts as well, just to to give them a little more higher higher level education would be the most optimal process.
Speaker: situation for us because we just, what I found is people just don't know what they don't know, you know, and a lot of times it's a long lacking legal literacy and digital literacy and just, you know, general little risk literacy skills in in in general. It's hard to get them to come in and try to understand that, hey, this this may look right.
Speaker: I think one of my co-presenters, I think some it says trust and verify, but he says he's just his he's dyslexic. He says verify then trust. you know and I think as law librarians, you know we're we're you know we're taught that from the the minute we started this profession, but to your pro se you know average person on the street, if it it looks really good, but It could be made up. It could not fit. Your argument could not fit. So I think there's responsibility there on our part. One thing I did want to ask about that you, you talked about the concerns, but what are people getting wrong about adoption? I think from in the law firms, what are what are your thoughts on that?
Speaker: What are they getting wrong? Yeah. I think part of it is they, assume that there are things that the AI tools can do that it cannot do.
Speaker: And I'll i'll give a good example of this. So yesterday i was showing a group how to create playbooks and the the whole idea with playbooks is you set up a standard for how you handle know things like contracts, what's acceptable, what's not acceptable, what fallback provisions you can do.
Speaker: And the the person was like, well, that's great. But what i want is the ability for the AI just to go into my document management system and pluck out the five things that that I wanted to to get. And based on you know this exact partner's thing, and I'm like,
Speaker: Yeah, that's not what this does, like right? Maybe maybe someday. but you know But that also helps in the fact that at least they're now asking these questions.
Speaker: So I think even though there may be an over assumption of what the AI tools can do right now, I think it's it's at least started the conversation because even if it's not the AI tool, there might be another tool that we bought you five years ago that works really well that you've never really used to do this problem, you know, to solve this problem. So it it has helped in getting the attorneys and business professionals at least engaged, even if they they come with assumptions that don't necessarily meet the facts.
Speaker: Hey, can I just say one thing quick on this? I'm going to take it a different, I'm going to take it a little different way. And thinking about our library friends that, you know, when you're looking at adoption, you know, what are you, what are you tracking? You know, what story are you telling to the firm about adoption? um You know, it's not just simply somebody's using the tools or someone's using certain skills or or things like that. I think it's it's sort of marrying that.
Speaker: and And maybe this is sort of working with, you know, your finance team or your matter management team to, you know, not only figure out like who's doing stuff and what they're doing, you know, you share that type of knowledge with them in terms of the pricing knowledge to figure out, okay, where are the real value propositions? You know, where is it really making the biggest amount of impact? Because sure, there's a lot of experimentation, but looking at like big ticket items where it's really changing things, it's changing the workflow, it's changing the timing. Like that's I think where the library can can really be important in terms of sharing the data that they collect. yeah
Speaker: Marlene, that reminds me of of a term one of my colleagues, Heather, hello Heather, introduced me to is thick data. It was the first time hearing that, but essentially, thick data is qualitative information derived from human observation, emotions, context, you know that you you know combine with the quantitative data. So that just popped in my mind. Collecting use cases. Yeah. Right. That's really important.
Speaker: if And now a quick word from our sponsor, LexisNexis. We'll be right back. Every great era of transformation has architects behind it.
Speaker: Today, information professionals are helping shape the next era of AI in legal organizations, guiding adoption, ensuring authority, and turning innovation into real impact. At LexisNexis, we're building AI solutions designed for that responsibility.
Speaker: Thank you.
Speaker: and tools like protege to help you move your organization forward with confidence we're proud to support double a double l and the all in
Speaker: to explore resources built for leaders at the forefront of ai innovation and connect with us at double a double l this summer lexius nexus private secure authoritative legal ai Yeah, it's interesting what you were talking about with the adoption in the firms, because I know sometimes that also depends on kind of the market that you're in. You know, I'm in Charleston, South Carolina. We have such a wide mix of firms in this space. So we have people that are adopting and have been using generative AI for several years. We've got some firms that are just starting to do it. So I know that, you know, the courses that we're teaching here, we've got everything from theory-based classes that I'm teaching to the skills-based classes. But one of the things that I do is I go into all the externship classes and I tell them day one, you've got to ask about that generative AI policy because there are some firms that don't even have that generative AI policy. And I feel like if you don't have that in your firm, things could go ugly really, really quick. So I want to ask you all.
Speaker: about that? You know, what's what's the ugly that can happen? You know, is it an ethics issue? Could it be a trust problem? You know, could this be a loss of like the originality when you're drafting? Like what what do you think the ugly? All of the above? well Well, I would say number one, if you do not have an ai policy for your firm, the first time that one of your attorneys gets in trouble for using AI,
Speaker: the first thing that the bar or the judge or whoever is going to look for is for that policy. And that policy may not save that individual attorney's the skin, but it will save the firm's skin. If the firm has set up the you know the policy is that you you know do not use commercial-based AI tools, you do not put any firm or client confidential information in it without ah being an approved tool, those sorts of things. That's the very first thing to look for. And if you do not have it, then you're on the hook for it as well.
Speaker: You shouldn't be using AI if you're not if you don't have if you don't have a policy, you know in an education plan, you you should not be letting your people use it. Yeah, the worst policy you can have is no policy. Yeah. I wonder how many librarians, like that was their first project was to go and research how to draft an AI policy. The other thing is like, if you don't have this governance policy and, you know, give people the tools that are actually going help them in their jobs and teach them how to use them, like they're going to go find tools to do it. Like everybody knows it's out there and, you know, people will like, know, we keep hearing about the shadow AI problem and, you know, it's a real problem. Because, you know, you see it in the news. And so if you don't want, mean, that's another risk in addition to the ones that you were talking about, Katie. For people who don't know what the shadow AI problem is, can you explain that to some folks who maybe are unfamiliar with it? Sure. Yeah. So shadow AI is essentially, you know, you may have tools, you may have governance, but people... for whatever reason, are using unapproved tools to complete their job. And they may be putting in information that they shouldn't put that could potentially be exposed and can potentially expose your company or your firm or your organization.
Speaker: Yeah. And that that spun out of what was called shadow IT. t which was if you're not giving your your attorneys or business professionals the right tools to use, they may go out and find and use the tools that they need and just not tell you about it. And so that's why it's in in the shadows. Same with ai Just to expand that just a little bit, one of one of the things that I'm seeing, and I'm i'm trying to coin this this phrase,
Speaker: is the next issue is going to be what what I'm calling a shadow UX or a shadow user experience. And that is as a vendor, if you're not giving the person the experience on how to use your tools in the way that they need to be using your tools, they will find workarounds to, to have, you know, the AI or automation do what,
Speaker: they need to to do without your approval. and so Can you perhaps give us an example? not Obviously nothing that you've done, but perhaps something you've heard someone else doing.
Speaker: No, I think a a lot of it is, you know, the the web is built for people. It is built for, that's why, you know, you get page one and then, you know, click for page two or, you know If a spreadsheet comes back on a web page or a table of format, you know it may be broken down because it's built for how you interpret the information. And you know it could be that I've set up an AI agent to go in and do a web fetch to grab information for me because it takes me way too long.
Speaker: To grab the the pieces of information. It doesn't mean that your information is any less valuable. In fact, it's probably more valued. And I would you know pay for access to this, but the way you have it set up is your metrics are based on me spending a lot of time you know, getting around on, on your webpage or in your database. And that's actually terrible user experience for me. Because now I've got, I may have an agent that's going out and needs to just pull, you know, one or two pieces of information.
Speaker: And I can't be spending an hour on your website, trying, trying to you know get that information so that I can then manually go and pull it off somewhere else. So. That's that's the type of experience that I think you're going to see people moving toward in next few years.
Speaker: oh yeah it It's super interesting you say that because I was talking to another lawyer a government law librarian, and we've noticed how our website metrics, like web page clicks are going down.
Speaker: And we're attributing that to people using generative AI to better search our information, you know, from a Google search, you're automatically going to get Gemini. And it's, we're seeing that trend go down over the last three to four years, which, again, it's back to framing that conversation.
Speaker: As Marlene talked about, Hey, website clicks are going down, but where their people are able to find information better. They have a better user experience, UX experience. I really like like that concept.
Speaker: Marlene's going to counter. I love it. love it. Fistica. Let's go. I'm adding on. I'm adding on. It's like, since this is the ugly part of the conversation, that's what I wanted to talk about. So totally agreed. Like less friction, more likely people going to use it and more likely they were not turned to something else.
Speaker: And I will ask for Greg's help here because I don't remember the term, but we were at the the Texas Trailblazers conference a few weeks ago and there was one presentation And what was it called about the, not exhaustion, was like AI something. And that basically people people are getting, he'll come up with it. people people are getting People are getting too used to it. Like they get too comfortable with it.
Speaker: And they therefore they don't actually check it. cognitive displacement. ah Something like that. We'll look it up. Basically where people are not even,
Speaker: checking the the results. It's cognitive offloading. That's it. cause like it's like it's too like Thank you. at Cognitive offloading. It's so good. It's so simple. The answers are so, you know, sound so good.
Speaker: but it's like, I'm just not going to bother to check. So I guess that's, that is a bit of a counter into things about like, I mean, if if something is super, super easy that we still have to sort of build in the, the, the, the human in the loop check checkpoints, you know, into that process.
Speaker: Yeah. I think that kind of lends itself to all these hallucinations right that we have seen you know earlier it was mentioned with the website i think the last time i did check that website we were at 1300 something worldwide but it was at over 800 for the us because the us are the worst offenders so fi that that yeah we're we're yeah we got we got some terrible go us with the with the not wanting to double check those number one nation Exactly. we We love those faulty filings.
Speaker: yeah So this is kind of similar to the Geek & Review and you're a crystal ball, but this is and this may not stick. so but We're trying it out though. It's our first one. Yeah, it's our first episode. so Beyond the model, beyond the model, but looking ahead, what developments do you see shaping the next phase of generative AI and what should law law librarians start doing or stop doing now to be prepared?
Speaker: I think we're going to see a lot more work, grunt work being handled by ai agents. um I think you're,
Speaker: I think if you're not jumping in and at least learning what those can do, and and I would highly suggest getting hands-on, know, go go pay the 20-buck-a-month license for a chat GPT and and start playing with Codex. Go pay the 20-buck-a-month for Claude and start using Claude code or co-work and just on your own personal things, just time testing things out and setting up agents to see how those work and how they can handle the routine tasks. Because i i think you're going to see more and more uses of that.
Speaker: And I know a lot of people think, oh, well, that's that's going to mean you're going to need me less. But I think it's actually going to require, and I was, I was listening to a podcast on the way in this morning, you know, that, uh, we still need an adult in the room.
Speaker: We still need a human in in the loop. Um, we need someone that can guide direction and set strategy going forward, but we don't necessarily need someone.
Speaker: you know, that that does all of the initial grunt work in order to just get to step one. i think, you know, we're gonna look at things as an AI first infrastructure and figure out how that works. And the people that figure that that out are going to be much, much more valuable. And there's going to be opportunities that will open that are just not, that we don't even know about right now. So if you think the work's going to go away, I've got another story for you.
Speaker: i I think ah on the contrary, you know, is the word processing, you know, didn't mean that, you know, that to creating documents went down, it actually shot up. So, and you hear a lot of it with the, I think it's the jet, ah Javon's Paradise or Javon's Paradise. or Paradox. Sorry.
Speaker: Paradise sounds better, though. Paradise. We're back in heaven. Yeah, we're back to the good. The easier it is or the less it costs to do something, the actual demand goes goes up. So we'll we'll see how it goes.
Speaker: So my take on the question is, if you have the ability, and this kind of goes back to Greg saying, you know, buy, you know, if you don't have the ability, go buy something and and you make sure you use it and get those skills. But start doing these individual agent builds like, you know, there's there's there seems to be a rise in.
Speaker: you know, these these sort of quick builds that that anybody can can make to to solve like discrete problems. So become very proficient at that because I think people are gonna want that and they're gonna be, you know, many of of our patrons and customers that aren't, don't you know, either don't wanna spend the time or just don't wanna do it. So I think that that you will be a true resource there if you can do that.
Speaker: Because attorneys are starting to build these things themselves. And so if you can be a help there, I think that's a great, a great idea. and they are they are, I can, I can validate that. Yeah, i know. same Same here. So also like look at start getting involved with discussions about workflows. And I don't mean like workflow in a particular tool. I mean, like the process from start to finish of a matter.
Speaker: And, you know, if you can start getting into those conversations and start mapping that out, that is something that is happening now where, you know We're looking from start to finish as to you know how AI is impacting something and how is that going to change what the process is.
Speaker: And you know i think that is going to be a big growing area. And also, I think for so for some of these things... It's not just the AI, like there's going to be an automation component. And I know we've had like automation for a long time. But so if you know if you're a you know if you're a power automate person or you know whatever tool of choice to to sort of work on automations, you know what I see is you're going to have a combo. You're going to have like you know this automation component and then...
Speaker: you know, have ai being, you know, drafting something or composing something or or analyzing something or whatever it is, you know and then you're going to have these stop points where, you know, people start looking, but, you know, you're going to start building these things to sort of flow.
Speaker: And again, I think that is a great area for for those who are so inclined to get involved. Great. Thank you both so much for joining us on our very first podcast. I feel like we didn't break anything. you know I don't know if that's a good thing, a bad thing, ugly thing. feel about good about yourself so um we We haven't had to stop recording yet. so so Exactly. No, I know. This is where I'm i'm reeling you out.
Speaker: But I do want to thank you all for joining us today. And i hope everybody listening is also a joining enjoying our very first All in podcast.
Speaker: Thank you. This has been a lot of fun. Yeah, thanks. Yeah, this has been amazing. And again, shout out to the podcast OGs for being on our our first show.
Speaker: And I guess I would leave, and this is since I started off with a y'all, I'll leave a shout out to our people across the pond in Europe. Bye y'all. We'll see you on the next episode. If you know, you know the joke. Okay. Yeah, exactly.
