Efficiency and Ethics: The Dual Impact of AI in Law

Well, thank you so much, Daniel and Jared. That was a wonderful introduction. I definitely sound way more important than I feel. So thank you.

Well, we'll see. We'll see.

We'll see.

Yeah. Thanks. Look, I've been wanting to do a podcast for quite a long time. I love at the end of the year, the podcasting app that I use does the year in review, And it's absurd how high I am in the percentile ranking of people who consume podcasts by minutes.

It's a little bit of a problem of how much podcasting I listen to. So I've been listening to them for a long time and thought, you know, for for a long time that I'd like to do one. And so the opportunity came up to do that when we were trying to build out and bolster our AI practice group at my firm Barnes and Thornburg. So it's been fantastic.

It's really a great opportunity to talk to people that are in a variety of businesses and industries, different sizes of businesses, and just talk to them about, I think, kind of the real practical implications and problems that they face regarding generative AI because, you know, I'm sure we'll talk about it, but we're still somewhat in early innings here. And there's still, I don't think, a convergence on best practices or best tool for everybody. And so I think a lot of people are trying to figure it out. And so we try to talk about it on a practical level.

And, you know, I think the idea is we can go as deep into the weeds as we might want to with a particular guest, or we can keep it very high level and talk about it from kind of the c suite or even the board of directors level of the big picture. So my hope is it's accessible and hopefully entertaining. I do not want it to be, you know, bogged down on on legal issues or legal precedents because I think those aren't terribly interesting to the general public, even if they might be interesting to myself or other lawyers. So it's definitely designed to have a broad applicability and appeal.

I'm a very systematic thinker. And what I mean by that is, you know, when I was in law school, I would see the things that connected kind of the dots between torts and contracts and procedure and why it all kind of made sense and where everything has to kind of click into place. And if it doesn't click into the place, it would just kind of bother me. And so then when I started practicing law and doing business litigation, a lot of these cases are going to be pretty document and fact intensive.

Like, sometimes a dispute that I'm handling might span, you know, three, four years of facts, maybe more, sometimes eight to ten years or even longer. You know, one example is the opioid MDL that I was part of. You know, that's something that the plaintiff's alleged started decades ago. And so it would continue into the present.

And so there's a lot of information and a lot of facts that in isolation or in the abstract don't seem relevant at all or of marginal relevance. Like, why would we care about that? But then when you start to kind of look at it systematically the way I do, you do start to see the patterns, not only in the things that are openly there and above the table, but the things that are missing and the things that are under the table. And that's where, you know, I think technology really helps me.

And that's what I'm super excited, you know, to bring it back to generative AI. I'm super excited that we're seeing those developments in technology because it's gonna help people like me, you know, do even better with those skills.

Yeah. I think that's, the way I would describe it as kind of a moments, you know? And what I would tell you is that not to be too dramatic, but, you know, as I'm going through my career, I've had this kind of like pain where I'm like, oh man, if only there were a tool that could help me do this faster or, oh my gosh, I had this flash of inspiration, you know, that feeling that you can get when you're in the flow, nothing should take you out of that flow. And unfortunately, you know, like the current state of tech, you know, five years ago, you know, two years ago, maybe even today would take you out of the flow because you're like, oh my gosh, I gotta do so many clicks.

I gotta go and find that thing on the kind of old systems that we have, or I gotta go to our cloud storage system and hunt and peck for this. Or, you know, even like I was telling Jared, even our main email software that people use has become difficult to find things. Those are things that take you out of flow. So, I would say that there's been a couple of moments where the first one would've been when one of the major large language models kind of, you know, went viral in the fall of twenty twenty two.

One of my buddies who is a serial entrepreneur has sold companies off. He's, you know, in his, like, early thirties still. And, you know, he's like, hey, man. You gotta check this out.

Like, ask us some questions that you would ask a first year lawyer. I'm like, alright. And it was really impressive. Right?

So these are things that I've been following for a long time. You know, AI again, the funny thing is I think we would all agree that AI is only what's the new and shiny thing and things that used to be considered AI, we don't call it AI anymore. So it's something I've been kind of following for a long time, but the first moment was that. And then I was using it since that point.

So follow twenty twenty two. And then the other big one that I like to talk about was a tool that was released by another company, another lab that I'm sure you've seen that can kind of create a podcast out of written sources. And that made me, I was sitting in my office and one of my colleagues and teammates from the AI practice group sent me a note and said, check this out. This is pretty cool.

It was a non lawyer too. So we have non lawyers in our group too, which is really nice because you get that diverse perspective.

And I got, I don't know the hairs on my neck raised. And I was like, this is so good. This is so cool because one of the ways I like to process information, going back to your question about how many podcasts and everything I listened to is orally. Right.

And so to hear it and see the documents come to life, because this is basically the best retrieval augmented generative AI. So rag AI I've ever seen the context window is big enough that you can actually use it. You can give it diverse sources. That was the second huge moment.

I'm like, oh my gosh. We're getting so much closer to what I had envisioned and wanted for so long that, oh my gosh. If we had these tools that would keep me in flow, that would help me even flow faster. We're starting to get that.

So it's very exciting to see these things come online. And I was having lunch with another lawyer yesterday, and we were kinda chatting about this because what he does, what his practice is he does a lot of general counsel work for other law firms. So if a lawyer gets in trouble, their is the line or the law firm's gotta be claiming as them, that's part of what he does is defends them. Right?

And so we're talking about AI and we're talking about the ethics of that. And, you know, the tension that I recognized is that the law in general, right, is trailing. The law is almost always behind whatever is happening in the quote unquote real world. Okay.

And so you have this profession that tends to operate with a lag time. And so it was like a trailing indicators, but the breakneck speed of innovation when it comes to Jira AI is pretty unique even compared to other technological evolutions or innovations we've had. And so that's a big mismatch where you have gen AI stuff. That's developing like every week in an arms race sort of way.

Like for instance, you know, one of the major models just released a huge improvement to its image generation capabilities, way better, Really amazing. And they're adding, like, a million users in a week, which is incredible. Well, if you think about it, that came a week after a competitor released its upgraded image generation model. And it was after that same competitor had released a new frontier LLM that arguably was doing better than some of these other ones.

And so then my point is, and like I've told people is if you feel like it's hard to stay on top of it, you're right. Because this stuff changes, I would say at least on a weekly basis, maybe on a daily basis and the law, you know, it takes a while. You gotta have the right case. It needs to actually go to some sort of ruling or even a trial, then it needs to go on appeal.

So people have to, you know, actually invest in that. Then you have to have a ruling that comes from that. And then maybe you have some precedent that would guide you.

That's a long process. And so that's one of the tensions I see. I mean, there's other ones too, but that first one is, you know, how do we as a profession look to embrace and harness this tool? Because by the way, our clients are right.

That's the thing that's always driven me kind of crazy as a practicing attorney is, Hey, there's this tool. All of our clients are using it. Why are we not? Well, you know, we have to vet this and vet that I'm like, I understand, but there's that always that tension of that catch up of, you know, the clients and the companies are doing this stuff.

So they expect, I would think their consultants and their lawyers to also do it. So, yeah, that's one of the tensions.

Yeah. That's a good question. It touches on a lot of things that I've also had conversations with people about and, you know, cognitive skills of knowing how to look at, like, let's just use an example, right? Find me the best case from X court about Y issue.

When I was in law school, I graduated in twenty ten. We had a really awesome, beautiful library. I worked there as part time guy and, you know, I would study there all the time and we would go and have to learn how to go to the books. But of course, at that time, twenty ten or two thousand and eight, when I started, we were absolutely using the major online legal research services.

Right. So they wanted us to have both. And, you know, and when I first started, my first law firm, we had a really nice library and, you know, the senior trial partner who mentored me was like, hey, you can do that. I'm gonna be in the library if you need to come ask me questions, you know, looking at the books.

And so I would go and still, you know, apply that same skill and go look at the books and do it online. But today, I guess if your question is, are the younger attorneys, first year lawyers, second year lawyers going to be increasingly using generative AI as their first resort and maybe not have the skills of using the online legal research providers? I suppose so. I guess the real question for me, like, what kinda gets me a little worried maybe or concerned is what I think ultimately it comes down to is clients are paying us for judgment.

And judgment comes not only from knowing things and knowing how to learn the things, but having seen this movie before. And I think, you know, that takes a while. There's probably a good reason why most law firms have a seven, eight, nine, ten year partner track because it does, whether you like it or not, take a long time to build up that judgment. Now my question is, you know, what's gonna happen if we get in a situation where it's such a no brainer and the tools are so good that nobody is going to be, you know, doing things the quote unquote old way.

Are these attorneys that are coming up, are they going to build the requisite body of knowledge that will give them the judgment that we ultimately are selling? That's, I think, a question that, you know, all lawyers are gonna have to think about, all law firms are having to think about. And, you know, frankly, it's probably a bigger question for the entire ecosystem, including law schools, you know, state bars. I think it's something that the entire profession has to kinda consider, and I don't have the answer.

I think there's a lot of people, a lot of smart people who are thinking about it and have probably different viewpoints along a spectrum of ideas. But I think it's something that it's good that we're talking about here in second quarter of twenty twenty five because it's something that should be considered. The human touch, I think, is gonna be important. I think just to use an anecdote, I had a mentor back in twenty fifteen, and he made a big point of buying personalized stationery and then writing people notes in pencil and then mailing it to them.

So they're lawyers, clients, judges, whatever. And he said, Nick oh, you know, so we would always say, Nicholas Sarokhanian. He would always give me the full thing. Nicholas Sarokhanian.

And note is really a thoughtful thing, and it's especially thoughtful if you get one with pencil. I'm like, okay. Alright. So I started doing that too.

And I think that's an idea of how to preserve that human touch. I think people are gonna still maybe even put a premium on having that in person connection or at least what we're doing, you know, over video conferencing or on the phone. I think there's gonna be premiums for that, but it's hard to predict all of this. You know, the other thing is I have six young humans at home too.

So we have six kids. Our oldest is thirteen. Our youngest is three. You know, I was walking back to the house with a couple of my older kids and I said, Hey, dad's gotta go down to Texas to do this speech about AI.

And they're like, cool. What are you talking about? I'm like, well, I'm kind of figuring out I'm using one of those LLMs to help me write the speech and write my slides. And they're like, and they're still in middle school, you know, and they're very conscientious.

And they're like, well, isn't that cheating, dad? Isn't that, like, plagiarizing? I'm like, no. It's not.

But that's a good question. And it's interesting that you thought about that. Right? I said, I guess not because what I've come to see is that right now, the current state of Journey of AI is really good for at least two types of people.

Okay. Maybe more. Right. But at least two. One is the expert and the other is the explorer, and I think you can wear the different hats.

So for instance, I'm an expert on only a few things. Right? I'm an expert on the legal practice that I do. But even within the realm of law, there's certain niches or expertises that, you know, there's other people that are gonna be more expert than I am.

Right? But then outside of that, I'm not expert on hardly anything else. Right? You know, I'm not an expert on how to fix the engine of my car.

I'm not an expert on how to build the treehouse, you know, that I built. But it's You know, you can look things up. And so I think it's really great for those explorers who need that quick primer on something. Right?

You don't have to necessarily trust it. You don't necessarily have to take it to the bank, but it's a good way of getting a starting point. And if you're an expert, you can take the output from the Jnerve AI tool, and you're expert enough to use your judgment, going back to that concept.

You're expert enough to say, okay, that's not quite right. That's right, but not super persuasive or relevant. And you can kinda hunt and pack through it and say, okay, that's good. So is it plagiarizing or cheating to use the LLM to help me generate a forty five minute speech about stuff that's already in my head?

I don't think so. All it's doing is speeding up the process because, you know, six kids, busy practice, all this other stuff. In order to sit down and really, like, write that from scratch, I'd probably have to lock myself in a room for at least two days to try to do that. You know what I mean?

Try to bring it back to your original question about the human touch. I think it's gonna help facilitate those things because it's gonna help us speed up some of those other, maybe rote work or low value work that, you know, because we're already expert, we can expedite, you know, and then maybe free us up for those kind of in person touches that can't be replicated. That's the, you know, going to lunch with somebody, taking somebody to do a ballgame, going to whatever.

You know, and I think you can take another step further, Nick. So one of the things that I'm hearing you say is that if you're wearing the hat of the expert, that it really helps in upskilling you and being able to do more substantive work.

Yeah.

Oh, yeah. Yeah.

Yeah. Yeah. Yeah. Yeah. A thousand percent. You're right.

Yeah. Yeah. I mean, well, that's the other thing. And going back to the other question we had about, you know, the younger lawyers getting that training, you know, motion hearings just, you know, for folks who don't know, they're pretty rare depending on where you live, what state you're in, whatever.

But, you know, it's not like the good old days, quote, unquote, that I heard about from the white haired trial partners. Right? So given that dynamic, when you do have a motion hearing, it really pays to have that experience, you know, the multiple jury trials and arbitrations and appellate arguments, like having all of that really helps because you might have fifteen minutes. You might have sixty minutes.

You might be in a all day slug Fest. It doesn't really matter, but you have the skills and judgment and wherewithal to read the room, to know what's landing, what's not landing, read that body language, understand what's behind a question, and know the best way, the most persuasive way, the most precise way to make those points in the correct sequence at the right time, in the right way, with a soft touch that I think that's the benefit of hiring somebody with that experience. Right? Otherwise yeah.

I mean, I guess in theory, you could even have a non lawyer. Right? God forbid.

But you could have a non lawyer read a script from an LLM with the same content, but not knowing the judges, not knowing how to talk to a panel of twelve people maybe, or three arbitrators or whatever, and not knowing what to say, when to say it, how it's landing, and what's behind the questions. Yeah. That's a total game changer. So, yes, the human element for things like that, you can't you can't overstate how important that is.

So look, if generative AI will help people, you know, like me and other people, just anybody, right? I think it's a universal human thing that there's certain things that are higher value. If we can get more people to that sweet spot where they really are thriving, I think productivity goes up. I think personal satisfaction in your job, which is incredibly important to human dignity that goes up as well.

So that's what I'm hoping for, whether it happens, you know, we'll see, but that's what I'm hoping for.

Yeah. That's another one of those really big questions that lawyers are grappling with. And I think it's a great question to think about deeply. What do you do with those nine hours?

You know, that's the flow thing. It's just, if I could take out all the roadblocks between the idea I had and deploying or implementing that idea or that strategy, or looking for that insight that's buried in the documents or in those giant Excel spreadsheets that, you know, are always exchanged in cases, that'd be amazing. Right. Then I would actually get way further along on my plan with those extra hours.

So, yeah, I'm not worried about, oh my gosh, I wish I were able to bill for ten hours. No, no, no, no. Like I have a real life example that I used at that talk I did. And I said, look, I have a case.

All I need to figure out to move forward, to tell the client this, you know, for their go no go that, look, these guys took these amount of people worth this amount of dollars within this twelve month period where they promised they wouldn't. Right? Straightforward question. The devil's in the details and looking and trying to peer into all these excels.

And I said, look, I'm not a financial analyst. You know, I'm very good at Excel, but I'm not like, I don't live in Excel. And so what used to happen, my workflow would be I'd spent eight, you know, for the sake of this example, eight to ten hours hunting and pecking through Google and YouTube. Right.

And just trying to hopefully find somebody that will help me answer a little part of the question that I have. And, you know, nowadays you can just even anonymize the headers, you know, and say, hey, assume I've got to an LLM. Assume I've got these headers, a, b, c, d. I just want the formula will help me do this and answer this question.

And, you know, that entire thing took less than twenty minutes. You know, there's some trial and error to make it work. But am I upset that I didn't get to build that? Absolutely not.

I'm rejoicing because now I can get so much more done in the day to actually solve the problem. And then I can say, guys, this is how much money that these guys took from you by doing this bad stuff. Go or no go. And they say, go.

Okay, great. Now we go. So it's so much better. So the question about the value billing, you know, I think that's something that a lot of really smart people in law firms, like my firm has got fantastic legal pricing people, legal operations people.

I welcome the day where the vast majority of our clients are looking at it that way. So if at some point we get to there, that would make a lot of sense. I'd welcome that. But, you know, at the same time, perfectly happy to do hourly rather than, you know, ring my hands about, well, I wish we could have charged.

No, that's not at all a problem. You gotta play with these tools. So you have some baseline familiarity of what it takes to use them and what's more effective. I mean, it's kinda like talking to somebody who hadn't used a major search engine when they were coming online.

Like, how do you learn to use them? You start using them and then you learn. Right? So that's my number one advice is that if you're in law school, you're in college, the bottom line, short answer to your question is start using it, do it in a responsible way, do it with things that don't matter, like, you know, music history or what's the best way to iron a shirt or whatever and try different tools and see what you like.

Yeah. And, you know, and the first thing that comes to my mind, of course, is ABA formal opinion five twelve, which is the first one that kinda codified for all of us. To me, when I read, I'm sure to you as well, Nick, I'm like, gosh, this is a no brainer, Right? It is literally citing all of the rules of professional conduct that we should all frankly know, but your firm, how do you lead this AI vision when we're talking about go ahead and just get into it, start using it, use it in a responsible way, but also be cognizant of the fact that you may potentially pierce the attorney client doctrine if you inadvertently put something in there and violate rule one point six c, the rules of professional conduct, etcetera.

Yeah. That's a great way of putting it is there's a balance. And just to be clear, you know, when I say just start using it, it's things that, you know, I say it kind of in a flippant way, but the way I'm trying to explain it is do it with things that aren't truly trivial. Right?

Things that literally don't matter. Like take a picture of what's in your fridge and tell me what to cook. That's a good example. But when it comes to professionals in our firm and we have a policy that covers not just the attorneys, but literally all employees.

Right? Because they're important. Their ability to do these things are very important, but also as lawyers supervising them, you know, we have to make sure they're not doing things that would traverse the ethical rules just like the lawyers are bound. Right?

And so what we have is a pretty robust and I'd say conservative, moderate, but more conservative than liberal policy in that we have a white list approach of tools that we say are okay to use. And we have a few default permissible uses, but anything coming close to client confidential information, and we follow the ABA policy pretty closely on that, is something that we do require client consent, written, informed consent, advanced consent from clients that we're gonna use it. And is there a tension there between wanting to, you know, innovate and try things? Yeah.

But the safest course, given how seriously we take our responsibilities to our clients in the profession is if it's gonna be anything related to a client or for an engagement for a client, we gotta get that. And so we've been doing a really good job of doing that. And then, you know, the other thing is firm, confidential information, financial information, whatever, same thing. You know, if it's gonna be anything like that, you know, you need to get the permission.

You need to make sure it's being done in a proper way. But if it's for things that are more mundane or things that are like completely, you know, idea generation or whatever, then there's a little bit more of a flexibility there. So it's something that we are proud to have and, you know, we're always looking at it and considering whether tweaks need to be made here and there. And frankly, it's gonna have to as things change.

Right. And we, again, going back to our earlier conversation about the tension between legal standards and the innovation of these tools is, is there. So it's something that, you know, we're taking a flexible and realistic approach to.

These things are gonna change, as you noted daily, weekly, monthly. We are in an unprecedented time in the legal arena and more broadly for everybody else across the world. And your insights have definitely helped our listeners today. So thank you so much for joining us.

Yeah. Thank you. And if anyone would like to connect or, or find me, you can find me at the Barnes and Thornberg Law Firm website, www.btlaw.com and click on artificial intelligence. You'll see my face there. Otherwise, find me on LinkedIn or just shoot me a note. So thank you so much for having me. Can't wait to continue following your podcast and, you know, kind of figure this out, navigate all this innovation and change together because it's definitely exciting times.

Thank you so much. It's it's great. Thank you guys. Thank you.