Transcript: 

Adam: Welcome back to Reed Smith, Disputes and Perspectives. I'm Adam Massaro, a partner in the Denver office of Reed Smith. I have a new special guest with me, Ross Guberman, today. Ross, introduce yourself, please.

Ross: Hey, great to be with you. I'm Ross Guberman, as you said. I'm the founder and CEO of BriefCatch, which is a legal editing tool. And before that, I actually did practice for a couple of years in another huge firm, although not in the Denver office, unfortunately. I'm a big fan of Colorado. And then for many, many years, I did workshops all around the world on different parts of legal writing for many of the big firms and most of the circuit courts, so judges and lawyers of all stripes. And that was kind of my sandwich between practicing and getting into legal tech. Plus, I've written a bunch of books.

Adam: I still think Point Made is one of my favorite books. When I was an associate, it certainly brought things full circle for me as just a little teaser. But that's only one piece of the many things we're going to talk about today.

Ross: I actually just finished a new edition that was much overdue. And many of the lawyers I profiled had gotten scandalized or passed away. So it's a nice, fresh list of lawyers and cases. And of course, lots of new techniques, given all the advances with tech and AI.

Adam: So to reset, of course, you're certainly well-known on the legal writing. And I do want to touch a little bit on that. But we're also now melding tech and legal writing, which is going to be the main thesis of today. And that's what I want to talk to you about.

Ross: Great. So that's my new passion. And it's been several years now.

Adam: Well, let's jump right in to the concept of AI and litigation. And I will present some theories I'd like to hear your thoughts on. One, say that AI is going to be the death of legal writing. And it's all going to be done by other people or other machines. But I don't think that's true. Where do you feel about that general concept? We're going to go deeper after that.

Ross: Yeah, luckily, I don't think it'll be the death of legal writing. I think you're going to start to see an enormous gap in skill level between the best writers and the weakest. So if you're someone who doesn't like writing, isn't really into it, unfortunately, even as early as maybe the first year of law school, AI will become a crutch. You'll be able to get away without really developing your skills. And that's a sad tale. I hope doesn't happen. But on the other hand, AI, if used wisely, if used sensibly, can really make you a superior writer, can let you experiment with your style, can let you learn from others and meld other people's styles with your own, and try different approaches strategically, persuasively, tone-wise. So I think the motivated legal writers will be better than ever. And it's going to be really exciting to watch. So yeah, not the death of legal writing, but I also don't want to be a Pollyannish. I, too, am concerned about the younger generation being deprived of all the good kinds of struggles you really need to become a better writer, because ChatGPT and Claude sometimes just make it a little bit too easy.

Adam: So for the writers that you expect to get better, how do you anticipate they are or will be using AI to hone their craft in legal writing?

Ross: Well, one thing a lot of litigators and appellate advocates are doing, and even some courts are doing in writing opinions, is using AI on purpose as a sort of skeptical reader. So you can do a draft of a motion in Lemonade. You can do a draft of an order if you're a district court judge. And if you're brave enough, and you got to be a little bit brave, your prompt is something like, if you are not inclined to grant the motion, if you are not inclined to favor my ruling, what would be your three biggest, most cohesive objections? And you can use that kind of a process to really, really refine not just your wording, but your actual message and strategy. So that's a really exciting use of AI that you can really do right now with just general LLMs. So that's one really great way to kind of preempt a lot of counterarguments that often come a little too late or came a little too late in the pre-AI era. Another thing you can do, and this takes a little bit more confidence and skill with AI, but another thing you can do is you can actually feed it a lot of your most recent work product as you embark on a new type of filing or a new letter to a client or what have you. And you can try to get it to help you figure out patterns in what you already did. Because of course, you don't want to have a basically glorified copy and paste job. But you can actually train AI to look at patterns in your own writing style. And again, not just on the wording front, but even the way you make arguments, the way you handle cases. And you can use that to really inspire yourself as you draft something new as well. So that's another way, again, to have better work product and to improve your skills. And then I'd say last, we have a tool, there are many other tools out there. If you use the right kinds of tools that don't take over for you, that don't draft a legal document from scratch, but give you some agency and maybe challenge you to think of better ways to sound even more like yourself and maybe even teach you something along the way, those kinds of tools can make you a better writer as well.

Adam: Going back to the harsh critic, and that's something that I'm completely in on, especially at the partner level, because I'm pushing work down. And so I oftentimes, I get the execution, but the harsh critic oftentimes is the hardest situation, especially at my level. When you approach that concept, I understand conceptually, yes. We say, LLM, give me counter arguments, but I think we can take it a step further, and you may have already done this, but I think you have to somewhat develop the persona of the other side. What is that veneer in the sense of how sophisticated are they? Who are they in front of? And to me, that seems where there's an openness that I don't know the technology is there yet, but how at this point where the tech is, can you create that harsh critic that's actually reflective of your judge, your adversary, or somebody that's gonna actually be thinking of the issues the way that the end decider effect is, to me, that's the real interesting question.

Ross: Yeah, that's a really great question. So let's take the example of like a client, right? So usually you don't think of a client as a harsh critic. I'm sure your clients think you walk on water, in fact, but clients often have questions or concerns about draft motions and briefs, right? Or have often differing views on things like settlement. So if you simply said to ChatGPT or Claude, hey, pretend you're my client and you just read this draft motion that I wanna file, what would you say? You know, as you suggested, you're probably not really going to get the kind of feedback that's gonna be really actionable. The interesting thing is, and this is why humans do matter, if you wanted to ask that question, I can guarantee you, you've got something in mind, right? Like I know my client likes me to use really fiery language that I know is gonna piss off the judge, but they're gonna make me do it anyway. Or I think the client's gonna wish, you know, we talked about damages before we talked about the merits. You probably know more, you certainly know more than the LLM, but you probably know more than you realize. You've gotta put that context in there as well, right? So to say like, what would a client think or even what would a judge think? As you suggest, you'll get some good output, it will be better than nothing. But if you really add what it is that you're concerned about, then it has what it really needs to go on. And then actually is fully capable even today of giving you the kind of reaction that the human you have in mind would likely give you in real life.

Adam: You know, it's interesting, I think the one nice thing about judges that have been on the bench for a while is we can look at their writings and historically we're even predicting what they do by looking at historical writings, maybe ineffective, but still nonetheless trying so. And so to me, the use case there is more simple, we can dump in all their rulings, regardless of they're on the specific issue and get some solid predictability. So I like that. But what happens when you have a brand new judge that maybe just been appointed to the federal circuits or at the state level? What are your thoughts kind of where we are today or where we can go on how we can start to predict how that type of judge who we don't have a track record may start to ingest and interpret issues?

Ross: Yeah, so in that case, unless you know something about their previous career, you know, were they a big law partner? Were they a public defender? Were they at an interest group? Otherwise you don't have much to go on. So I do a lot, actually I'm the one who trains all the new federal judges on opinion writing. So I don't know if I have any influence that would help answer your question, but if I could make a generalization, most new judges, and I know a little bit more about federal than state, but I've worked with quite a few state judges as well. Most of them, if you think it through, this is kind of what you would expect. Most of them tend to be very much by the book. In other words, they will talk more about the cases than they will 10 years later. They will like cite, they will deal with every fact in the record. They're gonna be more methodical. And actually, I think I looked into this once. They actually do have on average longer opinions than judges, not necessarily older, but judges who have been on the bench for let's say 10 years or more. So that might help you if you're giving, again, the LLM context. Is there any case, is there any doctrine or case that I skipped over? Or are there, another thing that happens when people go from like what you are, like a big law partner to judges, when they actually have to read the fact section and they weren't the litigator, then they realize, I read this fact section and I have a lot of kind of basic questions though at the end, like kind of skipped over a lot of things that I really would like to understand to figure out how these parties got into a contract dispute or what have you. So what I would say in those cases, what people want is, I'm skeptical of, they wanna know like, what's the magic word that this judge secretly really likes? Or like, what's a phrase that the judge really likes? I don't really think that's gonna get you far ahead. You wanna put yourself more in the mindset of the reader in that case. And that's probably a pretty sure bet. If they're a new judge, generally again, they're gonna want maybe a little bit more methodical writing, a little bit more methodical reasoning. And if you give that context to the LLM, it might give you a sense of where you're kind of skipping over steps or being a little bit too cursory or pat with the case, a lot of the facts.

Adam: But digging deeper on the training of new judges you do without getting into any level of detail, that's unnecessary. I'm curious, what are some of your takeaways and observations of that process from the outside?

Ross: It's interesting. I mean, I've been doing it for a while. I think I started like, I don't know, 2004, 2005. Maybe actually a little later, 2009, I think when Obama, Obama's first term, that's about when I started. It's actually changed. And it's changed in ways that probably won't surprise you. So your typical new newly appointed federal judge nowadays is very well aware that people on social media, X, LinkedIn, read opinions in a way they never did even 10 years ago. There are, as we know, there are death threats, there's violence on, unfortunately on the extreme end. So they're actually much more aware of themselves as writers and they're much more aware that they have a much broader audience than just the counsel in the case. So that's all good for me because it means they really care a lot about their writing. So that's one thing I've noticed has actually changed. The other one though, is that even though most federal judges have really distinguished backgrounds, when I started, for example, I was nervous because I thought, my God, I'm going to speak to a circuit, new circuit court judges about opinion writing. Like, they're gonna be like, well, who the hell are you? You know, they all knew my name, but it actually is the opposite. They're actually some of the nicest, like most curious audiences I have, not because they lack confidence. They definitely have confidence, but very few of them have ever written an opinion. And it's actually much harder than I think a lot of litigators realize, right? They think the judges have it easy, right? They don't have the clients, they don't have to deal with the initial putting together of the facts, but they're actually quite open to any help I can give them because it is actually quite overwhelming. And then of course, obviously, they're also very well aware of the power that they have, you know, that the opinion is just a means to an end too. Yeah, so they're actually honestly like a really, really great audience. And then the last thing I'd say I've noticed, and this is not telling tales out of school, because of course I say all that, but they do have by that point their own writing style, like they don't need me to tell them about the passive voice. What I've noticed is kind of the same thing I used to notice with litigators, which is if you say, if I say like, this is what you should do, you know, be skeptical, but they really wanna know things like, hey, how does Chief Justice Roberts write, or Scalia or Kagan, or, you know, some of the better known circuit court judges, or as time has gone on, I've tried to show them how district court judges write. They may not be quite as famous, but they obviously have many more opinions and orders than the circuit court judges do. So just as many litigators have, you know, people they look up to litigation-wise, writing-wise, so do a lot of judges as well. And they're actually quite curious, right? They wanna know like, what exactly did Scalia do? You know, maybe not in dissents, but in majority opinions, or like, why do people think Kagan's style is so great? They're actually quite interested in any, you know, concrete tips I can give them.

Adam: Since we're on the topic of judges, I think we're seeing a greater awareness and adoption that using AI for brief writing and development is a net positive, and the scaries on balance are reducing. But that doesn't address the judiciary using AI. And I'm curious what your take is there, because I could see even from a practitioner where you're like, I wanna use this tool, but I don't know that I want the judge to. But what are your thoughts on where we're gonna see the judiciary go first, and then we can go deeper after that?

Ross: Yeah, I mean, obviously when the topic comes up in the legal news, the most common thing you hear about is some dramatic hallucination story and a judge getting angry, and lately even sanctioning lawyers. That can give people a really false impression that judges are against AI. I can tell you one thing, they're not against AI for themselves. I mean, some of them are, right, just as some lawyers are. But, you know, other than maybe like the Supreme Court and some of the, a few of the circuit courts, I don't want anyone to get mad because they think they work hard too. The truth is most judges have really, really overwhelming dockets. And in the state courts, it's sometimes out of control. Around the time of COVID, they've started getting falling behind and they haven't really caught up. So the truth is I understand why people are nervous about judges using AI for decision-making, but putting that aside for now, because I'm not sure I agree with all the criticism of that, the truth is you'd want the judges to use AI for things like really looking at the record, really digging into the deposition, excerpts, you know, the case law. They're using it in those ways too these are not ways that appear in their opinions, right? They're using it just as litigators do in preparing motions, right? They're often putting the party's brief side to side and asking, I kind of teach judges to do this. A lot of them don't know how to do it. I show them how to make a chart that shows, you know, what the plaintiff says and what the defendant says and what the reply brief says and so forth. That's good. These are all things we should be happy about. And then another thing to know is that because of Gen AI, you have a lot more pro se filings now and they're not written in, you know, on pencil. I mean, they look like normal motions and briefs even if the writer doesn't have access to all the normal tools. So because of the increasing number and also the increasing length of pro se filings, AI is a way to give them, you know, give them their due in court as well.

Adam: When it comes to judges using AI in the decision-making process, I track all the things you say because I don't think that the norm is the judges over-educated. We all know that they have finite time. So to me, I definitely see that unbalance. My greater concern is you not understanding my arguments because you don't have enough time versus understanding and disagreeing. I can live with the latter. The former is tough to handle. But back to that concept of how they should use it. On some levels, they don't have the resources that private firms do, things like that. Where do you see the path for how the judiciary will get greater access to AI usage in a way that's both responsible and beneficial?

Ross: So there is a lag, as you say. So if you go to a big legal tech conference, not surprisingly, most of the products are for practicing attorneys. There are more and more for the courts. There are a few that are specifically for the courts, some of which I've seen seem quite impressive. And then there are many like ours, BriefCatch that work for judges and not just lawyers. So there is a lag. They also, of course, have security concerns, but so do firms like Reed Smith. I mean, in some ways, by the way, it's actually easier to sell the courts than to firms because although they have security concerns, one thing I've learned is when the judge says, I want this software, there isn't really anyone to say no. Like they get it. They get it pretty much days later. So there are also some judges that are sort of on the forefront of using AI in the judiciary. And I've noticed the last judicial conferences I've spoken at, that's now a standard topic, right? You'll have a judge, not someone like me, but an actual judge who's a leader in the field and the judge will walk them through some use cases. And the last time I saw that, they looked like they were watching like an amazing movie, like they weren't even moving, they were so fascinated. So I think it's actually really overall a very good development for some of the reasons we discussed before. We need to be very realistic, as you alluded to very diplomatically. We need to be realistic about what really happens in the courts. They are not pouring a glass of cognac and reading your motion for summary judgment, like it's a sonnet, right? And many people in the state courts wonder if the judges have read the briefs at all. So if this helps them process the record, process the party's arguments, all the better.

Adam: When it comes to some of the use cases you've seen referenced by judiciary individuals at a, for example, conference, I'm curious, substantive use cases, what have you heard of how the judiciary is using AI?

Ross: Well, I mean, without obviously naming any names, there are judges who, and I want to preface this by saying as a check on themselves, not as a way to skip a step, a check on themselves will feed the briefs and the record and say like which party has a stronger argument. But as far as I know, and I can't speak for all of them, it's not like they're just cheating, right? Or delegating, they're curious, right? So if they think they're gonna go in the defendant's favor and AI says the plaintiff is right, they don't change, right? But I think this is good, right? It's a little bit like what we were talking about the litigator saying, hey, okay, I'm in love with my own argument or at least my client is, but what would you push back on? So there, I've seen that kind of use, which I actually think if you explain it properly is very good, not bad. There are also many courts trying in different ways, like we have a product that we're launching that does this too, to verify, not just that the cases actually exist, but what you say about the cases is accurate. I mean, traditionally clerks do that, but again, outside the kind of lofty echelons of the federal courts, the circuit court judges with multiple clerks, we don't, it's not always possible to rely on clerks. So they're using it to verify statements about authorities, also to check the record and to check facts, right? For judicial notice purposes and the like. So there's a lot of that going on. And then, they're also using it in ways we talked before to look at their, sometimes if they're a new judge, I'm just making this up. Like if they go out of the Seventh Circuit, they might wanna know like how would Posner or Easterbrook write this. So they might feed some precedent, writing precedent, if you will, and say, hey, can you make my order sound a little bit more like that? Or they're doing it again with their own work as well. But yeah, so lots of different use cases. And again, you do have the same divide you have with lawyers. It's not based just on age. I know plenty of judges in their 70s and 80s who are using AI and plenty in their, well, I don't know that many in their 20s and 30s, but let's say early 40s, late 30s, have no idea what AI even is. Yeah, so it's fun to watch. It's fun to watch. I'm glad to be a part of it myself.

Adam: I think the reality too, right, is regardless of the judge's vintage, judges that have clerks, it's inescapable that they will be using AI. It's appropriate, they should be, and under the right circumstances. But you can't not expect that the budding stars coming out of law school that are very sophisticated are somehow gonna be unaware of this piece. So I think for outside practitioners, I think regardless of what's the inner workings of the court, it seems as though you should assume it's being used. And I think it's a positive because it should elevate your game and you should be considering how that is gonna change the potential outcomes because I think if you're not thinking multidimensional like that, then you are at risk that you're gonna be behind on the current approach to the decision-making process.

Ross: Totally, I mean, I think you mentioned the federal circuit a little bit ago. That's the circuit that rejects briefs, I think, at the highest rate for not complying with their zillions of rules. So whenever I hear one of these stories about a federal circuit brief rejected, that's the first thing that comes to mind. Like you could have just fed your brief into ChatGPT and then uploaded or linked to the federal circuits long list of rules for briefs. And you would have known before you even filed it that you didn't comply, right? The font's not the right size, you don't have the right order of the required sections and the like. And what I would guess is maybe some of the lawyers that don't do that, all they know about AI is case hallucinations, so they think it's all bad, which is of course not true. It's a really great tool. And they're depriving themselves of things that could really help them and help their clients. I mean, by the way, I just wanna say, I don't approve of all these ridiculous formatting rules that all the different courts have. I don't think it's fair and to expect the nation of lawyers to track all those. So I'm not saying it's good, I'm just saying it's reality and it probably doesn't look good to clients to see that a brief is rejected for non-compliance.

Adam: It's a fair point on all levels. So thinking a little bit more about where we go from here from a judiciary perspective, as we look forward and you've seen sort of the arc we're headed down, where do you expect to see AI and other uses amongst the judiciary going forward?

Ross: Well, I'm gonna say something controversial, but that doesn't mean it's wrong. I think it's not gonna be that long where you have a right to opt into an AI judge. And I know for some reason people hear that and they freak out, but I really don't get it, right? I mean, I do get it and I don't get it, right? These are the same people who think a lot of judges are biased. They think, as you said, the judges don't have time to read all the arguments. I think you'll, I don't know who will be first, right? I don't know which type of lawyer will be the most open to having an AI judge, but I think it's probably gonna come sooner rather than later. So I know that's bold, but again, I also, I took a Waymo recently. I would have never thought I'd be in a driverless car and here I was and enjoyed it. So I think the other thing is you will have, you'll probably start having, I hope, less lag time before complaints and answers are processed, motions are processed. There should be a speeding up of the entire process. I think also I would guess, I mean, courts, you know, as I'm sure you know, given your background, of course, generally like settlement. They don't like cases to go to a trial. They don't think it's a good use of limited resources. I think you'll probably also see judges kind of pushing parties to use some of these tools predictively. Like, are you really gonna get the judgment that you want? I mean, are you really, really completely safe and there's no chance that you're gonna be found liable? I think you'll start having a push toward, you know, analytics and predictive analytics. And the other thing I'll say, although I can't prove it, is I don't even know this for a fact, but I'm sure judges also like, you know, early on in a case, like maybe they also use these predictive analytics tools to kind of decide whether this case is gonna be on their docket for a long time or whether it's probably going to settle. Trying to think, oh, the last thing I would say is I hope opinions and orders will get shorter and also more readable, right? Because there's a wide range of writing styles in our nation's courts and many, you know, many of the opinions will barely be understood by the council, let alone the parties and the public.

Adam: Yeah, that's an interesting point, because as you know, as you're so close to something and a complex issue, it can be immensely confusing because at some point you get so deep into it, you just don't know that you're not being clear, even with best laid efforts. So I definitely track that piece, and that makes sense to me. I wonder if on the predictive piece, so I get you a hundred percent, I think that's where we're headed, including for certainly mediators, they're gonna use it all the time. That's just a given, right? And then from there, I think you see sort of opt-in predictive analysis for early case resolution to sort of warm people up to it. And before you know it, I think you're gonna see a, effectively what we used to see like a, you know, a directive to send pretrial filings to a magistrate. I think you're probably gonna see something along those lines as well.

Ross: Do you think, I mean, I don't do this for a living, and you do, do you think parties would be more likely to settle if a court or a district court judge or magistrate judge said, hey, just so you know, like this is what's, you know, this is the most likely outcome. And this is kind of like the range of outcomes based on similar cases. You think they'd be more likely to settle than they would be if it's just a lawyer telling them to settle?

Adam: I think there are absolutely cases where you have an over-enthusiastic client and over-enthusiastic lawyer, right? Or you have an enthusiastic client and a lawyer that doesn't fully see the chessboard, especially if they haven't tried a lot of cases or cases of that subject matter area. So I think it's a net positive to both sides because it's hard to deliver advice to the client that is inconsistent with their expectations. And so I think it's actually gonna be a useful tool as a check for all parties, including the lawyer. And I could see the settlement value being stronger. The other flip side is you're gonna get more sophisticated clients, even when we talk about in-house counsel. At best, they handle one subset of the law prior to being in-house counsel. So I think it's good to have a check because litigation is both art and science. And so having a check on you by your client, sophisticated, unsophisticated, are all net positives because if I have a client pushing back on me, the arguments get better. I may not like it in the sense that if my expectation is I should know everything, which I don't think is right, but if that is your mindset as a lawyer, then having pushback is good. And ego check is a fantastic exercise because when you become a partner, you don't get grades. You don't typically get a performance review, which you should, but we don't give you that. So I'm a big proponent of ego check, and however it comes, wherever it comes from, it's either gonna make my case better and I'll fix the problem, or if I can't deliver you a sufficient answer, we probably should settle the case.

Ross: Yeah, it's funny because we're a venture-backed startup, I have Big Law representing us now. So I'm a client of Big Law firms after having been an attorney in one. And I can see it's a pain in the ass to have clients, I've learned through my own reaction sometimes. And it's all true, everything you said is true. You don't always wanna hear the truth. It's easy to like be frustrated with the law, the lawyer for not telling you that you're gonna get everything you want in the financing deal. Yeah, and I try to be nice. So you have my sympathy. And by the way, I did use AI in our last funding deal. I did use AI to predict that, hey, are we likely to get this provision agreed to or not? And it calmed me down, it settled me down a little bit when it said, no, you're asking for too much.

Adam: That's interesting, but I think it's a good check. And I think any more information is better. We can't provide absolute certainty, but I also think the clients deserve better than the fallback of, I can't tell you anything, or I'm afraid to tell you something. Like that's not helpful to the client. And my view, that's somewhat of a crutch for the lawyer. Like, yes, we have to hedge risk, but if we don't give you a sense of your risk reward, then we also are doing you an injustice. And I think attorneys that are too worried about being too wrong in the future, I think you're almost doing a worse disservice than if you would take a position, and we may not be right, but as long as we know it's an educated guess and not a certainty, I think we're all still better off.

Ross: Yeah, I mean, it's interesting because I've heard and seen myself that in law firms, partners and associates, that associates are much more open to criticism from AI than from criticism from partners, including even like a red line or a markup. And I wonder if that's also true of clients, right? There's something sort of, I don't know if it's the coldness that you know it's algorithms, or that it's not really a person. People are often better to take even harsh criticism from ChatGPT. Yeah, I don't know. I don't know. I think my gut tells me though, with my limited anecdotal experience of being represented by three different firms, it's also true for clients, right? You're more likely to say, okay, I admit it. I'm not gonna get my way when Claude tells you than when a partner in a law firm tells you, especially when you see the bills. You wonder, like I'm paying that much, I wanna get my way.

Adam: Yeah, there is a real objectivity to it, right? And I think that's helpful. You know, I had a chance where one of my favorite use cases is, I tried several cases last year, so I have full transcripts. And so like after the fact, you know, like no one's telling me I did a bad job if I won, which is ironically the least opportunity to learn. But I dump in transcripts now, I dump in depositions, and I don't tell AI who I am per se in that sense, but I do ask for feedback. And it's funny to see the critical feedback on areas that I agree with, but I think it's hard, right? Like no associate's gonna tell me I did a bad job on that cross or I left something on there that I could have gotten. But I think if you want to evolve, which is the best part about being trial lawyers is you keep getting better and the cases get stronger. But there's an evolution there where like, AI is gonna give you more objective cost-free feedback than you'll ever get in any other world that we live in. And I think that's something that no one does as much as they should, but the tools are absolutely tuned to do that right now.

Ross: Yeah, I've done something similar with like an email chain negotiating some point. And when things went south, you can actually ask AI like, what was the comment and who made it? What was the real culprit? Sometimes it's me, right, they say it's me. But I like it when they say it's somebody else, but it's always correct, right? Whether it says it's on the culprit or somebody else in the chain, it's always, I would say it's backed up by logic. And again, it doesn't, you know, whatever it's saying, it doesn't have any, it's not neurotic, it doesn't have an agenda, right? It's not kind of worrying about your reaction the way humans do. And it can, you know, again, it can be, but as you were saying, it can be really, really, really helpful. And that's a really great example of the story you were telling about the trial transcript. That's why AI might make some lawyers have better skills, not worse skills, right? Because you can get that kind of feedback that you wouldn't normally get in the normal course of business.

Adam: So I do wanna drill a little deeper on BriefCatch, which is one of your companies. So where do you see it as far as the key use cases you offer now versus where you're headed next on that software offering?

Ross: Yeah, so, you know, I have like the benefit and the burden of starting a tech company before Gen AI exploded. So we spent years taking all the different tips that I used to teach in these workshops and books and coding them. So we ended up, I think with, right now we have about 14,000 different editing rules and algorithms, but that's, you know, that's good. It's kind of, it's the struggle, right? I did it the hard way. Nowadays, you could skip all that, I suppose, and just say to ChatGPT, like, hey, revise this brief. So that's mostly where we are now. We have some really cool AI products. You know, we have one for blue booking. We have one for verifying authorities. We have one where you can get advice on writing based on my own speeches and articles and writing. But it's mostly a coded product to date. So that, you know, that's kind of the present. And then the future, as you can guess, right now we really focus on words and phrases and making things tight and consistent and precise, but it's always at the word level because coding is, you know, there's a limit to what you can do with coding. It doesn't understand, you know, your overall argument and so forth. So we're very quickly moving into much more, I'd say ambitious types of editing and feedback where it's looking at an entire contract and looking for conflicting provisions. It's looking at a brief and maybe recommending a reordering or restructuring or saying, hey, you know, you're not really explaining why this case is distinguishable. You're just sort of declaring it. So it's quite exciting because of all the power of these latest LLMs. We're really getting into the underpinnings, you know, the logical strategic underpinnings of a legal document and not just helping people write crisper, clearer and more precise sentences.

Adam: So on that point of looking at globalistic issues, but especially with contracts, so I'll use indemnification clauses, right? Which are just right for trouble. Is it indemnification? Is it advancement of funds? Is it just indemnification at the end? Like all this stuff that only comes to bear when there's a problem. Do you envision there's going to be an aspect of what you'll be doing that will also, not only just look at the contextual layers, but also here's the different potential cases that may be out there that, you know, when this comes to the litigator on the backend, this is what the litigator is going to do. But I'm going to fast forward ahead and basically hypothesize the litigator's approach. And then I'm going to try to counteract that issue back in time to avoid the exact litigation that used to just spin out of, you know, bad drafting.

Ross: Yeah, it's really funny you say that because that was my approach to teaching and writing about contract drafting because I really understand language and I understand case law as you do. So you can actually look at sort of the body of case law over things like indemnification, like does hold harmless really mean something separate from indemnify? You can look at body of case law and as you suggest, actually preempt a lot of problems. You know, one example is all the case law and like best efforts versus reasonable efforts or commercially reasonable efforts. Nobody in practice really has time to track all the client alerts and all the cases that come down not just in America, in Canada, England, that some American courts cite. So that actually, you're sort of reading my mind. That is one of the things we're working on with contracts. Ambiguity and lists, you know, many of them have reached the Supreme Court in statutes and contracts, you know, comma, issues with commas or lack of commas. There are a lot of things that AI can help you spot in a contract beyond just things like, hey, is this market or not market? I mean, that's important too, but that's not my thing. That's more for like a pure like corporate product to tell you, you know, whether your liability cap is too high or too low. At the language level, if you think about it, most contracts never get read again, right? They're not really, unless you run out of money and you're looking for a reason to sue for breach. But when they do, when they do, people are often trying to seize on some language that gives them, you know, an exit ramp. And I'm not embarrassed to admit, I found this actually really, really exciting to see how AI can kind of understand all that case law as a unit and spot problems in draft agreements.

Adam: I want to wrap up with a few final points. So I'm going to pose a thought process to you on legal writing. I'm just curious your take. So when I think of music, there's all types of genres that they all can sound good to the listener and each one can be good. Do you think that applies to legal writing or do you think that sound legal writing is sound legal writing?

Ross: Yeah, I mean, I think if you're really gifted, you can have a unique style, like a certain jazz saxophonist might be identifiable or like Wynton Marsalis playing the trumpet and you're like, oh, that's Marsalis. I think a lot of lawyers think they have those kinds of gifts and they actually create problems for themselves they don't need because they try to sound like a famous judge or lawyer out there and it feels very kind of fake or contrived. I actually think, I know it's not what people probably want to hear, there is sort of the right way to write in legal writing. I mean, again, obviously there's some wiggle room and I'm not saying people don't have their special snowflake styles, but I'm saying there is kind of a way of writing that you have clarity, you have conciseness, you have some variety in sentence structure, but you're not obsessing about it the way you would if you were writing a literary novel. You have smooth transitions between your sentences and paragraphs. And this is something that I said from the beginning when I first, even before I wrote my books is that really the problem is people treat it like an art and that makes everyone anxious. It really is just, it should be more of a science. Again, we're not expressing ourselves, we're not writing memoirs, right? Most people really don't need to have a distinctive style unless they're just trying to impress their spouse or their parents or something. There is a sort of, I don't wanna say garden variety because it makes it sound easy, but there is a sort of ideal way to write at least a prose document. I think contracts gets a little bit more complicated. You're not really trying to win a prose award or a style award in contracts, you're trying to avoid trouble, but in your traditional legal documents, motions, briefs, letters, court opinions, there is one main way to do it right.

Adam: Last question I have is when you get to the point as the drafter and you need to leave an argument on the cutting room floor, but it's very hard because it's your argument and you love it. How are you seeing AI help you get to the point of leaving something on the cutting room floor?

Ross: Well, there's always the footnote compromise, right? You put it in a footnote, you think maybe nobody will read it but at least it doesn't go to waste and people do do that, right? You can tell, you can tell somebody spent a lot of time researching, it was like a rabbit hole and they didn't want it to go to waste, so it's in a footnote. But I think actually in that case, that is a very, I haven't done it, so I wish I could tell you, but that'd be a very interesting use case. You could say, look, this is an argument that I personally found promising or at least intriguing. Is there any way you can see to get it into the body of the main argument section in a way that's not totally contrived? You might be able to help you find a compromise, right? Or it could serve as a repository for your great ideas that just need a home. And then maybe another day, another motion, another brief, it says, hey, remember that really clever thing you wrote but you decided didn't quite fit? Maybe you should give it a go now.

Adam: Well, Ross, I appreciate your time. And this is Adam Massaro for Disputes and Perspective. Ross, it's a pleasure having you on. I can't wait to see what BriefCatch comes out with next. And tell the audience, you mentioned on Point Made you're coming up with another edition. When will that be out?

Ross: I gotta ask Oxford University Press. I think they're gonna try to expedite it. I would guess if I remember correctly, let's say like six months. But we'll be distributing previews and sneak peeks on social media for sure.

Adam: All right, well, we'll look out for that. Ross, thank you for having, thank you for joining the show and we look forward to see what you come up with next.

Ross: Thanks so much for having me. I really enjoyed it.

Outro: Disputes and Perspective is a Reed Smith production. This podcast was produced by Shannon Ryan and edited by Julian Baughman.

For more information about Reed Smith's litigation and dispute resolution practice, please email [email protected].

You can find our podcast on podcast streaming platforms, reedsmith.com and our social media accounts at ReedSmithLLP.

Disclaimer: This podcast is provided for educational purposes. It does not constitute legal advice and is not intended to establish an attorney client relationship, nor is it intended to suggest or establish standards of care applicable to particular lawyers in any given situation. Prior results do not guarantee a similar outcome. Any views, opinions or comments made by any external guest speaker are not to be attributed to Reed Smith LLP or its individual lawyers.

All rights reserved. 

Transcript is auto-generated.