Latest / Brains & Bots / Will AI Replace Lawyers? Bilal Sayyed on the Future of Legal Careers
Transcript
- speaker-1: Hello, welcome to Brains and Bots. It's a podcast brought to you by the National Association of Scholars. I'm Lila Nourakish, your host for today. We are going to talk about the future of legal education today, which is a topic that's very close to my heart, particularly because I'm a lawyer by training, and I'm also very curious what the technology is bringing to our life and profession. Our guest for today is Bila Sayed, who is an antitrust lawyer at the Hogan Loves and Kadvalader, as well as he is an adjunct faculty member of Scalia Law School and Arizona State University. He teaches antitrust law as well as cybersecurity and European competition law. Formerly, he led the Federal Trade Commission's Office of Policy Planning as director. Let's bring in our guest. Bilal, welcome to Brains and Bots. speaker-0: All right, Lilla, thanks thanks for having me. I'm excited. Excited to talk about this topic. speaker-1: Yes, me too. I just told in earlier to our audience that I'm super excited because as a lawyer also enthusiastic about technology, you know, seeing what AI is doing with our profession is pretty impressive. So what would you tell to a student who is worried about their future? Is AI going to take away our jobs? speaker-0: yeah, I would I would tell them to be worried about it. it doesn't necessarily mean it's gonna take away young lawyers' jobs or senior lawyers' jobs. But you know, the the technology is getting to a point I think it is at a at a point where it clearly can do better work than first year associate. and so students in law school right now that may be graduating in a year or two or three, should realize they're stepping into a a a world, you know, that's much different than even three three years ago. you know, clients, clients of the firm, many clients, need all clients, many clients expect us to use AI for drafting not only of simple documents, but even first first drafts of, you know, what you might call complex documents. And so, you know, the basic research skills, the basic summarizing skills that often form the backbone of the work of a junior lawyer, that that stuff's going away, right? I mean, I used to I used to myself s participate in meetings and take notes, you know, phone calls or meetings. I used to have people do that when I was more senior. Well, that's going away. You don't need someone to take notes on a on a, you know, Zoom call or Teams call that can do it for you. you You don't need a young lawyer to summarize documents, to summarize cases. y you've move much more quickly into the thinking and analyzing phase of a legal problem than than you did before. And you know, the models are sophisticated enough to identify patterns that I think in cases that that lawyers have a tough time doing. and even if you can do it, it takes it takes time. So now none of this means all legal work is gonna be handled by models and and machines and such. but it's a much, much different environment, as I said than even three years ago and frankly, even Even one year ago, it's a much different environment. speaker-1: Does it mean that firms are hiring less junior associates? speaker-0: I don't think firms short answer is I don't know the answer to that question, right? at least systematically. But I I I don't think firms have as yet figured out how much work will be you know sort of machine driven. where where in the in the stack of attorneys that work, you know, will will take the biggest the th the machine drafting and research will will sort of take the biggest bite. but firms I think are thinking about that, and I think it will lead to less hiring, a different type of hiring, and greater greater expectations, right? I mean look some of the skills you know, some of the skills of analyzing cases, putting together first drafts of briefs Those are not really the province of junior associates. Those are things that, you know, senior associates do. And those th that that workflow is at risk is at risk as well. speaker-1: I see. Where do you you know, if if AI is going to change and I it already did change the basic research and the fabrication of first draft, then where can, you know, new opportunities arise for fresh graduates in the legal field? speaker-0: Well, I think there are a few places. one of course is just to get better at at analyzing cases yourself, right? get better at it more quickly. learn how to leverage, you know, what what is a very super smart companion to your to your benefit. I find most documents I produce using AI require I I must say I'd say substantial revision for points that you know maybe the lawyer maybe a lawyer believes are superior to to then what's generated. So you know, there's still a role for expert practitioners and young people just have to become expert more quickly. And of course the the AI tools c can help you do that. Can help you do that, right? This this is not all that it's not all a competition between young lawyers or senior lawyers and and you know machine machines models, you know, AI, GAI. Course in other places learning how to use the technology in a way that's much more efficient than you know, somebody just taking stabs and iterating to you know, to a final work product. I think that does require still learning the law, learning what in a sense what prompts, what questions to to to prompt research with or drafting with, explaining you know, I mean law is advocacy as well as research and and explaining your your advocacy and your position in such a way that you know the the first drafts or fifth draft are are helpful. And you know it's it is well I think and and so those are areas, you know, learning the technology is important. When I when I was a student, I was a student nin nineteen ninety three to nineteen ninety seven, they were still teaching us to use, what is it, the West Key Digest system. I never used that w in practice, because we had Westlaw and Lexis, right, which had their own sort of systems. But it was useful to understand how to use the tools that, you know, thirty years ago existed in Westlaw or Lexis. And I think that's the same it will be the same here, right? Learning how to use the tools you have, make maximum use of them. learning how to in a sense create your own agents to to to do to do research for you or, you know, other other other things they can do. So that that, you know, and and that's a different different education as well. I mean, that that becomes a more practical hands on requirement. it requires a willingness to be sort of innovative and among the schools, it requires the schools to be innovative, right? And to move away from, you know, no, we're gonna And obviously they don't do this, but everything's gotta be done in pencil and paper, right? obviously that's not what anyone does. But law schools have to figure out how to how to both teach the law and prepare law students for a career in law, which are two different things. Of course, that's always what they've had to balance, right? Teaching the law and preparing somebody to be a practitioner. but I think, this is a significant change. and, you know, so far I think most schools are struggling struggling with it. And and that's not really a criticism, because it is very easy if you allow students to use AI to create work product. you really don't have to learn the material. You can in a sense purchase high quality plan, right? And you will absolutely get better output than you would do on your own as a first year. but it may still be it may still be wrong, incomplete. and you know if you haven't learned much about the law, you you may not catch it. You may not catch it. speaker-1: So does it mean that law schools, besides teaching the rulers and the basic lawyering skills, now also have to teach some technological skills, you know, that are di very different than searching in databases like Westlaw or other data sets. And I'm asking you as a as a practitioner who also teaches, because you know, you have you have seen both sides as well. speaker-0: Yeah. So I you know, I don't I don't know that the schools ever taught you know well okay, I I should be careful. I don't know that I ever learned in school how to use Westlaw and Lexus as as examples to their not only not to their fullest potential, but anywhere near their fullest potential. And so I don't think the law schools as presently construct or staffed are in a position to teach students how to use either you know you know sort of general models or law specific models, right? I think they do need to do that, but they're gonna have to get l likely gonna have to get different people to do it, and they're gonna have to struggle through developing a curriculum to do it, right? Because it is not it is not just it is not just well you you have to apply both and you have to you have to a student has to see I think and a as do instructors have to see how much how much better their work product can be using you know GAI you know various models and such and that is you know that's a that's almost the process of taking a a paper, taking a memo, and just iterating to a final work product. And you know, most law schools are not well, most law students are not prepared to spend, you know, I think five weeks on a single memo figuring out how to improve how to improve it from an early draft, either con construct it on their own or you know with with with with help from from a from a tool or a model. So I you know, I this was a failure on my part in two respects. to the extent I learned how to use tools like Westlawn Lexus, it was from the representatives of those companies. and you know a as a as an instructor think I failed and have failed in in providing opportunities for students to learn how to use again Lexis or Westlaw in the in in the dis you know in the discipline I teach, right? It I would have liked to have done a a much more project based work with a representative you know, drafting a memo, drafting a complaint, drafting a brief, using at you know, what were at one point relatively as compared to now, rudimentary tools and now, you know, both both again, this is not an ad for Lexis or Westlaw, but you know, both of them have AI built into them, right? And you need you know, students need to spend And instructors need to spend a lot of time working with those tools, I think, in their classroom preparation, in their in their work product. you know, I I've tried to move away from exams in my classes because exams I mean, you don't get an exam when you're a practicing lawyer, right? Nobody that's you know, no client comes to you with a series of random facts and says you have three hours to just tell me everything you remember about, you know, what I told you a few months, you over a course of a few months. I've tried to move towards and I don't like the idea of writing papers because no junior associate writes a 30 page paper. well, almost no junior associate writes a 30 page paper. You know, I tried to move towards writing whatever, five page memos on a particular question. And that's what you're gonna do as a lawyer. Maybe it's in an email, maybe it's not really a memo, but you know, it's the same idea, right? Convey convey information in a relatively short form. well, if you if you cannot if you do not know how to use AI to start that process, to do the research, to do some of the drafting, you're gonna be well behind your your better prepared or differently prepared colleagues. So you know, I may have lost the thread of the question here, but but schools I I I doubt that schools are prepared to prepare their students for what they're gonna face next September if they're graduating. And I doubt The students who are starting jobs, you know, now are prepared for what what will be expected of them, not you know, be because clients are expecting it, right? We've already in the I've been practicing about thirty years I think in the what is it, in the last ten or fifteen years I've joked maybe it's true. I won't I won't I I think it is true. Sometimes it's hard to get clients to pay for the work of first year associates. Now that may be because the rates are high, d depending on, you know, w what sort of type of law firm you're at. But it could be 'cause the rates are high and they're quite they they can be inefficient. And n none of this is a criticism of them, right? it's your first day on the job, your first year on the job. now there is a clear alternative to a first-year associate. if I want a memo on predatory pricing, if I needed a memo or an email on predatory pricing, as an example, as you know, I'm an antitrust lawyer. you know, I have no reason to go to a first-year or second-year associate and say, I know the law on predatory pricing. I don't have time to write a memo. You go look at the cases and write a memo, right? I can type in a short prompt and either you know, a sort of a a a a product developed for a law firm can just draw an existing work product and create that memo, right? Or I can say, go go out in the wild and write me a memo on, you know, Predatory pricing, right? You don't have any client information there, you just tell whatever, ChatGPT or one of the anthropic models. I need a memo on on predatory pricing, or I need a memo on the ninth how the Ninth Circuit treats predatory pricing. That that can be done you know, in fifteen minutes. I'd have to check it. I might, you know, have some questions. But that that that's a day's work for an associate that can be done in a half hour. speaker-1: Exactly. And I think it puts law schools and also graduates in a very hard spot because they not only have to learn the rules of the profession but also have to understand how to become an expert so much quicker and faster and provide those answers you are talking about. And I don't I I agree with you that the the schools are not doing a very good job to prepare the next generation of lawyers. At I wanted to ask you, do you teach differently since we have AI? You mentioned that you you asked for memos instead of papers and regular exams, but how do you prepare for classes? How do you you know, allow students or not allow students to use AI? What what changed for you? speaker-0: So a number of things and I'm still I'm still working through it my myself. but I do want to say that the the tools can help people learn more quickly, right? I d I don't think this will necessarily lead to over time, I don't think it'll necessarily lead to less work. I think what it's gonna lead to is an expectation of of better work, right? I mean, you I'm no expert on these, you know. mathematical problems that have, you know, bedeviled the world of mathematicians for a hundred years. But right, it was it was OpenAI that dumped about five hundred solutions to, you know, insolvable math problems the other day. And what does that mean? Does it mean all math problems go away? No, it means we now we now have a colleague who, you know, we We we can start thinking about harder things, right? Different things, harder things. Now, maybe, you know, maybe it's a little bit of a race that AI will keep it up. So how do I teach differently? First, I am I'm trying to learn how to use AI to produce material for class, right? I am trying to learn how to how to take new cases, long cases, and basically excerpt them down for students to read in the same way that it took I would say multiple iterations of textbooks to iterate down to you know the case excerpts we read in in case books, right? So that's one way. Another way is I was never this is a silly example. I was never very good at at diagramming things, you know, diagramming legal problems or diagramming diagramming sort of factual connections. And I I use AI to to create diagrams, right? To illustrate relationships that I was otherwise expressing, you know, orally and in in words. I find it very helpful. I I use AI to help me think through the assignments I'm gonna give to students. What have we read? What should I, you know, can let's craft some let's craft some assignments that draw on what we've read, right? and and see how the students apply, you know, what what we've learned, right? I mean I bef before I moved away from giving exams. I I used AI to generate my exams and what I would do is take either my syllabus or the cases, drop them into a model, and say, Okay, this is what the students have read, hopefully. I need three problems and the problems should focus on this, that, or the other thing and create a fact pattern that's, you know, not too easy, not too hard, whatever, right? you know, now these are all sort of simple things. you know in a way they're simple. Most of teaching is sort of getting the simple things right. but you know I I try to use it to identify differences across cases that you know I might have students read, that maybe I can't maybe I wouldn't find, right? so it's it it you know, it's it's all intended to be done in a way to make me a better better teacher, you know. I I've always been I've I'm I'm somewhat critical in the in the writing I you know in the written work product I get from younger lawyers. Again, it's not meant to be a criticism of them. It's hard to write. It's hard to write well if you don't actually know, you know, the law, right? What's important, what's not important. I've taken not my client related work. but other other things I've written and tried to identify you know what's what's a s sort of successful structure to my to you know to a to a written piece of written work product, right? now I know you don't you I think I think schools try to do that in their legal research and writing classes, but I'm not sure they give students particularly good instruction. And And I think in particular what seems to be lacking is the ability to structure a memo, structure an argument. And I find that, you know, by and and I think firms will find this using, you know, what I'll call proprietary information, proprietary work product, but in my case, just using my writing, using the writing of others effective writers. How how do I structure an argument, right? What guidance can I give about a an effective argument? and and so I use AI to help me to help me do that. To help me do that. Sometimes it works, sometimes it doesn't. but I think it it it identifies things that I hadn't identified, or maybe it identifies them more quickly. it looks across You I don't I don't just feed it my work, I'll feed it other people's work. I'd say, It would be great if all law students here's here's my view. It would be great if all law students could learn to write like, you know, Chief Justice Roberts, right? I think he's a very good writer. so why not take, you know, ten opinions by the Chief Justice and rather than reading them, and I mean you should read them, but put put Put them into, Chat GPT or Ligora or whatever tool and say, show me the commonalities here, right? And I think that's can be pretty effective. Now you should have your own you should be able to, you know, press press press on what comes out. you know, you should you should know yourself what good writing looks like. But I I I think that's I think that's effective. And I think it's more effective honestly than having a second year or third year, you know, law student grade the papers of first year law students and give comments. I mean, again, none of this is meant to be n you know, I mean it is being critical, but none of it, you know, it it it th these tools are so much more powerful that you know they they really can make the teaching easier and clearer, you know. And I th I'm I'm still learning how to do it. I mean we're you know, you're always learning, but but I think I think it's gonna be much, much more it'll lead to much more effective teaching if people are open to it. speaker-1: Yeah, that's a great point actually. And I I noticed something when it comes to my work as a researcher. I oftentimes drop in the main arguments I'm working on and I ask AI to criticize it, to tell my shortcomings and my blind spots and you know the strongest counter arguments so I can prepare, you know, what may be the criticism when an article is published. So I think it's really useful and it's kind of the same thing, just in in another field. speaker-0: Yeah, I I look, I think everything we write we should put to a test. And you know, you may have done this. you can write you can write yourself a prompt, or you can have, you know, ChatGPT write a prompt that basically says, you know, review this as a hostile reviewer, review this as, you know, the other s the advocate for the other side. and you know, you y you can write something that you think is very good and then and then you put it to this test and you're like, my god, you know, I'm never gonna and and and you can just do it over and over. And you know, you have to use judgment. You have to use judgment in what you react to, but it it it can be quite effective at highlighting inconsistencies or mis cases that you didn't cite, maybe you didn't cite them for a reason, but but you you can get feedback on on the limit you know, the limit you know, the f your failure to actually address an issue in some other case. So Yeah. No, it's it's speaker-1: important. What is the line between, you know, using AI to learn and become better through it and avoid it and just do everything in the old school way? So is there a balance between the two? speaker-0: Well, that's a tricky question. You know, I think I I I don't think it's that you should only use AI after your first year of law school or, you know, th as an example. Students who are lazy and I'm I was a lazy student, may just use the technology to skate by. But if you use the technology and think about the output you get, test the output, in a sense, argue with your AI companion, you can get better right away, right? So it's it's it's the engagement with the material that that helps you learn, right? It's so I I've struggled with this. I in one class I taught, I said you can use AI start to finish on your memos. and then in another class I said, I don't want you to use AI because you don't know anything about this topic. And that's n you know, to that that's not a good response. The right response is to figure out how to make students engage with the work product from AI, right? So if somebody, if somebody, I haven't done this yet, but I will. You know, you can write your memo using AI, but then I will surprise call on you and you need to come and explain that memo to me. And if you can't explain it, you know, you've done the assignment, so you'll get whatever grade you get, but I will know I will not come to you with my legal problem, right? Because you don't actually know the answer. So I think that that I think is is maybe a way to think about it. You know, I'd heard that some schools were moving to oral exams, right? So people so student you know, so students couldn't I I hate to say it this way, but you know, cheat, right? I think the better ideas. Maybe the better ideas they need to learn how to use it, they need to engage with it, they need to explain it, you know, and and that and and that requirement that they come unprompted, no notes, and answer the question may be a way to see if they did engage with the material and didn't just you know, do a few iterations of a of a of a memo through either a, you know, proprietary system or an open system or whatever. So I th there is no there is no balance. I mean, I didn't don't I don't balance my use of a spreadsheet anymore. I use a spreadsheet to calculate things. I don't I don't think, well maybe today I should use a pencil and paper, right? Or bring out my abacus. or even bring out my, Texas instruments calculator. If I have to calculate something, I use a spreadsheet. if I have to do research, I I use Lexus and Weslaw or or comparable products and and I'll use AI, you know. I don't I don't I don't balance anymore. I just engage with the material. I don't take work product. Well that's nice. Pass it on. I engage with it. And that's what students have to do. Banning them from using AI. I understand the the instinct there. but it's just it it's just silly. You know, it it's absolutely silly. have to learn how to use it. And and and you can't learn how to use it if you aren't allowed to use it. speaker-1: our first episode was with Jared Horvet, who is a neuroscientist, and he pointed it out that AI is an expert tool and in order to be an expert you have to have basic knowledge of your field. So you can't really use it if you just expect it to, you know, draft answers for you. I mean it's going to draft answers for you and as you mentioned, it's probably going to be much better than what you would come up as a freshly graduated lawyer. But if you can decide whether that answer is accurate or not, or you know, how to prompt as another thing you you mentioned, then it's probably going to be pretty obvious that it was not done by you, but by the technology. and I think it it might be a different challenge for universities in this ban or not to ban or you know, full green light versus red flags when it comes to AI. So how should professors assess the students' knowledge? You mentioned that some professors are moving to oral exams. I'm coming from a country where all the law law exams in in law school were oral exams, except for the statistics and the economics ones. And it was quite good because, you know, we needed to learn how to speak, how to talk comprehensively. But I also saw the other side as a professor that it's a lot of work to sit there all day and interview all the students and it's it's really a lot of time and effort. And and anyway, so the question here is how to keep academic integrity, you know, make it sure that the students have the knowledge, but also develop the skills necessary in this new world we live in. speaker-0: Well again, it's a good question and I think teaching may need to change. I mean the fact that the fact that it's a lot of work for an ac you know, for a professor is not a reason not to do it. Maybe a reason either to change your teaching or change or get more professors, right? y but I I guess I it is It is entirely possible to learn all the material that you learn in law school without going to class, right? Here's a book. I mean, what do we what do we do in class? What do we do in law school? Here's a book, or here's some cases, read the cases, I will ask you questions about them, or I will lecture about them. None of that strikes me as very efficient. Right? I I the textbooks are incredibly in good textbooks are very good, very high quality, right? I cannot cover everything that's in the textbook in a class or in, you know, in a semester or two semesters. If somebody actually reads the book and understands what's in the book, they don't, you know, they don't need to come to class. Now they come to class 'cause there's attendance requirements. Anyway, what does this mean? It means a lot of what professors do is probably not necessary, right? in that, you know, read the case, read five here are five cases, here is a problem. Write me a memo, and then come explain it to me. Right? And so it's a little bit less of a lecture type approach, more of use the tools you have, right, to to learn the material, to engage with the material. And then professors evaluate your knowledge in a way that's a little bit different than, you know, cold calling on people in class, whatever, eight times a semester. Right? I mean the you know, I I I think students learn best by engaging with the material, engaging, let's say in the class with the professor or reacting to what their classmates have said about something, right? Not listening to me lecture or anyone lecture. I mean, there's some very good lectures that I probably pay to listen to and all. But you know, I think w I'm an adjunct, right? I don't know how much effort I will make o outside of the fact that I enjoy this and I'm thinking about it, I don't know how much effort I would make to, you know, become a better teacher in the AI age, right? But if I'm a full professor and want to make you know, be a full professor, I'd start thinking about how do I how do I change to to this to this change in change in technology, change in tools, change in expectations. And you gotta let go of some of the things you're doing and step up new things. And I would I would try to get much more engagement, much more. You gotta come to class having read the material, right? You don't need me to summarize the material anymore, right? You can create a summary yourself, right? And you ought to engage with that. You ought to read the cases, you ought to engage with whatever material machine or model creates for you. But I'm here to a more to evaluate now. Did you get it right? Do you understand it? I don't mean that in a you get an A, you get an F type thing. But y you know, you need to have done the work. It's a lot easier now. I mean it was always kind of easy. You just read. I mean it's not the most strenuous of, you know, jobs being a student. can be stressful, but it's not strenuous. come to class and I'm gonna assume you've read everything. Answer my questions. Tell me, right? Do you you know, let me see if understand it. Not only you, but everybody. And that's they have to do way more of that than get up there and use the same notes they've been using for five years or ten years, maybe update it a little bit because, you know, there's a one or two cases every year you wanna work in. But it requires a much more active type of teaching than than, you know, somebody might be used to or than or that's what's expected, right? We sometimes we pick schools because they have great lecturers. Well honestly you know, is anybody you know, the the level of you you can get a lot now from all these tools that you you use. what you need is somebody to tell you and and you people may feel more comfortable with this, having you know, having an a having a professor tell you, yes, you do understand it, or I don't think you understand it, right? And you can of course you can have the machine tell you that too. You can say, grade my paper, whatever. And you're like, I missed a lot of things. Anyway, so lot, lot more work for It it it you know, this this is I'm not I'm not really a chumperian, right? I don't think that's real economics. I think it's ideology, right? But I mean this is creative destruction, right? This is creative destruction and people need to adapt or they will be swept away. I you you may have more questions, but I wanted to make sure I made one tell one story. You have to be an idiot to believe that AI cannot do the work of a junior associate because law firms large law firms are spending hundreds of millions of dollars on their technology. And you know, large law firms are basically collections of lawyers who basically evaluate the work. of their, new lawyers, younger lawyers, colleagues, whatever. And they're basically saying, we're putting our money on the on the you know on the machines, on the models, on the technology, on the equations, whatever. and if, you know, you want to step into a world where, you know, you believe that that that This is not going to replace young lawyers, you just are mistaken, right? Now, you know, right now nobody's wheeling in the computer and, you know, Chat GPT is making arguments to the to the bench or to the jury. You know, so I think if I were a law student, I'd think about what is it that, you know, the machine, the model can't do, right? Even if they write me the script for my opening argument or closing argument, I better learn how to deliver that in a convincing way, right? So I don't know, maybe they should put I why not take a few acting classes? I had a co I had a few colleagues who took acting classes. They were trial lawyers and they took acting classes because it was, you know, how do I convey this story, right? Well, I think that's gonna matter more here. because now it is you know, where is it that law is more personalized? You know, is it in the writing of a memo or in the oral briefing of a board or a jury, right? Well, it's I think right now it's those areas and those are those are things students should should work on. And it's more than it's more than the type of advocacy, I believe, training you're getting at a law school. I absolutely believe law students should take act it should take acting classes because you know you you want to learn from somebody who conveys emotion, right? the importance of this story that makes you feel something. So anyway, always think think think a little differently. Think about, you know, break down everything a lawyer does and think about where is the technology gonna swamp me and where is the human element and how do I maximize the the strength of that, you know, strength of my ability in that human space. Anyway, okay. I can't remember speaker-1: touching up on this topic, my question here is where would you use AI in the law firm world and which are or I ask the opposite. The question is where would you, not use AI? Is there any part of the workflow that you want to keep for yourself? speaker-0: Well, look, I think it's all myself, even if look, if I give a junior associate an assignment to write a memo, I I don't feel diminished because of that, you know. So it's I I in some ways I don't want to keep any work for myself. I want I want to you know, in a sense, get the right outcome for the client at the, you know, fastest quickest your quickest, cheapest, best way. And so you use all sorts of tools to do that. I think if if what I really liked is doing the research, you know, s and not the advocacy, oral advocacy, the and I'm not even talking about the business development side of this, but you know, the the advocacy before a court, advocacy, before a judge or jury, advocacy before a board, or not even advocacy before a board, but conveying a conclusion, an analysis, then I'd feel a little nervous, right? Because then I'd have to move into something I didn't like to do. But but if you do like that, or even if you don't like it, that's what you ought to get good at, right? I mean you should also get good at using the tools. But least right now, we're not rolling in you know, an Alexa or whatever, to make arguments to the court, right? People want to see a human, you know. And you know, as long as a when a judge is a human and the juries are humans and your board is humans, they probably want to see a human being convey, you know, whatever the information is or that that that that they need to take in to make a decision. So I guess that's the stuff I'd want to keep for myself, but I don't really mean it that way. I mean that's that's where everyone has to develop, you know, to a greater degree than they did before. I mean, it used to be it was, you know, you're a tax lawyer, I don't you're one of X number of people who know the tax code, you know, from A to Z or from page one to a billion. Now I you know, that the stuff you're looking for can be found you know, by a program, right? So look, there's still a benefit to knowing things, a lot of benefit. But you know, figure out where AI cannot be used and and in a sense that's what I I keep that for myself or people keep that for themselves because that's where the tool is not effective, right? Go with the most effective tool. And sometimes the person is the most effective tool, right? So that's what you keep for yourself. speaker-1: Talking about skills, there are a lot of skills that are underestimated in today's age. And you mentioned taking acting classes, but also just being able to deliver an oral presentation is really a skill. And during law school, as I mentioned, all the exams were oral exams, preparing us for the future, you know, standing in a courtroom and convincing the judge and everything. And then I graduated, I started to work for a law firm. And for two and a half years I felt like a droid because I was sitting in front of the computer implementing d data and changing stuff in existing contracts and I felt like, okay, well I'm not talking to anyone here. So why so the profession the the rules of the professional somewhat changed of course because there are so many settlements, so many you know, so many procedures that are not taking in person, assistance at all, but you know, happening through forms and and different, you know, programs and so so but there are still litigation, there are still negotiation where your other human skills and emotional intelligence and so on and so forth are needed. So we still need to be human, right? speaker-0: Yeah, yes, yes. And you know, the point of my reference to the acting classes was to make someone a better oral briefer, right? I mean he and and it's not that look you can watch you c you can watch a TV show on law, right? You get about at best the thirty seconds, maybe fifteen seconds of a closing argument and it sounds so great. Well Nobody does a 15 second or 30 second closing argument, right? So you're not taking acting to learn how to do a 15 or 30 second closing argument. You're learning how to how to how to reach, you know, an audience, right? You still have to bring your legal skills to the actual argument, right? The you know, no no writer is writing you a 20 minute closing argument that pulls in all the evidence. It's just you've got to you use you use that type of talent, you know, to reach to reach people. So yeah, it it it it that there's what what is the there's a book by the the what is it? The Messy Jobs. I think it's called Messy Jobs by a European economist, which basically says, You basically want to do jobs that require if you're worried about your, you know, future, employment future, you want to do jobs where the human element is a necessary element. and I think that's true in law, right? Because law is not just here's the memo you wanted, right? It's it's explaining things. to to judges, to juries, to boards, to, you know, some large, you know, differ different types of clients. And it's sometimes a memo is all you need, but often it's not. Often it's not. So that's the human element. So I think that's r you know, for a while, I I don't think that human element will be overtaken. Now I've seen some cool robots doing stuff I couldn't imagine they could do and you know, maybe people will accept oral argument from a robot, but I think that's a ways away, you know. But it seems like it's coming. speaker-1: I would love to see Swiss CPO in a courtroom, maybe at the Scotus, it would be fun to see that argument. speaker-0: It's that would be very interesting to take a couple of a a robot lawyer, let's say, and you know, put them through one of these, you know, practice jury things that we sometimes do before litigation, right? To evaluate the strength of our case and how how s some other twelve strangers will react. Just take take take a robot is probably an unsophisticated term, but take a robot that can move like a human, and then maybe you can put facial, you know, facial expressions on and see how they do in front of a jury, making you know, convincing a jury of their you know, a as compared to a human, right? Which which one is more convincing and if it's a robot, then I think the legal profession has to think about a lot of things, you know. speaker-1: I would love to see that experiment if if there's any. you mentioned that clients expect law firms to use AI and you know, I wanted to ask what does it mean? And when you know the assignments are getting faster and also, you know, more probably more efficient or better, what does it mean for from a billing perspective or fees perspective? Does it save money to the clients or it doesn't change that on their end? speaker-0: Well, the short answer is I don't know, right? When I I've come back to practice in the last three or three or so years, with the hope that I never have to deal with client billing issues again. you know, because of course I love my clients and and all, but you know, the administrative part of legal practice, including bills, is often not often always the least fun aspect of it. what does it mean? Look, they it the the hourly fee structure, you know, is probably gonna change. I mean it it ha it has changed, it is changing, you know. There are all sorts of alternative fee arrangements that law firms do, whether they're small or big. but I think it's just gonna continue to change and you know in some ways I mean the the again, the amount of money law firms are spending on this technology is only possible because at least large law firms, you know, have have the money to do that. and at some level somebody does have to pay for it, you know, eventually, right? So I I'm not I'm not sure how firms will how billing rates will change over time, but you know the marginal cost of a A brief, even if you go through a number of iterations, is gonna be quite low as compared to using, you know, l you know, human lawyers to do it. And that's gotta affect the the the the the billing. on the other hand, you know, it's it's always been a bit tricky to charge people by the input, right? The amount of hours you spend on something rather than the output, right? The quality, whether you win or lost. And, you know, as as it becomes harder and harder to in a sense, bill by the hour, i it it it's just gonna have to change, right? But somebody has to cover these these costs. mean, you probably saw maybe I circulated it, because it was, I think, in the spring of this year, Kirkland and Ellis, you know, said they were gonna spend five hundred million to develop their own proprietary AI. Now that sounds like a lot of money. I mean it is a lot of money. but they're not the only ones spending that, right? And the f the firm I'm at, Hogan Lovell's Ked Walleter, they say publicly that that they spend about five percent of their annual revenue on technology. Well, you can look up the revenue of this firm. It's about three billion dollars. And without a calculator or spreadsheet, five percent of three billion is a hundred and fifty billion. Right? So, you know, that's not all AI, but it it's you know, pe people are spending money on this. So it's you know, it it's it's there may be a bubble in, AI stock prices, I don't know, but this technology is pretty powerful and it it's not you know, no one's gonna stuff it back in the bottle. speaker-1: Yeah, that's that's a great point actually. And I wanted to ask you, since I know that you also teach cybersecurity, that the reason probably that these firms are developing their own AI is because you can't just put client data in ChatGPT or Cloud or any you know, system that is also used by others. but are there any other safeguards law firms and lawyers should follow when it comes to using AI? speaker-0: Well, I'm sure there are a lot. And you know, I think one reason my the work I the I use AI mostly for you know what I would call non-confidential, you know, where I can use public information, is to, you know, avoid limit those types of questions, right? What is the risk of sharing client information with, you know, the cloud? Right. Now all our stuff is in the cloud, right? but, you know, the same you know, the same protections law firms use with respect to just general data security, you know, preventing cyber attacks and you know, hacking, you know, will apply to whatever AI system we use, right? I mean, if we may not I'm I'm not commenting on the firm at all. A firm may use something like Lagora or Harvey, which are, I guess closed models. Well, but but the communication is still leaving the firm, you know, to the cloud. And and all general, you know, data security practices have to be followed. When you are sharing client information confidential business information, trade secrets, whatever, financial records, whatever. you have to be in a sense even more secure. And so you know, you have to have the clo first, you have to have the client's approval. At least you should have the client's approval before you use AI, right? Because that that's just a a of a potential new vulnerable vulnerable point. you have to think about do you mix client work product, right? So some law firms will take their existing I'm I'm just this is let's say it's just hypothetical, right? Because you may you may want to use all the work product you've developed over however far back your electronic files go and maybe you'll scan in your paper files. Who knows, right? But there's a lot of good intellectual thought. that goes into the work that law firms do. And do you do you take the work for one client or any client and mix it in some, you know, aggregated, body of all your work product and then use that to generate work, you know, to generate advice to to somebody new. Or do you keep everything segregated and, you know, you done a lot of work for Company X for fifty years. They'd say, look We expect you to make use of the work you've done over the last fifty years and not reinvent the wheel when we ask a c a question, you know, that that might otherwise allow you to reinvent the wheel. so there's that, you know, how do you treat client material? if you I mean, you you're not supposed to mix client material. How how have we thought about that, you know, in the context of these you know, internal proprietary models? So there's a lot. There's a lot. to think to think about, some of which you know, some of which can lessen the utility of your existing, you know, fifty years of well, let's say thirty years of electronic, you know, files, right? So those are those are things. I I think I think outside of the cyber security issues, probably the ethical issues are are significant, right? The conflicts issues, the conflicts issues. and you know, people are thinking about them and you know, on on conflicts, particularly mixing, pe people are very cautious, you know. So but yeah. speaker-1: You specialize in antitrust. if you were starting law school today, would you stick with antitrust or would you find something else more interesting? speaker-0: Well, you know, the the sad thing is I don't know of anything more interesting in the law. So after thirty years, I'd probably say since I've been happy doing antitrust, I'd probably say do antitrust again. I mean I I think there are there are other parts of law many areas of law are interesting, right? but of course if if I picked up an IP problem today, It probably would be very interesting, but I'd be very frustrated, right? And I wouldn't put enough time into it to get to the interesting part, right? So for me, antitrust is good for I'd say two at least two reasons. One, you know, can do you can have antitrust can involve you in transactions, it can involve you in litigation, involves you in investigation, and involves you in just counseling. You know, it It touches on other areas of law, because they may impact the antitrust laws. So I can learn a little bit of IP law, I can learn you know, health you know regulatory law. so that's one reason, right? It it's broad. You can have a broad practice, right? We have people in some law firms, you know, they have a litigation group, they got a corporate group, they got an investigations group. Well, you know Antitrust touches on all of them, right? So I like the breadth. I like that it is it is a universal sort of discipline or application is universal across different types of businesses, right? Most b even industries that are regulated, face you know, have have antitrust issues or or haven't fully displaced antitrust, right? So I learn about all different types of businesses, why they do what they do, right? So if you like economics and, you know, how goods get to market type thing, well, antitrust is is great for that. So, you know, can I'm sure my friends who have done securities law for 30 years would would say securities law is great for various reasons and IP law and general contracts law. it the it it the it it's the problems that are interesting. And in all areas of law you can get interesting problems. so I wouldn't make the change. but I I you know but I because I've been happy. Other people I I I look I I one thing that attracted me then I trust was there like three statutes Right? You know? Sherman Act, Clayton Act, FTC Act. And they're not in general, they're not very prescriptive, right? Do this, don't do that, not a lot of parts to them. And so I didn't and and and and they are often, not always, at least but often now, it's sort of an economic argument. I didn't even have to read a lot of cases, right? When I was a young attorney, I was like I didn't I I needed to know What was the business reason for doing something? And that's what I needed to understand, not what does the law tell me about this? Because the law kind of followed the reasoning, right? Is is there a pro-competitive reason for this? Probably okay. So let me understand the reason you're doing it, right? If there is no pro-competitive reason for it, let's think how who does it harm? Harm somebody? All right, that's probably not okay under the law. I mean, that's a rough, you know, summary. But the basic point was what did attract me as a young lawyer was I d I didn't have to have this large body of knowledge to do a large part of my job. Now, over time, there are small questions that are legal questions and you can go down, you know, and go very deep on them. So do need to know lots of law. But when you start, you know, you can feel comfortable if if you if you have the willingness and or the ability to learn about business, you don't actually need to know a lot of law when you're a junior lawyer. And that and that also can give you some comfort, right? and and that was another positive thing. I was able to grow into learning the law. over a pace that I found comfortable. speaker-1: Wow. Thank you so much for sharing that and thank you for coming on Brains and Wats. It was fun to talk with you. speaker-0: All right, thank you. I hope I I hope this is interesting. I I I i it pe people should think very hard about how they're managing their legal career, particularly as as students. and they should be doing it right now. speaker-1: That was Brains and Bolts Episode Three with Mr. Bilas Sayed. If you like the content, please consider subscribing to our YouTube channel, following our Spotify and Apple Podcast channels. If you would like to participate in our work, consider becoming a member, as the National Association of Scholars is a membership organization. Thank you very much. See you next time.