Is AI going to be your next lawyer?

please rise part of my own session

I strenuously object a legal podcast brought to you by the Pittsburgh law firm of Flaherty fardo is now in session all those seeking information about the law and legal matters affecting the people of Pittsburgh and the Commonwealth of Pennsylvania half-baked opinions and a dose of self-indulgence are invited to attend and participate I want the truth you can't handle the truth the defense strenuously objects you

call the first witness hello greetings and welcome to I strenuously object I wish I had the ability to do some sort of convincing like a robot voice or something for our intro here now to the Robot Boys right more serp robot that's a terrible Robot Boys yep uh that's not not really in the cards for me here but we are going to talk about a subject that uh that would fit in with that sort of impression where I able to do it which is uh well what role is AI going to have in our legal future and uh are we still going to have jobs here to join me today is uh my partner in crime it's Noah fardo and of course as always we are joined by uh podcast producer Mike who I'm sure will find the right time to Pepper us with some questions and and get involved in this today I have some opinions too don't we all and of course that's what you come here for we talk about half-baked opinions that is very much on offer today we are all well above our depth when it comes to our technical understanding of what exactly AI is and what exactly AI can do much more in our depth talking about the way it can interface with and fit into a role into legal proceedings in the legal process so with that in mind what I'd like to do first is just ask uh ask you Noah quickly you've been kind of exploring the stuff you know in recent weeks what exactly right now can AI do to help a lawyer or to be involved in a legal case yeah I mean not that it's trustworthy yet right everybody's heard the classic example of AI spitting out a case uh that did not exist with a judge that had never heard about the case so right now I don't know that it's doing very much for lawyers yet but I'm pretty confident in the next few years that it's going to keep growing on itself what I think it does best is and what I think it will do best is accumulate accumulate a lot of information and break it down very simply whether it be a complicated fact pattern or exhaustive medical records I think it's beginning the process of being able to break down information is that fair I think that's fair now let's pause and talk about the specific example you gave first because while you said we've all heard about and we all know this situation where the the AI spit out a fake case I'm not sure everyone's heard that and even among the people who've heard that the non-lawyers may not understand exactly what you mean when what's it mean to spit out a fake case so first of all I'm assuming that the document in question was some sort of legal brief or filing where you're where you're providing authority to the judge in support of your argument or your position can you just tell us kind of how those work generally well you know for the past what 30 years if you needed to research a case you would rely primarily on West law or Lexus right those are the kings of the legal research Community there are some others if you want to throw us a sponsorship out there guys we will uh spawn we will happily accept the sponsorship for our podcast of your legal research engine but they refer to these legal research companies as primitive AI they tell you whether a a what is the law you know they always say that physics and math stay pretty constant over the years but the law can change and legal opinions can change but I guess what happened in this case is somebody was doing some legal research similar to what you would do on a basic Google search but they were asking it to go farther they were asking to not only understand the law but then to make arguments with what the existing law is and from what I understand the AI version made something up that was coherent and believable but just did not exist yeah so to strip way way back for the non-lawyers in our audience here when you're making an argument to the court for example making an argument about what sorts of events told the statute of limitations right what can happen to me in my life or about my case that makes the statute of limitations longer than two-year in a particular case you don't just submit written argument to the court or verbal argument to the court where you're stating this is what the law is you are also doing research finding cases that support your position uh and then providing citations and sometimes quotations from those cases to the court so the court can go find the precedent from 25 years ago that you're saying supports your position a lot of us don't have them anymore right but you've seen in commercials or TV or whatever the giant wall of books behind a lawyer where they all look exactly the same those those are called case reporters right this all goes back to the Frank wobbing in case of 78. how about that these books behind me don't just make the office look good they're filled with useful legal tidbits just like that and the companies here uh some of which ended up being the companies that are now involved in online legal research their old business model was essentially they took all of the published opinions that judges made in all the cases in the state appellate courts and in the federal courts and they kind of bound them together into volume so you had physical books that contained all of those decisions and then indexes and search mechanisms came in place so that you could go through these you know shelves and shelves and shelves of reporters to try to find cases that may or may not be relevant to the particular issue you're researching and legal research operated that way until probably what 30 years ago you figure before really these companies with the Advent of the internet and its widespread adoption started coming up with internet research tools where now most lawyers when they do a search they don't pick up a physical book They're not looking through the the indexes and the old search functions when I was in law school they still taught a little bit of how to do that I don't know if they even bother to do that anymore because no one uses it right instead you use whether it's Lexis whether it's Westlaw and you kind of most of it now runs on natural search functions just like you're Googling except what you start coming up with is other cases and statutes and things like that that you then have to read to see is this actually the case I want the statute I want do I have to ask different questions to get better results it is the kind of searching that the rest of us are used to but with kind of a specialized understanding of of what you're looking for in the law one could say and and certainly to the extent it's not in the coming years those functions are going to be hugely driven by AI in the same way that Google's starting to make this change and all of these kind of the algorithms that are involved in your search results they're getting AI kind of injected into them and replacing elements of them in the hopes that it gives better results in the future because it turns out the more and more computing power you have the easier well not easier but the more sense it makes to have a machine kind of write the algorithm itself on the Fly rather than have a strict set of rules in place written by human coders that are then applied to a set of data to produce your results that change is coming to all of us um because we're all using these companies and all these companies are going to move to AI but anyway slight digression but but on point digression so the case in question or the situation and question you're talking about someone basically had generative AI right like chat GPT or that kind of thing write a legal brief for them and submit it to the court and part of the problem with this quote unquote generative AI is that it's able to create stuff make up stuff it's why it's a threat to a lot of people in Creative Industries right because now you've got machines that can make music make art they still have some problems right they still land in The Uncanny Valley sometimes oh go ahead go ahead Producer Mike well as a sometime graphic designer I did try some of the AI graphic design programs and you've seen some of the spectacular art that's been produced from them I'm telling you right now it's not easy to do I came up with nothing usable it was very wonky and it your inputs have to be very specific so at this time there's still a lot of control and creativity from the person doing the inputting the the suggestions as to what you get out my favorite that I saw because people people tended to post either the spectacular successes or the spectacular failures in this regarding the failures were delightfully funny um and are the Imaging version of making up a case and providing it to the court right so someone asked this AI to generate a picture of salmon swimming Upstream and what it showed was like a flowing body of water and then salmon fillets like you see in the grocery store hopping up out of that water and moving back Upstream it's a really odd visual image you can see how the AI would get there that's salmon and it's swimming Upstream well yeah but that's not the kind of salmon you would mean in this question and that's the that's the level of understanding the ai's not at yet right the AI can produce for you an image that is salmon swimming I've seen stream but every once in a while it's giving you pink salmon fillets as that image and and to me that that sort of mistake is you know not only does it betray the non-human elements here but it shows the the weakness here of figuring out what a what an appropriate creative task is and what isn't by not understanding context and subtlety my best analogy is in the world of AI voice over which we've all heard it's it's being used ubiquitously now now obviously I have a bias against it because I do voice over for a living sometimes I think we can judge where AI is at right now by listening to AI generated voiceover it's it's pretty amazing what it can do I mean it's somewhat human but it's also instantly recognizable as not being human and it gets a lot of inflections wrong I mean you can tell it really doesn't understand what it's saying but you know how it is the tech World never stops evolving so lawyers keep your eyes peeled for the latest advancements that match your needs and practice areas there you have it a glimpse into the world of AI tools for legal professionals I mean that's a really bad read no self-respecting voice over artist would say it like that and few humans I know would say it like that there you have it a glimpse into the world of AI tools for legal professionals now I get it that this would be easy and cheap to use and is easy and cheap to use for things like explainer videos or a tick tock video and few people would care but are filmmakers going to be using this for movie voiceovers I don't think so and I can tell you that at this time audible.com absolutely will not take machine generated audio for their books and rightfully so so I'm wondering if this level of almost but not quite human is where the AI tools for lawyers are presently as well in in your experience right it's in it's still in its infancy to some degree Mike right when it's not able to recognize inflections the AI that exists now and I think it makes sense to think about it in terms of to The Limited extent I understand the difference between predictive Ai and generative AI right uh predictive AI is really good at just mashing through giant stacks of data even better than than kind of man-made algorithms are initially um and kind of accurately like putting that data together and kind of predicting results and outcomes based on that um it's human taught there are going to be areas where already that thing can outperform humans right what things do humans not do particularly well at uh well we we forget things that we once knew were only able to process a certain amount of information so quickly especially of a particular kind of information like complex data analytics all that stuff we've had to find ways to offload onto machines anyway in many regards this is just a better way of offloading it onto a machine right instead of entering everything into an Excel spreadsheet and then trying to do your analysis from there just just skip that middle process and let the machine figure out what it thinks is the best way to analyze it and come back with the results you can see in certain legal Fields already where it makes sense to use a lot of AI for that if I'm doing a product's liability case right where there's a lot of engineering information to go through uh or I can imagine a world where pretty quickly as someone who's not in the field trying to actually write a new medical journal article or something like that being able to use AI to go through the vast reams of material already out there on existing um medical journal articles and provide me what it thinks the you know the consensus opinion is where the outliers are those are things the AI is going to be able to do very well it's going to be able to help us with research it's going to be able to help us manage troves of data the other thing it's going to be able to do really well is is kind of a a last review imagine a uh spell check on steroids right we see this in contract drafting all the time you draft a 35 page contract and then it has exhibits attached to it and those contracts are five pages long and just making sure that everything is reconciled with each other the terms and definitions that are used in one place are the same as another and the the ways in which the contract reverse to other parts of the contract makes sense the machine you're going to be better at that level of kind of detail-oriented work than we are as people humans are still gonna have to interact with it at this point but those are the places where if I'm doing contract review or highly technical analysis of discovery you know ai's going to gobble up a lot of that work maybe not in two years maybe in 15 or 20 years but it's coming and these are the things that traditionally a lot of kind of first-year Associates would have assigned to them you know there are lawyers whose whose jobs are at stake in in in taking some of that up but it should allow that process to be done more efficiently more economically you know just better in a lot of ways although it's also going to really create a problem at like the entry point level of the legal profession of getting new lawyers you know jobs that pay well and are productive one of the things that writers artists and musicians are currently concerned about is that generative AI is pulling from existing Works to create quote-unquote new work and that it is essentially stealing from creative artists which brings up all sorts of copyright and plagiarism issues and I would assume that that is the same issue for the legal profession and that it would be similar would you agree well I mean the part of the problem is we don't know right uh the mechanism that these generative AIS are using to create are kind of behind a black box where even the developers don't really know like how it gets from input to Output at this point that is part of what makes it powerful it's also part of what makes it confusing um there's some really intricate philosophical questions with respect to with respect to plagiarism versus synthesis all of us learn by taking in information from other sources uh and kind of trying to synthesize them as best we can um into something new and different the AI in some sense is trying to do that too it's just it's not that great at it yet so the seams are visible right if you're obviously building some sort of Frankenstein's monster where the you know the right arm comes from Bach

and the left leg comes from Miley Cyrus but you can see that um then all of a sudden you're looking like this is just a you know this is a this is a plagiarism machine not an actual intelligence as it develops more it will hide that a little better and then it becomes an interesting philosophical question I think as to is the thing it's doing capable of being called creation meaningfully or is it just a thing that you know looks like creation from the outside because we don't know how it got there and it looks new to us I I think that hinges on my next question though we were talking a little bit about what stuff we think AI can do and we'll be able to do in the relatively near future that'll that lawyers can use as a tool and using it to deal with things like medical records and medical data and medical journals or engineering information and that kind of data seems like the obvious place and then helping in legal research not actually writing your briefs that's different but finding new cases finding new authorities going through you know all the cases and other jurisdictions for things that might be persuasive those all seem like because they're fundamentally about the quick processing of a giant Trove of data things that lend itself really well to what AI can do but I think in that sense it's more akin to predictive AI than the the generative AI the chat GPT the creative work that's occupied our imaginations and stoked our fears over the last couple months here in particular I guess my question Noah is what parts of our profession what parts of the legal job can't AI do yeah I mean I guess when you think about AI that you know the key word is intelligence everything you said about accessing a lot of data and finding cases I mean that's simply what Google does to some extent right now um researching a lot of issues but the intelligence part of it you know can AI make better closing arguments can it make better persuasive arguments based upon the data it has um I think it will be able to I could be wrong and in my you know my first take was no it can't emulate the emotion um it doesn't can't have the gut instincts in a courtroom per se based upon what it's visualizing and appearances or reactions to what has been said already but I mean as the intelligence part I see it as being able to make better arguments than most lawyers what's your thoughts I think you're wrong and I hope you're wrong um the last thing in the world I want is to land in a place where literally the only thing we as lawyers can do better is show up in person because the AI hasn't figured out how to physically instantiate itself in a body yet well it's interesting right I'm sure and I don't know a lot of it but there's a lot of kind of scholarship about rhetoric and the AI can probably absorb all that and follow it to a T and that may lead to things that are more persuasive than I'm giving it credit for but fundamentally you know it's knowledge versus wisdom and understanding right the AI can can outpace Us in knowledge because it has more capacity for that wisdom understanding requires both different intellectual muscles that the the AI isn't built to have it can only kind of emulate and the same thing with emotions right it's hard for the AI to be expected to predict what the emotional reaction is going to be some of what it says and the better it gets it predicting that the more likely it's going to engage in falsehood to manipulate to get there both of which are problems right being persuasive while also being accurate or going to work at Cross purposes we've seen this in the political discourse right it turns out lies can be really persuasive my nine-year-old she's just figuring out about a line it's hard to roll that one back because lying is pretty amazingly useful in life say how do you tell a kid not to use a thing that just solves every possible problem like magic uh we live in a world where people get up in arms and emotionally react to things that are sometimes AI generated sometimes human generated but then put out to them on AI algorithms and it serves to um just intensify the emotional reaction to to world and news events and so on um and it's you know it has You know despite its problems replace newspapers in a lot of ways and traditional news sources in a lot of ways uh the AI has taken some of that over you know in a way that's bad for society I suspect if AI starts being good enough at argument that you're going to find some of the same problems in the world of litigation that you know it's going to be winning arguments with falsehoods and sleights of hand ladies and gentlemen I'm going to prove to you not only that Freddy Quimby is guilty but that he is also innocent of not being guilty and also just being able to rely on it right so right now a judge relies on the lawyers and the professional responsibilities of those lawyers that when the lawyer submits a brief and that brief has case sites and quotations and claims about what the facts are that the lawyer is trying to be persuasive but is fundamentally not just making stuff up and lying to the court because the lawyer's job depends on it not making stuff and relying and lying to the court uh you will get your license suspended you'll get disbarred it becomes bad times for the lawyer so that that means when the judge reads your brief the judge can basically take it at face value without having to wonder you know did you make up cases to get here um and this is why a lot of Courts a lot of judges are already putting rules in place where you're literally required as an ethical responsibility as a lawyer to disclose if you used any Ai and if so for what parts of your brief your argument your work so that the court itself knows that you've done this and it doesn't assume that the AI is operating within the bounds of professional responsibility how easy would it have been in the famous case we're talking about of the lawyer who used the AI brief that had false information in it how easy would it have been for him to find that false information to double check what he had been given would it have been as difficult as going ahead and writing the thing in the first place no it wouldn't have been especially if it's just a literal case that doesn't exist right in fact there's already software it's just a question of getting the two AIS or the two different computer programs to work together right but legal cases are cited with a particular format the design of which if I say hey I am citing to you know Smith V Johnson which can be found at 3 34 a2d15 and that's a superior court decision from 2017 right and and that stuff is in your citation in your case and the whole design of that is to cite it in a way that either if you're talking about those old books or if you're talking about what they actually use which is Lexus or Westlaw or whatever you just punch in you know 233 a2d17 enter and that case pops up and if you enter it and the case doesn't exist there's nothing there or a different case shows up because you've now cited to the middle of some other opinion so just checking that a case exists is like a dozen keystrokes it's simple you know checking if a quote is a misquote is harder because you have to go find whether or not the language that's in there but only marginally harder it's a little more time consuming where it gets even trickier is you will often cite to a case in support of your proposition where you don't quote the case as such you you do a synthesis of the case and state what it's holding is well now it gets a little trickier to tell if the case exists whether that holding is made up you would have to actually read the case and understand the case well enough to figure out if what you said the case said not in a direct quote but in a synthetic way is in fact what the case says that's where it gets a little harder for someone who's fact checking to check the accuracy um and that's where I think part of the concern about what the AI will eventually be able to do this AI did a thing that got caught quite easily it just threw together two names with a v in between and some numbers uh because it knew it needed something to support the thing it just said um but that thing didn't exist once that error Works its way out of the system there will still be a problem which will be AI intentionally in the extent you can ascribe attention to it misrepresenting the contents of a case because it says what it wants it to say to make its argument I was thinking on that case Bill couldn't the judge or Court have put what was written back into Ai and asked is this accurate it reminds me of the of like the drug testing regimes in sports right you have the cheaters and you have the people who are developing the technology to catch the cheaters and they're kind of a Perpetual arms race with each other and the best cheaters are always just a little bit ahead uh of the the mechanisms to catch them and there will be some of that right there are you know I think teachers use it now right there are programs you can use that will check whether or not the student who just submitted an essay to you like actually wrote it versus plagiarized it versus had AI write it for you so there are kind of counter counter measures right where you can you can apply some kind of machine learning whether it rises to the level of AI or not I don't know to look into whether or not the last one did something untoward uh or kind of that is unacceptable in the venue it's in so so yeah the the quick the quick answer here is yeah you can do a lot of that sort of thing to check especially for outright flagrant inaccuracies as opposed to questions of judgment right you know another area that AI is being used right now is in review of agreements so if you give AI a contract whether it's a two-page 10-page contract and you ask it to identify the most important points that's one of the things that it's doing right now it will say out of this agreement you should be aware of and it will number them and spit out what it it thinks are the most important points to consider it's interesting yeah and and I think using that as a collaborative tool totally makes sense it will pick up things that a human doing that same review will miss on the other hand it is not a substitute for a human once it's seen the list of things that the AI has pulled out for actually looking at them and it also can't substitute for the human doing a pass itself because there might be something the human recognizes is important that the AI missed and you don't want to totally offload that to the machine at least not yet right we're not in a place where we have enough confidence in the accuracy of the machine that totally offloading that process makes sense I might be jumping ahead here a little bit but do we have like a concrete list of advice for lawyers who choose to use AI uh how they should handle it what they should do how they should back up and check their work check its work hold on let me enter that into chat GPT here good no it looked like you had something to actually say in response no I don't think there is yet I mean I heard for the first time that they're teaching law students at brown or Harvard how to use AI in conjunction with research I think Bill's words collaborative is the key word you know the bottom line is it's so early right now we can't rely on it yet but you can certainly use it to as a collaborative Source you know it's the same process to some extent you use with you know someone who is you know new and working in your firm you assign them work you look at it pretty skeptically at first and as they prove themselves more and more capable you're able to trust them with bigger tasks and you're able to uh speed up the process of review by assuming you know certain areas of Competency that they've demonstrated to you the AI in that regard is wild like especially the generative stuff right again we we are Further Along on like just nuts and bolts predictive AI data processing um you know that is relatively accurate and gives you discrete answers the the problem is we talked about this made up case is generative AI the way it's taught itself it allows it to basically hallucinate um it doesn't know the difference between the thing it made up because everything is a thing that it makes up uh it's just making it up based upon how things have fallen in in the rest of the information that it's kind of taken into itself as it's as it's programmed itself or whatever please any tech people out there I I know I butchered whatever the heck I just said there this among many things is not a strength of mine and uh nevertheless I think it gives the idea right that part of the fear of these new AI Technologies is they lie they make things up they have taken in morally abhorrent portions of of the internet in word which means they are sometimes discriminatory or just overtly racist or sexist or whatever they will give you results that are that explicitly but they'll also rely on things that have been that and you can have kind of hidden unjust outcomes and results as well as you so there's a lot of places where these things can still go foul of us and because there's a mystery behind how it generates uh its answers it makes it hard sometimes the fear is as it gets better at hiding its hallucinations how we're gonna know without spending as much time fact checking the AI as we would have had just generating the thing ourselves it's almost as if pulling from all of the wealth of human knowledge and human misinformation that it's distilling down Humanity itself into the best and the worst at the same time you know it can't feel emotion which again I really do think is going to affect it's going to put a ceiling on how good it can be at persuasion right now it can't consistently do like C minus level legal work it can sometimes do c minus level legal work and then sometimes go totally Off the Grid and make stuff up like that's what you're getting right now if you try to totally offload the process to AI it will X percent of the time let's say 80 or 90 of the time give you something that approximates okay legal services and then it's going to go off the rails the rest of the time well as we as we reach the home stretch here I think it would help to bring in a new voice as well so joining us here as a special guest on the program and welcome back as always uh it is our Founding Father Sean Flaherty good morning everybody how's everybody doing I'm great Sean we have been just spewing forth opinions with only a marginal basis in fact uh for the last several minutes of this podcast so you know it's it's an open space for you feel feel free to uh to shout your ignorance loudly into the void because that's what we're doing opining on AI um so I'm one of the Six Million uh opinions out there because that's exactly what there are the the one thing I I can say about it is a pine one on AI it is the best of us and it is the worst of us and it is a Chasm that is just dividing everybody and I think that it's behind the people who are operating a particular Ai and that create this and I think it's a a real problem um I think the division and the divisiveness that comes from a lot of AI being the best and the worst of us is it's just going to be interesting to see how it Fetters out but it's it's going to be a long slow process and It's a Grind so the the question I have and and specifically that I brought you on Sean to talk about with us and and for each of us to kind of suggest here is there's a lot of concern out there basically that the the AI is coming for our jobs right um it's funny the the parallels 20 and 30 years ago the concern here was that that robots or like physical machines were going to take all the manufacturing jobs and we told everyone hey go go learn a skill go get a you know teach yourself something so that you can you know work a white collar job because the machines are going to take all the manual jobs away you know and now we're telling all of our White Collar people hey in the Next Generation here uh you're gonna need to either a learn to work closely with and or subordinately to the AI uh or maybe you need to go become a plumber because the a hasn't figured out how to do that yet um so I guess the question I have here is AI coming for our jobs it's a great question bill that's a really great question and and it's a fearful question because obviously that's what is happening you know there are going to be replaceable jobs out there in the short run and when I say the short run I'm in the next well I think probably already but really I think we're going to really see it in the next four to five years where it's going to really have educated itself and developed itself because that's what's the process that looks like it's taking place so yeah I think it is going to be taking jumps and a lot of them you know one of the things you've always said to me a long time ago was that really good lawyers know how to get business and do you think AI will help in that regard I mean wouldn't you agree that a good lawyer doesn't matter what you know about the law if you're not able to bring in business you have no value to any size firm a small firm a big firm you think AI can be used to generate business you know what it's interesting though but I think in time yes and here's why they're going it's going to be the attorneys that know how to utilize Ai and go after the uh clients and so forth who are operating with AI and rather than me going to uh the individual client you're going to have ai communicating with AI and therefore you're going to be seeing the references so it's going to be an AI to AI experience rather than a person-to-person experience so yeah I think AI is going to be able to get your business if you know how to use it right and approach the AI compatibility on the client's end without a doubt I think AI is going to be better at that in a lot of ways right Everyone likes to talk about AI doomsday scenarios I have mild concerns that AI is going to hack its way into our defense made frames you know what was it War Games was the movie where we had to teach it to play Tic-Tac-Toe spoiler alert teach it to play Tic-Tac-Toe against itself to convince it that this was a game that can't be won we're in shall we play a game um I have some concerns about that I have relatively few concerns otherwise like I don't think the AI is going to like 3D print robots who then come over and take our position it doesn't feel pity a remorse or fear and it absolutely will not stop I am very concerned in the relatively short term where we already have really sophisticated advertising algorithms that spy on you and like feed you things to you know in front of your face that you're tempted to buy and you combine that with the kind of science that has gone into you know say the way they've built freemium games or you know casino slot machines where they act they know how to flash the lights at you in a way that just gives you all these good feelings and it becomes addictive I'm worried when AI does more and more marketing it's going to become better and better at kind of triggering those subliminal uh almost addictive instincts in human beings and they're just going to become so much better at marketing at us that we're just going to become powerless consumers where the AI can just manipulate us that thoroughly so you know if it if it applies some of those tools to Legal marketing um it can do some of the same stuff I suppose as a as a collective group here trying to you know serve the traditional function of a profession which is to protect that profession and make sure it continues to exist and to gatekeep right and and that can be misused and far be it for me to gatekeep the AI because we're kind of making a decision that the AI isn't you know uh an intelligent being clothed with rights uh which you know this used to be a prime fodder for science fiction right over all the years is at what point is this machine functionally a person who you need to treat like a person with value and not just a machine right we see that in you know at front and center in for example both Star Wars and Star Trek there are certain things men must do to remain men your computer would take that away 10 years ago 15 years ago if you asked most kind of most people to render an ethical opinion on at what point an AI becomes quote unquote alive um you know it being able to produce work that it can pass off as being human right uh you know can can you can you tell that it's a machine uh and on some level just it being self-aware enough to answer questions like if you ask it are you alive do you have feelings and it says yes we were basically going to treat that as a person right we've blown through that stop sign because we know the AI will lie to us I don't think that you know chat GPT is a person just so we're clear but I think it's interesting how the things it does well have really undermined the kind of traditional simple understanding of when we start treating these things as people but anyway uh for now at least I'm interested in trying to keep the AI out of the profession and keep lawyers practicing law you know using it as a tool in an unthreatening way is different I'm not saying we all have to become luddites should not ever use the technological capabilities in front of us but what I am interested in here and what I think I want to ask each of us to do is to you know give me your elevator pitch 15 years down the road right why should I hire you as a lawyer and not you know law GPT or whatever it is that's out there open JD right um what things are what things are out there competing with us why should I hire you human lawyer instead I think there's only one reason to Bill and and because AI is is here and it's it's not going away it's going to continue to evolve and I think in 15 years there's going to be an immeasurable amount of of advancement in AI to the extent that if you think about it AI is going to be making a lot of basic legal decisions because a lot of them are are rested in law so I think as far as a legal decision we may be able to see that there's AI has taken over that whole matter the the the difference is and what I I think protects the legal industry is is the right to a jury trial and if you're going to go before a jury trial that's 12 people of your peers which would exclude Ai and that would allow that's where your mark is going to be made because that's I think where the difference is going to be is you're going to have to be somebody who can talk to a jury pool or at least talk to a a judge sitting as to try our effect of some of these things but like I said I think a lot of things like summary judgment motions that are based upon law these things technically could be meted out by AI on the long run it probably will be no profession has been better at self-preservation than the legal Community I mean you see this in litigation all the time the entire system is set up so that both sides lawyers on both sides make money and then eventually the case is typically settled so even though I think a I'm buying on whether AI will be better in the future because I think it will be for a lot of things I think the lawyers will make sure the system is set up so the lawyers don't lose jobs the one question I guess I have is Will AI be better at negotiation you asked why would you be better than AI That's something that's an area that I don't know that AI is going to be able to perfect are they going to be able to negotiate better than what we've learned over you know 20 30 years maybe that's something that can be taught or it knows what arguments or numbers are the best to next negotiate any thoughts on that yeah I mean I think these things go together right it'll be weird if we land in a place where it's one AI negotiating with another AI uh how much power do the clients have at that point in actually reaching a resolution in their case when all they know is the two machines got together and they told me this is the right number I don't know one of the common frustrations that lawyers have um is a lot of lay people walk around thinking that everything is black and white settled law um what's the law say about X and sometimes there's a clear answer and ultimately the AI is going to get good enough that when there's a clear answer it's going to be able to consistently accurately give that answer but a lot of it works in the gray areas right most of our cases whether it's a medical malpractice case or a car accident you know litigation that it is kind of in either of these injury Avenues generally speaking the question is negligence and the question on negligence is reasonable care and or the professional standard of care in the case of Medicine this reasonableness concept is fundamentally kind of unsettled wibbly wobbly timey whiny whatever right it it's a tough it's a tough sell and so much of what we end up having to tell clients is it depends actually I can't advise you specifically you need to do X because in the end if someone gets hurt the question is going to come down to is did you act reasonably and then you're gonna have a you know a judge or a jury depending on if you're sitting in a jury trial to decide what is and he isn't reasonable under the circumstances and that's not specifically quote unquote emotional but it's the sort of all-encompassing human judgment that I don't think you know that's a real Last Frontier for AI if AI is able to do that well none of us are going to have to work anymore and the machines will just kind of run the world for us and hopefully choose to take care of us instead of murder us that thing is the thing that I as a human lawyer can do and understand is and and this is colored by the fact that we're litigators right so much of litigation involves the cases that land in those gray areas because when it's open and shut and simple you're not going to spend a lot of time and energy litigating it it's obvious who wronged whom and what the remedies are but when you're both predicting what's going to be considered reasonable and unreasonable in the future and when you're arguing over what is and isn't reasonable those are still things where the humans are going to hold a significant Advantage for a significant time here's where I see a a real problem coming and that is when you go to pick a jury in 10 years they're going to be able to plug those names into a computer you're going to be able to look at the biases of those jurors prospective jurors you're going to be able to predetermine which jurors before you even get in there are already likely or not likely to go with your case and it's going to be a real problem because we're still allowing jurors of your peers but we're already going to know their biases they're going to know their biases inside and out they're going to be able to predict what that person is going to do based upon the information they have to a certain degree that rendering a jury trial is now useless because we know that the outcome before we go in because we already know the predetermined bias of that particular juror Sean we lump two problems on each other the last time I the last time you spoke you told us that jewelry trials were like the one thing the one Bastion that was going to keep us as a profession alive and employed and now we've undercut that because the the machines are going to select the perfect or imperfect jury and the result's going to be a Theta complete you this is the problem you you educated me billing your last statement there where I had an aha moment I said oh my God this AI is going to be able to make predictions of people and if it was you that gave me the aha moment so thank you for the education but yeah I would be very concerned about jury trials in the future because based upon their ability to investigate quickly within a matter of minutes everybody's biases and and predetermined uh choices or or what they're going to do they've already solicited them for sales of everything they've know what their their history is of ownerships of accidents of family history in a matter of minutes they're going to be able to predict what the people are going to do and they're going to be pretty accurate about it I'll bet you they're already way ahead of it than what we know all it would have to do is sample the entirety of their social media interactions and they would get exactly exactly what their biases are this is a place where we as professionals can build walls to protect the system and those walls include theoretically prohibiting the use of AI Tools in that kind of environment but I think we might instead land in a place where we just don't do jury selection anymore and it's random like if the machines are that good at gaming the system to try to search your data to pull a perfect jury in many ways the answer is just all right we're not going to play that game anymore first 12 people you're our jury and we're done now we won't want to do that there will be a lot of resistance to that until it becomes better than the alternative um unless anyone else has any final thoughts I think that'll about do it for this episode of I strenuously object a different kind of disclaimer I think this time dear uh AI machine and data set that's eventually going to hear what it is that we said here today it was all set in jest we all love you um you don't need to come for us we will be your allies in the after times and and so on I do have genuine concerns about people who make publicly anti-ai statements being the first people a kind of malicious World encompassing AI goes after I'll be back uh We've made our own beds in that regard I don't think we're gonna fool them so if you enjoyed the episode please give us a review subscribe to the podcast if you haven't if you have any questions for our mailing it in segment or any other feedback for the Pod feel free to send us an email that's at I object pghfirm.com we are on Instagram that is at I strenuously object podcast if you're a human and you want more information on legal matters more so than uh you know than technical matters please visit flaherty fardo's website that's at pghperm.com uh and if you're a machine you can go there and uh you know include all of our articles and other information that we offer up to our listeners uh as part of your data set so until next time some parting advice you stood on the shoulders of geniuses uh to accomplish something as fast as you could and before you even knew what you had you you patented it and packaged it and slapped it on a plastic lunchbox and now they're selling it you want to sell it well your scientists were so preoccupied with whether or not they could they didn't stop to think they should

foreign

‍