Superjustice: Ben Alarie on AI, Abundant Legal Help and the Future of Lawyers

Speaker A: Often legal disputes have two dimensions, right? One is actual resources, actual behavior, actual effects, and then also how people feel about those things. And uh, often there's, there's the legal, the dollars and cents, the factual, material impact of certain things and then there's how people feel. Do they feel respected? Do they feel that they're being heard? And I think to be maximally effective, a system of super justice will attend to each of those dimensions effectively, will understand how to properly or most effectively present the findings to each of the parties, taking into account their emotional state, how sophisticated they are, the level of language that they want to interact at, uh, how much detail they want about the underlying legal aspects of it. How do you stay up to date with AI news?

Speaker B: Welcome back to Future Ready Lawyer. We are delighted today to welcome Professor Ben Ellery Osler Chair in Business Law and at the University of Toronto. New book Law in the Age of Artificial Intelligence, co authored with Professor Samuel Becker, explores how AI could transform law and make justice more accessible, dynamic and responsive. Ben is also co founder, uh, and CEO of Bluejay, a University of Toronto spin out firm M which he launched with law professors now being used by over 6,000 companies and rapidly expanding, funded by a 2025, US$122 million Series D round. So Ben, we're really delighted to have you along. Welcome to the podcast.

Speaker A: Thanks Mark. It's great to be here.

Speaker C: Great. So we just going to start with a general question about the book and the concept of super Justice. This is something that you discuss in the book. Could you tell us what is Super Justice?

Speaker A: Sure. I mean the title is intentionally provocative. The principal idea I think when you scratch away at uh, super justice is that we foresee a future in which justice can become super abundant. And so I think right now so many different things in our legal systems are bound by scarcity. And super justice really portends the end of legal scarcity. It imagines a legal system in which people's rights and protections become radically more available. Um, and so it manifests in different ways. One is legal assistance can become abundant. So if you have a problem, you need basic legal services. They will be available at a scale that no lawyer centered system can match. It also means that what many feel now is a legal labyrinth that's very difficult to navigate, can turn that maze like system into something that, like pathways that ordinary people can actually use relatively easily and can navigate maybe on their own. Um, and then finally like the most intuitive way of putting it might be something like justice can be much more like Public roadways. I don't think many of us wake up in the morning, hop into a vehicle if we're commuting to an office or to work, turn the keys in the car or push the button to get the car going, and we drive on a public road and we take it for granted. But it's a huge public infrastructure play that many of us take for granted that we appreciate or not every day. Maybe we gripe about the traffic, but it's there and the legal system is very far away from that. It's like a series of private toll roads with no maps available or difficult to figure out maps. And so we imagine a world in which super justice will be an expected, available and valuable everyday amenity like public roadways.

Speaker C: But that's a great analogy. Could you tell us more about how it would look like, let's say we achieve super justice and I have a legal problem with my neighbor. How would it, you know, look like if we to go through this super justice path?

Speaker A: I wouldn't want to be too prescriptive about it, but I can share some thoughts about how this might go down. So let's suppose it's, I don't know what kind of dispute it could be. It could be about a tree or it could be about a fence, or it could be some sort of common neighborly dispute. I could easily imagine both parties to the dispute agreeing, opting in to have their dispute governed by the super justice system. And whether it's a local government or a state or provincial government, or even a federal government, a ah, governmentally publicly provided system. It could also be a private system for dispute resolution. Uh, it doesn't really matter who the provider is, but the parties might say, okay, we clearly have a disagreement. We both clearly feel like we have some aspect of this, that we have some entitlement either to do the thing or to not do the thing or to be unharassed about doing the thing. Let's submit it to, let's uh, submit it to the system to give us guidance. And each party can make, can upload information, can give certain permissions to this adjudicator or dispute resolver and the system will work away on it, ask further questions of each of the parties and broker a resolution. It may not announce, oh, this person is right or this person is right. Instead, I can imagine mediating the dispute and explaining to each party the meritorious aspects of the other party's position, validating the meritorious aspects of each party's position and really resolving the dispute in the way that the most talented mediator alive, human mediator alive, might be able to resolve the dispute or ultimately, if something can't be mediated, can arbitrate and get win the buy in of the parties to the arbitration and say, okay, we're going to solve this according to prevailing law, and then provide a very rhetorically effective description of why that is the outcome as between the parties. Now, uh, I think the important thing, at least in my imagination as I scope this out, is often legal disputes have two dimensions, right? One is actual resources, actual behavior, actual effects, and then also how people feel about those things. And often there's. There's the legal, the dollars and cents, the factual, material impact of certain things, and then there's how people feel. Do they feel respected? Do they feel that they're being heard? And I think to be maximally effective, a system of super justice will attend to each of those dimensions, effectively will understand how to properly or most effectively present the findings to each of the parties, taking into account their emotional state, how sophisticated they are, the level of language that they want to interact at, uh, how much detail they want about the underlying legal aspects of it versus the Terrian aspects, like how to, like, actually, in a lot of ways, somebody can accept an adverse verdict from one of these systems, I suspect, if it's described in the right way, if it respects what's valid and how they were feeling about it, and can help solve some of the hurt feelings involved. Um, and also if a party wins, it can point out, uh, yes, technically you are correct, but there are some other things that would be very minimally costly for you to actually move forward with your neighbor. And so it could be like, it'll border on. It won't be just a legal determination. It'll be a very richly textured, highly contextualized way of, uh, resolving disputes, reducing the temperature in these kinds of neighborly squabbles. That's what I'm imagining. You mentioned in the introduction, Mark, that this book was co authored with Professor Samuel Becker, who's a dear friend. Samuel and I agree on a lot of things. We don't agree on everything, so I should caution and say everything that I'm sharing in this podcast. This is my opinion. Samuel may or may not agree with, for example, how I'm describing this, and I wouldn't want anyone to necessarily assume that I'm speaking for Samuel as well, because the motivation for this conversation is our shared academic project here. Super justice. And Samuel, in the event of disagreement between me and Samuel, is almost always right. So I'LL just, I'll put that on the record for the benefit of Samuel and for everyone listening.

Speaker B: Sabine, you've given us this vision, I guess, a utopian vision for the justice realm. Fully automated luxury, legal adjudication, solving of disputes. And you can obviously imagine a world where we devote all our resources to paying humans to provide those services. But we're talking about technological and artificial intelligence mechanisms for achieving this. What are some of the underlying technological mechanisms? Do we have them now? How have they evolved in the last 10 years to make this even a possibility, this dream of legal utopia or legal completeness to come true? And to what degree do we have to further invent these technologies underlying this vision?

Speaker A: Yeah, I think it's clear that to achieve the sort of super justice system that Samuel and I are describing in super justice, we will need a system that, that is fundamentally founded on super, um, abundant intelligence. So some might call it super intelligence. Clearly we need systems that are even more efficacious, profoundly more efficacious than the artificial intelligence systems that we have today. Having said that, Armin and Mark, the AI systems that we have today are profoundly better than the systems that we had even just a few years ago. So I started Bluejay, as you mentioned in the introduction, I started Bluejay with colleagues from the University of Toronto back in 2015. And we've been on the frontier of using AI and machine learning for tax research ever since. It is really in the past three years, and I would say, especially in the past one year, where we've seen the capabilities of AI for advancing the state of the art in tax research really begin to be very impressive. The underlying computational infrastructure is improving exponentially. We are having an acceleration of algorithmic improvements. So the methods for inference are continuing to improve. We're digitizing more and more information than we ever have before. These are the tailwinds powering the improvement in these inference systems. And I'll just give you some examples from the front line at, uh, Bluejay. So we started leveraging generative AI, large language models for powering Bluejay back in 2023, three years now, three plus years later, what we've seen is close to two orders of magnitude of improvement in the accuracy of the system. So if back in mid-2023, leveraging GPT, it would have been early versions of GPT4, um, we'd see like 90% accuracy in the tax research answers coming through, like, really good. And around that time, um, folks were experimenting with having GPT4. Take exam. And it was passing and it was, people were excited, slash scared, slash provoked by those kinds of findings now on Bluejay. So if it were 90% back in mid 2023, we are now seeing disagreement rates in the platform of less than a tenth of a percent. So that's a, uh, 100x reduction over the past 3ish years in the level of disagreement with, of our users with the results of Bluejay's algorithms. And there's a lot running behind that. So we've grown a lot more sophisticated in how we are, um, teeing up materials, retrieving materials to have the AI analyze in response to a tax research question that somebody's asking, um, there's an improvement in the frontier models, obviously significant improvement in the frontier models. We have a better level of understanding. We've processed more than 10 million tax research questions. It took us nearly two years to get to our first million tax research questions. The most recent million questions have taken under a month. And so we're seeing, uh, an acceleration in the use of the platform as it becomes more effective. And a lot of this is powered by the generalized improvement in the underlying models, but also paired with the know how that we are gleaning from all of this experience in building these systems and how users are using them and the kinds of tasks that people are pursuing. And so the kind of future system that we are talking about, something that is capable of satisfying the humane dimensions of dispute resolution as well as performing what you might call the more kind of technical, legally technical or technocratic aspects from a policy making perspective of the legal system. Both of those are going to be loading heavily on the relative abundance of intelligence and insight into what creates a dispute. What is the etiology of a dispute amongst parties and how do we properly, um, deal with those disputes and how do we defuse them, how do we lower the temperature in those disputes. It turns out that having abundant intelligence can be helpful with all of this stuff. And so I would argue that the single biggest driver of the possibility of building this kind of system, you described it as utopian. I would describe it as human centered and machine scaled. Because for me at least, utopia as a term describes something that's unachievable. I think this is achievable. Um, and so I'm reluctant to label it utopian. It may sound too good to be true today. I do not think it will be too good to be true in the future. And in the future, I mean, let's think long term, longish term, like a century, two centuries. But Maybe decades. I think it'll take some time. I don't think Samuel and I are unrealistic about the institutions that need to change and all the problems that will need to be solved along the way and the scale of the challenge in having super abundant intelligence in order to drive these systems. I do think we take the perspective that this is doable, this is achievable, and we want to put humans at the center of this movement to make sure that what gets built is really the kind of thing that we'll be proud to have been talking about, to play our small part in helping to bring about, as I suspect you, Mark and you Armin are also doing with this podcast, you'd love to play your part in helping to bring about a more humane AI driven legal system. So I think it really does turn on how effective is AI? And if we have really effective, super abundant machine intelligence that we can draw upon, how do we harness that to create super justice? Create the system that really does justice to the humans and keeps the humans at the center of the technology.

Speaker C: So, Ben, that's very interesting and it would be an amazing vision to address a lot of access to justice issues. And our understanding is that it's not only the cost problem when people do not access justice, it's also the complexity. They don't know even where to start. So an intelligent system that can guide them step by step would be. And when I say step by step, I mean correctly in the right direction, not hallucination. That would be a very amazing vision. One question that I have in that regard is that you were saying that like you're imagining the AI would be more effectionist and not only look at the legal problems, but also the tone of the parties and what they're actually struggling with. And uh, that's one of the things a lot of academics tell the students that what is valued in the age of AI is empathy and emotion that AI doesn't have. And one of the things that I hear a lot from judges is that AI does not have the capacity to exercise discretion. And Mark and I previously discussed about these two issues that we call a disagree with both things. We just want to know what's your vision about the human involvement? Because you say humans should be still at the center or in the loop in this system. Um, if an AI model can do those very specific, what we call human features, then what is left for to humans here?

Speaker A: Yeah, I think one of the, one of the core things is that humans will be responsible for monitoring, building and Monitoring the architecture for super justice. And if you are building, monitoring the architecture, values are going to be built into that architecture. You make design choices. You make architectural choices about how the technology operates that is going to reflect a certain set of values. In the book, we claim that those should be human values. Those should be values in support of human flourishing, in support of the kind of legal system that we would want. And I think I'm with you and Mark. I think I disagree with people who would make the claim that AI is incapable of exhibiting empathy or exhibiting an understanding of how people are going to take things. Understand things. In fact, I find that a lot of times my tone can be improved in emails by simply doing a first draft myself, understanding what I'd like to say, and asking for suggestions about how to improve the tone. I can describe the tone that I would like to strike. Sometimes it's very helpful to have an AI suggest different ways of putting the same message that, uh, actually maps on to what I, as an emotional matter, would like to convey. I'd like to put this in a way that will be received in the following way. How do I do that? And I can get some beautiful suggestions from an AI interlocutor about how to put things in the way that I intend to put them. And so I value that already. I think I'm not alone. I think a lot of people do the same thing, and it's extremely helpful in order to do that. It's as if I'm able to channel some level of empathy that I know exists, like it's capable of being done. It's just more efficient for me to find those words to express what I'm feeling. Leveraging an AI. Now, some purists, perhaps those for whom those words come more easily, well, poo, poo that and say, oh, this is, this is not the right way to do things. I challenge them. I'm like, if instead, I'm going to be misinterpreted as being blunt and uncaring because I actually do care and I do want to be perceived, as I perceive myself, to be. Leveraging those words. What is wrong with leveraging technology to help me define those words in order to convey precisely how it is that I'm feeling. I think there's evidence that blind tests, double blind tests, single blind tests involving patients and doctors, and having online consultations with a chatbot versus having online consultations with a medical doctor. Most patients preferred the chatbots. They didn't know if they're talking to a real physician or to a Chatbot, but they preferred the affect of the chatbot over the physicians in a blind test. I think that's the right kind of empirical evidence to marshal against those sorts of claims that AIs are not capable of exhibiting empathy.

Speaker B: So Ben, just in terms of the role of the human, this is a big hot button issue at the moment. All of what we have just been discussing there suggests that there isn't some special human role. And there's a lot of people worried about even what we have now in terms of the AI tools that are helping out with legal work or all sorts of other kinds of work. They are doing the jobs of say 10 humans. And then another human is just monitoring them, overseeing them. And so there's this big worry about what is going to be the role of humans given AI. And I think it's even more profound because we're talking about the AI today and we've seen that whether it's an exponential or even if just line goes up, and that is what we're talking about with the super justice idea. Super intelligence is the foundation for it in that kind of world. Why even say that the human needs to be involved at all? And then I guess the second part of that is how many humans do you see, like some commentators, Daniel and Richard Suskin, predicting the professions are going to contract in terms of numbers of people employed there? Uh, what is going to be the role in terms of the human labor market in the legal realm?

Speaker A: Yeah, I think this is very difficult to predict precisely how it's going to play out. And I think the articulation of the challenge, it's a real one. And I think now is the time for this conversation. This is part of the impetus for why Samuel and I authored Super Justice Now. Now is the time to put in place the guardrails, put in place through what we call algorithmic constitutionalism. The, the architecture that will ensure that human values are very relevant for the deployment of super justice in a way that prevents super injustice from emerging. Because you're right, at a certain point, if the systems, if the underlying technology becomes so powerful and those guardrails are action, the system is going to kind of, we can't predict precisely what will happen. The system will implicitly be making decisions, trade offs consistent with whatever its architecture is. And if we haven't done the work to ensure that we remain on a human centered path, it will be very difficult, if not impossible, to reel it back in and to reestablish that. So for those listeners who maybe don't have the intuition. I like chess analogies. So the chess analogy would be back in the 1970s and 1980s. No human grandmaster felt particularly troubled by chess software. The hardware wasn't powerful enough. The chess engines were not at all sophisticated. Any relatively strong human player could defeat any chess engine. This famously came to a head in the 1990s with the battles between IBM's Deep Blue and Gary Kasparov. And I think Gary is still salty today about losing the highly publicized matches with Deep Blue. And there's some controversy about exactly what was the boundaries of what was Deep Blue's reasoning and what were the human boundaries. I think with the benefit of almost 30 years of subsequent experience today, take someone like Magnus Carlsen, one of the strongest players ever to have lived. I think if we asked Magnus, Magnus, like, how do you fare against the top performing chess software on powerful hardware? I think he would admit, like, the best I could hope for is a draw. I cannot hope to beat these systems. That would be, that would not be a realistic hope. The best I can hope for is a draw. And so that's just, that's chess. It's a closed form optimization problem. They're a set of valid moves in any given position. There are clear ways to score the match. We know what criteria must be satisfied for a victory. We know what leads to a draw. Um, and the real world is not closed ended in the same way. It's not a closed form problem. But just take that setup. And if Magnus, Magnus is not in a position to second guess what that very powerful chess engine is going to recommend as its favorite move in a given position. Because Magnus would say, maybe that does match up with what I would suggest in the position. If he departs from the view, he would probably also say, yeah, that also appears as a rival to my favorite move. Um, he's not going to say, I confidently will override the system on that point. Because he would acknowledge there's a chance at seeing 40 moves ahead or 50 moves ahead, whereas I could only see 25 moves ahead using my experience and intuition and my internal model of chess. And so what happens when the best AI for law can see much deeper into all of the accumulated human experience, all the precedents doctrine, and is able to confidently recommend, no, this is the right resolution, this is the right policy, conditional on what you've told me about your preferences. And it would be accumulating preferences in an optimal way, given local preferences and going more broadly, I think it'll be beyond the scope of any human, any one human, to contradict that. Now, you can imagine different ways of policing at different safeguards that you can put in place. Auditing, having auditors audit other systems. And like how, for example, the accounting profession, there's, at least in the United States, I suspect it's similar elsewhere. There's a system of peer review amongst auditors, regulated auditors, and auditors audit auditors for adherence to the auditing standards. And so you can imagine a community of AIs auditing other AIs for adherence and certain disclosure obligations. I mean, with human justice systems, we have reason giving, judges must give reasons. And of course, those reasons are for public consumption. They may or may not accurately perfectly capture what led that judge's nervous system to a certain legal conclusion. And so we deal with all of these imperfections today in the legal system. But I think technology poses the possibility of magnifying them and raising the stakes for them. And so I don't have any pat answers other than it seems urgent to have this kind of conversation about this and to explore how do we build those guardrails into our existing legal system to ensure that it's robust to ever increasing machine intelligence.

Speaker C: If we assume that we are on the trajectory towards the super justice. What would be the role of education? Because this is one of the biggest questions in higher education. What are we supposed to teach students?

Speaker A: Armin, that's not an easy question to answer.

Speaker C: I know.

Speaker A: Um, I think. And you could ask any number of academics, I expect you'd get any number of answers to the question. It's one of the values of higher education is having free inquiry, academic freedom. The possibility of contesting any answer to that kind of question is, I think, an important value that higher education, um, advances, I believe. And certainly this was my experience as a law student at the University of Toronto. I think it's extremely important to focus on critical thinking, exploring different mental models, different ways of analyzing legal problems, social problems, economic problems. Um, I think there's. And so I think having a versatile approach to analyzing problems. It's a big upshot from legal education. I think the way that the University of Toronto teaches law and Samuel and I met at Yale Law School, I think this would also be true of the methods of legal education at Yale. It's very focused on multidisciplinarity. It's focused on thinking about things from multiple different angles. It's less about mastering the content of any particular doctrine, it's about mastering the tools of inquiry to challenge different proposed doctrinal solutions, to think creatively about different kinds of solutions and to be comfortable with the discomfort of hearing other points of view, respecting them, and finding within them the best version of those arguments, those positions, and then grappling with them honestly with the high degree of intellectual integrity. I think that's, that's, um, Personally, that's how I would answer the question. I also understand that other people would have different answers, competing answers. Professor Becker may have different answers. So I don't want to speak for him. And what he would say, I think in parallel. One thing that's become clear to me after 22 years in law school classrooms as a faculty member is there's also a process of intellectual maturation. I think that that comes through several years of academic study. So in North America, the usual thing that you would have an undergraduate degree. About 20 to 25% of incoming law students at the University of Toronto also have a graduate degree before beginning their law studies. Even still, like somebody, like many students grow up a lot intellectually during law school. And so humans are malleable, they're changing, they're plastic over the course of legally. And it's not just academically intellectually, it's socially as well. It's like contending with things like I was the most talented student at UH in high school, and then I was amongst the most talented students in my undergraduate class and maybe I graduated with the gold medal or the silver medal in my program. And now I'm surrounded by people who are a lot like me and how do I, like, how do I cope with that? If my identity has always been, I'm tops academically in my class, and now by definition the average student is going to be average and half of them will be below average. And how do, like, how do you grapple with that as a matter of personal identity? And so, uh, there's also that whole thing that people go through as law students and we've had a force, it's an anthropological fact that happens as part of legal education, at least at the schools I've attended. That's been a common thing. No, I think that's somewhat orthogonal to this whole conversation about AI, But I think it's a very healthy thing. Um, yes, you're talented. How do you deploy, how do you develop those talents, how do you cultivate those talents and then how do you harness them for public good and private well being for you and for your family, for your clients, for your firm, wherever you end up? Um, there are a lot of things happening over the course of a legal education that are not just mastering contract law Doctrine or family law legislation or what have you, legal theory, whatever, whatever really you gravitate towards in law school. There's much more happening to our law students over the course of that education and I think a lot of it is extremely valuable and should be protected. So it's a long winded non answer to your question, which is to say, I don't really know. I think there are a lot of things that are really valuable. I think humans, uh, need to, in the future, I think today need to learn how to adapt to new technologies. And how do you work with these new technologies? At Bluejay, we're making our technology available to tax law professors, tax accounting professors, free of charge. So if there are any academics who teach law classes, teach tax law classes, any listeners who are interested in access in North America or in the UK, you can go to bluejay.com academic access and we will, there's a form you can fill out, very simple. And we will get you access to Bluejay for use in your classroom because we believe that students should have access to the best tools to allow them to teach themselves what the law requires, give them the tools to interrogate the law. And just in the same way I said I love chess software analogies, here's one which is, you know, one of the best ways to become a better chess player is by leveraging chess software to get an analysis of a position. So you play a game against an opponent who's calibrated to your level of play, and then you run it through an engine and you get annotations and you can figure out where you blundered, why you blundered, look at different possible variations of the moves that you actually played, and then you get better. As a law student, I would have loved to have something like Blue Jay. As a tax student, I'm a tax law professor. I got there, I figured I did the work. I believe I would have learned more quickly if I had access to something like Blue Jay. To accelerate the pace at which I was able to explore tax law and ask the questions that I would inevitably frustrate my professors with in office hours or after class or at the break, I'd go, and what about this? What about this? What about this? And you know, they did their best to help me figure out these answers. I didn't expect them to have all the answers, but I knew that they could point me in a close, like approximately where I should look for, um, doing more of my own work. I wasn't afraid of the hard work, but If I had one of these systems so that after hours, in the wee hours of the morning when I was reading this stuff, I could have poked around and accelerated my learning, I would have been way better off. And so we want to make that available to students today. So that's one thing I'm sure of is that we want to make these tools available to students. It's very interesting. Not everybody agrees with this. So, for example, I think recently at the University of Chicago, they announced that the entire 1L curriculum will be computer free, AI free, I think even like iPhone free, like smartphone free in the first year. And I'm like, okay, that's a different commitment to this. And I think, see, there are going to be a number of experiments about this over the next years and we'll see what emerges with respect to the use of AI and other technologies in the law school classroom.

Speaker C: I think you made very interesting points about the intellectual and social maturity. And that's something I believe is not as much valued in universities anymore because of. I don't know about the, the US and North America in general, but in Australia we have a lot of online degrees or online classes and the students find them much more comfortable. But in my opinion, you're missing out on something really important that for me, the campus life, university life was a big deal, was a big experience, and I, um, learned a lot about exactly the social maturity that you were talking about and intellectual maturity, interactions with others. And I think, uh, in the age of AI, that's something that is more important than before. And yeah, we are not talking about it enough and we are not encouraging universities to hold onto that.

Speaker B: There's this book that's coming out from Daniel Suskind, which is called what Should My Children Do? I think that's the title and it's on everyone's mind who have got, uh, kids or worried about the future generation. The last question I have is, you know, what is the lawyer of the future going to look like? Do you have a vision? Are, ah, they even called a lawyer? What's the same about what this person is doing? Is that something you've thought about?

Speaker A: Yeah, I think it's difficult to make very confident predictions about this. I think it will be true that we will still have people whom we refer to as lawyers. I think it's important that we have dispute resolution processes. I think it's important that they, that there are humans in charge of these systems. I think lawyers play an incredibly important informational role, of course, in representing clients today. I think one of the things that, that has been lurking on the margins of this conversation is also that lawyers play a very important social role vis a vis the clients and supporting them, giving the clients confidence that there's somebody who does understand the system is on my side, will listen to me, will explain what's happening to me, has my back, is my representative in these proceedings, whatever the situation, the legal situation may be, and is like a layer of a friction reducing layer. Right. And so maybe a lubricant for individuals who are subject to legal processes, whether they're criminal or whether it's the family law context, um, or elsewhere. And it's really important and clients I think will be loath to lose the equivalent of that lubricant in, especially in situations that involve life and liberty. So, uh, in family law disputes in the context of criminal proceedings, insofar as criminal law is still a thing, we need a whole other podcast to talk about the future of criminal law. I have my views on that. I think we're going to see the perpetuation of lawyers for a lot of those social reasons for the client, well, being kind of dimension of it. Having that person who is the master of the technology, if nothing else, master of the system within which that technology is operating, having that advocate is really valuable. Having that person who can vouch for you and can represent you optimally and give you confidence about the system. I also returning back to the initial example that we were talking about, the neighborly dispute. There may well be things where you don't need an advocate because the system itself is able to reduce the friction, reduce the heat sufficiently, but in really high stakes circumstances, I think we'll continue to see lawyers. There's also a more, I guess, cynical answer to the question, which is we have self regulating professions. Lawyers are self regulating. I don't see lawyers banding together and deciding let's get rid of lawyers. This has all been a very nice thing that we've done for many decades or centuries. But you know what, let's wrap things up here. Let's not do this any longer. Uh, I don't see that happening. So I see a very high likelihood that things do evolve, that there are different interests involved and the public interest is going to evolve alongside the technology. There will be people we refer to as lawyers in the future, if only because, and this is the cynical motivation, if only because the self regulating bodies of people will make sure that there are still lawyers absent a very strong populist surge that reminiscent of Shakespeare here. Let's not have lawyers anymore, but I don't see that.

Speaker C: So thank you so much, Ben for coming to our podcast and having this amazing and interesting discussion about justice and how it could look in the future and the importance of having these kind of conversations to make a decision, how we want the justice system to look like in the future, when AI will be a big part of it, and making it more humane and not letting it just to happen, especially not letting private companies like OpenAI and Anthropic to make those decisions for us, I think is very important. Thank you for your time.

Speaker A: Thanks Armen. Thanks Mark.