AI, the Unauthorized Practice of Law, and the Future of Legal Fees

Episode Notes

Transcript

Damien Riehl [00:00:00 – 00:00:35]

Right now, if you have a dispute that’s less than $75,000, good luck getting that resolved because no lawyer is going to take that, even on contingency. But if you can have a bot decide that in 20 minutes without all the legal fees, maybe you could bring that down to $50,000 or $25,000 or even $10,000. And that can, for consumers, that’s the difference between, you know, $75,000 whether you live in a house or not. Everything that we as lawyers do is based on words. We ingest words, we analyze words, and we output words. And it turns out that AI does all 3 of those things at superhuman speed and a postgraduate level. It’s good for society, access to justice.

Phillip Vaden [00:00:36 – 00:01:17]

AI isn’t just coming to the legal space. AI is already here. But what does that mean for legal professionals? What does that mean for my fee structures? What does that mean for my daily life? Am I going to be billing by the hour? Am I going to have a flat rate? Is there going to be a subscription model? How am I going to handle this? Well, we’re going to talk about that and a lot more today with Damien Riehl. He’s, uh, one of those rare people who genuinely lives at the intersection of law and technology, not just as a spectator, but as a builder. Damien is a lawyer and a technologist with experience in complex litigation, digital forensics, and software development. He’s been coding since 1985. The guy’s amazing. All right.

Phillip Vaden [00:01:18 – 00:01:56]

Today he works at Clio, where he oversees the integration of large language models with a corpus of over a billion legal documents to provide deterministic, reliable AI solutions to global firms. His work doesn’t just stop there though. He chairs the Minnesota State Bar Association’s AI Committee, which oversees an AI sandbox specifically designed to promote access to justice. And he does so much more. The guy is a polymath and he is truly amazing. I welcome him to the show, and I know you’re going to learn a lot from him because I sure know I did. I’m Philip Vaden, partner manager at InfoTrack. This is the AI Sidebar.

Phillip Vaden [00:02:05 – 00:02:06]

Damien, welcome to the show.

Damien Riehl [00:02:07 – 00:02:08]

Thanks so much, Philip. Really thrilled to be here.

Phillip Vaden [00:02:08 – 00:02:21]

So I want to pick your brain about a lot of different topics, and we’ll see how much we can cram in. What we can’t cram in, I will, I will be begging for you to come back and talk more about it because you know so much about so many topics.

Damien Riehl [00:02:22 – 00:02:22]

That sounds great.

Phillip Vaden [00:02:23 – 00:02:51]

Yeah, the first thing I’d like to touch on is your work with the Minnesota State Bar Association’s AI committee. And specifically around, this is a question I’ve gotten quite a bit from legal practitioners, is around their ethical responsibilities with AI use and what they need, how they can use it ethically and what that means for them moving forward. So you You’re an expert in many things, and this seems like one of them.

Damien Riehl [00:02:51 – 00:03:56]

Yeah. So let’s think back to January of 2023. And, you know, ChatGPT had just come out a few months earlier. Everyone was freaking out, just saying, oh, what does this mean for the practice of law? And so right around that time, the Minnesota State Bar Association has been saying, oh, we probably need to think about whether the AI is committing the unauthorized practice of law. Because, you know, ask ChatGPT a legal question, and it would be happy to give you an answer. So anyway, so I saw, and right around this time too, back in January of 2023, a few bar associations have come— came out with some knee-jerk reactions saying, yes, you are prohibited from using AI in your practice, and gave what now in ’26 seems like a very embarrassing group of proclamations, and they didn’t age very well. So in that mindset, when the Minnesota State Bar said, we need to think about the unauthorized practice of law and the extent to which AI is doing it, I raised my hand to maybe say, I’d love to chair this so I wouldn’t be embarrassed by my home state. And I don’t want them to be essentially like, those other states that I won’t name that have now embarrassing proclamations.

Damien Riehl [00:03:56 – 00:05:04]

So they then asked me to be the chair of this committee. And I quickly said, yes, we should be thinking about whether it’s committing the unauthorized practice of law. But I also said, to what extent can AI help the access to justice gap, the 92% of legal needs that are unmet because we lawyers are too expensive? Because yes, there are maybe harms, but there are also great good that we can do, because we have not been able to pro bono our way out of that access to justice gap. And maybe AI could be a way to be able to help that gap. So since 2023, our committee has been thinking about both the pitfalls, that is, the extent to which unauthorized practice of law, and maybe ethical problems that might come about, hallucinated cases, et cetera, but also the benefits of the AIs and the ways that it can help both the practice of law for lawyers and also the consumers of law. That is the people that legal aid helps. So we’ve been trying to balance both of those things, both the benefits and the detriments. And by the way, that’s what the ethics rules say, that they rule— comment A to rule 1.1, I think it says that we as lawyers, our duty is to be competent both in the technology’s benefits and maybe pitfalls.

Damien Riehl [00:05:05 – 00:05:07]

And so we’re really doing both of those things with the Minnesota Bar.

Phillip Vaden [00:05:08 – 00:05:52]

Oh, that’s great. And one of the things that you’re also— you’re a builder too. I mean, you’re building all kinds of applications for various things. Things involving legal technology and with the Catholic Church and all kinds of great things. One of the things I was curious about with access to justice, that focus, is how much you guys are looking at leveraging the ability to build quickly and for laypeople or sort of pseudo-coders to actually put out products to be able to help even accelerate groups like Legal Aid and provide them with additional tools to be able to represent more people And, you know, kind of maximize their time. I just was curious what you’d seen on that front.

Damien Riehl [00:05:52 – 00:06:24]

Yeah, I think that everybody in the world has a build versus buy question. Build versus buy is a question that law firms have to think about. Like, we can build a lot of things as law firms. Legal aid can now build a lot of things as legal aid. And corporations can say, we as a law department can build a lot of things ourselves. So everyone has a build versus buy question. And the question is what, and my friend Oz from Oz Ben-Arim says that it is both build and buy and rent. There’s some things that you’re always going to want to rent.

Damien Riehl [00:06:24 – 00:06:59]

You’re not going to be able to build, for example, yesterday’s case from this obscure jurisdiction or yesterday’s statute or yesterday’s regulation. Those things are very hard to build. So you’re probably always going to want to rent that. And so really, whether you’re a legal aid organization for access to justice, or whether you’re a consumer, that is maybe a consumer of legal aid, you should be always saying, okay, what are the things that I can build myself? And what are the things that I cannot build that I’m going to need to rent and pull from others? And so I think that everyone, law firms, legal aid, and consumers, all are thinking about build versus buy.

Phillip Vaden [00:07:00 – 00:08:15]

Oh, that’s great. I guess from another angle too, just thinking through The tools, I’ve seen several tools that are coming out, some that are a little bit more established that are specific products. And I guess this would be in the rent and buy categories, groups like Courtroom 5, and I’ve seen one pop up called Pro Se AI, different products like that. And this has been, and you were discussing this earlier about the unauthorized practice of law and kind of walking that line of— you know, where do we cross the line? And I’d like to hear some of your thoughts on that. Like, you know, how do we avoid crossing the line with some of this technology into UPL or unauthorized practice of law? And where is it, where’s the safe space as people are thinking of building those tools? Because I know that there’s a lot of lawyers out there that are, you know, going out and vibe coding, you know, like you and I are doing. And if they see an opening there, like, hey, people are asking me to do these cases pro bono, I just don’t have the time, but I could build a tool over a couple of weekends that may help them with landlord-tenant cases or something like that. What would be the advice you would give them for avoiding crossing over into that unauthorized practice of law?

Damien Riehl [00:08:15 – 00:09:24]

So yeah, I think everyone has been saying, you know, if it’s the unauthorized practice of law, you need to take it down. Because the question is, how are you going to take down OpenAI? Or how are you going to take down ChatGPT or Gemini or any of these? And what is the difference between legal information and legal advice? And has AI just gotten rid of this line? Because legal information historically has been, well, the cases, the statutes, and the regulations, that’s legal information. But if you try to take your client’s facts and apply them to those cases, statutes, and regs, now that is the legal advice. But it turns out that AI is happy to apply your particular facts to the cases, statutes, and regulations. So this line between legal advice and legal information is pretty much obliterated. And if you think about historically, we’ve been thinking about treatises. Are those legal advice or legal information? They’re pretty clearly legal information, right? They’re commentaries on where the law is going. But if you take your particular facts and say, use a search engine, say Google, and be able to say, of this 1,000-page treatise, here are the 3 pages that apply to your facts, does that flip it like some kind of quantum particle from legal information to legal advice? No, that’s still legal information.

Damien Riehl [00:09:24 – 00:10:10]

So is it any different if you go to Google and say, what are the elements of breach of contract under New York law? Is this legal information or legal advice? Pretty clearly legal information. Everyone should have access to the cases, statutes, and regulations that we’re at. Then if you ask the AI part of Google, what are the elements? And it gives you the elements of breach of contract in New York. Is this legal information or legal advice? Still legal information, right? Because especially if you put a disclaimer saying, this isn’t legal advice, consult your lawyer. And is Richard Poo creating an attorney-client relationship by creating legal advice on his website? And the answer is no, of course not. So really this line between legal information and legal advice is just going away. And it turns out that the big folks don’t care about the unauthorized practice of law. Google is happy to be able to give you legal advice.

Damien Riehl [00:10:11 – 00:11:09]

OpenAI is happy to give you legal advice. You know, there’s all sorts of AIs that are happy to give you 200 pages worth of cases, statutes, and regulations. And in this way, is this really any different than treatises? And the answer is probably not. It’s just taking what’s on the internet and giving you just the particular legal information that applies to your particular facts. And so then this is now all along windup to answering your question is how can you be able to, as a legal aid provider, be able to not have to worry about giving legal advice, but only giving legal information? And the way you can do that is maybe turn a question for legal advice into legal information. And the way you could do that is, you know, I’m going to play a game with you. And for our listeners. And the game is, is it legal information or is it legal advice? And so number 1, is marijuana legal in my state? Is that a question asking for legal information or legal advice? And I would say that that’s information, right? Because it’s just, you know, the cases, statutes, and regulations say whether it’s legal or not.

Damien Riehl [00:11:10 – 00:11:57]

But then, can I possess marijuana? That is my facts, right? That’s a little trickier, right? Can she, can I tell her whether she can possess marijuana? That’s also trickier. Now, should I possess marijuana? Now, this is pretty clearly legal advice, whether you should or should not do something. So now let’s imagine if the legal aid system says, I know that you as a user have asked 2, 3, and 4, but I’m not going to answer those. I’m just going to give you legal information. Now, if you turn that legal advice question into a legal information saying, I’m not going to give you legal advice, but here are the cases, here are the statutes, here are the regulations, and here’s what those cases, statutes, and regulations say. You’re not giving legal advice. You’re giving legal information. So that’s way number 1 you can get around it.

Damien Riehl [00:11:58 – 00:12:21]

Way number 2 is through empathy. Because you could, if you go to a doctor, doctors have been trained to say, when I ask my doctor, doc, what should I do? What should I do? What’s my medical advice? The doctor will say, well, I’m not going to tell you what you should do, but if I were in your position, here’s what I would do. And the reason for that is because if the doctor tells you what you should do and then bad things happen, if you die, they could have a malpractice lawsuit, right?

Phillip Vaden [00:12:21 – 00:12:22]

Yeah.

Damien Riehl [00:12:22 – 00:12:41]

So the way that they avoid that is to say, well, this is what I would do because what I would do is not advice. That’s just empathy. And so you can imagine an AI doing the same thing. I’m sorry to hear that this is happening to you. I, as an AI, would consider doing this. I, as an AI, if you’d like more legal information, I’m happy to give you legal information. So number 2 is empathy. Number 3 is just giving statistics.

Damien Riehl [00:12:42 – 00:13:11]

Statistics is not advice. In these situations, 80% have this result, or with your facts, 75% are dismissed. But if you have this other fact, then dismissal is only 10%. This is not legal advice. This is legal information. Statistics are not advice. So if we think about what if all the systems were that are, whenever you’re asked legal advice, the system just turns that into legal information saying, here are the cases, statutes, and regulations information. Here’s empathy, here’s what I would do, and here’s statistics, none of which are legal advice.

Phillip Vaden [00:13:12 – 00:14:11]

Thank you for that. Yeah, that, that’s, uh, you made something that, uh, I feel like a lot of people are wrestling with, uh, really clear. Uh, and then in terms of, of legal professionals, we’ll pivot back to, to them for a minute and, and talk through, um, you know, as we, as we think about the, their ethical requirements, attorneys’ ethical requirements. I’m a, I’m a still licensed attorney, former practicing attorney. To stay abreast of what’s going on, what do you feel like, and is there something coming out, is Minnesota maybe leading the way on this saying, hey, for your continuing legal education credits, you know, I think every state has a certain number of ethics credits that you have to do. And then I’ve seen several states, you know, moving towards technology-specific continuing legal education credits that they have to have. Do you see that kind of zeroing in even on, hey, here’s, you have, do a technology class, but specifically around artificial intelligence use and how it’s done responsibly and ethically and all of that.

Damien Riehl [00:14:11 – 00:14:50]

Yeah, I think continuing legal education is having a bit of a moment because, you know, where 10 years ago you’d read the newspaper while somebody talks about, you know, what happened 6 months ago in employment law or patent law, right? That has now gone to the existential question is where is the legal profession going to be as AI becomes prevalent with every single consumer that is— and every single corporation is bringing a lot of this work that they used to spend on external lawyers and bringing it in-house. And now, you know, I just talked to a Fortune 20 company that said that they’ve already built 160 different workflows that they used to spend on law firms that they brought in-house. And now they don’t spend that on law firms anymore.

Phillip Vaden [00:14:50 – 00:14:50]

Oh, wow.

Damien Riehl [00:14:50 – 00:15:45]

And so I think CLEs have kind of turned from, you know, yawn, let’s think about the updates to the law, to my God, where is our profession going and what’s going to be happening? So as the AI becomes the tip of the tongue of every CLE on the planet. And with just cause, because everything that we as lawyers do is based on words. We ingest words, we analyze words, and we output words. And it turns out that AI does all 3 of those things at superhuman speed and a postgraduate level. So as every AI, every CLE talks about AI, justifiably, it goes from a, you know, what’s happening with the law to what’s happening to the business. And so I think that when you say, okay, what is the CLE that is talking about both the practice of law and the business of law? What are you going to talk about? I think that that is one thing to talk about it in a classroom setting. Another is to do it yourself. That is to be able to do the AI.

Damien Riehl [00:15:45 – 00:16:13]

And it’s almost like saying, what’s the best CLE to take to learn how to swim? And there’s no CLE. You can just, you swim. Right? And it’s only by swimming that you can actually do the thing. So in the same way, like what’s the best CLE you could take to learn about AI? And there is none. You just do it. You use the chatbots. And then when you see the failings with that, then you use the agentic AI that is there. And if when those don’t work, you use the law built for legal and you do that too.

Damien Riehl [00:16:13 – 00:16:23]

So really, I think doing is much better than learning about. And so I think learning about maybe is a good entryway into doing, But people should be doing a lot more.

Phillip Vaden [00:16:23 – 00:17:19]

So I’ve spoken with several practicing attorneys that I practiced law with or went to law school with over the last few months, just ’cause I was curious. I was seeing statistics come out saying, hey, you know, 78% or something like that of legal professionals have access to AI tools. And that always struck me as kind of like, I don’t know if you’re an Office fan, but there’s one episode where Michael Scott is asked if he’s read this book and he says, read it, I own it. And it’s like, well, that didn’t answer the question. And so I’ve wondered if, Hey, you have access to it. Are you actually using it? And so going in and talking to the attorneys, it seemed like the divide I was seeing was for transactional attorneys, they were leaning on it already at a pretty common, like it was a pretty high rate. The litigators seemed pretty hesitant. And some of it was around fear of mistakes that were made.

Phillip Vaden [00:17:20 – 00:18:23]

with, with AI platforms, you know, years ago, um, that became big news stories. Some of them are still happening where people aren’t even, they’re not putting a human in the loop. They’re just going to trust that magically their case has a bunch of great citations that are directly on point, which is like not very common whenever you’re, you’re drafting your, your, uh, your briefs and motions and everything. But the, um, uh, yeah, but I, I’m, I’m wondering if, if there’s a specific task that for, for an attorney, for maybe even specifically a litigator, if they’re, if they’re kind of gun-shy about using artificial intelligence generally, if there is a specific, uh, task you would say, hey, this is sort of your, your, you know, gateway into a broader use of AI, um, whether that’s, you know, setting up timelines or just, just drafting the first, you know, first, first section, you know, for first internal brief on a particular topic, anything like that. But I don’t know if you had any sort of like, this is a good sort of like gateway AI drug or whatever for you guys to go and sort of dip your toe in the water to learn?

Damien Riehl [00:18:24 – 00:19:19]

Yeah, I think what you’ve just asked really hits on the head of what it is, what does it mean to be a lawyer? Because really, when you think about legal advice, that is applying facts to the law, right? But a lot of legal work, to be frank, doesn’t need the law. That is, if you’re doing a transaction, if you’re doing an M&A deal, how much are you really referencing yesterday’s case? Yesterday’s statute and yesterday’s regulation versus what are the business needs of the business, my client, and the business needs of the counterparty. And then of course, we all know contract law, right? That is pretty well settled. And so we’re really not applying yesterday’s case for contract law as much as we are trying to effectuate the business deal to go forward. So I think transactional lawyers have learned quite well because they don’t need yesterday’s case, yesterday’s statute, and yesterday’s regulation. ChatGPT out of the box does quite well. Claude does quite well. And you could put what’s called harnesses around those foundational models.

Damien Riehl [00:19:19 – 00:20:12]

And that’s what Harvey, Lagora, and we as Clio provide a lot of harnesses around the workflows that are needed to be able to do a particular workflow. But then of course, you can also, there’s now open source competitors called Micah OSS, and there’s Laverne AI that also put a lot of harnesses around these things. So really, transactional lawyers are saying, gosh, we can do a lot on the build versus buy spectrum. We could just build a lot, which is, I think, why transactional lawyers are dipping their toes in much more than, say, litigators, as you say. Whereas litigators, you know, we’ve all heard about the now over 1,000 cases where lawyers have cited hallucinated cases. And so there’s a gun-shyness, and the litigators are like, gosh, I don’t want to be that lawyer who’s on the New York Times or this 1,000 list. But there’s a way to avoid that. And the way you can avoid that is to read the damn cases.

Damien Riehl [00:20:14 – 00:20:15]

That is an easy way.

Phillip Vaden [00:20:15 – 00:20:15]

Yeah.

Damien Riehl [00:20:15 – 00:20:59]

And I clerked for the chief judge of the federal district court and the chief judge of the state appellate court. And I can’t tell you how many times I read briefs where the proposition and the case itself don’t match up, that people cited cases for the wrong proposition all the time. And that’s just bad lawyering. And so right now, what I think with hallucinated cases, that’s just bringing to the fore all of the bad lawyering that we’ve always had for decades. But if you read the case, if you do go as easy as clicking on the thing and realizing, oh, that case doesn’t exist, really, if you cite a hallucinated case, that’s just showing what a bad lawyer you are. So this is not an AI problem. This is a bad lawyering problem that we’ve had for decades and is now just brought to the fore. So that’s point number 1 with litigators.

Damien Riehl [00:21:00 – 00:21:10]

Point number 2 with litigators is you need to have a tool that has yesterday’s case, yesterday’s statute, and yesterday’s regulation. And if you have that, then the rate of hallucination is actually less than humans.

Phillip Vaden [00:21:10 – 00:21:10]

Wow.

Damien Riehl [00:21:11 – 00:21:47]

There’s a double-blind study where they said, hey, humans summarize the document and then machines summarize the document. They mixed up the results and then they gave them to humans to review. And it turns out that humans hallucinated more than the machines. That is, the humans made more factual errors than the machines did. And by the way, that was back in 2023. And over the last few years, the AIs have gotten much, much better. But how much better have the humans gotten? So if you have yesterday’s case, yesterday’s statute, and yesterday’s regulation, and you ask the AIs to be able to reason with them, it is going to hallucinate less than that first-year associate or that fifth-year associate or even that junior partner.

Phillip Vaden [00:21:48 – 00:22:00]

Wow. That is amazing. Yeah. Yeah, seeing those, those cases come through and like, you guys, you didn’t shepherdize, you didn’t key cite. Well, you didn’t even read it. You didn’t even read it. Yeah. Yeah.

Phillip Vaden [00:22:00 – 00:22:10]

It seems like an easy first step. And, and it is. It’s funny because it’s, you know, it’s like, well, you know, I could easily, you know, just as easily fall into that trap. It’s like, only if you don’t do your job.

Damien Riehl [00:22:11 – 00:22:12]

So that’s right. Read the damn case.

Phillip Vaden [00:22:13 – 00:23:07]

Yeah. Yeah. Yeah. And then one of the things you mentioned was, was these, these, uh, sort of legal-specific models that are, that are coming out. And, uh, we saw in the news, uh, fairly recently in these last couple of weeks that, uh, that Kirkland is, is spending $500 million, uh, on building their own sort of Kirkland-specific version of like Harvey Legora tool. And so that they’re, they’re not going to put out in the world. How much do you think this is going to happen with the, the top law firms that they’re basically going to say, let’s try to capture all of the knowledge we have today in our firm, which is substantial, and, and then, uh, but then hold that close? And do you think that that’s going to be able to be competitive versus, versus, you know, um, Clio’s product, Harvey, Legora, these, these different ones that are pulling in from different firms around the world? Curious as to where you see this heading.

Damien Riehl [00:23:07 – 00:24:24]

Yeah, I think that, uh, this is going to be, uh, one of the biggest challenges to law firms, uh, going forward. Is to be able to say that a lot of the moat that any law firm has is in the brains of its lawyers. That is, there’s a— I heard from someone say from a law firm that knowledge, our value as a law firm walks out the door every single day. Every time somebody leaves as a lateral to go to another law firm, that’s stuff that’s stuck in that lawyer’s brain that can’t be used anymore by this law firm. So the real question is, as as law firms seek to keep that value, how do they keep that value? How do they take the things from that lawyer’s brain and put it into an artifact that is now able to be kept by the law firm, whether it be Kirkland& Ellis or any solo small law firm? How do you keep that knowledge from walking out the door every single day? Because once you take it from the lawyer’s brain and put it into an artifact, then you can use the AI to be able to do it. So really, Every lawyer loves the word, it depends. That is, and really, what if any law firm would collect all of the it depends and be able to say, okay, well, in this kind of deal, it depends, go this way, but in that kind of deal, it goes that way. Or in this kind of litigation, it goes this way, but this judge hates it when you do that.

Damien Riehl [00:24:24 – 00:25:22]

What if you collect all of that stuff from the lawyer’s brains and put it into an artifact? So when that lawyer leaves, then the firm still retains all of the it depends. And you’re counting what has previously been uncounted. But then the real question is, if you’re a firm, you say to your law firm, hey, put your brain into this knowledge base because I want to keep the value. But the lawyer says, wait, I’m thinking about jumping ship to your competitor. And my brain is my moat. Why would I give up my moat to you for free? Especially when I’m thinking about jumping ship to the other because I’m giving you value. By putting my brain into your artifact, law firm. So how are you going to exchange that value? Are you going to pay me for this exchange of value? And this question is going to be the question that successful law firms have to think about, is that as we think about, we need to retain our competitive advantage, that is our moat as a law firm, we need to make sure that our knowledge doesn’t walk out the door as people lateral out.

Damien Riehl [00:25:23 – 00:25:35]

Then if that’s true, how do I exchange value to that lawyer to say, take your valuable brain, All of the, it depends, and put that into our corpus of knowledge. What’s the value exchanged in money for that?

Phillip Vaden [00:25:36 – 00:25:36]

Yeah.

Damien Riehl [00:25:36 – 00:25:37]

Yeah.

Phillip Vaden [00:25:37 – 00:25:49]

And I can see the employment contracts probably changing too. Like, hey, by getting paid by us biweekly, we’re paying you for that, for your brain and everything you’re putting out. Maybe.

Damien Riehl [00:25:49 – 00:26:10]

But even then, how does a deal go well or poorly? That’s not proprietary to this law firm. that’s something that I can take with me to my next law firm, right? And this is just basic competition. So yeah, I think that what stays proprietary and what is portable with the lawyer, that’s going to be the question over the next few years. Yeah.

Phillip Vaden [00:26:10 – 00:27:04]

And I did have several attorneys that I was just having lunch with, asking questions about their use of AI, feelings about AI. And several of them expressed some, like, you know, I mean, they did it in kind of a joking way, but there was a seriousness behind it. about, you know, hey, is this going to, like, I think, is this thing going to replace me in my, my job? And I can’t imagine that that’s on the horizon. This is my own personal point of view. I can’t imagine that’s on the horizon just because, um, people need, number one, I think that companies in a lot of situations, they’re paying the attorneys to be responsible for the case. And it’s going to be a lot harder to hold, uh, an AI agent responsible. Particularly one that’s an external one because it’s like, well, you know, Claude’s not selling itself as your lawyer. It’s like, hey, I can answer some questions about these things.

Phillip Vaden [00:27:04 – 00:27:29]

So I think some of it’s just about responsibility. And then you’re, you’re going to, if particularly for litigators, you got to go to court and we don’t have, you know, holograms showed up in court. So I think that, I think some of those things are going to, are going to be pretty far out. But I imagine you’ve thought about this and probably more than me. So I’d like to hear your thoughts on where you see that. Coming over the next 5 years, 10 years?

Damien Riehl [00:27:29 – 00:28:18]

Yeah, I think that we as humans tend to focus on things like scarcity. That is, is AI going to be like a tsunami that we can all run in front of for a while, but it’s just eventually going to crash over everybody, right? But I think we’re kind of, I vacillate between scarcity and what I’m now on team abundance, that maybe we’re going to have more work than we ever have, not in spite of AI, but because of AI. So this Parkinson’s Law, we all know the concept, even if we don’t know the words. If you give yourself a month to do something, it’s going to take a month. But if you give yourself 2 weeks, it’s going to take 2 weeks, right? So work expands or contracts based on the time allotted. And I think legal work is the same. My friend Darth from Ford Motor Company, Darth Vaughn, he said, hey, do you realize how much I’m regulated and how much work I’m not giving you because you’re so expensive? I have plenty of work. And number one, you’re going to use AI.

Damien Riehl [00:28:18 – 00:28:51]

If you don’t, You’re fired. And number 2, because you’re using AI, I’m going to give you way more work. All the work that I haven’t given you in the past, this I’m going to give you. So this is called Jevons paradox. We thought LED light bulbs were going to save us a lot of energy, but they’re so cheap, we just leave them on all the time. Light bulbs are too cheap to measure. So what if legal work were too cheap to measure? What if every time, every time I call my lawyer, gosh, it’s $1,000, forget it, I’m going to risk it. What if $1,000 shrinks to $100? Maybe I’ll call you every day and maybe I’ll call you 15, 20 times and then I’ll spend actually more money than the original $1,000.

Damien Riehl [00:28:51 – 00:29:25]

So really this Jevons paradox is kind of like the 1980s AI. We had amazing AI in the 1980s. You know what it was called? It was called the spreadsheet. And this amazing AI, you know who it freaked out? Is all of the accountants that used to take a week to do this ledger. Accountants said, you know, this AI is going to put us out of work. Because it can do in minutes what we took a week. But the clients, when they figured out, wait, it’s not going to take a week, it’s going to take an hour, they say, run scenario 2, run scenario 20. And now we have more accountants than we ever have, not in spite of the spreadsheet, but because of the spreadsheet.

Damien Riehl [00:29:26 – 00:30:00]

This is Jevons paradox at work. So maybe we’re going to be at team abundance. So that’s point number 1 is that Jevons paradox might be an answer to these people worried about maybe this is the AI is going to go away. Point number 2 is that there’s still going to be taste that lawyers have that the AI will not. And this is best explained through an actual example where I’m here at my house in Minnesota. My next-door neighbor wants to build a fence literally 6 inches from my house. So while I was giving a talk to the Arizona Supreme Court, they said, yep, we’re going to break ground tomorrow. And so I vibe-coded a complaint saying, no, there’s adverse possession.

Damien Riehl [00:30:00 – 00:30:41]

I have an adverse possession claim. I also have an easement. And here are all the ways, reasons why you need to be able to work with me To be able to make sure that the fence doesn’t go 6 inches from my house. So anyway, so this is now, that was about last October. It’s now come to a head in the spring and I ended up filing a lawsuit against my next door neighbor. And I used my tool that I’m building with my team, Vincent, and I had about 5 different versions of the complaint and I got about 10 different versions of the temporary restraining order that is to keep them from building the fence the next day. So even though I have access to this amazing AI tool, and even though I litigated for about 15 years, I still have a lawyer. And my lawyer’s friend, his name is Jake.

Damien Riehl [00:30:41 – 00:31:15]

He’s been a good friend for 25 years, and he’s a great lawyer. And so even though I had access to this AI, he said, Damien, file the complaint, but don’t file a temporary restraining order. I said, wait, I’ve got a gun. Why wouldn’t I shoot it? I’ve already vibe-coded a bunch of versions of this. It’s ready to go. He said, no, no, no, no. He said, if you don’t want the fence to be done, the TRO, you have to jump through a bunch of hurdles. And something the court’s going to say is like, hey, is this really irreparable harm or can you just tear out the fence later on? So what if you lose the temporary restraining order? For the rest of litigation, you’re going to be the guy that the judge ruled against at the beginning.

Phillip Vaden [00:31:16 – 00:31:16]

Oh yeah.

Damien Riehl [00:31:16 – 00:31:22]

So, but if you just file the complaint, what are the odds of your neighbor actually building the fence in the middle of an active lawsuit?

Phillip Vaden [00:31:23 – 00:31:23]

Yeah.

Damien Riehl [00:31:23 – 00:31:52]

Next to none, right? So you get the results you want without having to have the risk of losing the TRO. Now, that is stuff that’s stuck in Jake’s brain. That is, I’m a litigator for 15 years and have access to all the cases, statutes, and regulations. And I feel like I’m a pretty good lawyer with pretty good AI, but I still didn’t know that insight from Jake. And so really the question is, how much of those insights like that are we going to continue to provide to consumers? and corporations that are going to be worth Jake’s paying Jake money and everybody like him.

Phillip Vaden [00:31:53 – 00:32:43]

Yeah, that’s great. Yeah. And that is, uh, that, that’s the thing that, that always felt good when I was practicing law to be able to, to give somebody some really good strategic advice that it’s just, it’s just from experience that you, you kind of get, get down to that, that level. Um, yeah. And then, so you mentioned, um, uh, you talked a little bit about the, the pricing. It’s like, hey, if it’s $1,000, I can’t, you know, $100, Let’s talk. Um, I, I did wonder about the, the different fee structures, and I know there’s been talk, you know, for 15, 20 years about going to kind of a flat fee structure, uh, versus the, the standard billable hour rate. Um, and I’ve wondered if there’s, if there’s another one, and, and there’s, this has been around for a while, is sort of a subscription model for, uh, for legal, um, assistance.

Phillip Vaden [00:32:43 – 00:33:14]

And I’ve wondered if, if that may actually end up getting supercharged with, with even mid-sized law firms. Um, saying, hey, you know, you have, uh, a list of disputes that come up. If you pay us $100 a month, uh, anything like that comes up, we’ll, we’ll take care of it. Have you seen anything like that expanding or is, is it, uh, is it still kind of delineated between sort of, you know, hourly fee structure or the, um, you know, or, or the, uh, the, the standard kind of flat rate for, uh, for the, for a case?

Damien Riehl [00:33:14 – 00:33:58]

Yeah, I think that the billable hour, we’ve talked about the death of the billable hour for decades. And so far that death has been greatly exaggerated and I don’t think billable hour is going to die. But what I think is going to happen is the current 90% billable hour is going to— and 10% flat fee, subscription fee, those kind of fees— that 10% is going to grow to 20%, to 30%, to 40%, to maybe hover around the 60% to 70% flat fee and similar fees, and maybe 30% to 40% billable hour. Because there’s probably— and that’s the 30%, 40% billable hour is still being charged by the accounting firms, the Big 4, because they say we’ve already systematized 60% to 70%. And they realize they can make more money with the flat fees than they could with the billable hour. But there are still some unknowns on the billable that they still charge billable because there be dragons in lots of things.

Phillip Vaden [00:33:58 – 00:33:58]

Yeah.

Damien Riehl [00:33:58 – 00:34:27]

So I think that the death of the billable hour is a misnomer and that really shapes our conversation. And I think it’s just a question of how much we’re gonna allocate it. But with the billable hour, there’s an old joke that the slowest lawyer wins the race because if you charge, you bill 10 hours for a thing that I bill 1 hour for, you make more money than me. And that is true, but it’s only true for hourly. It’s not true for contingency fee. It’s not true for flat fee. Also not true for subscription or success fees. Also not true for corporate law departments.

Phillip Vaden [00:34:27 – 00:34:27]

Yeah.

Damien Riehl [00:34:28 – 00:34:53]

Also not true for government lawyers. For all of those categories, they love efficiency because they have more work than they have time. And so really, if we get away from the billable hour mentality to go toward these contingency, flat fee, subscription, success fee, and be able to say, oh, now I’m not incentivizing hours because you show me the incentive, I’ll show you the outcome, said Charlie Munger. If you wanna incentivize hours, if you want me as an associate to bill 3,000 hours, man, I’ll give you hours.

Phillip Vaden [00:34:54 – 00:34:54]

Yeah.

Damien Riehl [00:34:55 – 00:35:35]

If you want to incentivize efficiency and quality, I’ll give you efficiency and quality. And so when you think about billable hours, you just think about how much money did I bring in last quarter or last month, right? You think less about how much did that cost? What was my overhead? What was my associate, right? Because revenue minus cost equals profit margin. But if you then go over to the flat fee side and you realize, oh wait, with AI, I can shrink my cost, which then increased my profit margin. This 12-hour task is now 5 minutes. I’m going to make way more money. And in fact, I can make so much more profit margin that I can tell my clients, I’m going to give you a haircut below what I usually charge you. And the client says, wait, not only do I pay less, I get the certainty of flat fee. Man, I’ll take that every day of the week.

Damien Riehl [00:35:36 – 00:36:16]

And then if the client doesn’t know what that top line is, you can actually push that top line up and charge more money. And then it just becomes a negotiation with the client to be able to say, what’s that top line revenue going to be? So we as a law firm, we, you know, I worked for Robbins Kaplan, a big law firm that did antitrust work. And often these would be over price fixing and bid rigging. And whenever there’s price fixing, almost inevitably one of the competitors would drop their price. And then what would happen next is they would steal the customers from all their competitors. And then almost inevitably, all the competitors would have to drop their price too. This is how price fixing cabals go away. And of course, we’ve had an hourly billing cabal as lawyers for a long time.

Phillip Vaden [00:36:16 – 00:36:16]

Yeah.

Damien Riehl [00:36:16 – 00:37:04]

But law firms are now AI native that you’ve got people jumping ship from big law firms and making AI native law firms. And they’re saying, I’m using AI and I’m billing flat fee. And they’re stealing all of the client work away from their competitors. And what happens next is, of course, everyone is going to fall into line. And the question is, do you want to be that law firm number 1, or do you want to be the law firm number 20 that doesn’t know why the phone stopped ringing? And so I think that we are now in this point where the biggest law firms in the world are thinking about how do we shift over to the flat fee world? Because not only are these competitive law firms, AI-first law firms, eating our lunch, But also the clients themselves are eating lunch. Darth Vaughn has said, hey, competitor, hey, law firms, you have a new competitor in town. It’s me as an in-house counsel, me as a corporate law department. I can do a lot of things.

Damien Riehl [00:37:05 – 00:37:21]

What can you do that I can’t do? And so I think that, you know, biggest law firms are getting lunch eaten by AI-native law firms. They’re getting their lunch eaten by their competitors, by their clients themselves. So what are the— how are the law firms going to respond? Flat fee seems inevitable and how fast are we going to get there?

Phillip Vaden [00:37:22 – 00:38:13]

Yeah. And I think the missing ingredient too for flat fee is AI is kind of the secret sauce that makes that really, really achieve comfortably achievable. Because you can run metrics on, hey, you know, standard contract dispute case, there’s this many filings on average and this many hours worked on average, et cetera. We come up with a rate internally. We may be able to go broader with some of the, the information that, you know, Lexus, you know, may have on hand about how much people are, are, uh, how many hours people are billing for certain matters. And then, then, you know, you’re going to be more efficient now. Um, whereas before it was like we could go to a flat rate, but it’s, it’s a little bit of a crapshoot whether or not it’s gonna, it’s gonna end up working out in our favor or not. Now it’s, yeah, it gets to get to your point, you know, it’s, hey, we can lower our fee by 15% because we know we’re gonna be 80% more efficient with this case.

Phillip Vaden [00:38:14 – 00:38:23]

So then everybody ends up Making more money and more money means probably more lawyers coming in and handling more cases.

Damien Riehl [00:38:23 – 00:39:12]

Because we’re building plaintiff spots and we’re building defense spots, which is shrinking the cost of litigation and increasing the volume of litigation, which is overwhelming our already overwhelmed courts, which are now using bots to be able to deal with that increased volume. So you kind of have a triangle of bots against bots against bots, which is probably going to mean Jevons paradox is well in play, where we’re going to have more litigation than ever. And we’re going to have bots both building the litigation and also interpreting the litigation. You might know the American Arbitration Association, AAA, has a bot arbitration, where you upload your documents, I upload mine, 20 seconds later, the bot gives an answer. And today, they have a human arbitrator reviewing that and getting it out the door. But Bridget McCormick, who is CEO of AAA, says that’s not going to be forever. Of course, her goal is to be able to say, get rid of the humans. And maybe that will help access to justice.

Damien Riehl [00:39:12 – 00:39:43]

That right now, if you have a dispute that’s less than $75,000, good luck getting that resolved because no lawyer’s going to take that, even on contingency. But if you can have a bot decide that in 20 minutes without all the legal fees, maybe you could bring that down to $50,000 or $25,000 or even $10,000. And that can, for consumers, that’s the difference between, you know, $75,000 is whether you live in a house or not. And we should help this, not just because it is good for lawyers, but it’s good for society, access to justice.

Phillip Vaden [00:39:43 – 00:40:02]

Yeah, that’s great. And you’ve given a lot of this answer already, so I think I know what your answer’s going to be, but I’ll let you do a summation of this. So do you consider yourself a techno-optimist or a techno-pessimist? And what’s your closing argument for your point of view?

Damien Riehl [00:40:03 – 00:40:50]

I think that I would, Your Honor, I would question the question because optimism and pessimism imply a degree of passivity that I don’t have control over where we’re headed. It’s just, it’s going to happen great or it’s going to happen horribly. And I don’t think that we have passivity. I think we have an active role as lawyers to be able to shape whether it’s going to be good or bad. And I think that if we think about the multiverse of all the potential futures that we have and all the timelines, There are a lot of timelines where things go horribly wrong. And you can imagine totalitarian regimes and authoritarian regimes having a panopticon of how the AIs are able to surveil us and be able to see us all the time. That is a lot of bad paths. But I would say that we, I personally, and we societally should be pushing toward the good.

Damien Riehl [00:40:51 – 00:41:38]

That is to say, how can we make this a better world than it was in the past? Not in spite of AI, but because of AI. That is, how can we be able to reach the 92% of legal needs that are unmet and have been unmet for decades that we haven’t been able to help? We haven’t been able to pro bono our way out of that 92%. But with AI, maybe we can. And maybe with Jevons’ paradox, you know, you could be able to serve as a lawyer more people than you ever have in the past. And maybe we can make more money than we ever have in the past, not in spite of AI, but because of AI. And maybe we can be able to have the rule of law. That is, you know, Agents are going to go out and do a lot of work in the world, and they’re going to break lots of laws. But what if we build the agents to be lawful by design? To be able to say, before you take every one of these steps, see if it violates US law, Texas law, or see if it violates EU law.

Damien Riehl [00:41:39 – 00:41:56]

And if it does violate one of those things, let a human know so that we could be lawful by design. So I think there are lots of ways that we could be, you know, I don’t like optimist or pessimist. I like to say, what is the future that we want to build? And let’s go ahead and push our societal frameworks that way and let’s build the systems to allow that to happen.

Phillip Vaden [00:41:57 – 00:42:14]

I like it. I like it. Well, Damien, thank you so much for the time. This has been excellent. I have learned a lot as always from you, so I really appreciate it. And I hope to have you back on to talk about what’s next because there’s always new stuff coming out. There is.

Damien Riehl [00:42:14 – 00:42:30]

Philip, I really enjoyed our conversation. And thank you for asking the important questions. I think we societally have to think, and we as lawyers, as, you know, the keepers of the rule of law, we should be thinking about these things and we should be thinking about how AI is going to fit into them. So thanks so much for having me on and thanks for having these good conversations.

Phillip Vaden [00:42:31 – 00:42:32]

All right. Talk again. Thanks, Damien.

Damien Riehl [00:42:33 – 00:42:33]

Thanks.

Phillip Vaden [00:42:33 – 00:43:01]

Thanks again to Damien from Clio, where he oversees the integration of large language models. The guy knows so much about so many things. forward to having him on again. I hope you enjoyed the conversation where we went from unauthorized practice of law and AI use to the largest law firms in the world and how they’re leveraging this great new technology to better represent their clients. Lastly, thank you very much for joining us. I look forward to seeing you again soon.

Continue Reading