How Technology Reshapes Law: Practical Guidance from Clio’s Solutions Champion Damien Riehl - S9E42

0:00The reality is this is happening whether we like it or not. What if you were to take 100 countries worldwide and turn those cases, statutes, and regulations into symbolic AI if-then statements? I think that we societally can run quite better because then we could be able to have the answer as to if you build a fence up to somebody's house, then what are the results of those things? And then all of us can be clear as to what the societal OS really says. Ignorance of the law is no excuse. What's the excuse of hiding that law behind a paywall and not giving access? But that's the problem. 92 % of legal needs are unmet.

0:33The solution is to use this AI sandbox that we're doing in Minnesota, saying that if you're using AI to help that 92%, then maybe you will not be prosecuted for the unauthorized practice of law. On today's Legally Speaking podcast, I'm delighted to be joined by Damian Reel. Damian is a lawyer with experience in litigation, counseling businesses, handling cybersecurity crisis, and managing intellectual property matters. He has represented Fortune 500 companies, startups, and nonprofits. He is also the solutions champion at Clio. As a technologist, Damien has spent more than a decade designing and developing AI tools in the legal practice.

1:09His goal is to build technology connecting law to the people it serves. So, a very big warm welcome to the show, Damien. Thank you so much, Rob. Really good to see you. It's an absolute pleasure to have you on the show. Really looking forward to today's discussion. But before we get into it, I do have an icebreaker question or two for you. Firstly, what is your favorite beverage and what is your preferred choice of footwear on a typical work day? Those are great icebreaker questions. I would say that on preferred beverage. So I like to drink sparkling water and bubbly is my preferred brand of sparkling water.

1:42If you're pressing me on a flavor, I'd say cherry bubbly is my favorite beverage. So that's number one. And my footwear, I tend to wear something that is easily slip on and offable. I travel enough that I go through security, airblind security enough that I say, okay, if I need to take off my shoes, I want it to be easy. So on a typical workday, I'm having loafers. There we go. Bubbly, sparkling water and loafers. With that, we can move swiftly on. And I mean, it's a hell of a journey. When we were doing the research into this episode, this could be a five-hour long podcast, given the accomplishments and the amazing work that you've done over the years.

2:15But to begin with, could you give our listeners just a bit about your background and career journey? Sure. So I've been a lawyer since 2002. I worked for a big law firm where I represented Bernie Madoff victims. I also represented Best Buy and much of their commercial litigation and also helped sue JP Morgan over the mortgage-backed security crisis of 2008. So I did high stakes bet the company litigation, but I've also been a coder, I would say a terrible coder since 1985. So anyone who works with me says I can't code with my salt. But these days with vibe coding, it's quite easy to code. So I've always had technology and law as part of my practice.

2:46And so in 2015, I pitched Thompson Reuters. I said, here's legal tech that can change the world. I think you should build it and hire me. And they were dumb enough to help build and hire me. So anyway, I worked with a bunch of people on a cool thing there. And then I went on to do cybersecurity, where the biggest thing I did there was that Facebook hired me and my company to investigate Cambridge Analytica. So I spent a year of my life on Facebook's campus with Facebook's data scientists and my former FBI, CIA, NSA people. And we figured out how bad guys use Facebook data. Did that for a while until I figured out that maybe traveling to the Valley isn't quite as much fun as it should be.

3:16And then I talked to Ed Walters, who was leading a fast case. and I said, Ed, you've got a massive data set. You've got cases, statutes, and regulations. I said, let's do some really cool things with it. So since July of 2019, I've been with Fastcase, which turned into Vlex, which recently turned to Clio. Absolutely. And I told you, listeners, there was a lot there to unpack. Highly, highly impressive. Just going back towards the early stage, what inspired you to pursue a career in the law initially? Because you studied music, I think, at North Dakota State University. So yeah, where did that inspiration for law come from?

3:48You've done your research. Yeah. So I was doing a choir conductor, at least I was going to be a choir conductor. And so I was conducting a Brahms piece as a student teaching, and two of my tenors started punching each other in the face. And I thought, oh, maybe I should go to law school. So the joke is that I broke up a fistfight in a classroom so I could create fistfights in the courtroom. And so, yeah. But I thought, so between that fistfight, I worked for the federal courts, where I worked for the clerk's office for the U.S. Bankruptcy Court. And I got to know a lot of the clerks who had just graduated from Harvard and Yale and those kinds of places.

4:19And I thought, gosh, they're not that smart. And so maybe I could go to law school too. So that's ultimately what I did. Yeah. And rightly so. Everything you've done seems to touch the gold ever since. You touched on before some of the cases that you've worked on. I mean, it's incredible. You've litigated council businesses, handled some of those cybersecurity crises, represented Fortune 500 companies, and on and on and on. What have been some of the most memorable cases, perhaps maybe less in the publicly available that you've worked on that have really helped shape you commercially, not just as a lawyer, but as a general business person as well?

4:52I would say that the most important aspects of my litigation experience were that it had a lot of drudgery. That is, you know, you think of, you know, you watch legal dramas and you think, gosh, I want to be able to catch that person in a lie while they're on the stand, or I want to be able to make that pivotal argument that makes the other side say, oh no, I lose, right? So that's really what the zeitgeist is with law, but really the reality is mostly just saying, you know, I'm just going to look at a screen and look at an email from somebody saying, I'm not going to give you my documents. And then I'm going to say, no, you're going to give me your documents, right?

5:23And fighting over that for a while. So really the drama of legal dramas really is really betrayed by the drudgery of a lot of litigation, which I was a litigator and advisory. So I would say that really with AI right now, we have an opportunity to get rid of that drudgery. So my friend Casey Flaherty says, we as lawyers do a lot of thinking, but we do a lot of thunking. That is a lot of drudgery. And AI can maybe do a lot of that thunking for us. So I would say that there were really amazing aspects of my career, where is advising clients and being able to say, you have options one, two, and three, and kind of guiding them through those options.

6:03Those were my favorite parts of litigating. And maybe with AI doing a lot of the drudgery, we could be able to maybe enhance and be able to do more of the thinking and a lot of the high level advising. Yeah. So let's get into that because I'm a big advocate for AI. You and I both sort of, you know, tech for good and, you know, removing friction and providing more fusion rather than confusion with technology. So as a technologist, you've spent more than a decade designing, developing AI tools, particularly within legal practice. Can you tell our audience about some of the AI tools you've been involved with building and which you're most proud of?

6:34Sure. So yeah, since 2015, I've been building. And so I can't talk much about the stuff I was doing at Thompson Reuters because of NDAs and that kind of thing. But I can say that while we were at Fastcase, while I was at Fastcase, Fastcase turned VLex turned Clio, we've built some quite remarkable things. And the reason we've been able to build those things is because we have the data. We have the cases, we have the statutes, and we have the regulations. And if you have those things, if you think about almost all of legal advice is applying your client's facts to the law. to those cases, statutes and regulations.

7:06So if you don't have the law, really just have facts, right? And any AI model can be able to analyze facts, but not every AI model has yesterday's case and yesterday's statute. So the things that we've been able to build have really taken advantage of those cases, statutes and regulations. That is to be able to, in Fastcase, we have several patents, had several patents on being able to visualize what the law looks like and who's been citing whom and which cases are the most authoritative. And then moving on to Vilex, they for 25 years had been creating even cooler visualizations and doing it not just in the United States, but in 100 countries worldwide.

7:41So being able to ask now, when comparative law comes into the fray with AI, now we could be able to say, tell me about the privacy regulations for online advertising in all 50 states in the US and in the UK and in the European Union and in Spain or France or Italy or Latin America, Mexico or Brazil and Hong Kong and Singapore, right? So to be able to do these multi-country jurisdictions, you need that data. So being able to use AI to be able to parse data that, to be frank, nobody else has, it's been a great joy. Yeah, absolutely. And I want to double down on data because, you know, it's well-bounded.

8:18People say it's the new gold, et cetera, et cetera. how do you know good data is good data what what should you go by for people that maybe are looking to acquire businesses or think about that to to get data what are some of the things you would highly recommend doing your due diligence so you're right that data is oil and you have private oil and you have public oil uh the private oil is what's you know the settlement agreements and contracts but then there's public oil cases statutes and regulations and i say public in quotation marks because they're really often not that public uh that is it's really hard to get all the cases.

8:49It's really hard to get all the statutes. It's really hard to get all the regulations because a lot of courts actually, even though it's quote unquote public, they make it very hard to get because they see people that are pulling that data as quote unquote data brokers. So they fight every tooth and nail to be able to keep people from acquiring it. So not only is it hard to acquire the cases, the statutes and the regulations, but often that data is quite dirty. That is, there are line breaks where there shouldn't be line breaks. And there are headers and footers at the bottom that you have to strip out to be able to get rid of those things, to make clean data.

9:22So I would say anyone to your question, anyone who's thinking about the quality of the data they're receiving, whether it's public data or private data, cleaning that data, you know, when I was at Facebook, the data scientists would say cleaning is about 80 to 90 % of my work. And then the fun part is the 10 % of the work. So, so, you know, VLEX and before that FAST case had been cleaning that cases, statutes, and regulatory corpus for 25 years. And that hard work paid off when AI came in. Yeah, absolutely. And now it's, you know, it's a superpower, isn't it? When you bring the two together. Very, very exciting.

9:56And that probably leads quite nicely on to Clio, because, you know, you are a solutions champion at Clio. What is your role in Hattel? Tell us a bit about it. As a solutions champion, like we get to, at Clio, do something I think has never been done before in the history of legal tech, which is connect the business of law with the substantive law. And that is to be able to say, if you bring in a new matter, we could be able to say, here are the cases, the statutes, the regulations that relate to this matter. And maybe here are some new complaints that relate to your matter. And here are some motions that have won before your judge.

10:26And to be able to connect that business of law that is making money with the law and bringing in new business with the substantive law, with the practice of law, those two things, bringing them together. It's quite extraordinary in that I don't think it's ever been done in the history of legal tech. And I think we have a chance to do it. You absolutely have a chance to do it. And that's why we're big supporters of Clio, obviously support the show. And, you know, I think it's just pioneering. And to your point around, you know, Clio generally, ClioCon this year was just incredible in terms of Jack Newton's keynote and just the level of sort of innovation that was pouring out from Clio.

10:59So for you, no two days the same? Typical day? Could you talk us through what that looks like? Well, my typical day is on the road. So last year, I think I was on the road 200 days out of the 365. So I was on the road a lot. And I think that there's an aspect of the, we are right now in a moment with AI and that everyone knows that everything we as lawyers do is built on words. That is, we ingest words, analyze words, and output words. And it turns out that AI does all three of those things at superhuman speed and a postgraduate level. So the typical day for me is going to the four corners of the earth, spreading the gospel of things are changing.

11:37And, you know, this is my friend Casey says that this is happening and it's either happening to you or it's happening by you. So my typical day is going to the four corners of the earth, saying to lawyers, this is happening. You need to be able to adapt or not adapt and your competitors will be in your lunch. But I think for those lawyers that get it, that I'm able to talk to, they're really excited about this because of that drudgery aspect we talked about earlier. We can get rid of that drudgery. Yeah. And, you know, we talk a lot about, you know, career enhancement on the show as well. There's a big play here by removing that drudgery for, you know, things that you may historically have missed.

12:13You know, being a lawyer in high pressure environments or going through paperwork that you may have to forego. And actually, tech for good could get that back. There's some moments you might be able to buy back time, which is the one thing none of us, you know, we all want to try and get more time, right, to do things that we want to do. So, yeah, we're big advocates for that. Okay. And really, and I want to, you just said something that you, AI catches things that you would have missed. And I really want to focus on that, that, you know, in data science, you talk about precision. Did it get it right?

12:39And then recall. Did you get all the things that you wanted to get? And I think the recall aspect is something people often forget. They often focus on the precision saying, was there hallucinations? but did it get all those cases that you needed to get and did it get all the statutes that are necessary a short anecdote that might be interesting to your users is my next door neighbor right over here uh they what they did uh is that they um they were going to build a fence uh literally right up against my house uh within within an initiative my house uh and so i used vincent to be able to say what are the potential causes of action uh and uh and it turns out that i've been walking that place for 15 years so i have an adverse possession claim i also have an easement claim.

13:15But what it caught that I wouldn't, after 15 years of litigation known, there's an aspect, a cause of action in Minnesota called boundary by mistaken line or something like that. So this is a cause of action that I wouldn't have pleaded had I been just using my brain. But AI actually gave me this extra cause of action. And then within 15 minutes, I had a complaint that had all the causes of action. And I said to my neighbor, like, listen, let's talk about a reasonable place to put this fence. Otherwise, this complaint is going to be on your front door. So anyway, so this is a way that we can be able to enhance and not just do things faster.

13:52People talk a lot about faster, but do it better. Get more causes of action. So doing the faster, better. There's an old joke that faster, better, cheaper, pick two, because you can't have all three, right? You can have faster and better, but not cheaper. You can have cheaper and faster, but not better. But really, AI maybe gives us all three. It gives us cheaper, faster, and better because it gave me that cause of action I wouldn't have known of before. I love that insight. And thank you for sharing that and giving a real practical example. And I think, you know, I heard from someone, I forget who it was speaking the other week about there's still the debate, isn't it?

14:23Human versus AI. And he said, let's stop having that discussion about AI replacing humans, whether it won't it, whatever. And actually talk about AI being more useful, you know, in certain parts, you know, AI can be a lot more useful. And you've given a great example there where things were missed, but the usefulness of that piece of technology has outperformed that particular, you know, where the human mind is. And so, you know, actually see this as a tool that can really help you sharpen your axe 10x. Yeah, that's right. Steve Jobs said that, you know, at its best, technology is a bicycle for the mind.

14:52And I would say that AI is not just a bicycle, but it's a motorcycle that is giving you even faster, better, stronger insights. Yeah, absolutely. And look, we're tech for good on the show. And I think it's great, all the things that are out there, you know, with due diligence, but the opportunities here are huge. But with opportunities, obviously, are threats. And so many see AI-generated content as a threat. So you see it as an opportunity. What's the biggest misconception about AI in the legal profession today? I would say that the headline-grabbing aspects are lawyers cited hallucinated cases.

15:27So that is, of course, the biggest thing. So generated AI, a lot of lawyers say, I'm not going to use generative AI because I don't want to hallucinate cases. But let's think about that a bit more. What if you have yesterday's case and yesterday's statute and yesterday's regulation? And what if the tool like Vincent actually shows you the text, the actual non-hallucinated text from those actual non-hallucinated cases? And what if it does that 100 % of the time, not 99 % of the time, but 100 % of the time? And your listeners might be saying, wait, wait, wait, wait, how can you say 100 %? I know about generative AI.

15:56And I can say 100 % because the text that Vincent shows wasn't generated by generative AI. The text was a database search, like we've been doing a database search for the last 40 years. So what if you add the actual non-hallucinated text from the actual non-hallucinated case 100 % of the time? Then that risk of hallucinations goes to zero because you can actually go to that case and the text of that case, copy and paste the text of that case, and then that mitigates the risk. So that's thing number one. Thing number two, my friend and colleague, Ed Walter, says, you know, a way to mitigate that risk is to read the effing cases.

16:33You know, if you actually read the case or, you know, the people that use ChatGPT with that hallucinated case, had they tried to find that case in a regular database, they would have known that it was hallucinated. But I think that what those hallucinated cases show is not that generative AI is bad, but that bad lawyers, bad lawyer. That is, they, you know, bad lawyers have been citing cases for non-existent attributes for decades. I, as a judicial clerk, know that I would read the citation and the proposition in the brief, and then I would look to the case, and that case would not have that proposition at all.

17:08So bad lawyers have been doing this for decades, but the hallucinated cases, what it does is really brings it to the obvious for, to say that people that have been citing cases for their non-proposition support have now are hallucinating cases. So bad lawyers aren't going to bad lawyer yeah there's a quote in that i'm sure we're going to use absolutely because it's so true isn't it you know ultimately yes this can speed up you know ai can speed things up it can make us more efficient we can pass on the games but we still need to make sure we're engaged you know as the human and like you say doing that due diligence and doing the job which ultimately we're putting our names on the tin for okay let's stick with ai and ethics because you know ai is becoming more and more embedded into legal workflows and it's only going to get faster and better over time.

17:50But what are some of the ethical considerations lawyers should really be thinking about and aware of? Yeah, I think that there's two aspects of this. One aspect is, you know, right now in the year of our lawyer 2025, we have to validate the cases and we have to validate the text from the dead back cases, right? So if you don't do that, you are arguably ethically perilous, I will say. But really, as we go forward to 2026, 2027, as the models get smarter and they're doing an exponential intelligence leap and they become exponentially cheaper, which they are. What used to be a$500 for a million tokens is now 50 cents per million tokens.

18:26So I would say that as intelligence goes up and as cost goes down, I think that our need for human diligence is going to decrease. Because right now we at Vincent have, you can think of it like agent number one, finding the case, agent number two, finding the text that is most relevant to those cases. Agent number three is being able to validate how well that text answers the user's question. Then bot number four is able to put it into the memorandum. And then we have bot number five that validates that this is a case in our database. So that's already five bots, agents, doing this work. So you can imagine as the large language models get cheaper going forward, we'll add bots seven and eight and nine to do that extra double and triple and quadruple checking to make sure that this is an actual non-hallucinated case and an actual non-hallucinated text.

19:11So I think that in 2025, we ethically need to be able to validate this. But I would say that as bots become cheaper, a lot of the work is going to be done by large language models. And so then as we go to that future, and maybe even on our present, everyone talks about, is it ethical to use AI? But we should be saying, is it ethical to not use AI? That is, if you're billing your clients 12 hours for a task that Vincent can do in a minute and a half, and you can validate it in 15 minutes? Are you overcharging your clients? And how soon before we actually flip to say it's unethical to not use AI?

19:49And it's so true, isn't it? And it's being about as client-centered as possible, particularly if you're trying to be a law firm. And that's a great example where actually where the industry does shift to actually, this is table stakes now, actually, that there's an expectation that we know this could be done faster, better, accurately within this timeframe. And actually, this is our budget for it. And it's on the law firm to make sure they're meeting the needs of the clients because the market will show no loyalty. As many people who have joined Richard Susskind's keynotes, he talks very openly and very closely about that.

20:17The market will show no loyalty. And I think it's a really exciting opportunity and time for law firms who really want to get on the tech bus and really see where they can go with this. And bus is the wrong term. It's a rocket ship, that's for sure. Okay, you do so much. But one of the worlds that I wanted to ask you about is your chair of the Minnesota State Bar Association's AI Committee. And as part of your role, you oversee an AI sandbox expanding legal services to underserved communities. So how do you see AI reshaping access to justice, especially for communities who cannot afford legal help?

20:49I would say that, you know, start with the problem, then we can think through the solution. So the problem is that 92 % of legal needs are unmet because we lawyers are too expensive. So that is a reality. And another statistic is that 80 % of litigants in our courts in, say, family matters and, you know, debt collection matters and those types of things, 80 % of them are unrepresented. So they are pro se, self-represented litigants. So with this 92 % of legal needs unmet and 80 % unrepresented, these people are getting clobbered with the law. Uh, that is, they are, they are saying, uh, you know, gosh, I can't afford a lawyer, so I'm just going to risk it and bad things happen to them.

21:25And so, um, if we say that, you know, ignorance of the law is no excuse, um, then what's the excuse of hiding that law behind a, you know, behind a paywall and not giving access to those, to those, uh, people. So that's the problem that we're, we're having. Uh, the solution is to maybe, uh, use this AI sandbox, uh, that we're doing in Minnesota saying that if you're using AI to help that 92%, then maybe you will not be able to do prosecuted for the unauthorized practice of law. And this is something that we've been working on since spring of 2023, is that we've been working on this. And really the reality is that people are using AI every single day in self-represented litigants.

22:03One of my court of appeals judge friends said, you know, after ChatGPT came out, my self-represented litigants briefs became much, much better. And it's true. And so we're building a sandbox so that legal aid organizations could be able to be in that sandbox and do this. But we're well aware that ChatGPT is not going to apply the sandbox. Neither is Gemini, neither is Claude, neither is any of the Microsoft tools that are doing this work. And so the reality is this is happening whether we like it or not. And so do we allow ChatGPT and Claude and Gemini to provide what is maybe legal information or legal advice?