AI vs. Lawyers: How Artificial Intelligence Is Reshaping Legal Practice

Hey, everybody. Welcome back to the Dominate Law Podcast, where we empower attorneys as entrepreneurs. I'm your host, Ashton Rogers, and today's conversation might be the most important one you will hear all year, because we are talking about the technology that is reshaping every single corner of legal practice. Yes, we are talking about AI Before we dive in, a big thank you to our sponsor, equa Marketing. Here is the reality. AI is leveling the playing field in law, big firms, solo practices, mid-size firms. Everyone now has access to the same tools. What separates the firms that win is visibility and trust in the market. Equa helps law firms build exactly that through proven digital systems that attract better fit clients, not just any client. If you want to see how your firm shows up online and get a complimentary $900 marketing strategy session with one of their senior strategists, go to www.dominatelaw.com/msm.

Now into today's episode, we are diving into AI versus lawyers and going beyond the hype. Is the billable our dead? What skills do junior associates actually need? Now, where does AI fall short and where does it truly shine? We'll walk through the four pillars of this transition capability, careers, economics, and ethics. So you walk away with a real blueprint for the future of your firm To help us think through it, I have two extraordinary guests. Bennett Borden, lawyer, data scientist, former chief data scientist at DLA, Piper, and now CEO of Clarion ai. He helps organizations turn AI from a science project into a deployed business asset. And Maura Grossman, research professor at the University of Waterloo, principal at Maura Grossman Law with a PhD in psychology and a JD from Georgetown. She's the scientific conscience of the legal tech world, and she literally wrote the case law on technology assisted review. This conversation is packed with insights you can use today. Let's get into it. And let's start with the state of the art. How effectively is AI currently being used in core legal tasks such as research, document review, contract analysis, and drafting?

Thank you, Ashton. It, there's a wide variety of options, um, and AI across law firms, um, because as we'll talk about today, a, uh, AI is an existential threat to how we have practiced law since the 17 hundreds, basically <laugh>. Um, and so AI, when effectively deployed, is accelerating core legal tasks, is basically what we're seeing, research doc review, issue spotting, draft generation, um, are meaninglessly faster and less explicit to create than they were even two years ago. Where it works best is reducing friction and cognitive load, not replacing legal reasoning. The danger is mistaking speed for correctness or completeness used. Well, AI makes good lawyers more effective, used poorly. It just makes mistakes faster. And so what we are seeing is a complete rethinking of how AI and the law and the practice of law, um, interact and how you, um, keep the strength of what it means to be a human in the, uh, legal practice, which is absolutely essential. Um, and yet 80% of what a lawyer does today is better done by ai.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And if a firm is still doing manual document review today, what is the single first tool or category of tool they should license tomorrow?

Um, first fire them <laugh> second. Um, there's, it's really interesting, and Maura, it, like it, as you said, is one of the founding scientists of technology assisted rebuke. Um, and over the years, since the 2005, um, federal rules first really introduced the idea of electronically stored information as documents. Um, we have gone from associates sitting in basements with giant boxes of documents or just looking at a screen clicking one after they're there. Yes, no, yes, no, until really where you've got tremendously powerful categor categorizing or, um, predicting the relevance quite effectively of, uh, a document. Now we're heading into what can gen AI do that even these algorithmic approaches couldn't do, and while it's just starting to come out and people are really afraid of it, that's where the world's going.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And you evaluate these tools scientifically, where do you see the biggest gaps between what AI is marketed to do and what it can reliably do in a legal practice today?

So, back when Bennett and I were baby lawyers a long time ago, um, we, I don't know if we met at Trek, but we met, it was, we may have met at a, uh, a scientific venue that was actually testing and evaluating technology assisted review tools. And we did the really hard work of rigorously evaluating these tools, and we took a lot of bunk from people because they weren't ready for these tools, and we were sort of out on the very cutting edge. I think what's really surprised me is how quick the adoption of generative AI has been without the same level of rigorous evaluation. So people were really afraid and really skeptical about technology assisted review, but now it's like, oh, Jen ai gimme some of that, you know? And they don't seem to, uh, really worry too much about, you know, the bonafides and it has it been evaluated properly. So that for me is the big gap between one was really a proven technology by the time it was adopted and, and the other were sort of in the wild west. Mm-hmm

<affirmative>. Mm-hmm <affirmative>. And why does the marketing hype specifically target legal? Is it because we are perceived as inefficient

<laugh>, uh,

Because we actually are inefficient, right. <laugh>?

Well, all the motives, the incentives are wrong because of the billable hour model, uh, which incentivizes more time rather than less time. Um, I don't think legal, I, I, I see the AI coming into every field. I, I don't think legal is particularly different. If anything, uh, finance, you're being able to predict two seconds faster than Bennett is a lot of zeros, you know, in profit. So, uh, I think it's coming into most fields. Healthcare, you, I mean, you name it. Uh, but I think, um, we, Lela has been traditionally very conservative and resistant to change. So this is, this is a little bit new to see people moving so quickly, uh, into a technology and sort of embracing it.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And Bennett, let's talk about the human element. Are there specific areas of legal work such as advocacy, negotiation, judgment, or client counseling where AI consistently falls short? Why?

Absolutely. We have to remember that AI is a probabilistic determination of what has happened in the past, and so is extremely good at some tasks summarizing, pulling facts out. There's just, there are amazing things that it can do in the practice of law or AI struggles is where law intersects with that human nuance, judgment, persuasion, values, accountability, advocacy, and negotiation depend on emotional intelligence, credibility, and situational awareness that just simply aren't reducible to pattern matching, which is what all gen AI is. So, client counseling in particular requires moral responsibility, and that is a human quality, helping clients decide what they should do, not just what they can do. Those are inherently human roles. When we, when I was back at big law, um, one of the things that we did was, uh, all the headlines were coming out, right? Like, oh, the big, all the knowledge workers are gonna be replaced, right?

And so we undertook a tremendous study. Um, fortunately or unfortunately, lawyers write down what we do every 10th an hour, right? And so we, um, have a massive record of what lawyers at least save that they do. And so we undertook this, um, analysis of more than 10 years worth of 14,000 lawyers time entry data, and did a massive analysis. And it turns out that conservatively 80% of what a lawyer does every day is better done by ai. But that 20%, that human judgment, that accountability, that that human connection with your client, with the other party, the negotiations, that, that moral responsibility, that is the key of being a counselor. And so this is not a matter of replacement. What I tell everybody all the time is think Ironman not the Terminator, right? <laugh>, you're not unplugging a person and plugging in an AI bot. That's just not how almost all knowledge working works. Instead, it is taking well-trained people that have years of experience acumen that they've developed and extending and augmenting their capabilities by surrounding them with enabling technology. That is the what, that is the real way to take advantage of AI and up being a lawyer.

Mm-hmm <affirmative>. I love the example of Iron Man was exterminated. It's not the suit that wears the man, it's the man that wears the suit, basically. Right? And then going along with that, is it that AI cannot do these things, or that we as humans simply prefer another human in these high stake moments?

At this point, AI cannot do these things. Well, um, the, the, I am a huge AI optimist. It is one of the most powerful technologies that we have experienced as a species. It will change everything about every aspect of our lives. And we are just beginning to see the be, um, the start of that tremendous advances in education, healthcare, and material sciences, pharmaceuticals. We're just seeing these tremendous advances, but it can also be used for tremendous harm, if not controlled well and provably that it's being controlled well. And so the, the main thing that when we are advising clients, whether they're law firms or elsewhere, is you have to understand what it is you do as a company, and you do stuff in a series of steps. Call the workflow. Well, you take that workflow and that, that they have steps in them, right? And some of those are better done by ai. Some of those are better done by human. It's important to distinguish those things, put AI in safely where it's supposed to be, and then handle the handoff well with really good quality control checks.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And Maura, let's ground this in reality. Can you share an example that illustrates where AI meaningfully improved outcomes or created unexpected problems?

So let's start with technology assisted review. I, I think there is nobody that will argue, uh, that, well, maybe there's some people, but there's no sensible person that will argue, uh, that that didn't improve the way that we were managing large volumes of data. I mean, if you have more than a couple of hundred documents, uh, it's a no brainer. It was many, many times, uh, more effective and efficient and really helped you sort of drill in and get to, um, what you wanted. So that was a very, very meaningful, I think, improvement in, in outcomes for search. I think what's been surprising to a lot of people, uh, is the hallucinations that they see with the, um, AI tools, even commercial tools can hallucinate the nature of the hallucinations are different. So if I am in a Westlaw or Lexus, it's not going to make up cases out of whole cloth that don't exist in the same way that if I'm using clot or chat GPT, that might happen.

But that doesn't mean it doesn't, as, as Bennett was saying, lose nuance, confuse the dissent for the, uh, main opinion. Confuse confused when the court is saying quoting defendants said, versus when the, when the court is agreeing with defendants. And, um, so I think people were very surprised, for example, when Stanford put out a study showing that even the commercial tools could hallucinate, you know, between 17 and 33% of the time, they were just different kinds of errors. Uh, so that's, and, and while there's been some improvement there, it's sort of a feature, not a bug of the technology. I mean, it, it, it's intended to generate new content. So, uh, it's like asking a zebra to get rid of its stripes. Um, so I think that's been created a lot of unexpected problems for, for both self-represented litigants who are using these tools and for solos who are pressed for time and, and figure this is really a lifesaver. And then they file a brief that's got four, you know, fake citations in it, and three quotes that were weren't in the case, and this kind of stuff.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And when it comes to this, how do we recover client trust after tech assisted error occurs? It's not just AI tech assisted errors can occur in many ways. So how do we confront that?

I think, well, part of it is the process you put in place to check in and verify, and, uh, obviously with, with a client and with the court, the, the best thing to do is to fess up and to apologize, uh, for your error, not to try to hide it or say, well, it was my assistant or the intern or whatever, that, that usually doesn't go over very well. So you take responsibility. You, I think the best thing to do is you figure out where your process went wrong and you strengthen your process so that at the, the next time, uh, it's a better process and you're more likely to catch the error. But we're at a place right now where this all of the, the outputs of these tools need verification, uh, particularly the generative ai, but even tar, uh, needs to be verified and evaluated properly.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And Bennett, looking at the talent pipeline, how is AI changing the skillset that junior lawyers and law students need to succeed today?

It is changing everything. Um, you know, one of the things that we do at CLAREON is work with law schools across the country and across the world on this problem. Um, because junior lawyer lawyers no longer add value just by finding information. AI can do that in a heartbeat, right? What matters now is framing the right questions, exercising judgment over AI outputs, understanding risk bias, uncertainty. Um, young lawyers need strong fundamentals, not weaker ones. And so, legal reasoning, ethics, writing domain understanding AI actually raises the bar. It doesn't lower the bar. The, one of the biggest challenges that, um, law schools have, much less law firms is how do we train lawyers for the future? Um, and we all learned, like Laura and I learned by being those associates in the, you know, the conference rooms with banker's boxes stuffed to the ceiling, right? And that's how everybody learned. You learn by being thrown into the trenches. Well, the trenches for the most part don't exist anymore. And so law firms have to, and law schools have to reevaluate how do you get domain expertise and experience in so that these younger lawyers can exercise judgment. Um, and that is a much more intensive, interactive process with more senior lawyers than it's ever been before. And so, to me, this is one of the biggest challenges that law firms face is how do we educate young lawyers

Mm-hmm <affirmative>. And if you were hiring a junior associate today, what specific question would you ask them? Just the AI readiness,

You know, to me, you can train, teach anybody the law, right? And so what it really comes down to is it's a personality type and a character type that I particularly look for. This is a very disruptive time. And disruptive doesn't mean necessarily bad, right? Um, when change happens, change can happen for the very good. Um, but there are some people who don't like change. They want, especially lawyers, man lawyers. We like straight and fast rule. There's a rule, there's a doctrine. There's a case that I can point to. And when we're, especially in AI right now, that ain't gonna happen for a couple, a number of years. And so people who are curious, courageous, but skeptically, courageous, you know what I mean? It's, that's the kind of associate that I am looking for. And just like, you know, writing skills used to be like preeminent.

If you were on the law review, um, you were like, oh, man, you were in, right? I don't need good writers. I need good editors, right? Because very often, these first generation, first drafts are really quite good, but I need someone with a very skeptical pencil, um, to make sure, like all the things Morris said, that the things are right, but then are they right for my client? Are they right for my position? Are they right for my case? Right? And so it's a very different set of mental skills than it used to be when Maura and I graduated. Mm-hmm

<affirmative>. Mm-hmm <affirmative>. Maura. Now, on the flip side of that training coin, what traditional legal training or early career experiences risk being lost or diminished due to increased automation?

So, I'm gonna disagree with Bennett just a little bit on the writing point, because at least for me, writing is thinking and thinking is writing. And, um, I do my thinking while I'm writing. And so if somebody does my writing for me, I'm often locked into their architecture of how they've structured the argument and whatever. I worry, I see it in both my computer science students and my law students. Uh, they are losing the ability to think critically, and they're losing basic writing skills because they don't need to. And then often they feel, well, I really don't have to learn all that much because I can always go look it up, that, you know, 24 7, the answer will be there. And, uh, that's very dangerous to sort of outsource all of your thinking in that way. So I, I try to force my, my students to do it. For example, this semester, uh, I am requiring all my students to keep a journal. And so they can't have their, there's no way that they can have, uh, a large language model write their journal about what went on in class and their feelings about the discussion and the reading, and so on and so forth. So, oh,

Smart.

I, I worry about some of those very, very basic skills that we sort of took for granted because, uh, they were just part of everything we did. I also worry a little bit, uh, about the need for mentorship and whether these young people are going to get the mentorship they need, because now you can go practice with a generative AI tool. You can do an interview, you can pre prepare for a deposition. And a lot of what we learned was through apprenticeship, because there is strategy, uh, and, and nuance that a lot of these tools miss.

Mm-hmm <affirmative>.

Mm-hmm <affirmative>. That's a great point.

Yeah. And if juniors don't do the grant work of document review, how do they develop the intubation of what matters in a case?

Yeah, they don't <laugh>.

I, yeah, I, I would agree. I don't know that you need to sit in warehouses or like, you know, and, and go through millions of documents, but you need to be able to issue spot, and you need to be able, look, so much of, of litigation is telling a story, and you have to find the documents that are gonna tell the story you wanna, uh, tell, and that support that and, and, uh, dispute the other side's, uh, story. So you have to spend some time doing that. For me, I, I, I sort, I didn't love document review, but I loved, uh, particularly white color work where you have to put toge, it's like a puzzle. You have to yeah. Put together all the pieces, um, to, to build a story.

Mm-hmm <affirmative>. I could not agree more, more. And, and that is, look, Docker view tar, most people do tar, they do it pretty well, um, depending on the firm. Um, we're just in, in the era of Gen AI review, which is not that different from tar, but it's different enough that the, and we don't have agreed upon quality control measures. We don't have, a lot of them can be transported, right? Like random sampling and null sets, sting, and all the stuff that we've learned in TAR is completely applicable to Gen ai, right? You're just getting to the set of documents that are supposedly relevant in a slightly different way. But once you have those documents, you've gotta know those documents backwards and forwards. That's where everything matters. Um, even though it's only 1% of everything you've looked at, um, if you're lucky. Uh, but that's just, like Maura said, that's where the story is. That's where you convince the, the tribunal or whomever you're trying to convince. And so that is utterly quintessentially human. Now, AI can help you find patterns and find a, you know, Hey, this is what I'm looking for. And what about this? Is there more of this in there? Like it's, it's really understanding that AI is an interactive tool with you. And so, but it's your creativity in your brain that's going to get the information that you want out of these tools.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And then Bennett, let's talk money. Now. How is AI reshaping law firm economics, particularly around billing models, leverage and profitability?

This is one of my favorite topics, <laugh>. Um, I have been saying for years that the billable model, it billable hour model is dead. We have done this since, honestly, since lawyers have been lawyers, um, <laugh>, and at the time, an hour, the time I spent on a topic represented the value that I was providing to my clients. It does not anymore. Right? It is. And so value-based billing is where things need to go. Um, AI is putting tremendous pressure on time-based billing and by exposing how much work can be done faster, um, without reducing its value, right? So how we leverage that model's really gonna change. There's gonna be fewer junior hours, right? It's usually a big pyramid, right? That tons of junior hours, and then they get extra order rate. You go to the top of the partners, um, but much higher expectations for judgment and quality, the firms that adapt will shift toward value-based pricing.

So almost everything we do at my firm is a task-based, um, phase based, outcome-based something, right? It's just because there are times when, like going into a negotiation, going into a right, there's times where an hour is an absolutely correct, uh, measure of value, but not much anymore. And so the problem is that the billable hour model is so utterly entrenched. Uh, it's how everybody, um, from this most senior partner to the most junior paralegal is judge. It's, and it, it, it's foolish to do it that way. One of the main reasons I left big law to, to start a law firm that is AI based from the ground up, is because trying to get that billable hour culture uprooted out of big law, it's, it's not worth the fight of the next 10 years of my life.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And then for some deeper insight, does AI finally kill the billable? Or does it make the billable, uh, incredibly expensive?

It's not gonna make it the billable hour, more like, that's like, who, what do people are gonna charge $5,000 an hour? Like, it's ridiculous. No,

We're, we're gonna have two legal worlds. So the, the people right now who go to the firms, like we used to work at the Tels and DLA and, and places like that, they have the $900 an hour or the thousand dollars an hour to hire that senior partner that they want to do their work. And because it's a life and death matter, and they, it's, and it's a corporate matter usually. So it's not necessarily out of their pocket, it's, it's from the corporation and so forth. The rest of the world can't afford that and is going and is going to move into the AI world where 80% of the work can be automated. If I just need a lease on a $1,200 apartment or something like that, I don't need that $900 an hour lawyer. I can get it off, uh, the internet or one of these systems, and then I can get Bennett to spend 10 minutes of his time just checking it, uh, or one of his associates. And so we're gonna have these two worlds, I think, of the super duper experts that people will continue to go to. And then I think the rest of law, which will be much more commoditized, I, I think,

And you're seeing huge pressure come out. So just today, there was an article that came out in Forbes, um, where a large company, a large, um, uh, audit advisory firm said to their, their own auditor, their, you know, their, their auditor. Hey, what kind, what savings are you passing along to me? Because, uh, now we know you're using ai 'cause you're touting you're using AI everywhere. Um, and in fact, uh, one of the world's largest companies, like in the AM law, I mean, and the, uh, fortune five, um, just sent out new guidelines for their outside counsel, um, that says that they will no longer pay for certain kinds of things. So here, here's some examples, would that we will not pay for document summaries, deposition summaries, summaries of court opinions, contracts, any kind of summaries like that. If we ask for factual summaries, we expect AI tools to be used first, draft summaries, summaries done manually are no longer billable. Um, document comparison and analysis. Um, so finding differences or missing clauses, or pulling key terms from contracts we will no longer pay for that is a massive chunk of what young associates do. So the pressure is coming and it, it's the firms that are going to adapt, that are going to rise to the top. Um, this is one of the most beautiful leveling technologies. So there's this world of big law versus everybody else, um, is, is gonna start to get eaten away at really quickly.

Maura, speaking of clients, what new expectations do clients now have regarding speed, cost, and transparency because of AI enabled legal services?

Well, I, I think Bennett hit the nail on the head. Uh, their expectation is not that the law firm is going to be able to do more and make more profits, but that the law firm is going to pass through some of the cost savings. Uh, so they, they expect the work faster. Uh, they expect a reduction in, in price, no reduction in quality, uh, and, uh, and their firms, to be honest with them about the use of, of these tools. I, I would agree with Bennett, uh, largely that I would be surprised if we don't, I don't know if it'll be the death of the billable hour, but it's in trouble. Uh, yeah. I, I think there's, that model is gonna be very, very, very hard to sustain in this new, uh, in this new environment. And I think clients have, you know, lived with every year the hourly rates going up. And I think we're now at a, a place where clients are going to start to bring a lot of this work, uh, in-house, that, that like, review this, this, uh, contract and tell me what provisions are missing. They're gonna be able to do that right in-house and, and take that work right away from the law firms.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And it's not just us using ai, right? So how do you handle a client, like you said right now, who runs their own legal question through chat, GBT and argues with your advice? <laugh>?

Um, I, that's, that is a, it's a tough question. It's, uh, I talk to doctors about this also because patients are doing the same thing. They're, they're, uh, getting their health recommendations and then they're arguing with the doctor about, about a treatment. I think, um, eventually you have to, what the AI can't do, is it, it doesn't really understand their, the emotional aspect. There's so much emotional aspect to family related matters, even to business matters. I've seen people do incredibly illogical things because they wanted to smash the Smither Marines out of the other person, and they didn't, you know, really care how much it cost. So I think some of our role is still gonna be that counselor, uh, psychotherapist hand holder, uh, person who, who helps the client, um, work through some of the issues that get in the way of, of productive resolutions, uh, of, of matters. And, and I think you'll have have to point out to them, you know, when the information is wrong. And sometimes clients will say, I, this is the way I wanna do it, and Yep. You know, that's the way you do it. And then you pick up the pieces afterwards if you have,

And you really approve your work, right? Yeah. Like everything, everything that we have a a, a literally moral responsibility to that the, the advice we put out, we think is correct, right? And so, um, if you do your job right, you could explain why you think this is, right. And then clients, look, lawyer's job is to identify risk and quantify risk, not to say yes or no, or do this way or that way, right? Like, and it's up to the client to decide what they want to do, right? And so we do what they tell us to do. Um, the nice, I presume that every piece of work product that I give to a client, they're running through Chad TPT to look for weaknesses, right? So I do that myself anyway, as one of our QC steps, right? So like we've built, um, our own kind of Jarvis, um, we call Claire, um, which has, is like our high mind, um, in the firm.

And, um, we have a put a lot of work into getting answers that we would give out of this with tons of quality control checks and tons of all the things that Mora mentioned, right? Um, but I just simply presume that my client is gonna ask to chat TBT or somebody else, um, Claude Di and I, whoever, um, you know, find the weaknesses, find the whatever, and it always comes up with something. Um, and most of it's ridiculous. Uh, but, so that's pretty easy to defend, but lawyers should count on the fact that their work is being kind of DoubleCheck now.

Mm-hmm <affirmative>. And Bernard, what are the most significant ethical risks associated with using AI in legal practice today?

Today, to me it's absolutely overreliance, right? So, um, overreliance, opi opacity, um, misplaced accountability, um, are the three biggest that I have come up with. Lawyers may trust outputs that they fully don't fully understand, or they failed to detect these subtle errors or of biases or nuance that, that Maura was talking about. And there's also a risk of eroding professional responsibility. Um, if AI becomes a shield instead of a tool. Ethics require that lawyers remain accountable. Every word we say to a court, every word we say to a client, we are personally responsible for. And so we have to understand that AI is a magnificent tool, but it is just a tool. And so, just like any other tool that I use in my practice, it is the quality control and the personal responsibility that it takes, I think is the most important thing that we've gotta focus on. And where the biggest ethical risk lie, like how, how many headlights have we seen of employers <laugh>? The best thing that came out of these new headlines is that at least we've all forgotten about the, the guide during COVID who had the cat face filter on at court. He is like, I promise I'm not the cat. We, nobody remembers him anymore. We now all we remember that, you know, it's all the court, the headlines about the, uh, guys who submit stuff with cases that are made up.

So I'll throw in two more. Um, one for less sophisticated lawyers, the folks who aren't working with Bennett, um, there are real risks of breaching confidentiality. So they don't fully appreciate that if they put in, uh, detailed information about a client or information client has given them into one of these open systems that, and they haven't set their settings to protect privacy, to make sure that the company isn't training on their data, uh, they could very well breach, uh, easily breach confidentiality. Uh, so that's one thing. I think there, uh, issues around bias as well. It's very, very important how you phrase that prompt. If I say my arguments are the winning arguments, right? That's a very different prompt than, you know, a more objective or neutral prompt or show. Tell me the arguments on both sides. And I, I think, um, some lawyers will run into trouble because they have, in essence, biased the system in some direction. And, and these systems are chantic. They're, they're set up to try to feed you what you want to hear. And, um, you can, you can go down the wrong, wrong path fairly easily, uh, just with prompts that aren't structured properly.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And like you talked about, confidentiality, how do we balance the duty of competence with the duty of confidentiality? Mm-hmm <affirmative>.

So, uh, they're not intention. I don't think. I, I, um, I, yeah, I, I don't see competence and confidentiality as, as, as, uh, you know, being one versus the other. I think being competent involved is understanding how the technology works, what you can put into that tool safely, and what you can't put into that tool because, uh, of either data security concerns or training that the system may train. Look, uh, Bennett has a system within his firm. He, if he doesn't know what he's doing, he could have two clients that are in the same business, uh, that are competitors. And while he may not have a conflict because they're not, uh, you know, directly involved in litigation or something like that, the information from one could filter into the searches that are done for the other client if he doesn't understand, uh, and have competence in terms of how he's managing his systems and, and things like that. So I, I don't know that they're necessarily intention. I, I, I, I think that that competence involves understanding issues of, of confidentiality.

Mm-hmm <affirmative>. Beautifully said. Yeah. And more on that, on navigating those risks. How should lawyers approach issues of bias, hallucinations, confidentiality, and data security when using atos? I know you spoke about this before, but can you go deeper into this?

So they need to be careful. I think rule number one is to either educate yourself or to associate with, uh, with somebody who does know the tools. I think part of the problem is, uh, and I think Bennett may have alluded to this before, is people run in thinking, I want me some ai, they, that's the wrong direction. The direction should be, I have this particular problem or, uh, stumbling block or whatever it is that I need to solve. So let me look at my current process and let me see what alternative processes that might be more automated that might help me with this and not, you know, not, oh, the guy down the block got X tool. I should get that tool. And that's what people often do. I think they, they need to think about their own particular use cases and, and, you know, areas of, of challenge and, and figure out what can I do to address these things?

And then they have to test it, because if it takes you six times as long to verify, then it took you to do it in the first place. You are not, you are not ahead of the game at all. And I, I think, uh, very interesting studies were done where they asked coders and the coders said, oh, I, I saved 20% of my time doing this. And when they actually looked at their time, they had spent more time, not less time. They weigh underestimated how much time they were spending. Uh, so I think, I think lawyers need to, uh, it's great that they're embracing these tools and, and I agree with Bennett about the curiosity and the flexibility, but there has to also be a little bit of skepticism and, uh, a care. It's not just getting me some ai, it's the right AI for the right task, um, with the right guardrails in place. Mm-hmm

<affirmative>. And Bennett, looking forward, what does a future ready lawyer look like in an AI driven legal ecosystem?

Yeah. So to me, a huge ready lawyer combines strong legal fundamentals with systems thinking and ethical judgment. And what I mean by that is they n need to know how to work with ai, right? This really is testing it, supervising it, explaining its role. Um, e especially transparently to clients like I, we are, like, we tout the fact that we are an AI driven law firm, right? Um, and so associates are less focused on producing volume, producing hours, right? If you, if you pay somebody for hours, you're gonna get hours. If you pay somebody for outcomes, you're gonna get outcomes, right? And so, what I think that a future Ready lawyer really focuses on delivering insight, trust, outcomes. In short, they're AI helps 'em become more human, not less

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And Mora, final strategic part. What strategic steps should law firms take now to integrate AI responsibly and sustainably?

Unless you're retiring in the next two weeks, this is not going away. So you can't just bury your head in the sand because, uh, as the same goes, you may not be replaced by ai, but you'll be replaced by a lawyer who was using AI responsibly and ethically and properly. Uh, so I think law firms need to be looking at their pain points, their pain points, not the guy down the blocks, pain points, their pain points, and figuring out what are alternatives? Where's low hanging fruit that I can automate, uh, without, you know, massive needs to, to check and validate and all of that. And, um, and experimenting, start piloting and testing on a small level. You don't have to have the big ultimate project. You can start with some small defined task and getting people comfortable, trying out different tools, seeing what works, what doesn't work, uh, just getting their hands dirty, experimenting on low risk stuff. And then as you gain more knowledge and more comfort, you move into, uh, bigger areas.

Mm-hmm <affirmative>. Incentivize that, incentivize that a curiosity, incentivize that innovation. Right? Um, and one of the things that I truly love about what AI can do is AI is best used on highly repetitive data-centric products, right? Or processes. If you look at, if you really study what, um, a particular group or sub practice group does just about, uh, this is the same thing that our, our analysis of the 10 thousands and lawyers, um, revealed is that every lawyer, no matter what their practice area, basically does six or seven things over and over and over and over again. That's more true on the corporate side than it is in the litigation side. And so they are, if you can, I really work closely with these brilliant, experienced attorneys who are making judgements about m and a documents, or about, you know, uh, employment contracts, whatever it is, right? And

Take their acumen and build it into the AI systems judgment. You take their judgment and build it in and prove it in. And you'll find like we, we experimented with many tasks and took some that were, that took 15 hours down to less than a minute, right? And then choose, um, product, uh, legal products that are the most easily flippable to a per task bill, right? There are some things that are just more easy, evaluate the contract, do this, right? Like, so find with the problems, like Maura said, your problems, 'cause you got 'em, where are you, where are you most inefficient? Where are your pain points? You know, that, identify those things and then work on them and figure out how to get AI to make the process more efficient, more consistent, more valuable for the client, and then easily explainable why you're charging for a fixed fee instead of by the hour.

Mm-hmm <affirmative>. Mm-hmm <affirmative>. And now since we are getting closer to the end, if you could give one piece of advice to lawyers feeling anxious about AI's impact on their careers, what would it be and why? What's your one minute golden nugget for them?

I would say throw yourself into it. Learn it, uh, take a online course or there are tons of free resources available now. Uh, I spend an hour a day reading an hour a day to keep up, and I can barely keep up with this stuff, but I know that if I spend an hour a day reading, I pretty much have, have a, a decent grasp on what is going on in this world. And there are lots of daily newsletters and things like the blogs that you can sign up for. Uh, but I would embrace it. I don't think I would pretend, uh, that it's gonna go away or pass over you. I, I, I would spend the time to learn it. Yeah.

And it's, it's the, don't try to compete with ai, try to compete on judgment, right? Meaning AI can do things faster, they can compile stuff faster, and you're never ever gonna win that. What you're gonna win on is judgment. And so lawyers who struggle will be the ones who define value based on the task that AI can already do, right? Um, lawyers who, those, those gunners who just put stuff out, right? Um, the lawyers who really thrive are gonna deepen their expertise, spend time, not just with the ai and it, and you should like, and for people who are like, well, I I don't even know how to get started, ask it, it will tell you how to use it, right? Like, honestly, the best thing to do, I'm a lawyer, I just have, what's the best thing I could do to learn how to use AI better?

It will tell you, right? Um, but diving into the legal area that you are in, understanding the nuance of the law, understanding where a, a doctrine sits at this point in different circuits, at this, at different times, right? It's that piece and then it's your human connection with your client. They don't call us counselors for nothing, right? Um, it's not just we are giving counsel, but sometimes it's because we're kind of therapists. Uh, but, but, so it's do those things that AI cannot do. And then when you combine that with the power of ai, you are Ironman or Iron Woman. Pepper Pots had a pretty cool Ironman suit too. Um, but that is, would be my best advice.

Now, that was a powerful conversation. We covered everything from the death of the billable r to the birth of the AI augmented advocate. A massive thank you to our guests, Bennett Borden and Maura Grossman for sharing their expertise and giving us a real honest look at where the profession is heading. If you take anything away from this conversation, take this as Mara put it. So well, you may not be replaced by ai, but you will be replaced by a lawyer who uses AI responsibly, ethically and properly. The firms that adapt are the ones that rise to the top. And speaking of adapting and rising to the top, a quick reminder about our sponsor, Equa Marketing, the same way Bennett and Mora talked about firms needing to evolve their systems, your marketing needs to evolve too. The clients searching for your practice area online are out there right now.

The question is whether they are finding you or finding the firm down the street, Quas team will sit down with you, audit your online presence, study your competition, and walk you through custom recommendations to make sure your firm is the one they find. That session is normally $900, and it is completely free for our listeners. Book it at www.dominatelaw.com/msm. If you found this episode helpful, share it with a colleague who needs to hear it, and do not forget to follow the Dominate Law Podcast so you never miss an episode where we help attorneys build stronger, smarter, and more profitable practices. I'm Ashton Rogers. Thanks for listening. Go out there, use these tools and dominate your practice. We will see you in the next episode.