Trascrizione completa di The Hank Green Problem for Lawyers: Why Disclosing AI Use is So Complicated | Nika Kabiri

that, good bedside manner or explains things really well, et cetera, et cetera, because they can't judge lawyering. 30 00:04:53.480 --> 00:04:58.969 They don't know what lawyering even means. That's really kind of the simple answer to that. 31 00:04:59.100 --> 00:05:11.020 It sounds a bit like there's a halo effect in the sense that y- you know, the halo effect is if somebody's pretty, I assume they're smart, right? If one characteristic aligns with something positive to me or- Yeah... 32 00:05:11.060 --> 00:05:15.020 negative in, in that sense, then I'll assign it to other things that are, that are unrelated. 33 00:05:15.360 --> 00:05:25.140 Are consumers doing that because they actually care about things like what flavor coffee is in the office, and do people talk nice to me? 34 00:05:25.180 --> 00:05:33.360 Is, is, is that a good they're looking for, or are they using that as proxies for other things that they actually are looking for? Proxies. They're proxies. 35 00:05:33.480 --> 00:05:45.810 It all comes down to the fact that humans, the human brain burns up so much fuel. It takes up 2% of our physiology, 20% of our energy. 36 00:05:46.260 --> 00:05:58.700 So it is by design, without us even thinking about it or knowing about it, it is by design trying to constantly be efficient. We could go do all the research we need to do to know what a good lawyer is. 37 00:05:58.800 --> 00:06:06.776 We could l- study the law. We could, like, become good lawyers ourselves and then match what we know with a lawyer. That takes a lot of energy, takes a lot of time. 38 00:06:07.256 --> 00:06:14.416 Brains don't wanna go through that, and, and we don't really wanna go through that. Even if the brain could do that, we would opt not to do that. So it's really just a substitute. 39 00:06:14.456 --> 00:06:23.436 It's the same as saying, "Well, I don't know much about public policy. I don't know much about this presidential candidate. What did he do in Texas? I don't know, but I sure would like to get a beer with him." 40 00:06:23.476 --> 00:06:29.956 It's like that. It's kind of that sort of thinking, substituting one question for the other. Yeah. Yeah. 41 00:06:29.996 --> 00:06:36.936 Well, uh, dear podcast listener then, I want you to pin in your mind, uh, because we're gonna talk about how you communicate about your use of AI. 42 00:06:36.946 --> 00:06:44.396 I want you to pin in your mind the fact that consumers are using these proxy signals. They're, they're taking one truth to mean a bunch of other truths. 43 00:06:44.516 --> 00:07:10.816 I did wanna mention that Thomson Reuters did a consumer legal needs survey, and they identified five things, five enduring factors, and I think you see this echoed in Clio's Legal Trend Report, that when clients are choosing a lawyer, they are looking at expertise, which the way they're using that word is sort of more specialization, what's your area of specialization; reviews online; licenses; certifications; and trustworthiness, another vague word. 44 00:07:10.996 --> 00:07:18.416 What are those third-party signals doing? Why do consumers trust things like reviews, licenses, certifications? Same thing. 45 00:07:18.556 --> 00:07:28.416 They can't evaluate the quality of a lawyer based on the right credential, on the right merits, right? They don't know how to do that. What other people think is the same shortcut. 46 00:07:28.916 --> 00:07:34.246 It's a data point that gives them a little bit of confidence. There's also this idea of the bandwagon effect. 47 00:07:34.336 --> 00:07:49.256 Now, I'm not saying people hire lawyers because they're on some bandwagon, but the idea behind the bandwagon effect is that, like, if other people are doing this, the brain just automatically, without us even thinking about it, assumes if so many people are doing this, then it must be the right thing to do. 48 00:07:49.356 --> 00:08:00.096 So if you see a lot of positive reviews for a lawyer, you assume, "This lawyer can't fail. All these people are saying it's good." That's why it matters. That's why referrals are also a big deal. 49 00:08:00.136 --> 00:08:08.116 Like, a, they still and probably always will be important for lawyers because it's a credence. Knowing that a friend will say, "Oh, yeah, I hired this person for my divorce. 50 00:08:08.136 --> 00:08:13.906 They're really, really good," or, "They're a friend of mine. They really do great. They do a lot of wills. I would call them." 51 00:08:13.916 --> 00:08:23.796 It doesn't necessarily answer the question of, "This is a good lawyer," but it does answer the question of, "Can I feel comfortable that I'm not gonna make a crappy decision? Like, I can feel secure." 52 00:08:24.276 --> 00:08:29.196 And that's really what we want. We just wanna feel good about what we do. We don't necessarily always wanna choose what's best for us. 53 00:08:29.376 --> 00:08:39.366 And I wanna highlight because we're gonna shift to talking about AI use in communication. What people are looking for is what's called signaling, uh- Right... to your point. This is the part that you can control. 54 00:08:39.436 --> 00:08:49.376 If you are a lawyer and you're trying to get people to trust you, it's really hard to control things like referrals. Obviously, you can control things like certifications and, and your licensure. 55 00:08:49.416 --> 00:09:01.126 But in general, what you're trying to do when you have this kind of good is send these signals to the world that says, "I, I'm not saying I'm a good lawyer," because you don't have the ability to measure that, but I'm trying to say to you, "I did hard stuff." 56 00:09:01.396 --> 00:09:13.076 Right? Like, it took some work for me to get this degree. It took some work. And consumers like costly signals. AI. The issue with AI for some consumers is that it's a cheap signal. 57 00:09:13.616 --> 00:09:20.216 It's saying to me that you took the shortcut. You took- Mm-hmm... the short route. Mm-hmm.

But the sentiment about this is all over the map. 58 00:09:20.306 --> 00:09:31.256 And going back to Hank Green, this is what was so fascinating to me about that Hank Green experience. Mm-hmm. The kids do not feel the same way about AI that the boomers do. 59 00:09:31.376 --> 00:09:40.836 Anytime President Trump puts out one of those AI-generated images, do you know what the kids call that? They call it boomer art. The fake art is referred to as boomer art. 60 00:09:40.876 --> 00:09:55.176 There's data that backs this up, data that says boomers are more trusting, 63% to 47%, in terms of helpfulness. So the baby boomers, 63% of them think AI is helpful, and only 47% of Gen Z do. 61 00:09:55.396 --> 00:10:08.016 80% of boomers think AI takeover is inevitable, where only 51% of Gen Z does. What's weird about that is that Gen Z uses AI way more than boomers do. 62 00:10:08.036 --> 00:10:16.416 What's the relationship, do you think, with that, this idea that, that familiarity is actually sort of breeding contempt with younger kids? What do you think the relationship is with that? 63 00:10:16.676 --> 00:10:26.046 I'm not really sure that it's the familiarity that's breeding the contempt alone, and this is just, you know, the research I've done in academia, in sociology. 64 00:10:26.116 --> 00:10:37.716 Like, there are generational or cohort-based traits that are, are driven or defined by the moment in, in history. When boomers were coming of age, they didn't have access to the information we had. 65 00:10:38.016 --> 00:10:47.655 The information they got about a product was from the salesperson in a store, not reviews, not, like, all this online. You know, you don't go to Reddit. You don't get that sort of thing. 66 00:10:47.856 --> 00:10:58.956 So you get a very kind of curated, very packaged, uh, pitch. Same with, you know, political candidates. You hear w- what they are about from their, from their mouths, right? 67 00:10:59.516 --> 00:11:10.996 So these days, you get Gen X, millennials, Gen Z, even Gen Xers are s- getting a lot of information from a variety of places. So they learn, they see like, okay, so and so said X. 68 00:11:11.156 --> 00:11:15.456 They were probably lying 'cause all this evidence shows... You know, they have access to that. 69 00:11:15.516 --> 00:11:30.376 There is an inherent lack of automatic trust that's built into these younger generations based on the fact that they have all this information available to them, and they've trained those muscles to be skeptical because they've had the luxury of being skeptical because they have access to that data. 70 00:11:30.636 --> 00:11:38.116 Older generations are just naturally more trustworthy of a lot of things. If you have a, a, an authority figure come in, they're much more trustworthy. 71 00:11:38.316 --> 00:11:47.946 Lots of research that I did at Avvo pointed to this in terms of, you know, older generations are much more likely to trust a lawyer and say, "You take care of everything. You just do the work." 72 00:11:48.396 --> 00:11:57.256 Younger generations are like, "Yeah, I kinda, maybe I wanna DIY some of it. Maybe I wanna, like, second-guess my lawyer." So it's a, it's kind of a generational thing. 73 00:11:57.336 --> 00:12:08.596 Also, you know, when you are raised, when you are AI native, when something is, like, native to you, AI is native, tech is native, you feel less wowed by it. It's not so like, "Ooh, this is cool. 74 00:12:08.696 --> 00:12:10.476 Look at what it can do," because you're used to that. 75 00:12:10.516 --> 00:12:19.728 But when an older, you know, an older individual like my dad, [chuckles] when he asks me to ask, you know, ask AI what this bump is on, you know, whatever, it's like this-... 76 00:12:19.738 --> 00:12:26.028 this thing Magic And he, yeah Yeah And he will believe it, right? So I think that's, that's probably a lot of it as well. 77 00:12:26.288 --> 00:12:36.248 Well, and you see, we just had Joshua Lenon from Clio on to talk about the Legal Trends Report, and they noted this weird thing that solos usually are faster to adopt technology, but not with AI. 78 00:12:36.258 --> 00:12:50.248 It's just, just this weird thing, and I think part of it, I mean, the popular perception, is that for people who perceive AI as a threat, including young people, including solo and small firm lawyers, if they perceive it as a threat, they might be less likely to like it. 79 00:12:50.258 --> 00:13:00.468 They might resist it more. How do people think about change, especially in technology? Does that change our perceptions if we perceive something as a threat? Can I be honest? 80 00:13:00.548 --> 00:13:07.788 I'm a little hung up on the fact that solo lawyers are less likely to adopt AI. I would love to know why, because that does sound really counterintuitive. 81 00:13:07.808 --> 00:13:19.438 And again, like when you ask that question, adopt AI, you always have to ask for what? I've done a ton of research. I do still a ton of research for Amazon, Google, Meta on, you know, adoption of AI B2B. 82 00:13:19.708 --> 00:13:28.808 The answer to my questions are always, "For what?" Like, AI for what things? Without that nuance, it may be hard to really understand what the question really is, what the problem really is. 83 00:13:29.098 --> 00:13:35.878 I don't know if that makes a lot of sense, but yeah. It does, and I guess maybe what we're pointing to is high stakes, low stakes, right? 84 00:13:35.928 --> 00:13:48.608 Where there are low stakes, uh, issues, low stakes decisions, like, you know, I really don't care if an AI figures out optimal power usage for my air conditioner to make it so that I save a little money. 85 00:13:48.768 --> 00:14:04.868 If I see the money coming back, that's fine, but if I have a very complex decision like how long should this person go to prison for, you know, carrying a gun in the wrong place or whatever, I have a different relationship with that decision because the

decision itself is different. 86 00:14:04.878 --> 00:14:22.728 The interesting thing, going back to sort of widespread and going back to consumer perceptions, taking the generational, uh, divide out of it, th- there was a study that showed that only 14% of participants in that survey felt that disclosing your AI use increase trust. 87 00:14:23.148 --> 00:14:36.988 Meaning, if I see a lawyer and they say, "Hey man, I use AI every day," only 14% of people say, "Cool, that's what a pro would do." And in fact, that study showed over the last few years an erosion in that number. It... 88 00:14:37.128 --> 00:14:50.168 The distrust used to only be 20%. It used to only be, "Hey, I use AI" would make only 20% go, "Wait, what?" Now it's 40%. And so there's this sort of widespread weird relationship with trust. 89 00:14:50.188 --> 00:14:57.248 Do you think that that has to do with the stakes and the decisions? I think it partly has to do with what we're hearing about AI through the media. 90 00:14:57.268 --> 00:15:08.168 I don't ever open my New York Times or Washington Post app and read, "Look at what Claude did today. Isn't it awesome?" It's more like, "There's a hack. It's out of control. Data centers are everywhere." 91 00:15:08.648 --> 00:15:21.268 So I think that negative press, it doesn't really help the reputation of AI. I think we're also into it long enough, it's not as novel anymore to where we've started to see the, the downsides ourselves. 92 00:15:21.468 --> 00:15:30.827 I, myself, when it first came out, thought it was brilliant and amazing, and look at all the things it could do, and I found myself getting caught up in the novelty of the tool as opposed to the effectiveness of the tool. 93 00:15:30.868 --> 00:15:34.228 Now, I rarely use it to solve any sort of problem. 94 00:15:34.368 --> 00:15:45.548 I prefer to read an article that's written by, you know, a, somebody I can trust, because you, you yourself are using it enough and you're kind of tired of the glitches you are noticing in it. 95 00:15:45.788 --> 00:15:47.328 That might be part of it as well. 96 00:15:47.568 --> 00:15:59.868 I mean, I wonder, like, and I don't know if I'm really answering your question directly or just riffing off something that I'm thinking about because your question made me think of it, but, like, trust overall, I mean, we are kind of less trusting as a culture. 97 00:16:00.168 --> 00:16:09.408 You know, just look back at the last five, six, 10 years. The trust in our leaders is eroded. Trust in business is eroded. Trust in our economic system is eroded as a principle. 98 00:16:09.518 --> 00:16:20.928 Um- Just the idea of experts in general has eroded... N- totally, which is, in some ways, kind of nice because some experts aren't really experts, and so you do need to kind of call them out. 99 00:16:20.938 --> 00:16:27.008 There are a lo- [laughs] a lot of people out there who claim to know it all, and they, if you really look at it closely, they really don't. 100 00:16:27.048 --> 00:16:35.388 But at the same time, there are some really, like Fauci, I mean, he knows what he's doing. He's an expert, and to kind of lose faith in his credibility is kind of problematic, I think. 101 00:16:35.547 --> 00:16:50.188 So yeah, it's, it's kind of a, it's a weird time. Part of me wonders if the disclosure about AI, and we're gonna talk about the data around disclosure, the disclosure of AI is bringing to the surface a decision- Yeah... 102 00:16:50.198 --> 00:16:59.168 that somebody previously didn't have to make. Previously, they would just say, "Oh, he's got the certification. He's a member of the bar. He's got the law degree." Yeah It's all third party. It's putting off- Yeah... 103 00:16:59.248 --> 00:17:11.368 on somebody else the evaluation. As soon as you tell me you're using AI and you're a purported expert, now I have to make the call of, okay well, how are you using it? What are you using it for? 104 00:17:11.428 --> 00:17:18.728 Is that the right place for it? What am I fi- You're just adding all this cognitive burden to a decision maker. It's again, what is AI being used for? 105 00:17:18.888 --> 00:17:28.888 There's research on legal consumers that says that they are totally fine if lawyers use AI for routine tasks. 106 00:17:29.028 --> 00:17:39.088 Like, if you want to use AI to draft a fee agreement, as long as the fee agreement and the terms are acceptable to your client, why would they care how it, where it came from? 107 00:17:39.108 --> 00:17:49.008 But if it's, "Hey, I used AI to come up with the strategy for your case," then there are a couple things that go on, and this is, you know, what research has found. 108 00:17:49.068 --> 00:17:58.068 The consumers start to think, "I can't trust you because you don't have confidence in you, because if you had confidence in your own way, ability to think of a strategy, you wouldn't rely on this tool." 109 00:17:58.088 --> 00:18:05.284 Suddenly, the trust is eroded. And not only that, but you set- You know, you set yourself up for a lower tolerance for error. 110 00:18:05.364 --> 00:18:16.374 Again, research shows that people are much less likely to trust an algorithm twice if it messes up once than they are a human. As a human, you can, you can make errors. 111 00:18:16.404 --> 00:18:26.904 As a lawyer, you can make errors, you can correct them, and you are more likely to be forgiven. AI, not so. The people aren't so forgiving. I was just watching an interview of the author of The Life of Pi. 112 00:18:27.144 --> 00:18:38.724 The interviewer and he were talking about AI, and the conversation turned to, like, you know, if you think about it, what would a reader of a novel want a human to tell them versus AI? 113 00:18:38.804 --> 00:18:41.524 And you gotta put yourself in the, in the mindset of the consumer too. 114 00:18:41.564 --> 00:18:54.064 If you were to hire a lawyer in a practice area you're not familiar with for a case that, you know, you

really think is high stakes, what would you want a human to tell you, and what would you want AI to do for you? 115 00:18:54.104 --> 00:18:57.494 They're just kind of different, different things. Yeah, and you saw this. 116 00:18:57.584 --> 00:19:07.224 There was an author recently who was self-published and did really well, and then became traditionally published, and right after, they ran one of these AI- Yeah... 117 00:19:07.234 --> 00:19:14.764 spy tools to find out if, uh, she had used AI to write the book, and they decided that, "Yep, yep, that totally happened." Yep. Uh-huh. And it was just a disaster for this person. 118 00:19:14.864 --> 00:19:25.544 You actually see in the data, this was in the Clio Legal Trends Report, that 78% of clients want disclosure about whether you're using AI, but only 18% of firms are doing that. 119 00:19:25.604 --> 00:19:40.384 And in the financial industry, 79% of investors said they'd be upset if they found out afterwards that the advisor had used AI, but only 33% of their advisors were telling them that they were using it. 120 00:19:40.464 --> 00:19:46.904 So, what is the relationship? Th- this disclosure thing I find fascinating because we don't disclose that we're using email. 121 00:19:47.064 --> 00:20:01.444 What is it about this thing that makes consumers feel like, "Well, I hired you as an expert, you better tell me when you're pushing the buttons"? Again, it's like disclosure about what? AI for what? 122 00:20:01.804 --> 00:20:14.404 I'm pretty sure that if, as an attorney, you told your client, "Hey, I am using, you know, Westlaw's AI-powered research tool," they're one of the clients that I work for or do research for. 123 00:20:14.604 --> 00:20:33.644 If you are using Westlaw's AI-powered feature and you are doing the research much more rapidly, you're com- you're not missing cases that you otherwise would miss, um, if you tell a client that and you add, "But at the end of the day, I am making the judgments based on what I'm reading and based on what I'm human checking," that's important too. 124 00:20:33.684 --> 00:20:38.774 Like, I'm using AI, I'm human checking it, and you better human check it. You can't just make yourself a liar, right? 125 00:20:38.884 --> 00:20:54.304 I would imagine c- consumers would think, "Oh, wow, you are actually using a tool that will make your work for me much more effective, but you're still maintaining the role of an expert," um, instead of just like, "Hey, do you wanna know that I'm using AI?" 126 00:20:54.324 --> 00:20:59.124 And then the imagination just soars. Let me highlight that because I, I think that this is a really important take home. 127 00:20:59.144 --> 00:21:08.144 It sounds like what you're saying, and tell me if I'm wrong, is don't just have on page four of your engagement letter, "We sometimes use AI. Have a nice day." 128 00:21:08.544 --> 00:21:17.114 You're saying map out for me a good sense of how you're using it. Where are you making decisions? Where is the tool? Be more open, not less. Is that- If you- Is that what you're saying? If... 129 00:21:17.184 --> 00:21:28.764 Yes, yes, if you choose to disclose. And again, like, do your clients need to know the tool that you're using as you do, you know, your case research? Do they need to know that? Did they ever need to know? 130 00:21:29.144 --> 00:21:38.724 Did you ever need to announce to them, "Hey, I'm not really going, you know, like 1990, what was it? '90 whatever, one or whatever. I'm not going to the library anymore. 131 00:21:38.864 --> 00:21:43.204 I've actually got this, you know, computer, this software tool." 132 00:21:43.304 --> 00:21:52.444 As long as the work is well done and the parts that a human needs to be doing in the eyes of the consumer are being done by a human, I think it's a question, right? 133 00:21:52.524 --> 00:22:00.834 I'm not saying I sit on any side, on any one answer, but are you really morally obligated to tell a client that you're using AI to, like, respond to emails? 134 00:22:00.924 --> 00:22:08.894 As long as the content of the email is human checked and accurate, I guess I wonder what you need to tell them. Let's go back to the, the cognitive burden of the consumer then. 135 00:22:09.304 --> 00:22:22.584 There's this line from the ABA formal opinion that says that if you proactively inform clients when it's important to the client decision making, that it will help them. I'm gonna give you a, a platform to rant on. 136 00:22:22.663 --> 00:22:28.613 This feels like it's coming from that old economist idea of everybody's very rational. Every consumer who's- Yeah... 137 00:22:28.613 --> 00:22:38.284 looking to hire a lawyer is doing all this weighing, and they're, they've got their pros and cons table and they're figuring stuff out, as opposed to the way consumers actually make decisions. 138 00:22:38.444 --> 00:22:51.834 I guess what I'm trying to get at is when we're disclosing AI use, we're not giving the consumer the tools necessarily to evaluate, "Oh, here's what I think is the proper use of AI. Here's the negative use of it." 139 00:22:51.844 --> 00:22:58.904 What we're doing is signaling. Again, we are screaming at them, "This is what we think is valuable. This is what we care about. 140 00:22:59.164 --> 00:23:08.954 We care about your kids, for example, so we're gonna put that in all the documentation that says anytime there's a kid decision..." And then the consumer can go, "Oh, I like kids." What is the relationship, I guess? 141 00:23:09.164 --> 00:23:16.854 Basically, I just want you to tell me I'm right, that- Yeah... this disclosure [laughs] is not about- And you're pretty... h- Yeah, it's not about hyper-rational decision makers. No. 142 00:23:16.884 --> 00:23:21.064 It's about the signals we're sending, about what- No, it's the substitution questions. Right. It's all that. 143 00:23:21.144 --> 00:23:30.924 I remember I did a study at Avvo that asked millennials what they look

for in a lawyer, like are millennials evaluating lawyers in a particular way? And they liked energy. They wanted energy. 144 00:23:30.964 --> 00:23:39.344 Like, because if a lawyer showed they had energy, then they would fight for them, they would, like, work for them, they would call them back. This has nothing to do with being a good lawyer. 145 00:23:39.364 --> 00:23:43.084 It has to do with being a good client services representative. 146 00:23:43.184 --> 00:23:53.724 AI is another one of those things that if used well, could signal efficiency, could signal I'm covering my bases, could signal I'm not gonna leave any stone unturned. 147 00:23:53.784 --> 00:24:03.804 If this is a, you know, like, criminal case, well, the prosecutor's using AI, but I'm not gonna use it, like, that doesn't signal, that doesn't signal well. But it's also, again, how you, how you signal. 148 00:24:03.984 --> 00:24:11.904 It's about how you package or message the use of AI and what you reveal regarding what you're using AI for. 149 00:24:12.332 --> 00:24:22.732 Anything that involves big thinking, strategy, huge decision-making, something that could really speak to the crux of your case. 150 00:24:22.972 --> 00:24:31.492 If you have a lawyer that says, "I let AI totally take care of that," they wouldn't wanna hear it. No lawyer wants to do that anyway, from what I'm learning and the research I've been doing. 151 00:24:31.692 --> 00:24:41.272 Lawyers want to own that stuff anyway, so I don't know if you would have a problem there most of the time, but it can be a signal of, "I'm doing the best for you, man. 152 00:24:41.352 --> 00:24:43.822 Like, I'm doing the best I- I'm using the best tools out there for you." 153 00:24:44.152 --> 00:24:54.452 Right, and it, it, again, to your point about the how energetic the lawyer is, that lawyer projecting energy is not because the person on the other side is gonna be this hyper-rational decision maker. No. 154 00:24:54.482 --> 00:25:03.952 They're fine with you taking naps if you get the work done. They're using that to try to figure out the substance- Yeah... behind it. I do have to say, though, billing is a problem, right? 155 00:25:04.092 --> 00:25:17.942 Like, clients who have been surveyed have questions when the bill comes. If, if there's been AI disclosure and you can't show me why that didn't decrease the bill, then consumers are asking questions. So I, I... 156 00:25:18.212 --> 00:25:25.832 There's probably a special relationship with disclosure and the bill. That is interesting. I did not think of that. 157 00:25:25.852 --> 00:25:38.232 But that does make a lot of sense because, like in my industry, in research, I do a lot of qualitative interviews. You know, some companies are developing AI moderators to replace what I do. 158 00:25:38.432 --> 00:25:46.062 I bet the clients would expect to pay less for that, and if they have to pay the same, they would be a little bit dismayed. Like, why? You know. 159 00:25:46.092

--> 00:25:57.652 So yeah, I'm trying to think of analogies in other industries, because lawyers are unique, but then they're also not. Truthfully, if you are using AI and it is cutting the time that you spend on a case, 160 00:25:58.572 --> 00:26:09.312 ethically, should you be charging the same amount? I mean, is that an ethical question? Um, you know, there are different m- methods or approaches or thoughts on cost accounting. 161 00:26:09.412 --> 00:26:25.302 A lot of people who talk about expertise say you're not charging for your hours or your time, you're charging for the end product, you're charging for the value you offer, you're charging for the years and years of training and education, your expertise, and it doesn't matter if you spend four hours doing, doing work for a client. 162 00:26:25.552 --> 00:26:35.372 If you charge what might be the equivalent of 30 hours of work, it's okay because that's what you're worth. Um, I mean, these are kinds of a lot of, of questions that lawyers need to ask, but it's a competitive market. 163 00:26:35.492 --> 00:26:46.212 People want cheaper lawyers. Well, and of course the problem is, like, most studies are showing that AI is not actually speeding things up for people, so there's this weird butting up against perception and reality. 164 00:26:46.332 --> 00:26:49.132 And our final question is, okay, so what should we do? 165 00:26:49.212 --> 00:27:01.632 Given what consumers are signaling they want, given what they seem to care about, and again, this is very market specific, you have different positioning that will give you different results, and there's quite, there's a lot of it depends floating around this. 166 00:27:01.792 --> 00:27:04.172 You're gonna have to decide how you talk about AI use. 167 00:27:04.592 --> 00:27:20.272 But given l- letting the consumer's perception and beliefs then lead what we do, there was a, a study of freelancers on Upwork, and there was this crazy mismatch between the hirers and the workers in terms of what AI was okay for. 168 00:27:20.352 --> 00:27:30.252 So the worker said, "Hey, if it's just email exchanges, just communication, who cares? Use AI all day." But the people hiring were the opposite. They were like, "No, communications are super important. This is our brand. 169 00:27:30.292 --> 00:27:33.922 This is, like, the, the thing we do." So there's this weird, there's this weird relationship. 170 00:27:34.122 --> 00:27:44.632 Given what we know about experts, given what we know about how people choose their experts, is there any guidance from consumer perceptions in terms of how we should use AI? 171 00:27:44.712 --> 00:27:54.312 Where does it go in our workflow in a way that we can maintain that perception of expertise? What makes you an expert? What makes a lawyer an expert? Is it writing an email? 172 00:27:54.512 --> 00:28:03.612 Is it, you know, organizing files in a way that makes the papers easier to find? Papers, I'm so old school. Files easier to find. 173 00:28:03.912 --> 00:28:16.022 Does it make you a lawyer or not a lawyer to use an AI powered research tool like Westlaw might have? What makes a lawyer a lawyer? And I think it really comes down to the judgment. It comes down to decision making. 174 00:28:16.072 --> 00:28:28.571 It comes down to, like, you know, knowing how to care for your client and deliver, even if they don't win, doing what a lawyer... What we know as lawyers, like, we know what good lawyering is. Just stay true to that. 175 00:28:28.792 --> 00:28:36.612 Yeah, I think the compassionate thing is to have a long answer but to give a short answer, right? [laughs] Like- Right... ha- have your values established. 176 00:28:36.912 --> 00:28:42.532 Go put some e-book somewhere on your website that says, "This is how we think about AI and its impact on expertise." 177 00:28:42.832 --> 00:28:47.492 But when you go to a consumer- Get a journal Yeah, yeah Journal Talk to yourself about it I mean, you know- Talk to your therapist about it. 178 00:28:47.512 --> 00:28:51.852 But when you go and you talk to consumers, make sure you understand that they're operating on signals. 179 00:28:51.912 --> 00:29:01.812 They are taking very short shorthands, so you just have to be really thoughtful about, okay, what am I signaling about my beliefs, about expertise, about the work that I do, about the value that I provide? 180 00:29:02.212 --> 00:29:14.712 If you can be conscious of that, if you can be cool enough as an agent to do the work to have a long answer, but smart enough as a communicator to- Yep... have a short answer, you'll be okay. 181 00:29:14.812 --> 00:29:22.392 Uh, Nika, I appreciate that consumers are super weird people. [laughs] Lawyers- We're, uh, the people... are super weird people. We're pe- we're weird They're just, we're all just weird people. 182 00:29:22.402 --> 00:29:32.612 We're weird Uh, and I appreciate you getting us into the heads of us strangeo weirdos. Absolutely. One thing the AI will never do is be as odd as we are, so, uh, that's how- Thank, thank you. 183 00:29:34.312 --> 00:29:37.432 [laughs] That's how I know, uh, we'll remain, 'cause we're the weirdos. 184 00:29:37.572 --> 00:29:49.772 Uh, Nika, for people who wanna connect with you to learn more about Decision Science and the work you're doing, avoiding all that and writing fiction now, and writing in your journal, and processing the, uh, existential angst that exists in the world, what's the best way to connect with you? 185 00:29:50.132 --> 00:29:56.872 Just don't. Don't connect. No, I'm kidding. [laughs] The kind of work I do is so exclusive now, I don't even have a website anymore. I don't advertise anymore. 186 00:29:56.972 --> 00:30:06.762 I don't have a place you can find me, but you can find me on LinkedIn. So just connect with me there. I'm pretty, you know, unless you're sending me a sales e- email, I'm pretty- Yeah... open to connecting, yeah. 187 00:30:06.772 --> 00:30:14.002 Then ask me, I'll send you Nika's link. If you wanna send any sales solicitation business emails, let me know. I'll send it to you. That'll be in the show notes. Thank you. 188 00:30:14.032 --> 00:30:22.172 Yeah, Nika, thank you for joining us- Thank you... on the Red Line for Spellbook. Uh, for our, for everybody else watching, we will see you in the next episode. Thanks, Nika. Thank you