How Lawyers Can Use AI Without Losing Trust or Breaking Ethics with Christy Zlatkus

Speaker A: AI can be a force multiplier, right?

Speaker B: Yes. You gotta be really careful, and you can't take out your own brain. That's where a lot of people get themselves in trouble.

Speaker A: Have you used it for research and seen it give you stuff you didn't know that existed?

Speaker B: One of the things that I asked it to do was to craft a legal argument. It was interesting because one of its arguments was just like, no, you're so off. Ah. But it also did craft an argument

Speaker A: that I. Christy Zlatkis is the CEO and founding attorney of Z Family Law in Rockville, Maryland. A former associate attorney at Avery and Upton, she specializes in divorce, custody, and child support cases. Christy is committed to exceptional client service and leveraging technology to enhance her firm's operations.

Speaker B: As an attorney, you still have to read everything that your paralegal does. Your name is going on it. You have to make sure that it's right.

Speaker A: I want the AI to take the notes for me, and then I want to review the notes and decide what I want to use and what I want to throw away. What does authenticity mean to you in this realm?

Speaker B: I'm going to say it in three different contexts. Number one, making sure I have written policies and procedures. Don't you dare use AI without my permission. And.

Speaker A: All right, everyone, welcome back to the AI for Business Owners podcast sponsored by Syngen AI I am your host, Jeff Torello. As always, Syngen is an interesting consulting company. We are an AI consulting company, but really, we are a problem solver. And, uh, one of the mottos that we use in the company is that AI is not the solution to every problem. And so if you're looking for assistance or help or just want someone to talk to technologically, uh, give us a call. The answer may not be by an AI tool. In fact, it very rarely is. Okay, uh, in this episode, we're going to have. We're actually going to have quite a lot of fun. I'm going to let Christy introduce herself, but we have some interesting topics to discuss, and, uh, I think you're going to really enjoy this one. Christy, if you want, introduce yourself to the audience, tell us who you are.

Speaker B: Hi, I'm Christy Uzlakis. Um, I'm the CEO and founding attorney of Zee Family Law, where we don't write endings, we create new beginnings. And we are a family law and estate planning firm in Rockville, Maryland, in the D.C. metro area.

Speaker A: Awesome. I like the motto. I like that. That slogan's pretty cool.

Speaker B: Thank you. I appreciate it.

Speaker A: You're welcome. Okay, so Obviously, this is a, uh, you know, an AI focused podcast. How did you get to the point where your interest in AI got to the, uh, level where it's integrated or you're trying to integrate it in your business? Like, what was your journey?

Speaker B: So, you know, I've been watching AI for a while, and as an attorney, I have to look at things, you know, very, very skeptically because we have to make sure that we're maintaining client confidentiality. And also there's a lot rules when it comes to attorney advertising and making sure that we're following their rules. But I've also been fascinated by AI because anything that you can do that multiplies your impact is a powerful. I don't even want to say too, it's a powerful weapon. It can be an amazing thing, but it also can be a very dangerous thing. And so I have been really thinking about how to integrate it into my law firm. Um, and, you know, the first time it kind of whacked us in the face with a really good opportunity was, was last year. And we've been kind of exploring and adding things slowly, piece by piece ever since.

Speaker A: Let me add to that just a, um, I spent a lot of time working in the defenders industry, probably about 10 years. I was in Reston for quite a long time as the cto. So just in your neck of the woods, basically, the book I wrote in 2024 around AI, the phrase that I wanted to use and I stole directly from the DoD is AI can be a force multiplier, which is kind of what you just described, right? It can be a force multiplier for an individual, but it can also be a force multiplier for an entity, for a company, for an audience, et cetera. Do you have any personal benefits of AI that you could relate before we get to using it? Like, in other words, just for you on your own. And not necessarily like shopping. I mean more like helping you write or helping you think or research, et cetera.

Speaker B: You know, I personally really like ChatGPT. I know it's like the go to for everybody my age, but I have enjoyed it for helping me come up with like, catchy titles to things, or even as I was doing my own personal branding, which I guess is sort of work wise, but it helped me narrow down my ideas. I would tell it things like, hey, I want to do X, Y or Z. Give me like 10 names now give me 10 more names now give me, you know, 30 more names. And half the time I didn't even end up choosing in Fact, more than half what like ChatGPT did, but it's like having someone in the room to kind of popcorn some of those things, uh, you know, around for you. That's probably the first place that I started dabbling in it. Uh, and that was useful for me because it really did get my creative juices flowing in a lot of ways.

Speaker A: That makes sense. I think it's a very logical tool as a sounding board. And one of the things to comment on what you just said, uh, one of the things I do think is relevant is that the AI is going to suggest things you probably wouldn't have thought of and you may dismiss, but that process in and of itself helps you think, oh, well, that, that doesn't fit for me, but what about that? And then you can continue to explore that. You can bounce back and forth with it. What I find, the tooling that I tend to use, I end up using multiple different models. And I use that on purpose so that, um, I can explore the various models and see which way they lean. Oh, yes, number one. But also you get two completely different answers. If you were to use, you know, grok and ask a question and then ask ChatGPT, and then ask Claude or Gemini, the responses are different.

Speaker B: And from my initial dabblings in ChatGPT, like, that's kind of where I've taken it, where you dabble in the different ones. I've also enjoyed trying to train them to sound like me. Uh, to see, like, hey, is this. What's the impression of this? But yeah, using the multiple different tools is a really great way. Again, you gotta be really careful and you can't take out your own brain. And I think that that's where a lot of people get themselves in trouble, where they just, you know, dump something into one of these tools and then, you know, take what it spits out and use that without any critical thinking skills. Uh, and I think that's really dangerous. But you know, how much better, whether it's an entrepreneur or somebody who's creative, if you have someone sitting next to you in a room and you're popcorning ideas back and forth off each other, you just, you get better ideas. And if there's not someone always there, especially if you're, you know, doing work at 2am M or something else, like use the tool.

Speaker A: I completely agree. In fact, even before the advent of AI, there was a concept in the software engineering world called pair programming. Um, and that actually has existed for probably 25 years or so. And the general idea there was, um, One keyboard, two people, and you're both looking at the code being developed, but you're collaborating on the idea, the concept of which way to go because there's always different ways to solve a problem. Right? Hey, this approach might be quicker or faster or easier. Oh, uh, and they would slide the keyboard over and your pair would do something and then they'd slide it back and you'd look and go. And that motif is the same reason why a team can produce more than an individual generally. Right. And one of the things that I think is going to come to fruition, it started last year, but I think it started in the niche. Like if you're an AI nerd, then you were probably playing last year. But the concept of agentic AI having individual agents and to uh, me it's the plural that makes the difference. So having a different agent that can do different tasking and maybe a uh, controller agent that herds the group of individual sub agents, if you will. And then that agent is kind of like your Jarvis from Ironman. Right. So rather than asking one agent to do everything, which it will fail at. Right. Asking one agent to coordinate a procedure or process, help, whatever, research, doesn't matter what it is. And then individual little sub agents go out and do little pieces which they're better at and they tend to not be as off the reservation with their responses when they're just doing one little piece. Right. Um, and then coordinating it all together, bringing it back to you and presenting. Here's what we, here's what we got. What I found, unfortunately your comment was pretty spot on a couple minutes ago. Folks that end up doing that often enough stop paying attention to the details pretty quickly and is where they get into trouble. I know you've probably seen various, uh, attorneys over the last two years or so submitting basically chatgpt written material with nonsense citations of cases that don't exist, et cetera, et cetera. Right. Um, and it's that level of detail, uh, that I think tends to get people in trouble. And that can happen whether you're just having it write an email or you're doing a research paper or whatever. I think I'm old enough to treat this like uh, going to the library and doing research and taking notes. Right. I want the AI to take the notes for me and then I want to review the notes and decide what I want to use and what I want to throw away and I'm going to rewrite the pieces that I want for me to sound the way I want it to sound and to make sense?

Speaker B: Oh, absolutely.

Speaker A: Have you used it for research and seen it give you stuff you didn't either have on your own or that you were like, oh, wow, I didn't know that existed.

Speaker B: Yes, I have. I have been testing a software, um, and it's not launched to my team yet, but I have been testing an AI software for lawyers that has a closed universe and it has done some interesting things. Um, you definitely still have to watch it. You cannot check your brain out. Um, but one of the things that I asked it to do was to craft a legal argument responding to two different, I guess, points. I'm trying to explain this to a non lawyer. So let's say there were two questions of law. So I asked it to craft arguments responding to these two different questions of law, and I gave it a small universe of information from which to do it. And it was interesting because one of its arguments was just like, no, you're so off. Like it's not even close. And this is where, again, it's an AI software for lawyers, like, that they're trying to develop to do this for lawyers. So the fact that it was so off that it was arguing federal law in a state case, like that was a big mistake. But it also did craft an argument that I ended up using later that I wouldn't have thought of. And again, it was kind of that similar thing that we talked about at the beginning of like bouncing ideas off of somebody else. So could it be really useful to like come up with creative ideas, a different way of seeing the problem? Yeah, but again, you gotta use your brain. You can't check your brain at the door. You gotta pay attention. Because if we had actually been using that software and not just me testing it, that would have been a big thing. But it didn't mean it had no value because then I ended up putting another argument in an actual thing that I created for a client that, yeah, that was a really cool perspective that I, I didn't see.

Speaker A: I've been thinking about. I don't think that what I'm about to say is finalized in my mind. So I'm probably not going to have it right. But what I've been thinking for a couple years now, if anyone's familiar with the Gartner hype cycle and the traditional curve of technology innovation that exists for decades, um, and if not, you can Google that phrase, Gartner hype cycle, you'll very quickly see the imagery. There's a peak of inflated expectations, a trough of disillusionment, and then eventually, over time, this plateau of productivity. Effectively, what that hype cycle means is that when something new comes out, it solves every problem on the planet, no matter what, at least as far as the marketing is concerned. And then, uh, as people start to engage with it, they really quickly realize, oh, it doesn't do half those things. It doesn't do 50, 70% of those things. But generally there are things it does and it does well. And so then after that trough of disillusionment where we get rid of all the marketing hype that wasn't real, we find the value and then we move forward and we use it in the best ways. What I think I'm speculating on is, uh, because the structure underneath large language models is essentially a probability engine. And every selection, the math, I'm, um, simplifying the heck out of this. But the math is basically an answer between 0 and 1, right. It's very difficult for you to be able to rely accurately on the results you're going to get from an AI. It's just not when accuracy matters. It's unlikely that the AI is going to give you the best answer all the time because of that probability problem. Occasionally you're going to get the absolute wrong answer. But I think the role of AI should probably be shifted so that it's not in line with making critical decisions or attempting to make, uh, I don't want to say life and death. You know, I wouldn't put it on a plane and have it fly it. Right. That's not wise.

Speaker B: Right.

Speaker A: Um, but I might put on a plane and have it suggest three different ways you could accomplish a goal. Like, I want to go around this storm. Which way do I go? What's the best? Well, this one will save fuel. Right. And then a decision maker has data. And so the idea that it's an assistant or that it, it's providing input and not making decisions, to me feels like the place it's going to end up slotting.

Speaker B: I think that makes a lot of sense. One of the ways that I look at it when I'm testing it, you know, from a legal perspective is, is it a good paralegal? Like, is it you as an attorney, you still have to read everything that your paralegal does. Your name is going on it. You're, you have to make sure that it's right. You have to train your paralegal, you know, all of those different things. So is it a good paralegal? It's still something you have to oversee and still something that you are the ultimate decision maker and your name's on the, you know, bottom line. But it doesn't mean that it can't do a fantastic job and get better and better and better over time.

Speaker A: Yeah, again, I don't think I'm totally. I think I've solidified my response to that. Uh, what I've been thinking for a while as I look around it does this really well. It shouldn't be over here at all. One of the things that I've seen attempted that tends to work okay is this concept of human in the loop. Hiddle is what the industry calls it. But the idea that the AI goes off and attempts to answer a question and brings the answer's response back to a human who reviews it and says yes, no, or even can say no, change this, say this instead, et cetera. And like the scenario where I see that be very valuable, uh, from a productivity perspective is we'll take like um, tech support or you know, a chatbot on a product sales website, right. Where a customer asks what's the status of shipping? Like the AI can go get that out of a database, but it asks something that's a little bit more non deterministic and the AI, uh, makes up an answer but sends that to an actual agent who's in, you know, slack or teams or whatever. But that agent is now looking at responses to 10 or 15 customers, but only occasionally not and not engaged with 15 conversations simultaneously. Right. No, that makes, it helps the scaling.

Speaker B: No, that, that makes a lot of sense. And that's. You talk about on a sales site and you know we're, we look at it certainly in our sales cycle. But one of the things that we have been really wrestling with and we have not implemented this at Z Family Law yet because I will not implement it until I'm satisfied with it. But customer response times, client response times. You know, when I first started Z Family Law and it's still in our policies, I told clients we will get back to you in 72 business hours. You will get an answer to your email in 72 business hours. And that was really good. Like that was really, really good. Like my competitors got back in like a week to two weeks at that time. Right. You know, you're, you're busy, you got, you have so many emails. Like this was a big deal. Um, and I made sure our clients knew it and that we committed to it. And I was on everybody. Well now I have clients complaining if we don't get a response in eight hours, one in five hours. And as much as I would love to provide that customer service at that level, if my attorney's in court all day, I can't meet it. Or, or uh, the attorney that is in there is in a mediation or back to back meetings, like it's, it's an expectation that they can't meet. So with everybody being used to chat boxes or chatgpt, you can get instant answers. How do we make these things go f without getting rid of the authenticity of talk? No human. And also certainly not giving legal advice. So one of the things we've been trying to figure out is can an AI assistant maybe say, okay, hi, I'm responding on behalf of attorney, blah, blah, blah, they are in court, whatever, look at their calendar, they can get back to you at this time. So at least the client gets an idea of when they're going to hear back. Or can I schedule you an appointment at this time? Being able to look at their calendar, um, and see that. And so at least have that response time a little bit faster, but not necessarily, you know, again, replacing the human or replacing the interaction, but maybe give something and then can it get to a point like what you're talking about, where if the question is just when is my next court date? When can AI, um, be good enough to look at our case management system and just respond to that and be a huge cost savings to clients because now they're not emailing their lawyer at 500 an hour, you know, going in and doing that manual labor of looking it up, blah, blah, blah, and responding.

Speaker A: Yeah, and I agree with what you're saying. And it's this, it's the structure that we were just describing together. The answer I'll give you about when can AI be good enough to do. Let me change it this way. In my view of the world, and I spent 30 years in software development, artificial intelligence is just a real complicated way to spell software. Because in the end all we're really talking about here is software. But the way I want to frame that is asking an AI to look in a database. Your case management system is, is a database, right?

Speaker B: Yes.

Speaker A: And pull out an answer where there's no inference required. It's literally go get X, bring back X, display X. That's already completely 100% reliable.

Speaker B: Yes, it is. And I should put little asterisks. But when is the software available to lawyers in such a way that we are confident that client confidentiality is maintained? And that has for me been the biggest stumbling block because I can see so many amazing things like one of the things that took us a while to implement and be implementing was these note takers. How much, how many of us all have an AI note taker like and that follows us around the Internet and is amazing and has made life changing things. But how long does it take me to find one and trust one enough that if I'm having a meeting with a client or a meeting with another attorney in my office where we talk about a client, that I am not breaking attorney client confidentiality inadvertently, where is the information stored? Where are those, you know, for lack of better word, zaps? Because I'm not a software engineer, but like things that go back and forth with the information and how do I make sure that one of the number one tenets of my profession, which is attorney confidentiality is maintained? And so I think in the legal field we are a bit behind and the software that we have quite frankly is a bit clunky in ways because they're making it for us. Um, and we're like I said, a little bit behind the times. But, but, and I imagine the medical field is the same way. Like so many great opportunities exist there, but we also need it with these safeguards that we are honoring the commitment that we have made to those we serve.

Speaker A: I would be curious. A lot of the AI note takers explicitly uh, want, in fact almost everyone that's free, not all of them, but almost everyone that's free, wants to get the information, to train on it, to use it at one level to make the AI note taker better. But I'll submit to you that once it can take the notes, what more does it need? Like what else are you training on and what are you looking for and you're mining that data to be able to sell it for advertising purposes. Uh, I think the world knows what's going on, but I don't know, maybe some people don't. You found a commercial entity that has the ability to guarantee your, the confidentiality or is it a tool that doesn't run as a cloud software? It runs basically locally.

Speaker B: So uh, there are certain levels of certifications and um, if uh, you ask me specifically what they are right this second I'm going to triple over my tongue. But I will, I can get them to you. But if you look up, you know, the American Bar association has done some really good uh, things that they have put out for guidelines for this. Um, so have certainly state and local bar associations. But what you want to be looking for isn't necessarily a closed universe, although it, that is locally based. Um, it can be cloud, but it does have to be a closed universe and have certain protections, um, available to it. As for the note takers, there are a few that have met m that um, that will not store a thing that will, you know, you can opt out of and make sure that you're being very careful that they're not learning all of your stuff. Um, that you are. That all saving of any of these things is saved in either our cloud drive. So we use Google, uh, um, so we'll have it directly integrated into there or in, in something that is kept aside and that is, that has been the way around it. But again it came later. It's certainly not the free version of the tools and you also have to be careful. And then, you know, separately from that, and this is more for once you decide to integrate it, there's also the training and policy and procedures that you really have to do with your team. I had one, you, uh, know, legal assistant who came up. Thankfully she was very, very smart and asked first, but she came up with this whole presentation about, hey, I know that we can use ChatGPT to go ahead and create these, you know, notices to clients and look at what I created, that can do it. And I looked at her, I was like, no, no, absolutely not. Like, let me tell you all the ways. And uh, you know, to her credit, she completely understood. But, but it really underscores the need for, you know, lawyers are afraid, hey, we're going to get left behind if we don't integrate AI. We're going to get behind. We will like go figure out your AI or you're going to get left behind 110%. But if you just unleash um, it on your team without training and policies and procedures, God help you with that bar complaint. There's, there's a local lawyer right now, um, who got himself in trouble. And you know, I, I don't know him personally, but I don't, you know, colleagues, professionally, LinkedIn, that kind of stuff. And he had a paralegal who he didn't know was doing it use AI to write something that ended up being an issue. And a paralegal he trusted and known for a long time. And so he had counted on them, you know, doing all the right stuff and signing off on it. Is he per. And now we gotta trouble. And he's doing a great job, you know, being honest and open and learning from me. But that's the fear that I have is number one, making sure I have written policies and procedures. Don't you dare use AI without my permission and, you know, treat, you know, only specific ones to put client confidentiality information and other things in. And number two, if you're going to use it, here's how you have to use it and here's how you can't check your brain at the door and all that. So, yeah, I know I'm making very specialized to lawyers that this is what I wrestle with all the time.

Speaker A: No, I don't, I don't have a problem with that because I think even if folks that are listening aren't lawyers, they can still respect the concept of privacy, uh, and security with their own content, their own data, or their customer's data. I still think it's relevant. I wouldn't worry about that. Have you done any exploration of being able to run private models, local models, personal models, instead of using one of the commercial AIs?

Speaker B: Yes and no. Uh, very. I am working with two right now, very, very small groups or AI companies that are trying to figure this out for lawyers. So one is the one I told you about that is doing a great job trying to figure out, like research and writing and that kind of stuff. Um, my firm is also using a AI company called Ajax, um, big fan of them, to do two things for us. One of them is to help us with our time entries. Um, so that my team is not spending as much administrative time writing in their time entries. And they're, uh, that's one of the things that it does exactly what you said. It presents them with options at the end of the day and then they, you know, go ahead and click it over and it helps with filing our emails because it is so important that, you know, all of our emails and client interactions are in our case management system so everybody can see it. So it is taken, you know, BS administrative time much down so that my team can be focused on delivering that exceptional client experience. And I'm super excited about that. Other tests that we're doing, um, I'm getting involved in this one company as well that is doing a similar version of, uh, legal research and putting together timelines and case and deposition like outlines. And what I love about this specific one is they not only keep it a tight body of information, but you actually give it the data. So you give it a bunch of PDFs of different documents and it actually is a link at the end directly to the source document or directly to the source case. And that I love because that is exactly what you and I are talking about. Here's some options, but you don't, you know, you keep your brain going and involved in it. So that's where I am right now. And then my next. I don't know if his next phase is in, it'll come after. But the next thing that I'm really looking for is kind of what I mentioned to you earlier, Jeff. How do I use it to enhance the client experience with faster response times and things like that without the frustration of I'm talking to AI? And if you don't mind, I'm going to go off on a little tangent here. You can always cut it out. But this is, this is one of my hard things right now because I have worked on our client experience so hard. And it is one of the things that is the most important thing, in fact, is the family law's, you know, phrase of the year is client experience. Client experience. And I have a client happiness coordinator on my team. Hi, Sandy. Um, Sandy is amazing. Her entire job is to check in with clients, making sure they're doing a good job, to check in emails, check in. Like, I want to make sure everybody's having a good experience and if they're not, I want to be able to catch an early course. Correct. And all of those things. People think Sandy's AI. I am not joking. I. She's been accused of being AI multiple times, but she is a real life human on my team, authentically reaching out with one of the biggest hearts that I know, because we know that our clients are usually going through a really hard time and we want to provide that extra support. So then again, how do you have that authenticity versus AI if people are going to be worried that it's AI anyway? And these are some of the things that we're balancing. How do we leverage it to, again, enhance the experience, but also how do we make sure that the personal experience that we are providing people don't think is AI generated. By the way, Sandy's real. She's amazing, is a human. You can meet her.

Speaker A: I think part of the reason that people have that impression is because the amount of hype on AI right now, again, that AI can do everything under the sun, it really is in everyone's face, all the time, social media, et cetera. So they just believe that that's the case. Let me ask you this question. So inside of there, and you've. You've said the word a couple times. So I want to see if we can explore this a little bit. Authenticity in this context, what does authenticity mean to you in this realm?

Speaker B: You know, I'm going to say it in three different contexts because I'm wrestling with it really hard in three different areas of my life personally. And then I talk about, like, my personal branding, representing myself on the Internet, and then in the client experience. So what does authenticity mean to me? If somebody is DMing me on Instagram, is it them or is it manychat? Am I talking to a human and building a relationship? Or am I building a relationship with AI? And I struggle with that because on the one hand, it can open a lot of doors. If, for example, you use my. I'm used. I'm not knocking manychat, by the way. I'm looking at using them. But, you know, if you're doing ManyChat, open the door. But then when do you take over so that it's you creating a relationship instead of your brand, creating a relationship with an AI, maybe even an AI tool on the other side. Um, same thing with I do and I will be very open. Like, I use chat GPT, clot, all those different things for my captions on social media because I'm not really good at knowing what are the key, you know, word searches of the day or different things like that. But I don't just let it post it, I edit it and tweak it and sound like me. And yes, I've been using em m dashes way longer than ChatGPT has been alive. Um, but people see the em dashes and they're like, oh, you use ChatGPT like, but I loved that. I'm a weirdo. I don't know. I've always written with them. But, you know, how does someone know it's me and not, uh, chatgpt? And how do I make sure that if I am using the AI, it's representing me authentically? So that's the personal branding side. In the personal side, um, one of the things that actually stood out to me recently is, do you read Reddit? I don't know, Jeff, if you like to hang out on Reddit as much as I do. Uh, I love Reddit. Um, I hang out in certainly the business owner side and a lot of different things with divorce and relationships because of what I do. And I saw this one post not too long ago where somebody got really upset because their husband used ChatGPT or some AI tool to write his wedding vows. And people were mad, big, mad people had big feelings, like, throw out the whole man. I can't believe he would do that. The most important thing and all I kept thinking about is how many of Us are really crappy writers. And as long as he or she, you know, I can't remember if it was a, uh. I believe it was the grim. But, like, put in what their feelings were and what the important things. So who cares if someone helped them rearrange it to something that makes sense, as long as they're the ones saying it and using it. And how many times before then did people have, like, their friend or their mom or their sister, like, help them edit their vowels so that it sounds right? And why is one so much, you know, worse than the other? So then there's that ethnicity in your personal life of can you use it and have anybody believe it's what you really mean or say? I've certainly used it in my personal life when writing condolence letters. I'm a big fan of handwriting letters. And sometimes I don't know the right words to say and I'll use it. You know, I'm certainly, again, not leaving my brain at the door, but bouncing ideas like, how do you say, I'm so sorry? You know, one recently I wrote, somebody's child passed away. Like, there are no right words.

Speaker A: No.

Speaker B: But an AI tool helped me get a little bit closer to write.

Speaker A: Sure. Yeah.

Speaker B: And so that in my personal life is what is authenticity or not if. And by the way, I'm not suggesting anybody go use AI to write their wedding vows, like, take it with a grain of salt. But it just. It was one of those things that really struck me with the authenticity of. We used other tools before. Why is this tool the one that takes away the meaning if it really does come from your heart? You just need help with coming up with your words because again, not everybody is good at that. And then I go again to the super professional Z family law one where that our having a big heart and caring about our clients is my brand. Like, that means the world to me. I had a client tell me yesterday that we are something special because we care about her as not just her case, but as what she's going to do after her case. We've talked a lot and I always talk a lot to our clients about, like, what does their future life look like and how do you keep that while also, again, leveraging the tools. And I know for me, some of the answers are simple. We're still handwriting a lot of notes. We're still like, sending things like that. I'm doing stuff to make sure that it's shown that it's not just AI generated, you know, fluff, that it's real but separately, again, I want to still leverage the tools because not just the fear of getting left behind, but also the reality of people need faster responses these days and we can do better using AI. So those are the different ways, again, authenticity. What does it mean to me? It means to me in my personal life that I really mean what it is, whether it's handwriting, that condolence note, or, you know, somebody saying their vowels, you know, and use the tool to help it. What does it mean in the personal branding? How do I leverage figuring out what all those trending hashtags are and things like that when I don't have time to do it? And there's a tool that can tell me in two seconds, but it doesn't mean that the caption isn't something that I meant and fed it to it and, you know, edited back and all of those things. And then what does it mean in my business? How do I show our clients that we really do care that Sandy is real, um, that this matters, and also leverage the tools to keep their costs down or to get them faster answers to things in, in a more efficient way. So that is me, the authenticity part.

Speaker A: Okay, I didn't expect that to be a simple answer. I welcomed the detail. No, it was fine. A couple comments that I can make there. I think there's, um, multiple sides to all these pieces, but maybe a little too much import. Put on a couple phrases said in a ceremony that is not indicative of everything that person's ever going to say, think or do is too much weight in one area, perhaps, but not everyone's going to see that the same way, and that's fine. I think that there are people out there who would labor over creating their vows on either side, uh, for a while. They would think about them for weeks or months, and they would make multiple copies and carry it with them and look at it regularly. And that's one type of person. Then there's another type of person that's not going to think about it much at all or is afraid of it and may very well be doing it the morning of. And the reality to me having, you know, I've been married for 30 years, so I can look back three decades to when this happened. For me, if you don't know who you're marrying, something else is going on.

Speaker B: Oh, okay.

Speaker A: If you don't already know that person and what, how they're going to treat it, I might suggest you wait a little. But again, I'm not going to solve every problem with that solution. But like you should know which way that person's going to go before. And you shouldn't be surprised dramatically if they came up with, hey, ChatGPT helped me because I suck at the wording part or whatever. That's just my opinion.

Speaker B: Obviously it doesn't mean anything, you know. And since when did ChatGPT become any different than a Google search?

Speaker A: Yeah, uh, one of the things I was going to say to you is if you've ever used Microsoft Word as an example and the grammar checker has suggested a change to you and you've said, no, I don't want it to sound like that. You've already used and rejected AI. Uh, grammar's not AI. But again, AI is just software. So you look at the grammar checker and I've been using the EM dash for a very long time and I'm very irate that I can't use it any longer because now it then looks like it's foolishness. Punctuation suddenly is bad. But as you're saying, if you're actively engaged and you get a suggestion and you're like, I don't like the way that sounds, that doesn't sound like me. I would never say that, or whatever, then I think you're doing the right thing. But if you just ask the question and then copy paste and you don't even look at the response, that's where things are going to start to get bad. Back to the authenticity for a second. I take things in strange directions. Sometimes you can't tell if you're talking to a bot or a real person. Is that going to inadvertently cause people to make bots to do their screening? So, like, your AI should talk to my AI, and if you can get through my AI's gauntlet, then I'll talk to you. And the opposite's happening.

Speaker B: I was thinking, what about that actually,

Speaker A: And I don't know that I think that that's a good answer. Like maybe for unsolicited, you know, spam, whatever. If my AI, uh, assistant can be like, you don't need to see any of those things. And that's, you know, we already sort of have filters in GML and other places. Right. But I think it's just going to become more intrusive.

Speaker B: I agree. And that's why we have been so thoughtful about using things like manychat and stuff where I don't want it to just be that inauthentic. But it doesn't mean it doesn't have an option. For example, having the ability to give a downloadable Resource. That's cool, that's useful. Everybody knows that it's not a human doing that. And that's not going to have the same AI like conversation. I also do like the screenouts of the inappropriate comments that some of us get at times. I don't need to see, I don't need to see that content that it's not helpful.

Speaker A: Yeah, it's doesn't add anything to the conversation value wise. No.

Speaker B: And it auto deleting that stuff too. Um, and that's not something we're doing, but I've, I've seen do. I do appreciate that. So that again, I don't like inappropriate content or inappropriate DMs, and to the extent that AI can help me screen that stuff out so that it doesn't, you know, affect people's day. Perfect.

Speaker A: You know, one of the downsides of the anonymity you can have online is people lose their social shield and they say things that they wouldn't normally say. Like it's, it's not uncommon for someone to think something untoward and be like, I shouldn't say that and just move on. It's not like they're evil people. It's just that sort of snarky comment shows up and you're like, oh no, that's not good. Or, you know, in the wrong company or in the wrong audience, fine. And maybe in private you might say that to somebody you know really well who knows that you're just goofing around and that's probably acceptable. But online, um, no one even has that filter online. They're like, whatever, no one knows who I am. I could just say anything and that's not good.

Speaker B: To a certain extent it's good and it's not good. I'm gonna push back on you just slightly on that one.

Speaker A: Push. Push. Yeah, that's it.

Speaker B: So again, the inappropriate, snarky and somewhat sometimes really cruel and sick comments. Yes. Uh, not helpful, all of those kind of things. But I do think anonymity, at least for research purposes, I have found to be really helpful in addressing people's, I guess, real needs and addressing people's kind of like real deep thoughts and beliefs. I'm going to use Reddit as an example because we already talked about it. Uh, and I like Reddit and I'm very open about who I am on Reddit. You go to my Reddit handle and you can find exactly who I am. It has worked really well. I found my EA on Reddit. I have found other attorneys. I found clients like, I Like it. But I'm talking about this one specifically because it is a platform that most people have anonymity. I am rare on there that I say who I am and I can do searches, especially with the new, uh, Reddit pro of what the conversations around divorce are. What are people really fearing? Like what are they genuinely fearing in their divorce? And then from a marketing perspective I can, you know, create marketing around that. From a lawyer perspective, I can create a solution around that. And I don't think these same people, I mean some of them certainly who have no filter, you know, can do whatever on um, whatever platform, but I don't think they're going to be going on Instagram saying those things. They're going on Facebook and saying those things because when it's not anonymous, sometimes people are ashamed or embarrassed about certain feelings or fears and people have a hard time being vulnerable. We all want to look perfect on the Internet. We all do. I mean I, I'm guilty of it as anybody and I try really hard not to be. I do think sometimes the anonymity allows us to have more real conversations.

Speaker A: I can agree with that. I can see what you're, what you're getting to and I, in this case, I kind of was headed down the path of if you wouldn't say it in front of someone hiding online to be able to say it, I don't think is. It speaks to more of what I think is your. And I don't mean this in case the things you're talking about, I mean more the hurtful.

Speaker B: Yeah, um, no, I agree with you.

Speaker A: Outrandous aggressive sort of commentary like, okay, you're just showing who you really are and then I'm not impressed. Right. The other thing that I think is important to sort of balance in there and I don't think this is everywhere, but sometimes you're talking to someone who's not even like 16 and you might infer that they're a 40 year old human being who you think would know not to say these things, but they might be a very young individual who just has no idea what that means yet, or thinks it's funny or doesn't understand the implications, et cetera. And the Internet sort of allows that, unfortunately without you knowing. And so how many times has that inadvertently been the case where it's not actually an evil human, it's just a clueless individual?

Speaker B: No, absolutely. Or unfortunately when you're getting into, and I certainly don't want to make it political generally, but when you're talking about bots out there that are saying kind of things to sway public opinion. And I say political. But even taking political aside, looking at one of the things that kind of opened my eyes. I don't know if you followed anything about, uh, Blake Lively's lawsuit with, um, her former co star. And regardless of what side you end up on, like literally regardless of what side, One of the things that was most interesting to me about the documents revealed is I didn't even know there were companies, companies out there that hired people and bots to go online and try to sway people's public opinion of others. Like I heard about it in the terms of politics. I've never heard about it in terms of PR before. And that was for me a uh, big eye opener. I had no idea or concept that that was a reality. And that again can be a very interesting tool and, or weapon. And in that case, I mean, it was a battering ram of a weapon. Again, regardless of who you side with.

Speaker A: Yeah, I, um, I think that problem is way bigger than, than most people realize. And in, and in a lot of cases there's a state sponsored level of, uh, influence happening every hour.

Speaker B: It is terrifying. I'd also like to point out there that it's kind of exciting when you think about how it could be used for good if the right people can.

Speaker A: Yeah, and actually I think it is being used for good in some cases. I just think one of the things that uh, the uh, algorithmic social media tools have uncovered through decades of experimentation on us. For example, there was a point in time when you would see in feeds a lot of cat and puppy pictures and those would elicit that same expression. You just, you just showed. Right. And you would share it and you would. Right. But what the psychology shows is that that reaction wears off eventually. The cute puppy picture doesn't get that. If you see it too often, it starts to. So you don't get the dopamine that you were getting when you saw that amazing cat that rolled over on its back and showed you the right. Okay. But what doesn't wear off and what never goes away is the anger side or the hate side that when you see that and it triggers you, it will trigger you every single time. Uh, and unfortunately the algorithms uncovered that the research supported it. And so now what ends up happening is especially at the propaganda level and I'm talking state sponsored now, that negative message always works.

Speaker B: Well, that's just terrifying.

Speaker A: It's sad. It's a function of sort of the way our brains are wired to a certain extent but there was an experiment done by outside, um, of the U.S. but it was done in the U.S. the people sponsoring the experiment were not in the U.S. and were not generally friendly, where they created two opposing, um, support groups for fictitious causes, um, one side, you know, going one direction, one side going another. And, uh, eventually staged rallies across the street from each other at the same exact time for both of those opposing views in order to see whether they could transition from a digital marketing, for lack of a better phrase, into real world action. And sadly, they were successful. And it's not fun. No, but the level of manipulation that's possible is scary. And, uh, if an AI starts to do that and starts to learn from it and starts to implement that, we're looking very, sadly, we're looking at an Orwellian sort of model where, uh, I don't think things are good.

Speaker B: And now you know why. At the start of this podcast, I called AI a weapon.

Speaker A: Yeah, it is.

Speaker B: I truly do it that way.

Speaker A: It could probably be the worst weapon that's ever been invented. Could be the best weapon at the same time.

Speaker B: Absolutely.

Speaker A: But it's gotta be used the right way. And I don't know, I don't know whether the. I would like to think that the good will outweigh the bad. I just see a lot of people making a lot of money with the bad right now, sadly.

Speaker B: Well, the good people and the ones who want to use it for good in my very limited world experience, have so many competing interests in other things and competing interests for their dollars. But thoughtfully investing it in AI and in getting the messaging out, I think is going to be really important, even if it's hard.

Speaker A: Yeah. What I'm hoping is that people eventually tire of some of the, what I see as repetitious silliness and, uh, and finally go, all right, I'm done with that.

Speaker B: But, you know, you know, I think it will. It's one of the things I'm doing really in again, my personal branding and marketing push, which is M. On a very, very much larger scale. I see AI people are going to be able to produce videos. For example, I, I went to an Aspire conference and I heard a bunch of different people speak and one of the speakers talked about how, you know, once he has an AI, you know, trained thing for doing certain videos, and I'm, um, I'm not talking about like a clone thing, but like an AI thing to do certain videos, like, he'd be able to produce, you know, 300 to 3,000 in a day, whereas nobody had the ability to record that. And I think we're going to see a ton of that. And the market's gonna be saturated because then I think it's gonna swing the other way because people are gonna realize what's again, going back to that word authentic and what's not. And I think as long as you're sticking to that authenticity, you build trust. Because one of the things that I think AI is gonna do is be a big trust breaker.

Speaker A: I think there's actually a capability. I'll get technical for a second just to explain what I'm getting to. There's always been a way for us to have secure email, but it's a pain in the ass for the average person to implement. It requires a lot of finagling that most people won't jump through hoops, et cetera. But the AI would be able to do that without even thinking, and the person themselves wouldn't have to be involved in that level of complexity. So there is a methodology that exists already around certificates and other things, signing, et cetera, where you could authentically produce content with your AI that you are signing that it would be very difficult for someone to copy. Not 100% impossible, but difficult. And you could use that. Like Sandy could use it to stamp her messages, like, look, this came from a person. This isn't an AI, et cetera. Like, there's a way to authenticate, et cetera, so that people receiving it can have some faith. It's the same thing essentially where, uh, you've been trained for 25 years more. When you see the little lock icon in your browser, it's okay to give them your credit card. Now, that isn't actually a true statement anymore, and it really hasn't ever been. But we've all been trained for that and now we now accept it. So when it isn't there, we at least go, wait, what's that? Good? From a security perspective, that's brilliant. I'd like to get to the same level of interaction where I'm looking at this snippet and it's unsigned, and I go, nah, I can't trust this. Or I'm looking at this, it's signed by this influencer, it did come from him or her, etc. And so I know it's at least realistic. And if they say something weird, okay, they said something weird, not an AI made them say something they didn't ever say, which is where we could go.

Speaker B: Yeah, no, that's a good point.

Speaker A: So I'm hopeful that we may get through. We're in a, um. I called it a trough of disillusionment before from the Gartner cycle. But literally, there's a lot of people throwing a lot of spaghetti on the wall right now with AI just to see what they can do and what makes sense. And some people are getting caught up in that, and that's not realistically what'll happen. It'll settle out in the next three years or so. That hype will start to come down. People will be like, okay, I've had it up to here with how AI is going to solve every problem. How about we look at what it can do, et cetera. But this concept of being able to certify content and have some basis in reality behind it could be a way to infer, um, authenticity, even for just human created content, where you're just stamping it as an AI didn't write this.

Speaker B: I had not heard that before, but I want to look into it and learn more. That's a new one for me, Jeff.

Speaker A: Okay. Yeah, the technology exists, and, uh, it is possible to get there, and it would require this technology on both sides. So eventually what happened has to happen is everybody has to have a bot. But when I say that, what I generally mean is an app in essence, Right? It could be. Let's say it's Apple's Assistant. I don't want to say the name. I don't want to trigger the name, but. Or it's Google's assistant, and it's on the phone and it's that app that does the encryption, decryption, certificating whatever, and does the signing. So it then says to you, hey, by the way, I got a text from Jeff, and it came from Jeff. It didn't come from a number masquerading as Jeff again.

Speaker B: That's really cool.

Speaker A: I had no idea that's existed for a while. So it can be implemented that way. It's not currently being implemented that way, but there is an opportunity to make that secure and reliable until we get to Quantum, and then we'd have to change the methodology, but that's okay.

Speaker B: Well, after this podcast, you might have to teach me more. I might have to buy virtual coffee, because this is interesting to me.

Speaker A: Okay, well, we can certainly talk about it. I can give you a whole bunch of details. Don't worry about it. All right. I think we covered quite a bunch of stuff. Is there anything you want to ask or anything you want to bring up?

Speaker B: You know, the big thing was the authenticity. And if you had been listening to this episode, I certainly would like to know what you think. Um, whether it is, you know, commenting wherever this is posted, or if you want to reach out to me, you can find me on most social, uh, media platforms. Christie underscore, esq. So Christie, esquire. Or send an email to hellofamily law.com But I want to know from your listeners how they feel about the authenticity and how they feel about, you know, AI reaching out to them. And is speed worth it? More than, you know, getting to talk to a human? Is there a mix in there and what would make you as a customer or client and not just of Z Family Law, but of any business out there, like, what makes you feel valued and how can we all do better? So that would be my ask for the audience. I want to. To know more.

Speaker A: Okay. I like it. Yeah. Awesome. All right, thanks very much. I appreciate it was a great conversation. Good talking to you.

Speaker B: Thank you, Jeff. I appreciate it.