E300 - Cat Casey, Founder, The Technocat, Chief Legal AI Futurist, Masters AI
[00:00] Debbie Reynolds: The personal views expressed by our podcast guests are their own and are not legal advice or official statements by their organizations.
[00:11] Hello, my name is Debbie Reynolds. They call me The Data Diva. This is the Data Diva Talks Privacy podcast where we discuss data privacy issues with industry leaders around the world with information that businesses need to know.
[00:24] Now I have a very, very special guest on the show, Cat Casey. She is the founder of the Techno Cat and chief legal AI futurist at Masters AI. Welcome.
[00:36] Cat Casey: Oh, thank you so much. And you forgot the most important thing. Like a 20 year friend.
[00:40] Debbie Reynolds: Yeah. Oh my gosh.
[00:43] It's so funny because when I do these intros, I'm like, oh my God. People just have to know we've like known each other forever. That's so funny. Well, yeah. So you are.
[00:51] What can I say about you, kid?
[00:54] You are someone who sets the pace or sets the tone for the way people are thinking about technology and loss. To me, your background story is fascinating. I love the way that you put yourself out there.
[01:06] You basically like planted the flag on the moon as far as I'm concerned around legal and technology. But give people a bit of your background.
[01:15] Cat Casey: I like to joke that I'm kind of like Sisyphus in sequins. And you'll notice I'm always a little sparkly. And I have been shouting in the void about legal, tech and AI for about 20, 20, 22 years.
[01:25] I came into tech the way most people do by studying existential philosophy. So dead German and angry. I was equipped to either be a barista or an AI evangelist. I'm glad I chose AI evangelist.
[01:37] I came around kind of across a wide array on the lsp,
[01:42] the litigation support side, and in the E discovery space. So I was on the providing review attorney side to all the way up to running a global tech program at Gibson Dunn and building out tech programs with PwC and KPMG,
[01:55] to being in the C suite of software companies. But all along I had this real passion for trying to make the dauntingness, if that's a word, the uncomfortableness people were feeling around technology a little bit less unpleasant to talk with them in kind of a plain English way about Eli 5.
[02:10] Explain it to me like I'm 5. What is this technology? What does it mean to me? How can I get literate? How can I level up? And that kind of led to my next chapter where I'm the content curator,
[02:21] face and voice of a 10 city masters AI conference and have my own company. So I've kind of been around the
[02:26] Debbie Reynolds: block as you're talking. I'm just thinking about this now. This is hilarious. So you and I, I think on LinkedIn, we have probably the most connections, like, similar connections. It's so funny.
[02:39] But we work in different kind of realms and domains. So it's funny that we know a lot of the same people or a lot of the same people know us.
[02:46] So the people who know me from Legal,
[02:50] Legal work, you and all these other thousands of people included, and then there's just people out in the world who know me about my other technical, wacky things that I do.
[03:00] But one of the things that has always intrigued me about legal is that,
[03:09] in my view, legal has always been dragged kicking and screaming into the future as it relates to technology.
[03:16] And so that's a hard job,
[03:19] and that is a hard thing to do. And I feel you do it with so much finesse. You know, at some point I was like, I'm done, I'm done.
[03:28] But you march on. But tell me, tell me about that.
[03:31] Cat Casey: It's funny, like, I didn't set out wanting to. I mean, I studied philosophy. I didn't start out wanting to be the voice of AI and legal tech, and I just found I was a good translator, I think, because my grandfather made virtual memory.
[03:42] So I just had laptops and computers in my life probably for 10 years more than most of my peers. Like, it just didn't seem scary to me. And so back in 2005 and 6, when Judge Peck was talking about tar and people were figuring out like, oh, is there an electronic to this discovery thing?
[03:57] I just kind of naturally fell into it. And it wasn't scary and it wasn't daunting. And I was like, well, if it's not scaring me, how can I make it feel the same for other people?
[04:05] And so as much as the legal community didn't really embrace AI and tech, in fact, sometimes I would go to happy hours and like, Cat, you're banned from two letters.
[04:14] The two A and I,
[04:15] I still got, I think,
[04:17] validation for every individual that I helped. Like, every time I helped make it less daunting for someone or every person who was like, hey, that was helpful. Like, for me, that filled my cup back up.
[04:26] Even if legal was at times rather Luddite in their views on technology,
[04:31] the ability to help people who maybe decided when they were 15, 16, 17, or an undergrad to go more into liberal arts and were uncomfortable with STEM and hard tech to help it make more manageable for them, and it was just.
[04:44] It's been rewarding So I think maybe that's why I've leaned into it. But it was a bit of a thankless task. But then I would get thanked.
[04:51] Debbie Reynolds: I had to find a new word for you. I don't think influencer is the right term exactly. So I feel like you really put yourself out there in a way that some people want to hide away.
[05:02] And then also you find ways to bring different people into the conversation about tech. So sometimes people are afraid to ask questions or they're afraid to engage, and you always kind of throw things out there.
[05:16] You're always trying to bring people in.
[05:18] And I think that's part of your success as well, even. It's so funny because you and I, you. You're the AI futurist at the Masters Conference, so we should talk a little bit about that as well.
[05:28] But I had been involved with the Master's conference for over 20 years, and Robert Childress, who since passed away,
[05:35] was one of my dear friends. And he was actually the person.
[05:39] They helped me get into privacy. I told him in 2014 I wanted to go into privacy.
[05:43] And he's like, okay, well, if you want to go into privacy,
[05:47] we'll. I said, that's all I'm going to talk about. So I don't want to talk about anything else but that. So he agreed. And he said, any city you want to go to.
[05:55] He would change the schedule for the sessions just so I would fly in like I fly into Denver for the day. And he'd say, okay,
[06:02] Debbie's session is that in the afternoon. He really, really helped me to the point where I had been doing that for a couple of years. And eventually PBS called me.
[06:12] We said, you were at the Master's Conference. You were talking about privacy. And that was kind of the launch of me thinking, well, maybe I should do this for, like, a job,
[06:22] since people are asking me about it. But tell me a little bit about your work with the Master's Conference for people who don't know. So we have a global audience,
[06:31] over 160 countries, over a million downloads. So people, I think, are really fascinated about some of the things that we do in these other verticals, like Legal. But tell me a little bit about your work at the Master's Conference.
[06:44] Cat Casey: Yeah, yeah, absolutely. I mean, I think I had a similar experience to you with Robert Childress. He had a knack for seeing something in people who maybe they didn't even see it, because he might have been literally, I think he shoved me on stage.
[06:56] He was five foot nothing, and he literally shoved Me on stage about 15 years ago. And from that I started speaking. And so about just under a year ago, and this was years after Albert had passed on and the show had been kept alive by a loyal group of friends through Covid.
[07:12] And then there was a changing of the guard about a year ago. And right literally the day that I started my company, I got a call from the new chairman, Mike Delawitz, and he said, cap, what would you do if you had 10 cities to build the conference of your dreams?
[07:24] I'm like, well, tell me more for one,
[07:26] then for two, which one is it? And so when he said the Masters, it felt like a full circle moment. And so what we did is, I mean, Deb, you kind of alluded to it, the Masters was pretty squarely like 90% in eDiscovery conference and 10% privacy because of your brilliance.
[07:41] And what I did was I helped kind of shift it from being just this very narrow niche of legal to across the whole horizontal of legal and focused on AI and not necessarily a specific tool, but all the skills legal humans need in the age of AI to level up their careers,
[07:56] to understand the tools, frankly, to do the job of legal. And then I switched it from very panel focused to interactive sessions. We had you on to do a wonderful prompt workshop.
[08:07] We do TED talks. And so what I've been doing the last seven months is redesigning the conference, which is going to hit 10 cities, to be kind of adapted to the shorter attention spans caused by TikTok and the adult learning best practices, so that it's a room you actually want to be in and you're interactive and you're learning about,
[08:25] frankly, I think what's a pretty foundational skill. Like, even if you don't necessarily want to use AI in your personal practice, as a lawyer, you need to understand the risks, the opportunities, the type of tools out there, or frankly, I think it's kind of malpractice.
[08:39] And so what we did is kind of build a day of content and you were a rock star to be at our Chicago one. And we started AI101. And by the end of the day, at the last one we did in la, we.
[08:50] I love it. It makes me giggle. We vibe coded like Legends of Zelda themed contract analytic tool with three different LLMs like crazy. But we, we help bring people along for the journey and then we have a whole digital ecosystem as well.
[09:03] And we're just really building a resource for the 1.29 million legal humans who realized AI is not going away. Like you And I are part of the 10,000 that knew about it for the last 20 years.
[09:14] But there's a large cohort of legal that is like, oh, now I need some. I need some guidance. Help me level up. And so it's been pretty exciting to be the kind of the brain behind finding the voices like yours and curating the content to bring something, hopefully that,
[09:27] that's really valuable for legal. Our next stop is in a few weeks at the New York Law School on the end of July, and then we're going to be in D.C.
[09:34] in September, Miami, Philadelphia,
[09:36] kind of sprinting through the end of the year. So it should be really exciting.
[09:39] Debbie Reynolds: Oh, that is really exciting. And I actually want to give a shout out to Michael Dalowitz. He's someone I've known forever. I think I've even known him longer than I've known you, which is hard to believe, actually.
[09:50] And so when he told me that you guys were revamping the conference and he was like,
[09:56] we need influencer people, people who are kind of known and stuff, he's like, I know Debbie. I'm gonna call her up and ask her if she wants to do this.
[10:03] So I was happy to do it. Spoke at State, Arthur Shaw. It was a great. I love the mix. I go to a lot of conferences, actually, I go to a lot of conferences that aren't privacy conferences because I'm speaking at different things in different countries.
[10:18] The conference wasn't put together in a way where it. It crowded people out that weren't in legal. And also I felt like you talked about,
[10:27] this is what's happening in legal, but here's the forces outside of the legal sphere that is having an impact on legal, and then also having people from different areas of technology and law to talk,
[10:42] literally talk and collaborate together. I thought it was amazing.
[10:46] Cat Casey: Oh, making me blush.
[10:48] Well, I, I mean, like you, I.
[10:50] I think my busiest year, I was on the road 230 days and did 70 conferences plus another 30 webinars. And so I'd seen the good, the bad, or just the too much of the same thing.
[11:01] And so it was kind of fun to be able to think about, well, what would. What do I wish I saw? What do I. What would I want if I was sitting in the room listening or on stage?
[11:09] Because, frankly, sometimes I would do a panel and I'd be bored of myself talking, so I could only imagine the audience.
[11:15] And so it's been exciting to build that and to hear from folks I respect like you, that, hey,
[11:20] it's a Weird and different approach, but it works.
[11:24] Debbie Reynolds: Yeah. I feel like when people,
[11:27] instead of them listening to people talk, them actually being engaged and not being afraid to ask questions or getting to know person who's at their table or person sitting next to them, I think that's.
[11:39] You're doing a phenomenal job.
[11:41] They come and blush,
[11:42] oh my gosh. And then so we did a session recently with the Practice Law Institute on AI agents and governance and it was very well attended. I think we had over 700 people.
[11:55] Cat Casey: I think almost. Yeah, 750, almost 800, which is pretty good. Yeah.
[11:58] Debbie Reynolds: And I thought it was pretty good. It was a fun session. I always.
[12:02] You and I have collaborated on many different things over the years. Sometimes I just run into you randomly. I'll just be at a conference and you're just there.
[12:10] Cat Casey: We, I think we've done it for a few, like ILTA evolves or something for sure. And it's like, oh, what are you doing next month? All right, let's do it together.
[12:17] Debbie Reynolds: Yeah. I'm like, hey Cat, what's up?
[12:21] Tell me your thoughts about privacy. So privacy kind of snuck in here, I feel.
[12:26] And for me it's been something that I worked on for a long time. It wasn't even its own industry or its own job. It was just a bullet point on a checklist that you needed to, to cover.
[12:39] But that really fascinated me. But tell me how has seeped in to legal and how it's connecting, especially as we're talking about AI.
[12:49] Cat Casey: I mean, even long before we were still in the era of safe harbor, way before gdpr, like in the discovery space, you kind of always had to think about like where,
[12:57] how and where you were protecting individual rights around data. I did a lot of work internationally and so there was just a whole host of consideration just because I happened to be in a very kind of a narrow niche of legal.
[13:09] But I think in particular, as the value of data has increased and as legal considerations around data have increased, especially with the proliferation of not just generative AI, but all AI, it's gone from an interesting asterisk that some people in the know knew about to a frontline consideration that,
[13:27] that I think is a top line discussion point at the business level with, with the senior leadership of legal in a way I probably wouldn't have anticipated,
[13:35] certainly not 10 years ago. And it seems to have like, it kind of felt like the privacy issues, they weren't resolved, but they were kind of understood. Like darsars were mapped out.
[13:45] People kind of understood the expectation with GDPR but then when you had this proliferation of kind of agents that were playing with different tools, different AI models, it kind of reinvigorated the discussion and opened up a whole bunch of can of worms that have been a really big top discussion point across legal and I haven't seen before.
[14:05] It's been really, really interesting.
[14:07] Debbie Reynolds: Yeah, I almost feel like the energy I'm getting now with all the AI stuff that's happening is the same energy that I got when people started using the commercial Internet.
[14:19] I literally meant, I remember going into firms before people had email or literally memos with the inter office mail envelopes and stuff like that. And so just to see how much technology changed.
[14:33] And then some people were like, I'm gonna again be pulled kicking and screaming into the future. And they weren't eventually. And so people got over it, so they started using the Internet, using technology more.
[14:47] And so this is kind of the next evolution of that. But I want to talk a little bit more in depth about AI and AI agents. And so the session that we did was about governance and what I think is especially agents.
[15:04] So when companies are using tools that can act on their own, that are goal oriented. Right. Not just task oriented,
[15:15] they're doing things that maybe people are doing things in ways that people may not anticipate or expect.
[15:21] I think people think almost like AI agents are like a human and would do things that they would do, but the technology may see things that you don't see or maybe do things, do things in a different way to potentially create a harmful situation.
[15:33] But I think part of that is that you can't really think about working with technology like AI. Like at Santa's workshop, like everyone has this narrow view of their little segment and then they don't know what impact all the things they're doing or all the things AI is doing on their organization or the things that they're doing with data.
[15:55] But what do you think?
[15:56] Cat Casey: You know what's really interesting to me is like the,
[16:00] there was initial pushback around just straight up AI post GPT. So in 2014 and then it kind of, it wasn't resolved, but people calmed down. But then like with the agentic layer coming up, suddenly there's a bunch more questions about the unauthorized practice of law.
[16:14] Follow on risk for third party tools that the agent might be accessing on your behalf and also just reimagining like not necessarily anthropomorphizing the AI, but thinking of it maybe less of as a tool and a more as a,
[16:29] not a human component But a component that you need to supervise that is acting and doing and making within your case team. And so it's been really interesting the last six 12, six months, probably the discussion points have, have really shifted because as much as when we went from AI to gen AI,
[16:46] we had to think about risk differently because it wasn't finding a pattern in data, it was the generation of something new while the agent is doing something new on your behalf.
[16:54] And it may create more risk. And if you're the legal human that is attesting or signing off on that risk would, then we're not going to hold the robot in contempt, but we might hold the human in contempt.
[17:05] And so it's created, I think, some extra layers of uncertainty, some extra layers of needing governance and guardrails and proactive constraints and a reevaluation of what does it mean to keep a human in the loop if we have tools that are by design operating without human input.
[17:24] So it's been a really kind of not challenging, but thought provoking discussion point. And then, I mean, that doesn't even touch on the privacy implications of if you have this,
[17:35] this agent acting in a semi autonomous, semi supervised way and accessing data and, or moving, interacting and adapting pieces of data, or connecting it to third party sources, what does that mean from a privacy standpoint?
[17:48] How do you build a safe sandbox? Like it's a sticky wicket, for lack of a better word. It's a complicated thing that's only getting more complicated because the next layer is super agents, which is an agent interacting with an agent.
[17:59] And how do we put a human in the loop and all that.
[18:02] Debbie Reynolds: Totally. So. Well, I want to throw out some scenarios. These are like in the news types of things.
[18:07] And I want you to tell me what went wrong. Okay,
[18:10] so there's an example of a woman who was an executive at a big company.
[18:15] She had AI agent doing some tasks for her in her email box.
[18:20] And at some point, I guess the agent decided that it didn't have enough memory or room to do what it needed to do and it decided to start deleting her mailbox.
[18:31] So she was trying to furiously stop the agent from doing it, shut up her computer and different things like that. So what went wrong?
[18:38] Cat Casey: You should never give the agent deletion rights,
[18:41] like period.
[18:42] And we've seen problems like that where it's been code or any number of things.
[18:46] If you treat the agent like,
[18:48] not as if it's a human, but like a human, you want to give it access controls, you want to give it Limitations on what it can and can't do.
[18:55] I think we talked about it. Like, you wouldn't give the intern the key card to access every file and every office on day one. Why are you doing that to an unsupervised or with limited guardrail agents?
[19:06] So don't let it delete step one, ever.
[19:10] Debbie Reynolds: Yeah, that's a good one. That's a good one. And then,
[19:13] even though this example is not about AI agents, this is about people having chatbots on their website. So there was an airline in Canada that put a chatbot on their website.
[19:24] It was. Someone had interacted with the chatbot about a bereavement flight where the chatbot said, okay, we'll refund you if you buy this flight. The person tried to get a refund.
[19:35] The company said, no,
[19:36] even though they chatted with this agent,
[19:40] chatted with the chatbot. And then it went to court, and the court said that the bot created a contract.
[19:48] Cat Casey: Verbal.
[19:49] Debbie Reynolds: Yeah, a contract. And that the company had to honor it. So what went wrong?
[19:54] Cat Casey: I mean, I think that's a pretty standard human in the loop issue. Like,
[19:58] if it was me architecting that chatbot for certain things, specifically refunds or anything that could be construed as a verbal contract,
[20:07] that's when you say, okay, I've identified your problem, and I'm going to escalate it to a human in X, Y, or Z, Q. And to be honest, the person on the other side of the chat bot may not like that.
[20:18] They might feel like, oh, well, I've been discussing this and breaking it down for 15 minutes. Why do I have to reinstate it? But that's a limitation of liability that you want to create by having that human re injected in the loop.
[20:30] I think.
[20:31] I mean, would that be kind of your gut on it too?
[20:33] Debbie Reynolds: Oh, totally, totally. Right, Right. So. Right. Because I think people just as a novelty, thought, oh, it's so cool. It's gonna save me so much time and so much money if I put a chatbot on there.
[20:43] But if you don't know what it's doing, you don't have constraints or you don't have anyone, like, watching over what's happening. You know, you can definitely get into those situations because a lot of times I feel like people are so enamored with the things that these things can do,
[21:00] and sometimes they just let them loose in ways they shouldn't. And it's hard. Right. So if you have a tool that's working for you, oh, well, a hundred. It's done.
[21:09] This thing A hundred times you don't know that it's going to go left on 101, but you really need someone there to be able to check that and make sure they're doing what you expect them to do.
[21:21] Cat Casey: Yeah, well, and the bigger the risk of it going left, the more,
[21:24] more times you want to inject a human. Like I think we were talking about like some of the recent regulations where autonomous algorithmic decision making, like in I think Colorado,
[21:34] like depending on what the decision's about. So life, liberty, pursuit of happiness, getting financial support, medical care, anything like that. The bar for having a human in the loop is sooner, more frequent and higher.
[21:45] So I mean, I think that's another thing you want to think about. Like for the chat bot, it's okay if you see the word refund automatically escalate it to a human.
[21:52] But if it's a chat bottom medical advice, the human probably is even more so. Like there's other considerations that kind of determine, even to a greater degree, how soon and how often you want to have a human loop.
[22:04] Debbie Reynolds: Yeah, that's true. And so we're seeing regulations like the EU AI act,
[22:09] even in the US at a state level. Things with like Illinois and New York around, hiring,
[22:16] Colorado around automated decision making. The GDPR had automated decision decision making already in their regulation before the EU AI Act. But I think a lot of them, they're pointing towards,
[22:28] as you say,
[22:30] making sure that a human is in the loop, especially when decision could possibly harm someone. And then there has to be a level of explainability as well. So I feel like people, and I want your thoughts.
[22:44] I feel like some people get tangled up in explainability.
[22:47] Like they have to actually explain what is happening inside the black box. And to me,
[22:53] explainability should be about what goes in and what came out and what you did with it.
[22:59] What do you think?
[23:00] Cat Casey: I completely agree because I mean for some of the black boxes, the engineers making them don't know how they work. I mean, they know generally, but they can't explicitly say when I put this input in, why was this exact string of letters and numbers and answers created?
[23:15] I mean, I think explicitly explainability of the process and an understanding of what the decision impacts. Like that's I think been the smartest shift I've seen because the initial regulations were about like, what type of AI is it, how is it architected, all those safeguards which are important.
[23:30] But I think the smarter move in assessing how,
[23:34] how much scrutiny a tool is subject to or how much regulation is, how does it impact the humans at the end of the day. And I think that's been a really smart move.
[23:43] Debbie Reynolds: Yeah, I'm very,
[23:45] very in favor of human centric types of architectures and also laws and regulation because that's where the rubber meets the road in my view.
[23:55] Like if you want to use a chatbot to respond to your email, like no regulator is going to knock down your door because you do that or you use it to create a pretty slide deck, no one cares, do that as much as you want.
[24:08] But if you're saying, oh well, we're going to use this, we're going to fire everyone from HR and they're going to do hiring,
[24:16] you're like, oh, wait a minute, wait a minute, this could be a problem. And then also I think one of the things that people don't understand because I've heard people say this, well, well, there really aren't any AI regulations, really.
[24:27] And I was like, well, but there are regulations around data,
[24:30] so it's still that data use,
[24:32] regardless of the technology that you use,
[24:35] are there? And then also consumers,
[24:38] consumers, they have rights, they get angry. You don't want this to be a situation where you lose a customer because you're doing something with AI that maybe the customer would not.
[24:49] First of all, they either they're not notified or something that they wouldn't like. What do you think?
[24:55] Cat Casey: Well, and I think, I mean, I operate squarely in the legal space. So there's the consumer protections, there's the, the data obligations, and then there's the legal ethical obligations. Like for a lawyer, there may not be specific AI regulations yet, but there's a duty for technical competence.
[25:11] Model Rule 11 has a further note that that includes the legal technology tools. There's a duty to supervise. Like you can't just let an intern go willy nilly, nor can you let an AI go willy nilly.
[25:21] And every time a lawyer signs something, it's their name on the line. That's why they've gotten so much trouble for the hallucinated cases. So the, there's may not be specific regulation for Gemini, whatever, Sonnet 4.6 and ChatGPT 5.6, but there's regulation for data, for the humans it affects and for the legal humans signing off on it.
[25:42] And to me,
[25:43] that's more than enough to get started and know exactly what you're supposed to do, at least in a general directional way.
[25:48] Debbie Reynolds: That's so true. That's so true. So what's happening in the world right now that you're seeing that's concerning you most as it relates to data or privacy.
[25:57] Cat Casey: See,
[25:57] so I don't know if it's concerning, but it's interesting. So the thing that caught my eye the Most was actually OpenAI hiring a former legal technology CEO to build the equivalent of a codex within OpenAI for legal.
[26:12] So that's different from like the direct connectors that Anthropic did. That's different from the middleware. That's, that's a foundational model which has more data than everyone even knows what to do with.
[26:24] That's specifically building a direct tool which is going to host potentially privileged confidential information that is designed to learn from said information. It raises like privacy concerns for me. It raises confidentiality concerns like the Hefner case.
[26:41] It's also,
[26:43] to my mind, it sort of shows that Legal might actually for once not be the Luddites being dragged up the hill, but we're the tip of the spear somehow, which is insanity.
[26:53] Like,
[26:54] it's fascinating to me. Like I think Anthropic was saying for their cowork, the number one users are lawyers. So in my space, the legal space industry,
[27:03] known specifically for not being tech confident,
[27:07] is now the tip of the spear for the biggest technological transformation.
[27:12] I mean, I was listening to the CISO at Uber and he said it was the most transformative tech invention since Fire. So that's,
[27:20] it's not worrying me,
[27:21] but it's a holy moly.
[27:23] That does kind of change the context that we are operating in to my mind.
[27:28] Debbie Reynolds: Yeah, absolutely. So, well, there's another case I want to talk to you about. And this has to do with self represented people using AI.
[27:37] Cat Casey: Give it to me. Bring it.
[27:39] Debbie Reynolds: Yeah, so there was a case, I can't remember the name of it right off the top of my head, but there was a case where a company was,
[27:47] a self represented pro se person was suing this company and the company was saying that the AI helped them too much in some way. And then they thought that the AI was doing unauthorized practice of law.
[28:00] And then I think they sue OpenAI, I believe as a result of that. So that case is kind of going on. I have thoughts about that. But what do you think?
[28:08] So I don't know. I'm like, okay, lawyers are using AI.
[28:12] Why can't pro se people use AI? But what do you think?
[28:16] Cat Casey: So there was a big push around this three years ago, maybe with a company called Do Not Pay, which was like an algorithm to use,
[28:25] actually have the laptop open and use it in court. They were saying they'd pay a million bucks if someone used them for a Supreme Court case that got shut down because literally it was a pro se person treating the AI like a lawyer, in that case, unauthorized practice of law all day long.
[28:39] But if it's just a pro se, so a self represented human that's using AI for legal research, AI to better ask questions outside of the courtroom, I think that's a much harder case.
[28:52] It's a harder position to have.
[28:54] Now if the pro se person is on their phone talking with Claude or Chatbot or whatever in real time with the judge there, or using an avatar. I think we talked about an avatar case in one of our talks.
[29:07] That might be a different scenario because it is truly using the AI to outmaneuver the lawyer who is opposing them, who maybe is not allowed to access the AI. But I don't know.
[29:17] Like it gets murky depending on the task, depending on the immediacy in the courtroom, and depending on whether the opposing side has the same access at the same moment in time.
[29:29] So I, I don't know, like it's, if it's just research and just building your case the same way that you know the prosecutor is, then go for it all day.
[29:37] If it really is a level up based on using a specific bespoke self representing tool in the courtroom,
[29:44] maybe that is upl.
[29:46] Debbie Reynolds: Yeah, interesting. Yeah, I agree there could be, I guess it's situational based on the situation at the time.
[29:53] But one thing that I'm still floored by is the cases. The hallucinated cases.
[30:00] 1600. Oh my gosh. And so some of these, some of these people, like they've been lawyers for 30 years and it's like if you're an antitrust lawyer and you've been doing practicing law for 30 years and then you see 20 cases that you've never seen before and they say exactly the thing that you want for your case.
[30:22] Like, I'll be suspicious, right?
[30:25] Cat Casey: So I'm of two minds. Like for a very long time I've been like, and I still believe this. Like if you're hallucinating a straight up case, that's not bad tech, it's bad lawyering.
[30:34] But the hallucinations are getting more hard to find. So it may not be a fake case.
[30:40] It may not even be. The case might be real.
[30:43] The quote might be real, but the interpretation could be wrong. Or I think I was reading somewhere like a quote from Roe v. Wade was used in that like product liability case.
[30:50] Like obviously you don't want it to be.
[30:53] It's not going to Work. Right.
[30:54] So I almost wonder if, like, the earlier hallucination. Hallucination cases were, like, all right, bad luring, slap their hand. And the more subtle ones,
[31:03] then we have to wonder, is it an issue of, like, automation bias, where people are taught to trust technology more than their own judgment? Is it an example of, like, the confidence collapse and, like, outsourcing of judgment?
[31:16] Is it something we need to be training for? So,
[31:19] like I said, I'm of two minds. I think the original, say, 1490 cases.
[31:23] Bad lawyer, not bad tech. But then as it gets more subtle, is that something we have to look at in terms of human predisposition to trust tech over their own judgment?
[31:33] And I don't know,
[31:34] for the partner, maybe slap his hand for the baby associate who doesn't know better?
[31:39] Maybe it's a little murkier.
[31:41] Debbie Reynolds: Yeah. I actually saw one case where the lawyer blamed the paralegal.
[31:46] Cat Casey: Oh, that never works.
[31:49] Debbie Reynolds: No, I'm like, that is such a bad thing to do. Right. Because the lawyer's always the supervisor, so it's like, you can't blame the supervise.
[31:59] And my thing is just read the case. Go back, be happy that you don't have to go into some dark library and get some dusty book. The fact that you can actually go on the Internet and look it up, if you have Lexis or Westlaw,
[32:13] you can literally look it up or have someone to look it up for you. Doesn't take that long. It's not that hard.
[32:19] Cat Casey: Well, and, like, it's not something you didn't do before. Shepherdizing, verifying, like, that was part of, like, your 9 to 5 as a lawyer four years ago. It should still be part of your 9 to 5 as a lawyer today.
[32:32] So it's. That's what I mean. I'm like. I'm of two minds. Like, if they were to click the link and see the case is fake, I'm like, dude, bad lawyering.
[32:40] If they were to click it and see the quote says,
[32:42] not even there. Bad wiring. If it gets a little bit more interpretive,
[32:46] I don't know if I would give them carte blanche, but I might be like, I got to think a little harder. So it's. And I think it's probably only going to get more and more subtle.
[32:54] The hallucinations as the legal tech tools kind of build guardrails. So I don't know.
[32:59] Verify, Trust, but verify. Verify.
[33:03] Debbie Reynolds: Yeah. Verify before you trust. Verify before you trust. Actually, yeah. So, I mean, is there anything else that you're working on that you're super Excited about right now?
[33:14] Cat Casey: Well, if anyone on here wants to know about AI and legal tech, I literally wrote the book on it. So if you go to Amazon and look up AI and legal tech and Cat Casey, I've got the book.
[33:23] Otherwise, no, like I'm, I'm really excited about some of the digital content and the upcoming shows for the masters. If anyone's not following me already, I'm Cat Casey on LinkedIn.
[33:32] If the background of the photo is neon aggressive hot pink, you have the right cat. Casey. Um, I write and talk about a lot of this stuff and I, I would imagine you will find Debbie and I on more webinars and podcasts and things together because it's about a once every other month thing.
[33:47] Cause we like doing it.
[33:49] Debbie Reynolds: That's so true. That's so true. Yeah, yeah. I love our mashups that we do and it's fun because I just think this is a fascinating topic. As a technologist, it's a very exciting time to be in technology.
[34:01] But yeah,
[34:02] now, if it were your wish, Cat,
[34:05] anything in the world, what would be your wish for privacy? Whether it be human behavior, technology or regulation.
[34:13] Cat Casey: It's part human behavior. I just wish the conversation would start earlier. Like, I feel like it's an afterthought. And I feel like as, as companies are bringing in new tools and technology, I wish that the privacy discussion was at the same day as the risk of a cyber incident.
[34:30] Right.
[34:31] Because I think as data becomes more valuable,
[34:34] these tools are going to be even more about aggregating and accessing it. And so as in house teams, as executive teams, as law firms are bringing these tools in, the privacy discussion needs to happen sooner.
[34:46] And I feel like that might just be a behavior shift of the people in leadership to realize like as much as legal needs of the seat at the table, privacy does too.
[34:54] Because the ripple effect of tools that are by design hoses to suck up as much information as possible could create a big privacy headache.
[35:03] Debbie Reynolds: Yeah, absolutely. So I think some of that is because people think of it as like a compliance issue or after the fact reactive issue where I'm more on the prevention side and the proactive side because I think that's where the problems start.
[35:20] So you can try to stop that snowball from happening. You're going to be way better downstream.
[35:26] Cat Casey: No, agree. So that is my wish. More people listen to Debbie. She knows.
[35:31] Debbie Reynolds: Well, thank you so much, Cat. It's so much fun to have you on the show forever. We've been trying to get together and we finally did it. So. Yeah. So everyone check out Cat.
[35:42] She's always doing amazing things, always engaging in conversation,
[35:46] and. Yeah, that's amazing.
[35:49] Cat Casey: Well, thank you for having me. I will say day. It should not take 20 years for. For round two.
[35:53] Debbie Reynolds: Yeah, totally. Yeah. Yeah. We got to do. We got to be better at this, that's for sure.
[35:58] Cat Casey: I love it.
[35:59] Debbie Reynolds: Well, thank you so much, and I really appreciate you being on the show. We'll talk soon.
[36:02] Cat Casey: Talk soon.
