Lawyers Using AI | Legal Lunch Byte | Episode 2

We are live now we’re still new to this whole live stream business, when we did the global law and business podcast during 2021 to 2023 we did 100 episodes, but we had the flexibility of recording ahead of time and then editing afterwards, so new medium, new time constraints, but also it’s great to be with you again. I think I mentioned on our chat earlier this week that I miss connecting with you because you and I are in different parts of the country, we have different practice areas, we sometimes work on same projects but a lot of times we we’re in different circles and so I’m always happy to check in and get your perspective, hear what you’re doing and get your thoughts on things that I’ve been thinking about as well. So maybe you can take a second and reintroduce yourself to our our audience today. Sure absolutely, and I agree it’s great to reconnect and to be to be doing this once again, although slightly differently and with more more maturity and experience behind, you know, our backs but yeah, so name is Fred Rocafort, I am a colleague of Jonathan’s at Harris Sliwoski, officially assigned to our Seattle office but working from from the the Sunshine State, and my practice is primarily focused on intellectual property and within that, particularly trademarks both domestic and international, but I do get involved in many other matters that are not exclusively IP or or don’t have an IP connection, especially if there’s an International Nexus. So helping clients establish corporate entities overseas helping them negotiate all manner of agreements for their overseas business activities, and also because I’m here in Florida and I’m a member of the Bar here, sometimes I’ll get pulled into Florida matters that might not have that International component. Excellent. So we decided today we would take our our next 10 to 12 minutes to talk about AI. It’s been on my mind; I mean you and I read a lot, the news a lot, so when ChatGPT was launched, it’s been over a year now I think it was, no about November of 2022, we started to hear about it, read about it, lawyers aren’t really known as early adopters of technology, I think that’s probably an overstatement, but I would say at our firm we we do tend to look at technology very early, because we are, we believe we are “Small But Mighty”. We know we’re small, but we believe we’re Mighty as well and I think technology is something that we like to leverage on our behalf. So what are your thoughts on that generally, on the quick roll out and specifically on our topic today: how are lawyers using AI to make a difference in their practices for better or worse? Yeah and I think that’s right, that’s a very important distinction to make. I mean we can talk about AI as everyone seems to be doing these days, in terms of its potential and all of the things that it can do, and in my own personal life, you know I’ve I’ve asked chat GPT or some of the other applications to write Christmas cards for me, things things of that sort, and it’s great, but there’s a a bright line between what we do personally and what some people are doing out there, even in the context of business activities. I mean it might be okay to just turn to AI and “Let It Rip” but for us as lawyers, that’s a different story and just a couple of days ago I read about what the Florida Bar is doing in terms of regulating the use of AI by its members, and just as an aside… I still get my Florida Publications sent to the office in Seattle, but as it turns out the previous owner of my house was a Florida Bar member, so I just get her Publications, you know it’s an arrangement that works fine, but you know, sometimes there’s a bit of a lag there, so this actually happened, it’s not a very recent development, although I just looked at it more more closely and that was very interesting and I know that you know, you and I had a conversation about this, but it was interesting to look at the scope really of what ethics rules are going to be looking at, or what the ethic rules are going to be looking at in the context of of AI? I know that the the the one topic that we addressed was was confidentiality and that’s at the top of the list, looking at the the proposed opinion that the Florida Bar is proposing but it’s it’s not the only concern. So you know, what are your own thoughts on this issue of of confidentiality, I mean are you what are you doing in terms of making sure that you comply with your own obligations as a lawyer when when using these these applications, assuming you are using them. I am using them and probably less than than some; I know our our partner Dan Harris uses it a lot and and it’s very fun because he is an extreme early adopter, so it’s been very fun for him to lead and then the rest of us say you need look at this, here are the opportunities to use. I’d say on the marketing side it’s it’s very interesting because I noticed that Dan will write a blog post for instance, a draft blog post on something I’m interested in or I’ve been researching and I’m pretty good about getting content in there but intros and conclusions are not my specialty I’d say, and so I noticed that for quite a while Dan would take my blog post because it’s going to be in the public domain in a matter of of hours or days, anyway he’ll take it to you know Claud or chat gpt and say: what do you think about this and then he’ll get a he’ll get a response about whether it’s kind of the flavor of the writing, and then he’ll say, all right, now add a conclusion to it and it will and in a matter of 10 or 15 seconds it will take my writing, it will write a great conclusion and then I’ll review it later and I’ll say Dan this is a great conclusion and he says well it was the AI that wrote the conclusion for you. And so it’s very interesting dynamics I think, for me as a writer, you know a lot of us as lawyers consider ourselves writers, but I think in our firm in particular, our specialty is taking these complex legal themes, breaking them down into into digestible pieces for our audience and that also helps me be a better lawyer as I’m doing that. So that’s one way that I’ve been using it in terms of our outward facing marketing. I’m a member of the Association of young International lawyers there’s about 4,000 young lawyers throughout the world aged 25 to 45 and I’m on the one of the governing committees for our upcoming Mexico conference taking place in a couple of months and the panel that I’m over is looking specifically at AI in M&A transactions, and so trying to be a good moderator and preparing the team and we’re still looking for speakers actually so in preparation for all of that, I took a contract, a bulky contract, you know say a 30-40 page purchase agreement, and knowing about the terms and conditions of, you know, anything you dump into the AI becomes AI property; I went through and anonymized the contract, initially changed all the names, all of the dates, everything that that could potentially point to any, you know, any sensitive information of any parties, and then I told AI to create a, what I call a closing agenda, and so it it goes through, it creates a punch list of things, you know what what has to be done now, what has to be done post closing, who’s responsible for it, and I took I think five iterations of this, where I wasn’t asking quite the right question, and so I had to keep asking and saying ‘Okay add this’ and I figured out at the end I was even asking for it the wrong thing, something that I call a closing agenda is not exactly what what the AI understood I wanted, and so it took a while for me to get to that to that point, but in the end I got there, and it still only took about 15 minutes after I had anonymized the document, so maybe all together, it took a little bit over an hour, that was my first foray into really making the AI be a workhorse for me, the way that I would use, you know use an associate or paralegal, to help me to take a really robust transaction document, and simmer it down into the the data points I needed to continue the transaction. And that’s a good segway into some of some of the actual obligations that are, well not only for Florida lawyers but you know some of the regulations, some of the rules that we can expect for other lawyers, you know in the same way that you would that you would work with a paralegal or a summer associate. Right, I mean, well the same sort of oversight that you exercise in those situations, you’re going to have to exercise it when using AI. Here’s a a good line from from the proposed opinion that I’ll share with you, and I’d like to, well, know more or less, what your take on it will be, which it will be to say yeah sure but maybe we can talk a little a bit more about that and here’s what I’m quoting when using a third-party generative AI program, lawyers must sufficiently understand the technology to satisfy their ethical obligations, and let’s be honest you know I think there are lawyers out there that don’t quite understand the technology behind email or or cell phones right? And now we’re asking them, you know to take it, but helpfullythe advisory opinion continues, this specifically includes knowledge of whether the program is quote ‘self-learning’ end quote. Use of a self-learning AI raises the possibility that a client’s information may be stored within the program and revealed in in response to future inquiries by 3rd parties, right so I like the fact that they’re not being cryptic at all, they’re they’re laying out the risk and asking lawyers to be be mindful of that, but I think that right there is, is really critical get to know the technology, I mean it’s tempting to start, I mean I know that I use all sorts of technology that is helpful to me without really understanding what what goes be behind it, but when it comes to to legal work, right, I mean we we do need to take a closer look and know what the implications of our use are. You know one of the first things I use used AI, probably, I don’t know six, eight months ago, first thing I did when I got on chat GPT as a lawyer is make chat GPT explained to me all the risks of using chat GPT. I said, tell me, I because I wanted to understand, and I was more concerned about, you know, the IM robot kind of scenario, and so I wanted to see how much chat GPT would reveal about its own sentience, right, I’m like, because I really didn’t understand, you know, the idea of artificial intelligence, and all the terms that are analogous to that, I didn’t really understand fundamentally how the learning process worked, and so I spent a couple hours one afternoon just asking question after question until I got booted off, because I was using the free service, right, they said all right, you’ve answered enough questions asked enough questions for today, you’ve got to wait until tomorrow to ask some more questions. I don’t think it’s because I was getting into the sensitive territory with chat GPT but it was really informative to keep asking and then I confirmed that with some, you know, some Google searches to understand from data scientists what really happens, how the algorithms are trained and like you said, is this a learning algorithm where it’s it’s open source, basically where anything any inputs that get sent to the prompt will will be analyzed and will be used in the future, I think I haven’t looked at the terms and conditions for all of the AI Bots, but I can’t imagine a scenario where they wouldn’t do that unless it’s a subscription service where part of what you’re paying for is you’re getting a walled off version of the AI software to use in a confidential way, right, which brings us to another point, right, that many of the the risks associated with, or some of the risks associated with AI, can be mitigated by the use of proprietary models, right, that you develop, that your company develops or that it acquires. I’m not a big Black Mirror fan, I should say I mean I know some people really are, I don’t know if you are, but I watched a few episodes and there’s one episode and maybe we can share later, I’ll track down the specific season and number, but there’s one episode I think that is worth watching and to really get a sense for what we’re dealing with, I mean you mentioned that you wanted to get a sense, you know, for how sentient, you know, the model was, and I don’t know if you’ve watched this one episode where this lady basically acquires an AI boyfriend, including a physical manifestation of the boyfriend based on her deceased, you know, real human boyfriend, but it was very interesting because you’ve got that 90% of the time when it just looks and feels like a real human being, but then there would be those moments when basically, you know, she at one point she tells him like, ‘go away I don’t want to see you again’, wakes up the next morning and he’s like, outside the house, and it’s like, ‘well I have limitations, you know, I can’t, you know, go more than x amount’…so jokes aside, right, I think there are some real real questions here and we do have to be careful, because at times it does seem like, yeah, like you’re dealing with this extremely competent Source on the other side of the screen, but it does have its limitations and maybe if I could bring our conversation to a close now, since we’re close to our time, the Bar Rules are are very interesting, non-lawyers don’t understand how important those Bar Rules are to us, especially with the ethical Bar Rules, you know, I have clients who sometimes will ask me to sign an NDA and I explain to them that my Bar Rules are much more stringent than any NDA they could put in front of me, right, and so the idea that that the Florida Bar I think has got to be ahead of the curve, in terms of dealing with these issues and helping lawyers to understand what their ethical obligations are, it really matters to us, as soon as they get encoded in rules, even proposed rules, it’s time for lawyers to pay attention to how we’re using it, you know we don’t want to get stuck with what happened in New York with the the AI hallucinating new cases that don’t exist and never did, and then lawyers getting sanctions for that, I think that we all cringed a bit in our seats as we read that, so it’s a good prompt to also look at rules and things like the cloud and so forth. Yeah, excellent, well, thanks Fred, it was great to be with you, look forward to catching up again next week and in the following weeks with you and with other guests that we’ll have and thanks everyone for tuning in today we’ll see you next time take care.