Avoiding Modern Ethical Mishaps: How To Develop a Strong Firm AI Policy | Justia Webinars

Historically, that process required getting either paralegals or some other staff to meticulously go through all the bankers boxes full of documents to find the pertinent documents for the case. Um, this no longer needs to be done with AI.

You don't want to rely solely upon AI, even if the output of AI, the case law and everything is correct. This is AI is a tool. It's not a substitute for you providing legal advice. Human oversight is always required. Again, a lot of hallucinations happen, especially when people are filing things with the courts and in case law. So, be careful there.

Welcome and thank you for joining us for another Justia webinar. My name is Nina and I'll be the moderator for today. Now, to kick things off, I want to go ahead and introduce our speakers, Desiree Martinelli and Jeff Lance.

Desiree Martinelli is a dynamic multifaceted professional currently serving as the director of marketing and analytics at Esquire Interactive. With a strong background in business and law, she excels in website development, branding, and internet marketing strategies tailored for law firms and attorneys.

Jeff Lance is an attorney, entrepreneur, and CEO of Esquire Interactive LLC, an interactive digital marketing agency for attorneys. He has given dozens of presentations and webinars and wrote written numerous public published articles about law firm marketing. He is also an ABA book author, Internet Branding for Lawyers, Building the Client Centered website.

The webinar presentation that you're about to watch will be approximately 1 hour and time permitting. At the end of the presentation, Desiree and Jeff will be taking audience questions. If you have questions that you'd like to submit throughout the webinar for the end of our presentation, you can do so by clicking the Q&A button in the control panel at the bottom of your screen.

Now without further ado, here are Jeff and Desiree to discuss avoiding modern ethical mishaps. How to develop a strong a burn firm AI policy.

>> Great. Thanks for the introduction, Nina. So, let's go ahead and get started. So, Desire and I are very excited about giving the presentation today and uh we hope that you'll all get some valuable takeaways for the presentation and things that you can actually do and take back to your law firm um after the presentation. So today's presentation we can see here avoiding modern ethical mishaps how to develop a strong AI policy.

So again as Nina said I'm Jeff Lance. I'm an attorney and the CEO of Esquire Interactive. My background is a lawyer. I was a corporate securities lawyer and also general counsel for aerospace and defense contractor. And uh co-presenting with me today is Desiree Martinelli. Desiree is our director of marketing and analytics at Esquire Interactive. She's also an attorney and prior to joining Esquire Interactive, she was primarily her practice primarily was focused on IP and business law matters.

So today we're going to be talking uh about the ABA model rules. You're of course going to want to consult with the rules and the jurisdictions where you practice. Some of the things and your rules may differ a little bit from the model rules, but just to let you know, we're going to be speaking about the model rules uh generally today.

So the presentation overview we're going to be doing the following. We're going to talk about what is generative AI. You know do you understand what generative AI is or what it does. We're going to talk about that. We're going to talk about how AI is being used by law firms. Then we're going to talk about some getting into some of the ethics matters. We're going to be talking about how the ABA bar associations are guiding our use as lawyers of AI. And in specifically, we're going to be talking about things like confidentiality, data, security, risk, and protecting client information. These are all the things that obviously when we're thinking about using AI are a really critical concern for us. We'll talk about how the AI ethics rule or how the ethics rules apply to our use of AI. And then we're going to shift into the um privacy the AI policy development uh portion of the program. We're going to be talking about the specific AI policy components, the things that you need to be thinking about including in your AI policy, and then lastly, how to develop a law firm AI policy. With that, I'll turn it over to Desiree.

>> All right. But before we jump into the uses for AI within the legal field, let's talk about what exactly is it. Well, according to the National Security Commission on AI, it is not just a single piece of hardware or software. Instead, it's this constellation of overlapping technologies that give computers the ability to solve problems and perform tasks that usually would require human input. Now, today we're going to stay focused on generative AI, which you see here. It's in a little red circle um on the graph on the screen, but um I just want to call your attention to generative AI um and that the fact that it overlaps other technologies. So, many of these other components can be um in a tool or contained within a tool that is generative AI.

All right, but let's talk about what specifically is generative AI and why is it so important that we focus on it? Well, it's a type of artificial intelligence that can create new content. That content can be text, images, music, and videos. It does this by taking in large data sets and then it learns patterns and structures from that existing data and uses them to generate original outputs. So, some of the most frequent examples that you probably have seen are things like chatbt, C-pilot, Claude, Gemini, and Grock. In all these cases, the big thing to know is that they are self-learning models. So once you put data in, they can put that data back out into future responses, which can lead to ethical concerns, which we'll talk about shortly.

So let's uh start by talking about how AI is currently being used by attorneys. Well, one of the ways that AI is being used uh now and has been used for a while is with respect to legal research and case law analysis. So, in the past, we've had uh tools like WLAW and Lexus that allow us to go in and do our the legal case uh case law research. Now, these tools and these programs are being um upgraded for AI like tools like Lexus Plus AI and Weslaw Edge. And these are basically accelerating the AI process or the legal research process in terms of giving us more information and being able to uh deliver output based upon case law analysis that we're doing.

Another related thing with AI and and case law is what's called predictive analytics. Predictive analytics is an example. If you have a case in a particular jurisdiction, maybe of a personal injury case, a car crash case for instance, in a specific jurisdiction before a specific specific judge. Predictive analytics can be used to try to get an idea in terms of how that that type of case might actually re be resolved if it goes to trial. So that's another way that AI is being used by law firms.

>> Right? So we wanted to give you some actual practical tools that you could try out and use in your practice. So one of these is called Torch. It's actually in beta right now and it's created by Darrow Labs. Um, and it's an extension you can add onto your browser that when you open either case law or contracts, it will go through them and analyze the document and provide you with a really great synopsis within seconds. So it can quickly surface precedents or related documents of large data sets. It also can summarize legal filings, contracts, and policies so that you can get to the core issues without having to spend hours going through case after case. And then it also tracks your research trials so you can save memos, analyze documents, and work folders and go back later and pick up where you've left off. Um, so let's take a look at this tool. So that you could actually get this tool, like I said, for free and add it to your computer um by going to the link at the bottom of this slide. You can see here I opened up a Supreme Court opinion and I popped on this extension and within seconds it gave me this really nice document overview. It talks about the parties, the procedural history, the outcome in the case and really gives you nice summaries. It gives you plethora of information essentially within just a few minutes so that you can quickly find out what this case is about without having to read through 50 pages. Um, so I highly recommend checking this out while it's in beta because who doesn't like free stuff, right?

Yeah, another way that AI has been used as attorney by attorneys is with respect to contract review and drafting. And AI can really streamline the contract analysis. It can do things like complete or or construct contracts from uh you know from nothing. You could talk you could say for instance I want to do a purchase and sale agreement uh for a particular uh sale of a business. And what AI can do is it can create the contract completely uh from scratch or it can be used to analyze the contract that you might already have drafted or maybe opposing council drafted. So what AI can do in this context is it could review the contract, maybe find things that maybe you didn't know to include or you didn't think about including. It could be things like maybe saying, "Oh, you really need to do some more representations and warranties for this particular sale and here's an example of some good representations and warranties to add." And right now there are tools like Jurro, Kira, and Legal On that can do these types of things that can do that can identify missing clauses and ensure compliance and also uh compliance and also um uphold legal standards. So the things like again uh representations of warranties, it can suggest putting those things if that they're missing.

Now litigation support and eiscocovery and document review are are really really critical things that AI can be used now. Um, in the past, if you were involved in a large case, you might have gotten dozens or even hundreds of banker boxes full of documents. And then historically, that process required getting either paralegals or some other staff to meticulously go through all the bankers boxes full of documents to find the the pertinent documents for the case. Um, this no longer needs to be done with AI. AI can sift through massive data sets and the discovery process much much faster than humans can do it. And the second thing to know about how AI can help in the document review process is with respect to what's called predictive coding. And predictive coding entails taking a small subset of the larger set of all the documents going through that with a person to specifically identify what documents or what types of things are going to be critical and what types of things are not critical. Once AI has been trained like that, it can uh actually do a really good job of going through a massive data set and it can actually learn as it goes in terms of what kind of documents are critical and should be the subject of more focus.

We're also seeing AI tools being used for billing, which is super exciting. So, these new tools can help firms actually estimate their legal fees more precisely by looking at historical data and historical billing. Um, what's really cool about this is it's going to give firms the opportunity to start charging flat fees for matters that were previously only build by the hour, things like litigation where there's a bit more uncertainty. Now, with these tools, because it can analyze particular cases that have previously been handled by the exact people at your firm, they're going to give you really, really good estimates of of where your costs will be. So, like I said, that's going to be advantageous because you'll be able to bring in more clients by offering those flat fees. Clients are going to be happier because they don't have to worry about, you know, the clock continually running and what their bill's going to look like. And in turn, it's just a better experience across the entire fur. This is just an example of one such um billing software. This is big hand. Um unfortunately this one is not free but uh this is a really cool tool that will predict case costs again based on that historical firm data. It also will build really complex budgets with multiple parties that are tailored for each matter. So you can put in attorneys, parillegals, support staff and really account for everybody in the pipeline that's involved on the matter. Um, you can also again improve fee estimates for transparency. So you can tell your clients what to expect and also know what you can expect and budget for and what you can handle. And last but not least, this supports alternative fees. So again, this allows you to go to those flat fees in some cases, whereas you previously haven't been able to, which can be advantageous for both clients or if you're general counsel and just need to have a better handle on what exactly those fees are going to be so that you can make perfect budgets. this case um this type of tool is going to be you know the ticket to doing so.

>> So AI is also being used in alternative dispute resolution mediation. Uh what can be done with AI is you can input certain facts into AI tools. You can even put in things like jurisdictions where a case may be and AI can take a look at that and if you have a good data set to put in, AI can be uh really helpful in terms of trying to predict what the outcome might be to give lawyers and also their clients ideas about what kind of settlement offers might uh might be made or should be made or if there if a case doesn't settle what the likely outcome might be if it actually goes to trial. So there's there's a number of different uh tools like that. One of them noted on the screen is Dourus, but there's additional um tools that are being developed all the time for this particular purpose.

>> Now, we're also seeing this technology being utilized for legal summaries and briefs. So, AI can draft legal memos and summaries, which significantly cuts down on both the research and writing time involved in that. Now, I do want to note here though, if you are using any type of tool for anything that you're going to be submitting to a court, um, human oversight is absolutely essential and critical to make sure that you can avoid AI hallucinations, aka making up fake cases or fake facts. Um, this has been where a lot of attorneys have gotten in trouble with the ethics rules. Um, so if you are filing anything, just make sure that you check every single case citation to make sure that it is a real case and it's accurate.

Now, we're also seeing movement on the compliance and risk assessment side. So, now these tools can monitor regulatory changes um and with laws like GDPR, HIPPA, and the SEC regulations. And it can actually flag potential violations before they happen so that you can get ahead of it and help your clients avoid liability. Now, in a similar fashion, we're also seeing tools being utilized for IP. Um, as an IP attorney myself, I've spent hours and hours rifling through patent claims to find out if there's infringement for my clients. And, you know, now I don't have to do that. I can turn to these tools and they will actually scan global databases for trademarks and patent infringements. And then again, they'll notify attorneys of potential IP conflicts before they escalate, which could be really, really helpful on either side um of representation in an IP.

>> So, let's talk for a minute about how the ABA and bar associations are guiding the use of AI in the legal practice. So, one of the things that I would highly suggest if you haven't uh read this opinion already is opinion ABA opinion 512. And I put on the screen in uh the yellow highlight this is a short link to get to that opinion and it is case sensitive. So you can see it's bit.ly/aba-512 and this opinion is about I don't know 14 or 16 pages long but it covers a wide wide range of topics with respect to AI things regarding uh fee uh billing and fees supervisory responsibilities whether you should be talking to your client and getting your client's informed consent about using AI as well as security and other types of risk. So, I would highly encourage you to read this opinion um as soon as you get a chance.

In addition to the ABA opinion, there's other opinions that have been issued by state bar associations. Um some of these are on the screen. Um these come from, you can see there's Florida, Pennsylvania, Philadelphia, DC, New York, Mississippi, and Michigan. There are additional um guidelines that are being issued by the bar associations. So, you'll need to make sure that you check with the bar associations and see what they've uh put out or what regulations um they've developed for the organiz or for the um areas in which you practice.

In addition to the uh to the opinions, there's also another thing called practical guidance that have been issued by the state bars. Um on the screen there's uh just an example of three of these practical guidance um initiatives you can see from California, New York and New Jersey. So these are all additional things that are going to talk about you know your use or your firm's use of AI and the things that you might might need to do or should be considering.

All right, let's talk about confidentiality, data security risk, and how you can protect your confidential client information. So, first of all, AI tools process and store data, which in turn raises the risk of unauthorized access, data leaks or breaches. So, if you are inputting any type of client confidential information into an AI system, you at a minimum need to know absolutely what security measures it has in place. um because if it has insufficient security measures, it could expose the sensitive data to both parties outside of your firm and maybe even people in your firm that are supposed to be walled off from a matter.

Now, it's important to mention again, this is the generative AI point of this, that some AI models retain and repurpose past inputs. And because of this, they can potentially disclo private or privilege client information to others when it puts out those new inputs. So, you want to be very very careful with what you're putting into these tools um before you upload it and also the security measures that uh they have in place.

Now, in accordance with ABA opinion 512 that Jeff just mentioned, ABA says you should consider at a minimum these four things before you adopt an AI tool within your firm. First of all, the likelihood of disco disclosure, unauthorized access. So, can somebody easily jump on your computer, get into the tool, and access that confidential information? If so, that could be a problem. You also want to think about the sensit sensitivity of the information you're going to be inputting into the tool. So, if I am putting in um the client's notes from the last consultation into AI, that's going to be a lot different than putting in a public case law. Um so, if you're only using it for case law and it's something that's public, well, obviously that's going to lower the risk. But if you're using it for confidential information, you want to be much more careful. You also want to think about the difficulty of implementing safeguards. If you can't use a tool without going through this step and this step and this step in order for it to be safe and good to use in your firm, then you may not want to do it. The reason being is because your entire staff, junior attorneys, parillegals, um legal assistants, all of them have to comply with these rules as well. And you want to make sure that it's not too complicated. If it's an eight step process, the reality is it'll probably break down and you'll probably end up with issues in the long run. And last but not least, the extent to which the safeguards negatively impact your ability to represent a client. So if you really um are be restrictive or you can't use a tool in its full capacity or it puts out bad output and you think it's going to have a negative effect on your client, you may not want to use it in that case either.

Now, let's look at a real world example um of of kind of this in action. So, let's say that you decide that you're going to create a pleading and you're going to use chat GBT. Do you know what happens next once you upload that information? Well, when confidential client information is actually inputed into chat GBT, the model can save and learn from that data to improve its future responses. Now, it's important to note here there are a few different types of accounts on chat GBT. So there is an enterprise/ business account. In that case, chat GPT does not learn from the data if you are working on an enterprise or business account. However, if you are on the free plan or the $20 pro plan, which a lot of people are, in that case, it does save and learn from your data. And previously, it would only keep your data for 30 days. But earlier this year, the New York Times actually sued Open AI and the outcome of that case was that they had to indefinitely keep all of the information, including deleted records on their system. So, this means essentially that any information that you were inputting through the pro or through the free version will be kept indefinitely. So, you want to be super super careful there. Um, one other thing I'd like to mention is that chat GPT processes data on thirdparty servers. Um, this can create a risk that sensitive information could be inadvertently shared with others leading to a breach under rule 1.6. Um, I don't want to scare you though because a lot of other programs and software that we use on a daily basis also store information on third party servers, but I want to make you aware of it because it does potentially present some risk, but it's not grave grave grave. So, just to make you aware of that.

Now, one little note here. So, I talked about how they will keep orbited will keep your um information indefinitely when you're on the pro plan. Um, but you could actually go into the settings and there is a setting that you can toggle off for learning. If you turn off the learning, then that's actually going to stop them from learning from your data. So, you want to try to do that whenever possible.

All right. So, here are a few other suggestions on how you can protect client confidential information. So after you've found out what tools you'd like to use, once you're actually using them, you can redact all identifiable information. So remove names, take out socials, take out case numbers. You can also avoid uploading entire documents. If you're just working on a paragraph, just upload that. Don't upload 11 pages, you know. Uh you can also use hypotheticals instead of your real scenarios. So just pick a situation that's close to your client scenario. Again, disable chat history. If this is available, turn off the learning and chat GPT. Avoid using public Wi-Fi whenever you're using these AI tools. So, um, airports, for example, are renowned for being, um, hacked when people are using the public Wi-Fi there. So, just be very, very careful if you're using any type of public Wi-Fi. You just don't want to be putting any sort of information in. Um, next, don't input client information just to see what it says. Okay? I am super guilty of this. I'm a naturally curious person. I love learning. So, whenever I get my hands on a new tool, the first thing I want to do is go play with it, right? And figure it out. But if you're the same way, just don't put in information that's censored client information. Pick something that's public record. Pick something that's not going to have any type of risk of getting you into an ethical dilemma. And last but not least, and the most important thing I can recommend is to have a written firmwide AI policy. You know, everybody on your team needs to understand what tools they can use, when they can use them, how they can use them, and it needs to be clearly defined so there are no questions asked. and we're going to talk about exactly what components you need to include in the AI policy in just a few minutes.

>> Okay. So, let's talk for a minute about how the ethics rules apply to AI usage. So, first of all, we have rule 1.1 talks about competence and the and the competence sort of pertains to a number of different things including being competent on the use of pertinent technology uh in the legal practice. So this is under rule 1.1 and 1.1 essentially says that lawyers need to be competent and stay informed about the benefits and risks of technologies. This is going to include AI. Uh in addition to that you want to to make sure that you understand the AI tools that might be used or might be available to use to represent clients. This includes specific uh components or aspects of using AI, including the capabilities, limitations, and the potential pitfalls of using AI. And then lastly, you want to make sure that AI produces accurate results. This is a really critical thing, and we're going to say this a few different times in the presentation, but even today, even after lawyers have gotten disciplined for not checking uh case law and things that have been produced by AI, these kind of things are still happening. So just make sure that if you're using AI, that's great. You don't want to rely solely upon AI. Even if the output of AI, the case law and everything is correct. This is AI is a tool. It's not a substitute for you providing legal advice in giving your critical oversight of cases and other legal matters.

So in terms of what we should be doing according to ABA opinion 512 to comply with rule 1.1 uh for competence we should read about generative AI tools or sometimes are called GAI targeted to legal professions we should understand generally what types of tools are out there what type of tools may may make sense for us to use and representing our clients we want to specifically understand things like the terms of use the privacy policy related contractual terms and also you know the potential ability or the potential risk of disclosing client information. The third bullet point is one that I think is really important. We need to consider the impact to clients of not using AI and a good example of that is the document review process. If we have a document review process in the past, we've had you know all the bankers boxes full of documents or electronic discovery and we've relied upon uh parallegals to meticulously go through all that information and maybe it might take months to do it whereas using a good AI tool it might only take a week or a couple weeks. We need to be thinking about that. We can't just simply say, you know, I prefer the old way because I'm comfortable with the old way and the old way is going to generate a lot more fees for us than the new way. That's not really what it's about. You need to really consider whether you should be using AI. Uh you can consult with others who are proficient in GA GI technology. You don't have to be an expert. And with respect to your learning of generative AI and other AI products, this is not like a static learning. It's not you do it one time and you never have to do it again. AI is going to change very rapidly. It already has been changing rapidly. So you're going to make sure that you're going to need to keep a breast with the changes in the technology and the different products and things that may be available. So uh lastly at the bottom of the slide uh this is what ABA opinion 512 says. Basically, you don't have to be an expert in generator of AI, but you do need to have a reasonable understanding of the capabilities and limitations of AI products.

So, let's talk about rule 3.3 candidate to the tribunal uh specifically with respect to AI generated errors and misinformation. So, rule 3.3A1 says that we can't make false statements to the court and also rule 8.4 for C says we can't engage in things like dishonesty, fraud, deceit or misrepresentation. Misrepresentation really is the critical one here. So we need to make sure again that if we're doing pleadings or we're making other kind of representations to the court, even if it's during uh a trial and oral representations that are based upon AI, we need to make sure that we actually know and confirm what we're saying before we we say these things to the court. So AI generated errors and misinformation. There are, as Desire mentioned before, there's been hallucinated citations. There's been fabricated uh case law. There's been a lot of these different things that have happened. So again, you need to make sure that whatever you're using is actually correct. You don't want to over rely upon AI. Again, AI is not a substitute for your legal judgment. And lastly, courts are increasingly requiring the disclosure of AI assisted filings and imposing sanctions for errors. So, this is kind of an evolving area of of the law and jurisdictions. So, just be aware that courts in your jur jurisdictions um either now or if not now, maybe in the the near future may be requiring you to disclose whether you've you've used AI with respect to pleadings or other documents that have been filed.

So, let's just take a break for uh one second for this really super important question. So, who is the best rock guitarist of all time? So, if you've thought now in your in your head who who you would think is the best guitarist of all time, I put the search into uh Google into Gemini to see what uh Gemini would say. This is who they came up with. one of the top 15 rock guitarists of all time, the late great Aussie Osborne. So, you're probably thinking to yourself, if you're a fan of Aussie, that hey, I you know, I I didn't know Aussie played the guitar. I've never actually seen him hold a guitar, much less play one. Well, if that's your recollection, then you're correct. Azie Osborne uh was a was a great in rock, but he was not a great guitarist. He did not play the guitar at all. Somebody who was a great uh guitarist was a guy named Randy Rhodess uh who is in Aussy's band. Um guy who's tragically killed at a very young age, but he was the guitarist behind uh hits like Crazy Train and some just like really iconic uh rocks uh guitar solos. So what's happened is AI when I put this this query in, AI went and spidered a lot of different sites. It found a lot of sites talking about presumably Randy Rhodess being the great guitarist and Randy Rhodess playing for Azie Osborne and then it came up it came back with the um conclusion that it was really Azie Osborne who was the guitarist and not Randy Rose. So if AI can make uh hallucinations or mistakes and produce false information like this for things that are really really critical like who is the best rock guitarist of all time then it can probably do the same thing with case law. So again, you just want to beware. So the takeaways here again review AI content for accuracy and legal validity and also AI is not a tool for your legal judgment. We don't want lawyers or give the impression for lawyers is you now no longer have to do the legal work and everything you can just give it all to AI because that's definitely not the case.

>> Let's talk about rule 1.5 and how AI is affecting fees and billing. So, we know now that obviously I'm sure we're all aware of the fact that we must have reasonable fees, but when we're using an AI tool, it can significantly speed up the process for obviously putting out work product. So, this kind of raises the question, well, what can I charge for? What can I not charge for? Well, according to the ABA, if you're building on an hourly basis, you cannot charge the time that it would have taken to complete that work if you did not use the AI. So say that a contract usually takes you 5 hours of of work and you can do it in two with an AI tool. You can't charge for the 5 hours. You can only charge for the two. Now also if you are operating with a fixed fee model, you may also um need to revisit your fees too. Um for an example, we'll go back to Jeff's example with the hundreds of bankers boxes in the war room. In that case, if it would have taken three months to go through all those boxes, but now you can go through it in one month and you're charging a flat fee, you may need to adjust that flat fee down to accommodate the less um the less time that's going to be involved in that matter. So, bar associations, there have been a few that have taken um additional stances just regarding certain specific points regarding AI. So under ABA formal opinion 93-379, lawyers can charge the time spent using AI tools, but they cannot inflate hours artificially. So again, this is the only bill for the time you're actually spending making the work product. Now under Florida Bar ethics opinion 24-1, lawyers may charge for the reasonable use of an AI tool, but they can't charge the client for learning how to use that tool. So, if you bring in a new um piece of software that you're really excited about, you start using it on your first client, you're not really great at using, you're kind of learning, it's taking you a little bit longer than it should maybe or longer than it would the next time. In that case, you're learning that tool and you can't charge that to the client. Now under the New York ethics opinion 2024-5, AI cost should be considered overhead unless they directly impact a specific client's case. So this would be again turning back to those bankers boxes and the discovery. In that case, you know, in that case, if you brought a tool in just to complete that project for that one client cuz you're like, man, this can speed up this process by 2 months, then you could charge it to that client because you're only bringing it in specifically for them. However, if you're using a tool across the board for all your clients, that needs to be considered as overhead if you're in that jurisdiction. So, one more note here. You want to make sure that you clearly explain in your retention or engagement letter how AI is going to impact with the cost and delivery of your legal services. And this can't just be blanket boilerplate terminology. The ABA actually calls that out. It needs to be specific. So, just make sure that you can mention all the tools that you're using, how you're using them, and how it's going to impact that case.

Okay, let's talk about rules 5.1 and 5.3 and ethical responsibilities and supervising AI use. This is probably one of the largest areas that we are seeing attorneys getting in trouble. Um, it's usually when a junior attorney has created work product that has fake case law in it. Then they pass it on to the senior attorney, the senior attorney files it with the court and bam, both attorneys end up getting in trouble. So I say this all to say that it is absolutely critical if you're in a supervisory role to ensure that your firm's use of AI also complies with all the ethical standards including accuracy, confidentiality, and non-bias. You also want to make sure that you have policy and trainings that are implemented within your firm. Um you have to ensure not just that attorneys know how to use these AI tools, but your parallegals, your support staff, and then everybody's on the same page. So you want to have these types of policies and training in place. So if anybody has questions, they have somewhere to turn to. Again, as I mentioned, supervising lawyers must oversee junior attorneys. If you are filing anything with a court, make sure you check every one of those citations. And last but not least, under 5.3b, you know, this obligation extends not just to junior attorneys, but also to non- lawyers as well. So you have to make sure that your front desk staff, your parillegals, like I said, everybody's complying. Um, and you don't want it to turn into a wild west where everybody's just out there picking whatever tools they want and using them for whatever and you don't know what security measures are going on because essentially this is going to lead to some very very bad ethical violations. So, what should you do to help help with this process? Well, we recommend assigning specific oversight responsibilities to a person or a committee um at your firm so that there's somebody that's designated to looking everything over and making sure that you have that system in place. Again, you also want to implement very very clear guidelines for when and how AI can be used. So, you want to specify the exact tools that can be used, what they can be used on. Um, in addition to that, you also want to regularly review your AI tools and the data that it's putting out to correct any mistakes generated by those AI tools. So, um, one little note here, large language models are known to be biased. Um, they're from historical data. So, if there's discriminatory content or something that has been biased in that data set, it can travel into the feature responses that are output by that tool. Um, because of this, you want to just make sure that you are regularly reviewing your data sets so that it's not tainting that work product that you're putting out in the future.

So, let's talk about rules 1.6 and 1.4 and specifically, do we have to disclose our use of AI to our clients? And the answer is a resounding maybe. And what does maybe mean? what what are the specific circumstances in which we might need to disclose AI to our clients? Well, under um ABA opinion 512, it talks about we need to disclose or suggest that we need to disclose our use of AI to our clients in these particular circumstances. So, first where there's the possibility of the disclosure of confidential information. So if you are using AI and you are putting confidential information into um the AI platform, you should disclose it to clients. You can also talk to clients about you know all the security uh things that have been done to minimize uh the chance that there's going to be any breach of that information. things like you've turned off learning, you've turned off or that you've minimized the chat history um and maybe that to the fact that you might be putting in um anonymized information instead of the client's actual information. The second uh way or the second situation which you may need to disclose your use of AI to clients is when the output is being used to make material decisions about case strategy such as whether to settle a dispute. So in this circumstance, if you're using an AI tool that's going to predict maybe an outcome for a particular case, um that's fine to use it. It's fine to give the that and it's fine to talk about that with your client, but you should tell your client like hey AI the AI tools that um I've used, they recommend settling for this amount and based upon my experience, I think that that amount makes sense. So next, you may need to disclose your use of competent or of AI under rule 1.4. This is this rule is just the general communications with clients about how you're going to serve them. So you may want to take a look at that rule and and ask yourself whether that rule might specifically require you to disclose your use of AI. And then next, if you're required under applicable ethics rules or opinions, these ethics rules and opinions are going to be evolving over time very quickly. And so just make sure that you look in the jurisdictions where you practice to see if there are any um new ethics rules or opinions that would address your disclosure of AI to clients. And then lastly, of course, if you're asked by a client whether your firm is using AI for their matter, you need to tell them.

So, opinion 512 focuses on informed consent and it focuses on these specific things with respect to how you disclose your firm's use of AI. It talks about disclosing your best judgment about why AI should be used. Um, this is an example or an example of this is in the document review process. You can talk to the client say, "Hey, it's going to take us 9 months to get through this document review." if we did it manually, but if we use AI, it's only going to take us one month. Um, clients are going to be happy about that, you know, because it's going to save them significant money. You should talk to them about uh the extent and specific information about risk. Again, you might talk about the models that you're using and the things that have been done proactively to maintain security and maintain um uh or prevent the disclosure of confidential information. you want you need to talk about to the client about what client information will be used with respect to AI. So for instance, as Desire mentioned earlier, you may do uh some very minimal information. You may anonymize information that's being put into an AI tool in which case the real information is not going to be uh subject to any type of disclosure if there's some kind of breach. You can also talk to clients about the benefits of using AI. You may be using AI for things like contract drafting that again could save significant time and it can also suggest things that maybe you didn't think about putting it into the contract. Um, the big thing that ABA opinion 512 says is that you can't get informed consent by boilerplate disclosure. And what I mean by boilerplate disclosure, I'm talking about something very simple that says our firm may use AI or will use AI for your matter. that's not going to be enough according to 512 if you have to get informed consent from clients.

So, if you do need to disclose your firm's use of AI, well, what's the best way of doing it? Well, for new clients, um certainly a great way to do it is just in the retention letter agreement. If you've had clients that have been existing for a long time, you may need to either update the retention letter agreement or maybe you want to send them a separate letter that talks about your firm's use of AI. And again, the disclosure needs to be specific. You can't just say something like, "We may use AI if you're required to get informed consent from clients about your firm's use of AI." So a few things that you can do, you can suggest or you can uh discuss possible uses of your firm's use of AI. You can say we may use it to create pleadings. We may use it in case review or or case uh research. Uh we may use in document review. However you may use it, you want to make sure that you're communicating that with a client. The next aspect is fees. As Desiree mentioned, you can't bill for the time it would have taken you had you not used AI. You should talk to your clients about that. Put that in the retention letter or whatever other communication you might be doing. Um, certainly they're going to be happy to know that you're not going to charge them the 10 hours that it took you 5 years ago to do something when today it only takes you an hour with AI. And then lastly, talk about why AI is being used. Talk about the benefits of using AI. These are things that most clients are going to be very happy and they're going to um understand that your firm is proactively making use of technology, not staying, you know, the way that operating the same way that it did in past decades.

So, let's talk about the AI policy components for a second and why your firm needs an AI policy in the first place. Well, if you don't have a an AI policy in place that's actually being followed, there are a lot of bad outcomes. You can have like client breaches of information. You can have people using their own uh favorite AI tools that may not be secured. Um this can risk um ethical violations. Maybe attorneys don't know what they need to do. Maybe they're not billing at accordance with what they're what they should be billing. they're charging the amount that it used to take them rather than the amount spending with AI. There's just kind of a lot of bad things that can happen if your firm doesn't have a strong AI policy. And the AI policy itself is going to establish firmwide standards. This is going to help everybody know in terms of what AI tools they can use and how they can be used. And then lastly, the um ABA there there's the issue of potential disclosure of AI use to uh clients. So certainly if

You're required to disclose your firm's use of AI, then you're going to make sure that you need to do it. It's important not that just like like we and um you all that are listening understand that it's important that this information gets uh conveyed across your firm firmwide not only to all attorneys but also to other staff members.

So let's talk about the AI policy components and then we'll move into actually um setting up the steps for developing your firm's AI policy. So these are things that we covered in the last uh few minutes. We talked about uh competency. We need to have that as part of our AI policy. Talk about what things need to be done and how lawyers should be um maintaining uh competency with respect to AI technology.

Oversight as Desiree mentioned before that's a really critical thing. We don't want to have cases in which junior lawyers for instance put together pleadings using AI and then the senior partner just takes it and files it without actually reading it or checking the case law or citations. It's really important also for confidentiality and data security uh clearly that we want to make sure that we're only going to be using AI platforms that are secure. And then the platform approval aspect is also really critical. Again, we don't want people just working on whatever their favorite AI platform is at or because that's going to introduce things like the possibility of platforms not being secure.

The other things for AI uh policy components that we need to build into our AI policy is accuracy and accountability. We should build in things like, you know, who's going to check the accuracy of citations and case law. Again, this also relates to the supervisory um requirements that we have to comply with. We want to talk about uh transparency and client disclosure. There may be for instance um requirements in your jurisdiction that you have to disclose again and get informed consent with your clients. Well, if we don't have something like that written in our AI policy, attorneys may not know that. Some attorneys may understand that and be putting uh retention letters that include all this information and some may not be doing it.

And then the last two um components are ongoing AI training and audits and billing and fee policies. And I'll talk about both of those things in just a minute.

All right, let's talk about some particular steps for developing your AI policy. So first of all, we recommend choosing an AI policy committee. So this committee is going to be responsible for your policy development, the selection of approved AI tools, policy implementation, and policy compliance and updates, including training, um any ongoing editing or review. So um with this committee, we highly recommend having somebody from your tech department on that committee if you can. If you don't have somebody like an IT administrator with in-house, you can turn to external consultants. So you don't have to be an expert again in this case. Um so this committee could be made up of something like outside of external consultants, people from each of the practice groups in your firm. Really a myriad of different different people could be on it depending on um what you see as the most the best for your firm essentially.

>> So let's get back and talk about AI platform selection. But before talking about this, what I really want to emphasize is it's really really critical for your AI policy to be in writing because again the things that we're talking about today, you'll take it back and you'll remember that. But if we don't have an AI policy that addresses all these things, there are probably going to be holes in terms of the knowledge of of attorneys that don't really know or understand the things that they need to do to make sure that they're using AI ethically.

And so the big part once you've got your committee chosen is AI uh tool or AI platform selection. And it's going to be really really important that this is clear in your AI policy in terms of what tools can be used and not only what tools can be used but how they can be used. You may have some AI platforms that are really good for doing uh creating pleadings. You may have another platform that's really good for contract analysis or document review, but you want to make sure that you don't have people either bringing in their own platforms, which you you definitely don't want to do, or using a specific platform for an a use that's not approved.

So one of the things that you can do is once you have your AI policy uh committee formed, they can work with technology professionals to understand things like the security and confidentiality concerns. You can also do things in in as part of the vetting of an AI platform uh called platform accuracy testing. So platform accuracy testing is if you take an area that you know really well, you might put in you might develop a pleading um and put that information into whether it's chat GPT or maybe it's another AI platform. Put that information that you already know in the AI platform and then see what the output looks like. Does the output look exactly like what you expected it to look like or does it look completely different? If it looks completely different, then you may want to think about using another platform rather than that specific one.

And then you also want to have as part of your policy regular risk assessments of AI tools to prevent ethical and security issues. Again, the AI is is rapidly developing. It's changing continuously. So, what works today and might what might be a good um platform today may not be a good platform later on.

Make sure Oh, I'm sorry. I'm sorry. Go ahead. That's right. No, you also want to make sure that you define what permitted uses the AI tools can be used for. So, um, in some cases, you may be using a tool only specifically for certain things like writing pleadings um, or updating contracts. But either way, regardless of how you're using these tools, you need to clearly define each of the specific tools that can be used and then define tool by tool what they can be used for. Because in a lot of cases, like I said, there will be different uses uses across your firm and you want to make sure that everybody knows specifically what they can and can't do.

All right. You also want to make sure that you're identifying processes for keeping client information secure. So, do not input confidential data into an AI system if you don't know about it. Security protections or if it lacks security protections. Um, you can do things like anonymize names, take out information, but you know, to be on the safe side, it's just better not to put that information in if you don't have to.

You also want to review AI vendors terms of use and data sharing policies. So, this is what the ABA says that you should do at a minimum before using a tool. you should have already reviewed them and understood everything about those two documents. And last but not least, you require encryption and secure AI platforms for any client related AI use. So just make sure when you go through those tools that you're looking to see how they're handling your data. Is it being encrypted so that it's harder to hack into? Um you want to be very specific and make sure that those types of mechanisms are in place to securely keep your client confidential information protected.

Again, when we're talking about billing, just a little reminder here, you want to make sure that you also mention how you can bill for AI use in your policy. You know, a lot of people aren't going to be uh just aware of the fact that they can only bill for just the time expended or that they need to lower their flat fees if they're unreasonable. So in that case you want to make sure that you clearly define how you handle both hourly billing and flat fees or if you have a hybrid billing model how that is handled as well. Um and again you can only bill for the time actually spent creating work product not for what you would have spent previously.

In step six we're going to talk about oversight and accountability processes. So there needs to be a clearly defined process for how you're going to be overseeing the work product that's generated by these AI tools. So you want to say like if it's this person gets the first review, this one gets the second review, you want to clearly put those roles and define what those roles are in your policy that everybody understands that reporting system. Um you also want to implement a reporting system for attorneys to disclose AI use in case preparation or client communications. So in some cases this may mean keeping a log. So you have a log where your attorney is using AI use and maybe they passed off to a paralegal then the paralegal puts their name that they use the AI on this portion of the matter and so on and so forth. That gives you the ability to go back later if you have to if you have to disclose according to court rules or one of these other ABA rules. Um it's going to give you the ability to track all that. So we highly recommend having that built into your policy as well.

So, let's talk about client disclosure, consent, and retainer letters. This is another thing that's really important to be uh included in your written firm AI policy. And the first part will detail when you need to disclose um AI, your your firm's use of AI to clients and get informed consent. And again, on the screen, you can see a couple of examples when it affects case strategy, legal fees, or client privacy. Those are things that you're probably going to want to communicate to your client. So, that needs to be included in your firm's AI policy. And the AI uh disclosure should be included in engagement letters or maybe a follow-up letter if it's been a longtime client to ensure uh transparency and to ensure that clients actually understand, you know, how you're going to be using AI.

And then lastly, the client informed consent must be obtained before attorneys input case details uh into AI tools. So, one of the things that I mentioned before is that under ABA opinion 512, it says you can't just give a general disclosure about like our firm may use AI. You need to be more specific so that informed consent can be done. So, here's a great place in your AI policy where you can come up with, you know, maybe several different examples of what that specific disclosure may look like. So, you might have three different ones as an example. One may be geared more towards uh document review, one may be geared towards uh litigation and preparing pleadings, and another one could even be uh more geared towards things like contract creation and analysis. If you put these disclosures, these uh sample disclosures in your um AI policy, then when a new client comes on board, an attorney knows first of all that they they may need to to make that disclosure. And then it's also going to help them and speed up the time is they can go and then maybe just take the disclosure that's more applicable to their representation of the specific client and just go ahead and take that and then maybe tailor it a little bit in the retention letter.

So another important thing is AI policy audits and updates. This should be included in your AI policy plan. You want to create an audit plan where you're going to periodically just check and make sure that people are using the tools that are approved and they're using it in the approved ways. Um, you can you don't have to audit everybody and you don't have to audit people constantly, but you might want to do it on maybe a semiannual basis where you just take a look at maybe a subset of attorneys and staff and just make sure that you know what tools are they using? uh are the supervisory guidelines that you put in the AI policy are those being uh followed? Um we suggest reviewing the AI policy at least annually especially because now things are changing so quickly. Uh the AI policy committee really should make sure that they stay updated on any any regulations and ethics rules that specifically pertain to a firm's use of AI. And then also the AI policy committee really should meet on a regular basis. Maybe it's a quarterly basis. So maybe it maybe it's not even that frequently, but they really should get together and then understand like evolving AI tools and evolving AI methodologies to see if the AI um policy needs to be updated.

Training and ongoing compliance are also really critical. Um, once you get the AI policy developed, you want to make sure that the that the staff and attorneys are all trained. They're made aware of the things that are important to them or not necessarily important to them, but the things that need to be do that they need to do in the AI policy. You want to make sure that you're updating it again to comply with new regulations. And then really, really critical is the firm leadership really must enforce the AI policy compliance. these things and any type of policy is only going to be as good as the enforcement of the policy itself. And I don't mean enforcement in a in a negative connotation, but enforcement in terms of this is our AI policy and we all need to make sure that we comply with it.

So next, let's talk about implementing and communicating the AI policy after it's actually been done. the AI policy really should be, you know, included with any other firm policies in the same place. Maybe it maybe you have a written binder, maybe you have all of this online in an internet or some other fashion, but you need to make sure that the AI policy is there alongside all the other from policies. Uh we suggest that when the AI policy is first implemented that you require all attorneys and staff to somehow sign off or otherwise acknowledge that they've read the policy. This is going to be really important and people really do need you really do need to make sure that people are reading it. And then when you have new hires, this is another thing as well. When the new hires come on board as part of the onboarding process, they should also um read the AI policy. And you may also want to have a a training with staff and or new attorneys. Talk about some of the things that are really critical depending upon, you know, uh which different person you might have, whether they're a staff member or an attorney so that they clearly understand, okay, this is the things that I need to do in the the a uh for our use of AI. I need to I can't bill for time that I don't spend. Um I can't I need I need to supervise junior attorneys. I need I may need to disclose um our use of AI and a retention agreement. All of these types of things, they're going to be they're not going to be helpful really unless they're reduced to writing. So that's what we would suggest that you consider in developing your firm's AI policy.

>> So we have covered a lot of information in a very short amount of time. We're very much aware of it. So I'm going to do a really quick recap and give a few recommendations very quickly and then we'll move on to some questions. So first of all, there are plethora of tools and resources out there that you can turn to if you have questions about AI. State bar associations, they have guidelines, advisory opinions, or you call into their ethics hotlines. You can turn to the ABA, consult the model rules, look at opinion 512, or review ethics opinions. You could also turn to external ethics council and case law.

So just recap of key points. So, you must must must must regulate AI's use within your firm by having a policy. Um, you also want to make sure that client confidential data is not being put into tools without you knowing what type of safeguards are in place and without those safeguards being adequate. Human oversight is always required. Again, a lot of hallucinations happen, especially when people are filing things with the courts and in case law. So, be careful there. Disclosure to clients. You need to know when it is mandatory. Um so AI use must be disclosed according to the ABA and engagement letters if it impacts your legal fees, case strategy or decision-making. So in that case, make sure that you're putting that in your engagement letter. As far as fees are concerned, you want to make sure that your billing is transparent. You can only bill for the time spent making the work product. And if you have flat fees, you may need to revisit those. As Jeff mentioned, AI policies and regularly regularly training your staff is going to be critical to upholding the policy and making sure that everybody's compliant. We recommend having quarterly audits um or quarterly training, I'm sorry, with your attorneys and staff so that they could stay up to date with all of the new advancements in those AI tools.

Okay, I'm going to leave you with one final thing. So again, AI tool is not a replacement for legal judgment. Thank God we're still all needed. Um, and a strong AI policy is absolutely 100% necessary to ensure ethical, secure, and effective AI use and legal practice. So, if you haven't developed one yet, I highly recommend doing so. All right, guys. Thank you so much for watching. Um, we are going to open it up to questions and answers, but if you think of anything or have any questions after we get off, um, I did put our contact information on this slide, so you can feel free to reach out to us via email or at the phone numbers listed.

>> Wonderful. Thank you so much, Desiree and Jeff, for that wonderful presentation. I'll go ahead and guide us into questions. As Desiree mentioned, you can always reach them via the emails that were on the screen previously. If you have questions today, you can also put them in the Q&A box. Um, Desire, Jeff, did you want to go ahead and take those questions?

>> Let's take a look. So, I'm gonna go and I'll read out what the questions are so everybody's aware. So, first question says, does a rule 5.1 and 5.3 cover inaccurate cases submitted by AI? For example, a case that is close to the real one but lacks a volume number and pages or other info, not a fake case, but a false one. Okay. So, in that case, yes. um you have a duty for candor to the tribunal too. So even if it's misrepresenting facts about that case, even if it's like a real case, but it's it's it's swapping out the law in those cases, it could be a violation, too. Um so you have to be absolutely careful about that. And if you're in a supervisory role, that's where 5.1 and 5.3 get come into play. So if you have a junior attorney that is using AI that's misrepresenting the law and you don't catch that, you could be on the hook.

>> Jeff, do you want to add anything to that one? No, no, I think he said it exactly. Um, a couple other questions there. Many one next question. Many states do not have comprehensive AI laws. What if they refuse to disclose the reliance on AI? It has errors. So, what do you think about that, Desire?

>> I think about that. Um, so that does happen a lot of times, especially since um, a lot of jurisdictions are just now making their policies or have been over the last like year or so. Um, so as that's happened, what we've seen a lot of times is actually courts are requiring that they disclose whenever you make any type of filing. So we're seeing that more prevalently. Um, but even where there isn't a requirement to disclose. I have heard about from my my uh my buddies uh that they've had cases where people have submitted things that were inaccurate. They didn't say anything about use of AI, but then they went in motion for sanctions um rule 11 sanctions based on them misreporting the law. and they infer that it's because they're using AI even though they haven't admitted to that. And then from there it you know they investigate or they'll go back to the other party. So um all that to say you definitely definitely want to stay away from um any issue like that if you can.

>> Okay, good. Uh so I'll answer this next question. What contract terms should firms demand from AI vendors to protect confidentiality and IP? So that's a really good question. So as as we talked about the the main things that you want want to understand from a vendor for a specific AI platform is how exactly is the information being used and so that gets into a few different things. Is the platform going to be learning? Can I turn off learning? If I turn off learning, can I be assured that it's not going to be out there? information that I put into the AI platform like for instance chat GPT the general um data or the general uh preservation is 25 days uh or 30 days I think after information is put in so ask the vendor what happens then does this information is this information like uh gone through and somehow uh electronically purged so it's not there anymore um where are the um where is the actual data being uh kept is it on servers here in the US or is it on servers uh someplace else. So these and other questions related questions are really the things that you need to understand before agreeing to uh to adopt an AI platform.

>> Great. Let's see the next question. In your experience, have you found legal based AI platforms significantly help with statutory interpretation?

>> Jeff, do you want to answer or do you want me to go for that one?

>> Yeah, go no. Go ahead, Desiree.

>> So, I found it's kind of so so to be honest with you and I haven't tried all the products out there like I've done Westlaw or Thompson Reuters um AI tools. Uh honestly, didn't have the best output with that, but all these tools are also being updated at the speed of light. So, things may have changed since I played with it. I do like Torch, like I said, that free tool that you can add on to the extension. It does help you get to a summary and at least get quickly to it. But, I think as far as statutory interpretation, you still have to rely on your legal judgment. And I think that's where it's really lacking and a lot of these tools are lacking. Um because I'll even put in specific case law and query about things that I know because I've already read the case and it will come back and misrepresent some of those things to me and misrepresent how the law actually is and I'll have to go and correct them. So be very careful there.

>> And you also just had a had a recent experience where you're writing a a pretty comprehensive paper where you said that there were a lot of citations and things that were just completely wrong.

>> Yeah. So I um I don't ever ask like when I go in and use AI, I'm never asking it to go find these resources for me. I like to provide the resources that I'm using and do the research first by myself and feed it into the tool and really get help that way. Um I did that. I was writing a legal article for a client um and working on it. I had a bunch of citations to the statutes and there um I hadn't been footnoting as I was going along. I was just making notes and so I was like, "Huh, I wonder how AI can do if I ask it to footnote just this one section that's based on just this one statute, right?" Um, out of the 10 um citations or footnotes it gave me, only one was correct. The rest were were completely inaccurate. So, just be very careful there. That's why I really caution you to make sure that you're reviewing this stuff because I use these tools maybe on a daily basis and I constantly see issues. So, you want to be very careful. Um, let's see. Can you give an example of how you would conduct an AI audit at your firm?

>> Yeah, sure. I'll I'll be happy to take that one. So, the AI once you have the AI policy in place, it's going to have things like the supervisory uh responsibilities. It's going to have the platforms that are going to be approved. It's going to have the different uses for each of the platforms like we talked about earlier that there may be some platforms that are use specific. Um and then as well as things like retention letters, uh fees, um probably again what we would suggest is you don't have to go and audit everybody at the firm, especially if you have a large firm that may not be that practical. But what you could do is you can take maybe a subset of lawyers, a subset of uh staff members, and then the the different things for the audit are going to be dependent upon whether they are an attorney or a supervising attorney or a staff member. If they're a supervising attorney, they're probably also going to be ones that are going to be sending out the retention letters. So, you're going to like say like what processes are you doing? You actually make sure that you audit or I'm sorry, review um any AI pleadings that have been created by junior attorneys. Let's take a look at your retention letters that you're you're sending out. are you including the specific clauses that need to be included to um provide informed consent by clients for your firm's use of AI? So, um it just at a general I would suggest like doing those kind of things. I think that the main thing sometimes with audits is it's not necessarily the the big thing with respect to who you're auditing, but it's the fact that you're doing an audit in the first place. And so when you're doing the audit, it kind of gives I think an impetus for other attorneys and staff members to say, "Okay, yeah, I really need to make sure that I'm complying with our policy."

>> And two things to that really quick, too. Just a couple of examples of this. So, a way that you could audit use. So, you could have your IT department if they're tracking everybody's use on their computers, what programs they have up, you could have um messages or notifications that will alert you when they go to certain websites or AI tools that are not on the approved list. Um, and so that way you're aware when people are accessing those um, and you can address that. Another thing I mentioned earlier was that the data sets can be discriminatory or have bias built into them. Um, for example, environmental law, say that you have a large language model and you put a bunch of data from um, a few different areas into it, right? Um, if that data was only from affluent communities that could afford to do studies and they don't account for the less affluent communities that have different needs, then what happens is that data output ends up being just biased and only accommodates for those people that the data is on. So, you want to be going in there reviewing this regularly. What data, who does this cover, what areas does it cover? Make sure that you're really, really getting um information in that's objective and not biased in those ways.

Okay, another quick question about uh supervisory liability and what what what oversight should you do? Um Desire, do you want to answer that question?

>> Yeah, you are absolutely on the hook for anything just like you normally would be um if you're filing anything with a court. So, candor to the tribunal, you have to have accuracy in your filing. So if something's inaccurate, you could be on the hook even if it's done by a junior attorney, even though they have been the one that um did the the AI usage, it's still going to put you on the hook. Um one of the most famous cases was the MA case. That was one of the first cases that actually handled this on AI. And in that case, a junior attorney actually um this guy had been working for 20 years, so he wasn't like a brand new attorney either. He wrote the the pleading, then he gave it to a senior attorney in a different jurisdiction who just submitted it. Um both of them were very experienced attorneys. both ended up getting in trouble. So, it's just to say like if you're submitting anything, um, you need to make sure that it doesn't misrepresent the law, doesn't make up case law at the core of everything just like you normally would do.

>> Okay. Last question is whether we have examples of of AI policy templates. We don't necessarily have any examples of an AI policy template. These are things that Desiree and I are working on for clients. um we can if you take I think the steps that we've uh presented in in the presentation I think those types of things should give you a really good overview in terms of you know a a starting point. Um we'd be happy to talk to you if you want about maybe developing a specific uh AI policy for your firm. Um the reason is that it's not just like you know other things in in the legal world um templates don't necessarily work for everything. So a lot of times there's certain customization that needs to be done that wants to be done. Some firms might want to have more obligation for their their AI policy committee. Some firms may want to have less. So really it's kind of a a work in process. You could take a lot of the things that we talk about as a starting point and then you can make it more concrete in terms of how you know how often the AI policy committee is going to be meeting uh how the audits are going to be taking place how the decisions are going to be made about what platforms are approved and the specific uses for those platforms.

>> Awesome. Well, thank you both so much for all of your additional insights and answering our audience questions. We really appreciate it. Um, and thank you all for submitting your questions. We really appreciate your participation as well. We want to thank you all for joining us today for another Justia webinar. As I mentioned at the top of the webinar, if you'd like to upgrade to Justia Connect Pro, it not only gives you an enhanced legal presence, it also gives you access to CLE credit opportunities through live and on demand webinars just like this one, so you can stay informed and compliant. Plus, you get an enhanced directory profile, AI powered Justia chat to boost your leads, and exclusive discounts on Justia Marketing Solutions. Take advantage of these benefits by going to pro.justia.com. Thank you all so much for joining and have a wonderful rest of your day. Thanks for watching. We hope you enjoyed this video and if you did enjoy it, please click the like button and subscribe to our channel for more videos on law practice and legal marketing. See you in our next video. [Music]