AI in HOAs - Powerful Tool or Legal Time Bomb

Artificial Intelligence is already inside community associations.

---- This video is not eligible for CEUs ----

YouTube Video: https://youtu.be/kRXYzKmLAzc

Managers are using it to draft emails.
Boards are asking it to interpret bylaws.
Vendors are embedding it into software platforms.

But here’s the real question:  Are we using AI wisely — or creating new liability without realizing it?

In this livestream, we’re breaking it down the way we always do — practical, real-world, and balanced.

We’ll look at:

• How AI can dramatically improve communication and efficiency
• What happens when boards rely on AI instead of legal counsel
• Confidentiality concerns most managers haven’t thought about
• Deepfake fraud, cyber exposure, and vendor accountability
• Whether associations should adopt an AI usage policy

This is not a tech lecture.

It’s a candid discussion about how managers and boards can use AI responsibly — without turning a helpful tool into a silent liability.

If you manage or serve on a board, this conversation is no longer optional.

AI is here. The question is whether we understand the risks that come with it.

Panel:

Cameron Leyh, CMCA, AMS • Ravenel Associates, Inc. • cleyh@ravenelassociates.com
www.ravenelassociates.com

Raymond Dickey • AssociationHelpNow.com

Dawn Becker-Durnin, CIRMS • Acrisure • dbecker-durnin@acrisure.com • www.HOAInsuranceSC.com

Valerie Garcia Giovanoli, Esq. • McCabe, Trotter & Beverly, P.C. • valerie.giovanoli@mccabetrotter.com • www.mccabetrotter.com

Content does not represent professional advice.

Transcript

SPEAKER_00 0:00

Everybody, thanks for being here. I'm gonna let the video catch up to the audio here. It only takes about five seconds, so I think we're there already. Powerful tool or legal time bomb. Who knows? We'll find out today, I hope. Before I have the panel introduce themselves, use questions, make comments. I've been doing the CEUs a little different because a lot of these are being posted afterwards for CEUs. And we also have been converting them over to podcasts, which people like. If you are here for CEUs, use your question feature or comments, whatever which one you have, and type in I need CEUs. That's how I'll know you need CEUs. Just type in I need CEUs. And that's how we'll basically work the CEUs going forward. With that being said, I'm gonna have our panel introduce themselves with two. Well, I'll start with Cameron.

SPEAKER_02 0:56

Hey everybody, thanks for being here. I'm Cameron Lay. I am the senior manager and director of training and development for Ravenel Associates.

SPEAKER_00 1:05

And thank you, Ravenel, for inviting us. It's much, much appreciated.

SPEAKER_04 1:10

Thanks for being here.

SPEAKER_00 1:12

Uh Valerie.

SPEAKER_04 1:14

Hey guys, I'm Valerie Garcia Giovanoli. I am an attorney at McCabe Trotter and Beverly. We are a full-service law firm for community associations. My office is in Columbia, where our main headquarters is located, but we also have an office in Charleston, specifically Mount Pleasant. Happy to be here and thanks to Ravenel.

SPEAKER_00 1:34

Hi, Dawn.

SPEAKER_03 1:36

Hi, everybody. Thank you again for having us. So I am the practice leader for our community association division within AcroSure, and I specialize primarily and only in community association insurance. That's not just property and casualty, but risk management, benefits, all sorts of neat and interesting things, which is why I'm here today. Very excited to talk about this topic.

SPEAKER_00 2:00

Please, audience, use comments, use questions. I don't use anybody's names, so you should feel free to ask whatever you like, and I will relay it. I could make a mistake, so maybe you don't send me anything. Too top secret. I thought I would get this out of the way. I asked ChatGPT, are you here to take over the world and rule the human race? Because I know Dawn is very concerned about that. I don't think we need to worry because ChatGPT told me, no, I only want to do stuff that you don't want to do. I don't want to take over the human race. So, Cameron, do you feel a little bit better about the future? I feel very relieved. If you believe ChatGPT, maybe it's an elaborate. I don't know if it's being honest. But I think this is really what we're talking about here today, right? To do the stuff that we don't want to do as humans for the most part. So let's start off. Um, are we using AI wisely or are we quietly creating a new liability? Valerie, um let's just touch base a little bit on this because I think this is a very critical point.

SPEAKER_04 3:04

Yeah, I would say it depends on who we're talking about, but I do see the use uh by many professionals in our industry as a liability. Um, and I say that because all too often clients have um decided to save on maybe uh fees that they may be charged by an accountant or an attorney or some other professional, maybe even an engineer, by resorting to Chat GPT and saying, hey, this is the answer. Chat GPT said it. Um and to go to liability, you know, the business adjudgment rule exists in South Carolina, um, and that protects directors from any personal liability, but they have to exercise due diligence. And I can guarantee you um uh consulting with AI is not uh considered due diligence because of the flaws that are inherent in AI. So I do think there's a liability out there, but I do also think many people are using it very prudently and wisely.

SPEAKER_00 4:01

And we're gonna touch upon that audience in detail. So we're gonna actually give you a lot of examples. So we're you're gonna get all the info I think you need in regard to that. Um, Cameron, I use AI all the time. I don't use it for my legal advice, I don't use it for my insurance advice, but I use it to help me research topics, sometimes generate questions. It's a super valuable tool. I think it's definitely here to stay, and I think people need to learn about it. Do you agree?

SPEAKER_02 4:28

I absolutely agree. It has a lot of pros. There are you know cons with anything, but um, I'm trying to use it more and more. Some of the managers in our office are better at using it than I am, and I'm trying to use it as much as I can and and learn it.

SPEAKER_00 4:45

And we're gonna talk about today on ways you can use it also, which will be really valuable to people if they're not familiar with it. Um, you definitely don't need to be scared of it though. Do you agree, Dawn? And then it it works pretty pretty easily, correct? I think it's pretty simple.

SPEAKER_03 5:00

In fact, I think you can the best thing about it is it prompts you to get a little bit more defined in the responses.

SPEAKER_00 5:09

A couple of quick comments I'd like to pull out. Let AI create your budget, but then it's gonna have to create your apology email. So we'll talk about that. You may need to jatGBT on the witness stand if you don't use it appropriately. And the last one is you can't lose the algorithm says so. Valerie, none of these three are gonna work really good in court, I don't think.

SPEAKER_04 5:33

Absolutely not.

SPEAKER_00 5:36

Comment, have you ever, whoops, have you ever used AI for association related work? Audience, please. I would love to hear your comments on that. Like, I don't um relay people's names, so feel free to be honest. Have you ever used AI for association-related work? And if you want to add on to that what you used it for, that would be great. All right, so I'm gonna start now. Who does use AI? Valerie, you said you're very familiar with it.

SPEAKER_04 6:02

Yes, absolutely. We've been more in a testing phase, and um we've created policies because with the law field, we are held to a confidentiality standard for all clients. So it's very um restricted on what we can actually put into AI. And so it's important for us, understanding that maybe staff and other attorneys might be meddling with it for everyone to understand what the limitations are for us as attorneys and a law firm. Um, but yeah, my my take, and I'm gonna talk more about this throughout the segment, is um it sort of is more of um an assistant type tool. The way I envision it is the lawyer has to do the deep think. I would not rely on Chat GPT to do that, not tell you what the laws are, what the outcome is, what an opinion. It's wrong about 90% of the time. If you use it in that capacity, it's good to just kind of maybe give you other ideas, but the deep think still needs to be done by the people, the humans. Um, I think it it it it like a secretary who's gonna type up something, that would be a great use of that tool. Um, maybe proofread something that the attorney has drafted to help catch those uh pesky errors. Those are um sort of some uses that we see as beneficial at this point.

SPEAKER_00 7:16

You're getting so far ahead.

SPEAKER_04 7:18

I know. I'm sorry.

SPEAKER_00 7:20

You know, you know, we have to keep people on the hook, right? For information later. Right, but you can't give them the whole story in the beginning. Go ahead, Dawn, but let's know where you're going, right?

SPEAKER_03 7:31

We don't get a pre um ask to know where you're going.

SPEAKER_00 7:36

No, and just so the audience knows, I don't like to do that because it's much more fresh, but it's it puts pressure on the panelists also, and they're good sports about it, but it's a lot better of a program when we learn things. I like to learn as we go along. Just some quick comments because I'm gonna keep up with them. People use them for the minutes, people use them to help with their emails. Um, I recently had an attorney give advice. I put the advice at Chat GPT, and the AI said the advice was not good, and it gave me any reasons why. Should I go back and question the attorney? Oh, Dawn. Okay, go ahead.

SPEAKER_03 8:08

I want to finish the question, but what oh, I'm sorry, I apologize. Go ahead.

SPEAKER_00 8:12

Okay. Recently I had an attorney give advice. I put the advice, don't get so excited, audience. She loves insurance. I she loves it. She gets so excited, she can't wait. All right. Recently, I had an attorney give advice. I put the advice into Chat GBT, and it said that the advice was not good. And it gave me many reasons why. Should I go back and question the attorney, or would the attorney not like that? I am gonna start with Valerie on this.

SPEAKER_04 8:40

Well, number one, I don't think it matters what the attorney likes. Absolutely, you should ask the questions. They could be very wrong, and the lawyer can explain why. But I see that as no different as, you know, you give a client advice, they go and talk to their neighbor who has a cousin who has a wife who's an attorney, and they say something different. I think it's certainly advantageous for the end user to go and communicate with the lawyer about, you know, these are some other perspectives or things that I've heard. Can you explain this to me? Um, and I think the lawyer would be happy to do that. I don't think they should be offended. And if they are, they that's a whole other problem.

SPEAKER_00 9:13

You know, Valerie, I think this you should definitely mention it because you want to mention it in the appropriate manner, though, right? Hey, you don't want to go, you're a moron, and Chat GPT told me something. But I feel like this is a good way to double check that you're communicating properly to the attorney. Maybe you're not. Can you run with it?

SPEAKER_04 9:28

Yeah, I would just say, well, I would two things. I would say that I you don't necessarily have to tell the attorney you got it from AI. You just say that this is some things that I have heard or read. Um, I think they'll gather if you copy and paste, you usually can pick up that it's AI. Um, but one other thing that I did want to say is they said that they put a legal opinion into AI. And I do want to forewarn that if that is not a private AI system, that you have thus disclosed confidential information. So in in in South Carolina and most states, the confidential confidentiality privilege is the client's, but they can waive it. And by disclosing an attorney's communication uh to third parties or the public, which is what you're doing with a free version of Chat GBT, you're waiving your privilege to that. And so, if worst case, I don't know the circumstances, if there was a lawsuit and that information was requested in discovery, you couldn't claim privilege because you've already disclosed it. So do be careful with what you are um putting into, especially the public versions of uh of AI, because you could potentially be waiving attorney client privilege.

SPEAKER_00 10:31

Can you do me a favor, Valerie? When we get to the confidentiality section, because I have a whole discussion about that. Can you you keep mentioning the paid version? I don't want to get into it right now, the paid version compared to non-paid version. I'm not familiar with that. So if you could kind of make a note on that and make sure you explain that because I'm I'm not aware of it. Dawn, what did you want to say before I interrupted you? Sorry.

SPEAKER_03 10:49

No, I I was gonna actually comment about how when you're saying easy setup, we're finding that AI is involved in not just you know the chat GPT, but it's a function in a lot of programs that people are using right now, including internal software.

SPEAKER_00 11:05

Right. And today we're gonna concentrate on mostly though managers using and board members using AI. Um, correct, Cameron, that's the direction that we we want to talk about today, right?

unknown 11:15

Yeah.

SPEAKER_00 11:16

Um accuracy. I'm gonna come out with a statement, Cameron. If there's no way it's dead nuts accurate. Um I use it a lot, I find a lot of things pop up, and it actually doesn't even tell you when it's not accurate. It doesn't like to admit it's wrong. But go ahead, Karen.

unknown 11:31

Right.

SPEAKER_02 11:32

Yeah, I mean, I've I've put some things into I always think you have to review. Um, I put I put a an engineer's letter in there to the other day to try to write a letter to a contractor and there was something that was blatantly wrong. Um, you know, they the the temperature of something was just totally off. Um I think you have to be aware that it does have limitations. Um like anything, really.

SPEAKER_00 12:01

You know, Valerie, what it what scares me is that when it makes mistakes though, it doesn't tell you that it's not sure. Have you found it's the it's the most confident person you've ever met in your life?

SPEAKER_04 12:13

Yeah, I hate that about it. Um, I would like for it to just say, I'm not quite sure, or turn on the deep thing so I can do deeper thinking. Um, but when you do correct it, it does take criticism like a champ. It's like, oh, you're right. I was so wrong.

SPEAKER_00 12:29

It does. I did ask it one time, how many times could I insult you? Would you ever be angry or ask you the right question? It said indefinitely. You could never hurt my feelings. Don?

SPEAKER_04 12:39

I don't have feelings.

SPEAKER_00 12:41

How about you? Um, yeah, it's wrong a lot, right?

SPEAKER_04 12:46

It is. Um I'm sorry, that was for Dawn.

SPEAKER_03 12:50

No, I was just saying before we started the program that I put in a CC and R as a master D and it specifically misquoted sections. And I went back and said, Don't you mean article 10 instead of article nine? And it went back and it said, sorry about that, and then it came up with a different article, even though I told it was wrong. And I actually pointed to the article that it was located in, and it still couldn't get it.

SPEAKER_00 13:15

What's interesting knows it's learning, right? Valerie, uh, it you're actually teaching it. Like my ChatGPT, I've used it so much since day one on the paid version, it knows a lot about me. Um, Valerie, do you agree?

SPEAKER_04 13:27

Yeah, absolutely. And I I always made a joke when we first started kind of testing it out just to see how we could integrate it into this uh firm to make us more efficient. Um, I felt like this bet was the sole reason it was existed because it was wrong about everything. And I'm sitting there telling it, no, look at this statute, no, look at this case, and I'm like, I'm teaching this thing, and I got a little frustrated um because it didn't seem to be helping me with efficiency, but it's definitely learning.

SPEAKER_00 13:53

All right, so let's go through some of the ones that came in from the audience and let's just I love the samples, they're they're a good way for people to learn. Um, and some of them I don't know the answer to. Minutes. Um, someone wrote in, I have it right the minutes. Um, Cameron, I'm guessing it's okay to maybe check the grammar of the minutes, but not the content, because you can go you can download transcript transcripts into it, but go ahead.

SPEAKER_02 14:17

Yeah, I mean, I usually I don't I don't like the way it spits out minutes. I feel like I have to clean them up. I feel like it puts more stuff in there than I would normally put into my minutes. Like minutes are not supposed to be a transcript. Um so it frustrates me a little bit when it comes to taking minutes. Um but I know a lot of our managers do love it for minutes. So it's a personal preference, I think.

SPEAKER_00 14:43

Dawn, I tend not to use it for minutes. I mean, I don't have to do as many minutes though, I think, as the average manager does. But what do you think? And we'll go to Valerie for the legal.

SPEAKER_03 14:52

Well, I like it after a meeting, especially a board meeting that is perhaps online, and I want to summarize what information we've been discussed.

SPEAKER_00 15:01

Oh, so use it for some to summarize things. Yeah, I find it to be excellent for that. If you you just gotta make sure it may not go ahead, Cameron.

SPEAKER_02 15:09

It's good for like action list items. Um, you know, Cameron's gonna do this, John's gonna do this, Courtney's gonna do this. Um, but uh that's that's not really the format of of minutes.

SPEAKER_00 15:22

Um you tell it something about each person, Cameron. Like you do you like, okay. I'll just pick like John Doe's really annoying Chat GPT. Don't give them that task. I don't want to be anywhere near him at the next event. Do you share like that stuff with ChatGPT, Cameron?

SPEAKER_02 15:38

No, I do not. It would be funny. Um maybe I'll try that.

SPEAKER_00 15:41

Okay. Um, Valerie, I mean, how about minutes? Where's how do you the legality of it?

SPEAKER_04 15:48

I see nothing wrong with allowing Chat GPT or other AI softwares to prepare your minutes. Again, I'd go back to it needs a level of human oversight and input. So you're inputting the information, it's essentially organizing again, using that typist example, a secretary who's going to type it up for you. Um, and then you're gonna finally review it before it gets approved and deemed the uh minutes for the meeting. The one thing I would heed caution on is um, and this goes to just keeping minutes generally, um, executive session. My instruction is typically not to include any minutes with regard to conversations held in executive session. That's the purpose of executive session, to keep it confidential. And so you don't want to um uh put that into a public AI software, nor do you want that in your minutes anyway. So that would be the one area I would be cautious with in using AI for meeting minutes.

SPEAKER_00 16:40

Valerie, can I ask you to hold another question if you remember, or maybe the audience can remember too. I want to know if this is how about this a discovery for the future? So I'll ask that when we get to the confidentiality section because that's another thing I never thought about. Dawn, the um the next one is somebody mentioned about writing emails. I'm thinking, like, sure. I mean, I use it to double check my emails and my grammar. How do you feel about that?

SPEAKER_03 17:03

You know, I still think you need to give it a look over because different punctuation can provide different emphasis and perhaps change how you wanted the email to actually portray itself, the tone, if that makes sense. I also think that the accuracy, depending upon the subject matter, needs to also be reviewed because we've seen that the AI can go in different directions depending upon how it feels the information should be portrayed. So accuracy is extremely important, but it's hard to double check sometimes when you are treating it almost like a Google search engine at the same time.

SPEAKER_00 17:45

The way I use it, it's great. I don't know if you guys find this. It asks me, do I want it to be very casual? Do I want it to be formal? It gives me a bunch of options to pick on how I want to set the tone of the email. Um, Valerie, do you also get those prompts or is it just me?

SPEAKER_04 18:00

Yes, I do. Um, and I think we're gonna talk. I don't want to jump ahead, um, but I think we're gonna talk about you know writing emails with AI and our our input on that, right? And later.

SPEAKER_00 18:11

I think so. Just make a note of it. And and if we didn't, I'll double check. I don't know if I went that deep into particularly.

SPEAKER_04 18:17

Yeah, no, that's a great option where it says, you know, even with uh uh like QuickBooks invoicing, I'll say, do you want to send a stern follow-up on this AR friendly? And you know, give me some more background and I'll draft it accordingly. I do think that is a pretty neat tool uh with an AI.

SPEAKER_00 18:32

What do you write? Go get them, make a miserable chat GBT. I hate them. Uh okay, let's see. The other one was this, you know, three people sent this. This one makes me nervous. I think it's gonna be a big disclaimer. RFPs. Cameron, yeah, man, managers, RFPs take a lot of work. I mean, how do you feel about RFPs?

SPEAKER_02 18:53

I have used it to write RFPs. I've used, you know, an RFP from maybe a different project, a different property, and given it prompts to edit it. Gosh, it is a huge time saver. Um, again, it does you can't just accept that that is good to go. You've got to review it, but it yeah, it saves a lot of time.

SPEAKER_00 19:17

RFPs, um, Valerie.

SPEAKER_04 19:20

I I don't have anything else to add. That's not something with very much confidentiality issues. Um, and I would I agree with Cameron, it's what it's very good at organizing that type of stuff and saving you time typing it yourself. Again, you got to give it the inputs though, and do the review.

SPEAKER_00 19:36

Yeah, I think what I was nervous about, Dawn, is you really got to review that really carefully because you're sending it out to get quotes. And I just don't know if I would count on it a hundred percent. I mean, it could easily write a hundred thousand square feet instead of like 10,000 square feet.

SPEAKER_03 19:50

Dawn, do you agree at all or I do agree because insurance, we use AI for a lot of information relative to underwriting, and a mistake can wind up in a real pickle in in our industry at least. And I can imagine putting out wrong information in RFP or RFQ could be an issue as well.

SPEAKER_00 20:11

But man, it will save you like it'll it'll do it in 10 seconds, but would take you an hour. Um to agree, Cameron. The time saving is huge, it's phenomenal. So um somebody mentioned it's all about the prompts. Yes, that's we that's how you teach um chat GPT prompts. Chat GPT knows a lot about me. I was joking before we went live. Um, the minutes, are you saying you allow ChatGPT to hear the board meeting so it can compile minutes? I think there's a couple different different versions of this. If the board meetings recorded, you can download a transcript. I think most people probably have it correct. It's its own inputted notes. Valerie, this is going to be a big one back to confidentiality again, right?

SPEAKER_04 20:52

Absolutely. Um, I think that again, when we go to what software are you using, you need to look into whether or not it's open to the public versus private. There are options for private, uh, but in either event, um, a transcript, it can arguably become part of the association's records. Remember, in minutes, we're we're not transcribing everything everyone said. You don't even have to say, but you know, this person took this position. And so that document could potentially be disclosed in litigation through discovery, um, whereas the minutes that are ultimately produced from that document, maybe that's the objective, um, or could protect uh, you know, some things that were said in a meeting uh better than a transcript. So I would keep caution with that. Um if you are going to use the transcript uh option in your you know live stream virtual type meetings, um, then I would dispose of it because it's not a record that has to be kept as part of the association's records. And so you don't open that door for a potential litigant to argue that. It is a record that needs to be disclosed.

SPEAKER_00 22:03

I use it for my live streams to convert. If I want to have an article about this live stream, I will use it. But what I don't allow it to do is quote people. I have a huge disclaimer because I don't I can't remember and I don't feel confident allowing ChatGPT to say that Valerie made a statement. So what I'll do is I'll have ChatGPT just summarize the article and I have a huge disclaimer at the end that says, nobody on the panel looked at this. It may be complete garbage. Take whatever you want from it, but double check the video for yourself. So I do not trust it 100% for transcripts. It does make mistakes. And for me, I think it would be I would feel awful if it misquoted someone on the panel. Um board meetings, board meetings, Zoom meetings. Yeah, I could transcript all this kind of stuff. All right, let me throw some quick ones at the panel. Rewriting tense emails. Anyone have a problem with this? Yes, no, or maybe. An audience, why don't you chirp in here too if you want to? I will start with Cameron. Yes, no, or maybe.

SPEAKER_03 23:03

Yes.

SPEAKER_00 23:04

Dawn?

SPEAKER_03 23:06

I use it all the time. I try and talk myself off the ledge.

SPEAKER_00 23:10

I know I I use it too. Sometimes I'm horrified by it tells me to calm down. Valerie.

SPEAKER_04 23:16

Um personally, no, but as a um practice uh in a law firm, maybe, depending on to whom the email is addressed.

SPEAKER_00 23:25

Well, Valerie, you're good at writing emails though, like that. So you do them right, you somehow you do it in a nice manner, but terrifying at the same time because you're an attorney. So um drafting violation letters. Okay, I'm gonna be curious. Uh Cameron, if you don't have an opinion, I don't want to put everybody spot, but how do you feel about this one?

SPEAKER_02 23:43

Well, we we have a kind of standard violation letter, so we don't really need to use that for violation letters, but I do use it, and maybe you're gonna get to this. I use it to to craft um architectural review letters in some instances.

SPEAKER_03 24:02

No, typically not, because our letters are gonna come from the carriers now, whether they're using AI. Our system, to be clear, is a closed AI system that we are using internally in most carriers as well. But I would say anything that has to uh be it could be used in court, we'd want to really make sure we're reviewing it and putting the information in versus uh an AI software.

SPEAKER_00 24:28

I don't think I would trust it for this one, Valerie. What do you think? Well, you would know.

SPEAKER_04 24:32

I think uh for some type of form letter, like Cameron mentioned, they already have their form. I think it would be a great starting point to get something together. I also think that um because our practice is so specific to um governing documents, that it may be beneficial to have a form for each client because some may have some very long violation process that has a lot of different things you have to address in a violation letter, whereas another community it's just sounds like overkill. So those documents, the governing documents are already a public record. Uh, you could input those into your AI software and ask it to generate a standard form violation letter to comply with any of the requirements in the governing documents. And then, of course, look at that, make sure everything is sound and it it lines up with the governing documents you've proposed, and then have a form letter. And then obviously you can tweak it per uh case by case depending on to whom you're sending it.

SPEAKER_00 25:32

I typed in chat GPT where you were while you I was listening to you, and I wrote, can I use you to write a violation letter? And it wrote, Absolutely, you could use me to write any HOA or condo violation letter, and I could tailor it, blah, blah, blah. Yeah, that's why you can't trust ChatBT. So uh let's see. Translating communications for diverse communities. This is like in regard to discrimination, maybe, or being tone deaf to other kinds of um communities that you may have and you may not realize it. I'm thinking, Cameron, on this, that this would be a good one. Why not? See what it says.

SPEAKER_02 26:11

Excuse me. Yeah, I definitely think um I haven't I haven't used it for that, um, but I I think that that's a great opportunity to use it.

SPEAKER_00 26:22

Don on Thursday we have a live stream. Modern woke issues board should not ignore until they explode. Um, Valerie's gonna be there, and so are you. I actually had a lot of fun researching this because I use ChatGPT and it taught me a lot. Um, how do you feel about using Chat GPT in this manner?

SPEAKER_03 26:42

Um, I think it's the issue more of looking at patterns for Chat GPT a lot. And so abnormal patterns can be an issue, right? And I worry about discriminatory practices, meaning it picks out something that it shouldn't or applies a certain practice um in it unintentionally.

SPEAKER_00 27:02

But Valley, this is a great way to use Chat GPT. You have a very diverse community. Why not throw it in there and see what it comes back to? You don't have to take the advice. How do you feel about it?

SPEAKER_04 27:12

I think I'm a little confused on this exact use. Are we saying that we're using it to basically create messages in other languages?

SPEAKER_00 27:20

No, I'm using it to maybe not install or or or make your, you know, it's really easy nowadays to offend people when you just don't have to. You know what I mean? I'm not talking about craziness with little slights, but why not? If you have this diverse community and you're not familiar with with you know a lot of the background, maybe run your letter through Chat GPT if it's a touchy subject and see if it maybe tells you a different way to write it.

SPEAKER_04 27:45

I think that it's just like anything, though, when it gives you the output and you're reviewing it, what are you gonna cross-check it? How do you know it's not telling you to say something extremely um um um non-politically correct or offensive? So I think that my concern there is that I'm always going to review any output from AI. Um, and if I don't know the answer, or I'm gonna go consult somebody that does. So that would give me some pause to use that and rely on it, especially because I don't have any information or experience to cross-check it against.

SPEAKER_00 28:16

All right, audience. So this is where you forgot everything I just said and you listened to what Valerie said. That's why Valerie's here. That's why I can't do these on my own. So this one, you you this is a cautionary one. People want to be careful about this one, right, Valerie?

SPEAKER_04 28:32

Yeah, absolutely.

SPEAKER_00 28:34

Thank you. Uh, summarizing long meeting minutes. I find it that it's awesome for this camera. It could take like a five-sentence paragraph and put it into one sentence for me, and it sounds great. Do you agree?

SPEAKER_02 28:45

I definitely agree. Super helpful for that.

SPEAKER_00 28:49

Anyone disagree with this on the panel? And the last one I have is AI can reduce burnout, improve your tone, and save hours when used as a drafting tool and not a decision maker. When used as a drafting tool, not a decision maker. Cameron, do you think that's going to be the big lesson of the day today?

SPEAKER_02 29:08

Yeah, I think that that really summarizes it. A drafting tool, but not, you know, we just keep saying, you know, check it, check it, check it. Don't just assume that what it spits out is is is the gospel, basically.

SPEAKER_00 29:22

Don, how do you feel about that?

SPEAKER_03 29:29

I agree. I certainly do, but I also, again, I'm a little hesitant in worrying from an insurance perspective about all those issues: bias, discrimination, copyright infringement, intellectual property, all the things that can save you time in the beginning, but later on, what's the cost as far as time, especially if you have to sit in court?

SPEAKER_00 29:49

You know, Cameron, before I go to Valerie, it's interesting. I'm noticing a trend here, right? So here I am, Ray Dickey, the guy that does live streams. And I I love it. Yeah, okay, use it for this. You're a little bit more open to using it. Then we go to Dawn, it decreases a little bit more. And by the time we get to Valerie, it's basically like, ah, maybe not so much. Right. Okay. So, Valerie, same same question. Um, how do you feel about using it as a drafting tool compared to a decision maker?

SPEAKER_04 30:13

I think it's great for that, but I would emphasize that uh subject to confidentiality, that you're not putting stuff in there that's going to potentially risk this disclosure. But I would say rough drafting tool, because the decision maker absolutely should not be a decision maker, but the decision to send that letter out, but it just spits out, you're letting it decide. So you need to decide what that final version looks like. So I would just emphasize a rough draft drafting tool.

SPEAKER_00 30:40

Um, I'm gonna go through these quick panel, just jump in if you feel uh a certain way about it that I'm off base. Creating first drafts of policies, we went over that. We just actually did the issues with it. Generating checklists, I would say, why not? Anybody on the panel have a problem with it? It's not gonna be your final checklist.

SPEAKER_03 30:58

I love checklists from that's one of the things I think is great, especially let's say hurricane preparedness. What are items that boards should be thinking of? Not only that, but checklists for safety, building inspection list. I think it's great, but again, you do have to review it and look at um your specific situation in a very um prominent light, meaning that can't use it as a broad stroke.

SPEAKER_00 31:27

I bet you do love checklists. You got a little excited there. You're a checklist person.

SPEAKER_03 31:31

I am a checklist person.

SPEAKER_00 31:32

Do you give your kids a checklist all the time to get things done?

SPEAKER_04 31:36

Let me I would I think it was gonna be a good time to throw in. I mentioned this to the panel before we got on, but I actually use an app. Um, it's called the operator standard, but it does help with professional and personal development. And what it does is it integrates an AI system that you input, you know, your goals or what you're trying to achieve, and it helps you come up with critical tasks and it gives you a task list that you can manually enter or use the AI feature to help you develop the task necessary to get to that goal. Love it, it's helped me so much tremendously in the last two months. So I just wanted to throw that out there because of the checklist topic.

SPEAKER_00 32:12

Chat GPT is a great motivator, it really is. I mean, you could type anything in there, um, and it's it'll never get tired of hearing from you, right? Like sometimes my wife, I've been married 34 years, she's tired of hearing from me, Cameron. ChatGPT, never tired, thinks I'm terrific. Every everything I say, it knows to tell me whether it's right or wrong. So, okay, preparing RFP comparison summaries. Cameron, I don't know. This one makes me nervous. I don't know. What do you think?

SPEAKER_02 32:40

You don't I I think that it's okay to use it for that.

SPEAKER_00 32:44

Okay. Uh go on.

SPEAKER_03 32:47

I'm a maybe. I really am. I think in a in a rudimentary way, yes, but I think you really have to double triple check because if it extrapolates the wrong information or has a bias, that may be a problem in your checklist.

SPEAKER_00 33:04

Valerie, I think this is all about like like I agree with Cameron that you can use it, but you just have to make sure that it's a draft and you're double checking it. That's right.

SPEAKER_04 33:13

I agree.

SPEAKER_00 33:15

I did ask ChatGBT, Valerie, and I wrote, I'm feeling down because Valerie is always right and I am wrong. And ChatGB told me, Ray, let me tell you something. This comes from watching you lead chapters, expos, retreats, live streams, and real world situations where 50 people at a time are pulling you in different directions. You're not always wrong. Thank you, ChatGBT. I'm not gonna get that from Michelle, my wife. That's great. I tell my wife she's gonna be like, ah, just tough it up, it's your job, you know, and you don't know as much as Valerie. That's that's wonderful. I love it, Cameron. I love it. Thank you, Chat GPT. Okay. Um Valerie, what would do you mind sharing the AA program that you were using for motivation? Or do you want to keep that?

SPEAKER_04 33:59

Yeah, operator standard. It's in the app store on your mobile device.

SPEAKER_00 34:04

Okay, you know about chat TP though, Valerie. It's gonna remember I asked that about you now, and somehow it's gonna like now I'm gonna be stuck with it until I tell it to get rid of it. Okay.

SPEAKER_03 34:13

No, what you're gonna see is everyone's smartphone has listened to this, and now you're gonna get advertisements for that application on your phone.

SPEAKER_00 34:21

It's true, it's gonna be scary.

SPEAKER_04 34:23

People are gonna I do need to uh I guess asterisk. It is a um a very new program, and I heard that it's like in testing phase and it might be closed to the public, but it'll reopen. But just keep your eyes open, peeled for it, operator standard. I don't know if they're accepting new members, but it's uh developing and it's an amazing tool, in my opinion.

SPEAKER_00 34:43

But you're putting a disclaimer don't let it don't let it run your life. Don't come back, don't come back to you if you don't like it. Okay. Here we go. Legal. Interpret. I'll start with Valerie on these. Interpret governing documents. I have used it to interpret complex paragraphs of legal mumbo jumbo, Valerie. I think it's done a good job, but I'm not relying on it.

SPEAKER_04 35:08

Yeah, I think that this is a maybe. Um I have no problem with allowing it to take a stab at it. It might take an interpretation that I haven't considered, um, but I would never rely on it. And I would go back to there's two points I guess I would make. One is um I don't see how it's any better than Adobe's OCR feature, which very often will miss a term that you're trying to search. Um, I don't know what other uh, because you upload it, it's gonna search it. It's using the same essential uh technology to pick up on terminology. And when I don't know if any of you guys use Adobe, but I've got the highest level of Adobe on our because we use uh PDF so much, that it sometimes will miss. You control F and look for you know uh fines and it misses it. So I would think that AI also could potentially miss relevant provisions. Um, in addition to that, I think that um, and I think Don will uh has a story about actually where it was citing the wrong sections. Um, and then the second part is it's not a lawyer, it doesn't know. Uh I would not rely on a legal opinion, which an interpretation of a governing document is just that. But again, I'm not against it. Governing documents are publicly recorded documents, you can upload them even if it's a public AI software. You can see what it says. Maybe it sparks some inspiration or some ideas or things that you want to look into, but it should not be the end-all be-all for an interpretation of any governing documents.

SPEAKER_00 36:34

Uh, Dawn.

SPEAKER_03 36:37

I look at this saying if I'm going to do this interpretation, but I'm not going to speak with the attorney, I think it puts managers and boards in a very precarious position. Because if they are wrong, their professional liability most likely will not provide them a defense or any type of indemnification. So I think it's fine for a simple overview and to kind of get a position to then speak with the attorney or other um professions to make sure you have an accurate result.

SPEAKER_00 37:10

Cameron, I use it like we just did a live stream on bankruptcy in Florida, and there was a one lawsuit. So in the past, I'd have to read the whole lawsuit and it would take me forever, and I knit it, I would never read it properly, right? I use ChatGPT and it did a really good job of summarizing it. But remember, I'm not the expert on my live streams. I can say anything I want because I have people like Valerie and Dawn and yourself in certain regards. How do you feel about using it? But I thought it was amazing, by the way.

SPEAKER_02 37:36

I think it can be helpful if you know you've got 70 pages and you don't even know where to start. It can help maybe narrow it down to the section and then you can read that section yourself. But I don't like what you put interpret because I feel like it's not our job to interpret the governing documents. That is an attorney's job. So I would maybe if you change that word to something other than interpret, I would I would be more inclined to say yes, or maybe, um, but I don't like interpret.

SPEAKER_00 38:10

Go ahead, go ahead.

SPEAKER_03 38:11

That's a great point. Interpret. And I think for finding something in the documents, that can be useful. And as Valerie pointed out, sometimes the find doesn't always work. But I get asked a lot of times, is this a common expense, the deductible? And South Carolina being the wild, wild west, every document is essentially different from each other. So we do have to read those documents. So to find a specific section, perhaps it's helpful. But to interpret as Cameron, no, I'm I'm I'm not on that.

SPEAKER_00 38:43

Um Valerie, did I I think we did I go to you on this?

SPEAKER_04 38:47

Yep. And I I agree wholeheartedly with Cameron. I think when I mentioned uh uh interpretation being sort of a legal opinion, I agree that that's something. I mean, obviously, if it says, hey, um, the floor is black, there there's no legal uh opinion necessary to interpret it, but we're usually talking about these ambiguous terms, which often exist in governing documents. And I do think that's within the purview of the association's attorney to provide that guidance and not AI. And I 100% agree with Don's point as well, where you're that's not exercising your fiduciary duty and um and and making an informed decision. You're relying on uh like essentially an upgraded version of Google to give you a legal opinion, it's not going to protect you from liability.

SPEAKER_00 39:29

It is amazing, though, how fast it works, right, Cameron? I meant you literally you throw a 70-page document in there and then you can ask it to fine-tune. It is pretty amazing. Do you agree?

SPEAKER_02 39:39

I meant I think it can be helpful to find, you know, specific things about rules and regulations or you know, a quorum requirement, or something, you know, small like that, that again is not really, you're not interpreting, you're just trying to find where it says something that's pretty black and white.

SPEAKER_00 39:58

I always feel like I use it in a manner that if if it's wrong, it's not going to have any major repercussions. But yeah, go ahead, Valerie.

SPEAKER_04 40:05

Well, Ray, you said uh, you know, putting a lawsuit in there and trying to summarize it. And um, while I wouldn't do that, obviously, as a uh defending or even being prosecuting a lawsuit, I think it could be helpful if I know what the lawsuit says. I've read it, I understand it, and then I want to put together a summary in layman's terms for a client. That could be a good way for me to save time and manually having to dictate that, or um obviously with my final review before it goes to a client. So that would be a good use uh for our practice.

SPEAKER_00 40:33

I find it amazing because I'll I'll keep asking it, dummy it down, dummy it down, dummy it down for me. You know, break it down to its most, but like I said, audience, I don't have to be right. I'm not an attorney, I'm not here to offer professional advice everyone else on the panel. So I have a lot more flexibility than Valerie would or Dawn Woodwood and Cameron also in her profession. Um, draft legal course, draft legal correspondence. I'll start with Valerie. I there's the word draft.

SPEAKER_04 40:59

Yeah, well, for the boards and managers, I would not be doing that. Just because you used AI doesn't mean you should be preparing legal correspondence. Uh, as far as from the law firm, I think again, uh AI is a good typist. And so as long as I'm giving it the inputs and the law and the decisions are made by an attorney licensed to practice, then I think it would work well uh as far as rough drafting anything, uh legal correspondence.

SPEAKER_00 41:26

Cameron.

SPEAKER_02 41:28

I don't know that this is really a question for managers because you know we're not drafting legal correspondence. That is the attorney's job.

SPEAKER_00 41:37

So Dawn, you're not gonna like this one, right?

SPEAKER_03 41:42

Yeah, I don't like that at all. I don't think that boards should be involved. I I will just give you an example. So I had some marine work being performed by a community, and I asked for a copy of the contract that they were going to enter in with the contractor, and they said, Yeah, here it is. It's already signed and taken care of. And I got it, and I said, Who wrote this? This is the worst contract I've ever seen in my life. And you essentially have given all of the you're indemnifying the contractor, you're holding them harmless. Who is the attorney who wrote this? I'd like to have a word with them. I got a phone call after the meeting, and it was the board president, and she said, Is this bad? And I said, Was what bad? And she said, We didn't use an attorney, we used Chat GPT and we compared it to other contracts and it had similar wordings, so we thought it was okay. And I said, please, please, I said, you know that they sign this, and this is now a legal document. I said, There's a you've just practiced law without a license.

SPEAKER_00 42:51

Boy, Valerie, I'll let you comment on it.

SPEAKER_04 42:54

I I think that that there's definitely an issue there, especially with regard to like obviously everyone can be pro se, but when you are a director um or a manager of an association, that pro se line is very blurred. So I would agree that you could potentially be risking uh a claim of unauthorized practice of law, which is illegal and in our state.

SPEAKER_00 43:18

Replacing attorney consultations, this would be disastrous. Does anyone on the panel think this is a good idea? Valerie, any additional commentary on that?

SPEAKER_04 43:29

You said replacing legal consultations.

SPEAKER_00 43:31

Oh, okay. Replacing attorney consultation if ChatGBG tells me it's all right.

SPEAKER_04 43:36

I mean Yeah, I think this goes back to uh I think the initial comment that I made using ChatGPT as sort of your professional, um, whether it be an accountant, an engineer, or a law or a lawyer, I would strongly advise against that.

unknown 43:50

All right.

SPEAKER_00 43:51

So, Don, I think this would be a good point. Um, and we're gonna go into some caution things here. Why don't you talk about the insurance aspects and really drive that home to people? Because that is gonna be Disaster.

SPEAKER_03 44:01

Well, again, the attorney should be providing the consultation. Unless we're going to see Valerie like this, you know, it's really doesn't replace a human. And also there's critical thinking that goes on in a consultation. And not all facts can always be determined ultimately from just putting in a one-sided biased opinion or request, meaning that an attorney will usually evaluate and ask more leading questions to determine what the appropriate scenario that should be applied from a legal standpoint. So, for example, it's not always black and white, it's gray, it could be dark blue, it could be navy. An attorney is going to use their critical thinking to find out specifically what angles to use, and of course, how to apply that legal thesis. They don't have coverage. And they won't. I mean, the other aspect, too, and we we're going to get, I assume, further into the insurance specifics of it, but we do have to be very careful. Your directors and officers coverage, no matter who your carrier is, will say that there is a professional liability exclusion. What does that mean? It means don't act outside of your scope of authority. If you're not acting as a board member volunteer, then you don't have the applicability of that coverage to defend you. If you act as an attorney and you consult with Chat GPT versus a professional, Valerie spoke about the business judgment rule. You're breaching your fiduciary duties and you're going to be personally open to litigation and costs.

SPEAKER_00 45:42

I want to do it now, Dawn. I think we hit a lot of it because it's so important. I don't even want to run the risk of running out of time about this. So if I just could break that down the most simple statement, if you use Chat GPT for legal advice or other types of professional advice, and you shouldn't have, an insurance company could use that against you and not provide you coverage. Is that what you're saying?

SPEAKER_03 46:02

Absolutely. It's called malpractice. And essentially, you are committing malpractice without the opportunity of having insurance to financially defend you. Cameron, we're going to transfer of risk, right? We're transferring financial risk. So insurance carriers, when they know that there's the opportunity for you to um skirt financial, you know, you don't want to spend the money on an attorney. Well, you're using Chat GPT. Well, the insurance is wise to that. We know. So we put those exclusions in place.

SPEAKER_04 46:35

I think Dawn reads the good, but she said the word malpractice, and I would just elaborate on that. That malpractice is committed by an attorney. Chat GPT and the board cannot commit malpractice. And the thing is, is that if I commit malpractice and the board makes a decision based on that, I have a policy to cover that claim against me and my firm. ChatGPT does not. And I think the insurance understands that that risk shifting, um, and you don't get that protection with Chat GPT legal opinions, in addition to all the other liability issues it exposes the board to as well.

SPEAKER_00 47:07

Cameron, this makes me the most nervous with people using ChatGPT. I'm afraid of a board member using it, trying to save a couple bucks, instead of making a phone call to the attorney, and then finding themselves not covered, and then just praying that they're going to be indemnified, which they should be, but you know, hey, you never know what could happen. Um, this part makes me really nervous, Cameron, and I want people to understand that. Um, do you agree?

SPEAKER_02 47:31

Yeah, I I mean I've seen it. Um, and I try to really put my foot down and say, you you cannot do this. You have to use an attorney. You know, I understand it's usually down to you know saving money, um, but it's not appropriate.

SPEAKER_00 47:48

You're saving money for all your friends and neighbors, and they're gonna be the same ones that end up suing you and not wanting to pay for the money.

SPEAKER_02 47:54

And then you're gonna end up spending more money, you know, yeah.

SPEAKER_00 47:58

So AI does not know your state statues.

SPEAKER_03 48:03

I just want to comment on Cameron uh that I like to say in Cameron's example, it's the price now versus the cost later. So boards are looking ultimately at the here and now, but not thinking at the big picture. And a lot of times I will say that the cost is triple, double, quadruple from what they initially thought. If they had just gone with the attorney opinion from the beginning, it saves thousands and thousands of dollars, millions sometimes.

SPEAKER_00 48:30

And I think here, I hope the panel agrees. We're huge advocates of Chat GBT. We're none of us here are against it. We're just scaring you and making you aware, like, don't go too far with it. Valerie?

SPEAKER_04 48:43

That's right. We just want you to use it prudently, and we're here to give that uh advice on how to do so.

SPEAKER_00 48:49

So let's do some questions. Uh AI does not know your stat your state statutes. Maybe it does, maybe it doesn't. But the big thing there, Valerie, is maybe it does, maybe it doesn't.

SPEAKER_04 48:59

It's the maybe, exactly. So when uh boards and managers are using it, they may not have the legal uh background or education to know whether or not it's citing actual law. Um, like I mentioned before, I have actually trained uh the AI tool that I had been using. And so it does know some statutes, but again, if it's an issue that have not actually trained AI, um sometimes it's pulling up bills that had never been passed, maybe a bill from 20 years ago that was not passed into law. And so me having the um background in law and understanding, hey, that never passed, then I'm able to self-correct. But I think um a board or a manager may not have the ability to do so and may take that um misadvice and apply it in their uh whatever circumstances they're dealing with, and that could be disastrous.

SPEAKER_00 49:46

I bet if your AI met my AI, you would make my AI feel so bad because you're asking your like really complex questions and having really in-depth conversations. I'm asking like how many calories in an Oreo? You know what I mean?

SPEAKER_04 49:59

We're gonna take over the world.

SPEAKER_00 50:01

Yeah, my AI got like that the short end of the stick here, I think, in the world. All right. Um, AI does not know your community's unique amendments. Cameron, you're not a lawyer, but I think you would probably agree with this one, right?

SPEAKER_02 50:14

Right. I mean, you could put all of the amendments in there, but again, we've talked about how it can miss stuff.

SPEAKER_00 50:21

So it definitely missed stuff for sure. AI may hallucinate case law or legal standards. Dawn, I'm gonna go to Valerie last on this, but I've read about this in the news and a lot of attorneys getting in trouble. I don't know what what they're doing, but go ahead.

SPEAKER_03 50:38

I'm really glad you brought that up. It's actually a terminology that's becoming more common is the word hallucinate. It's essentially that they've they've put together an opinion that is based on a multitude of different factors, but it's really not correct. Essentially, a hallucinate is incorrect, and so we're seeing that in a multitude of factors. And I think Valerie can speak more to that as far as the hallucination of the case.

SPEAKER_00 51:04

Valerie, can I make a plug really quick for people to use? Okay, people should use HOA and condo attorneys. You always should have before. This is why you should now. You want to, I'll use Valerie as an example. You want to know someone that knows the HOA and condo law. Now you got lawyers out there that really don't know it, and they're crummy lawyers. Crummy, stinky lawyers who are using AI to try to break into the HOA and condo world. You want to be very careful what lawyer you pick. Valerie. I hope you're not going to be able to do that.

SPEAKER_04 51:34

Yeah, I think that kind of broaches a more general topic. Uh, absolutely, uh, AI hallucinates. It comes up with bad statutes, bad case law, bad interpretations, legal opinions. Um, but I I think it goes to really what we're talking about is comparing AI and its error rate to humans, because we humans have error rates as well. I I like in at least the versions of AI that I have tested out at to a bad law clerk. Uh law clerks are law students that work in a law firm, try to get training and they help the um uh attorney and their rates are lower, so it helps keep the client bills low. But I've had bad law clerks that give me bad case law. Um, maybe they didn't fully research to figure out whether or not that case has been overturned at a later date. And I think that the same is is is um apparent in AI. And so to me, it's like we're balancing, okay, who has a lower error rate? Because that's what we're trying to avoid, right? Everybody's trying not to make mistakes, but we're humans, we're going to. AI is also gonna make mistakes. So I think the happy uh uh marrying of the human that knows that is educated and trained with the AI tool to help spot things is probably the best mix of eliminating and reducing errors. Relying on just AI is is just gonna lead to disaster because of the hallucination. And maybe it learns over time and maybe it's it becomes smarter than me. Um that would be great, but right now it isn't there.

SPEAKER_00 52:54

I'm just nervous though, Valerie, because before, if I called an attorney who wasn't familiar with HOA and condo law, they would have to outright lie or fake it. Now they can go on AI and they can fake it a little bit better. I'd be so careful now about trying to find a firm that has a background in HOA and condo law. This makes it actually harder, I think, to you see what I'm getting at?

SPEAKER_04 53:12

Yeah, I think you're right. Uh I think that that has plagued uh the practice forever than an attorney representing themselves as someone who can handle a case that they can't, maybe they're struggling. Um, but you're right that if an attorney is using AI in order to get their legal education and expertise, um, that is is very dangerous for a board or an association that's going to go and seek their consult. However, that attorney better have malpractice so that if they do give the association bad advice, the association has recourse via a malpractice claim against that attorney. But I I do 100% agree that you should be looking for professionals with experience in the area in which you're seeking legal advice.

SPEAKER_00 53:51

Reliance may increase fiduciary exposure. I think we've gone over this in depth at this point. Cameron, do you think we've hit that point well enough? AI can assist that AI can assist thinking. It cannot replace licensed legal advice. Uploading government documents into public AI tools. You want to jump on that, Valerie? That was one of the ones you kind of wanted to bring home, I think.

SPEAKER_04 54:14

Governing documents, well, at least the way that we use them are public uh records. So I think we're we're fine with uploading those. I can see no problem.

SPEAKER_00 54:22

Wrong one. Um yeah. Well, let me hit you hit let me hit you two, Valerie, because I want to watch the time and leave it time for closing thoughts, all of you. Pasting resident complaints into AI for rewriting, inputting sensitive litigation or delinquency information. Valerie, we're talking about some of the confidentiality. You want to jump on that?

SPEAKER_04 54:38

Yeah, so again, there's two issues there. One, you don't want to be inputting, uh, at least from the legal perspective or the board or manager, anything that's confidential into there because you're potentially waiving your attorney client communication um privilege. Um, in addition to that, if you actually get into a lawsuit and um that you have some long history of chat uh with any AI tool with a lot of information that you don't want disclosed in the court proceeding, the attorney could potentially request that in discovery. Now, can a litigant delete their history? It's it's not advisable and obviously violates rules, but I actually don't know if there's a way for an attorney to go straight to the actual AI system and say, here's a subpoena, we have a lawsuit going on, we want the chats involved with this person on this date about this topic. I don't know, that area has not really um um um uh I guess grown. And so I do think it's a concern, um, one that I try to heed caution with people with what they're putting into AI, um, because essentially, even if you delete it, there may be an alternative route for an opposing party or attorney to still get that information. So just be conscientious, just like writing your emails. I always say don't write anything in an email you don't want read out loud in court. I'd probably say the same for any um AI discourse.

SPEAKER_00 56:00

Valley, there was another item too. I was kind of saving. Did we cover it? It was about email or uh if you don't remember, it's fine. I mean, I we probably covered it.

SPEAKER_04 56:08

Maybe the audience will remember and put it in the chat.

SPEAKER_00 56:11

I want to fly through this one, Dawn, pretty quick, but I want to make sure we hit, I'm gonna hit three or four at the same time. Where is your data going? Who owns the output? Are the conversations stored? I mean, I'm getting here with insurance, Dawn, cyber insurance.

SPEAKER_03 56:24

Well, there was recently a really large lawsuit involving insurance agents. Some insurance agents decided to go to a different company and they were sued contractually. And what came up was that all of the insurance information, everything that the insurance agents were doing was recorded. Not only that, every time they, you know, went on their computer at work and they looked at a personal email, it took screenshots. It had the information. So I don't think people know ultimately, even the employees of the insurance agencies, that their data was going somewhere and someone could replicate it in the future. And so in this court case, they were able to extrapolate a lot of the details that people just thought was gone, right? There's always an electrical, electric, electronic or digital footprint that exists. In fact, it's very hard to erase it. So I think the concern from a privacy standpoint, from the insurance, is you have to be very careful. And some of our insurance carriers are actually coming out with AI exclusions that you absolutely need to pay attention to. And those AI exclusions will be in connection with data sharing and also privacy. So cyber insurance policies are extremely important as to the release of that data information. Most of you out there, including the management company Ravenel, your insurance policy probably has an absolute cyber exclusion on the general liability coverage. So it's important to have that standalone policy. In fact, cyber liability is becoming more important than general liability these days, with all of the aspects of copyright infringement. We have privacy and we have defamation. We also have the issues of uh, you know, people taking pictures and making AI generated type um scenarios.

SPEAKER_00 58:18

That'll be a whole other topic on what people can get away with. Okay, we are under two minutes here. We got a minute and a half. Uh, Valerie, uh, one other thing if you want to hit it for your closing thoughts or maybe not. AI policies. Do you think an association needs to have one? Um, but outside of that, why don't you bring us uh your uh you know let me have Dawn go then, Valerie, and then Cameron to wrap us up. Dawn, closing thoughts.

SPEAKER_03 58:39

If you're gonna chat GPT or or any type of other AI, we have open AI and then there's closed AI. Closed AI is typically what a company will use intercompany between themselves and not release. Across uses its own proprietary AI information. That way it's not disseminated outside. So I would recommend if you are going to use that, to start to use your own proprietary, closed AI to help it become smarter and more beneficial and to eliminate your risk.

SPEAKER_00 59:12

Valerie.

SPEAKER_04 59:13

I agree with Don on that, and I don't know exactly all of the different ones that are private, closed, you need to do your due diligence, but you asked about an AI policy, and I think that it would be prudent for an association to have an AI policy because again, you're dealing with a number of directors and managers, you all need to be on the same page on what is being what AI is being used for. Um, and so even if that policy says, to the extent you're going to put anything association related into AI, let's have a meeting about it to discuss it and agree so that one person isn't unilaterally waiving attorney client communication, uh, putting things into AI that could potentially be discoverable in a lawsuit. And so I do think an AI policy, and you can actually start with AI to help you draft that policy. And I'm sorry, I've run out of time.

SPEAKER_00 59:58

So Oh, sorry. You know, Cameron, I always think we're not going to have enough content, and we it's never like a problem. Cameron, why don't you wrap us up? Is there anything that we forgot to talk about that we need to come back and explore at a different time? Or what do you think about today's session?

SPEAKER_02 1:00:11

I don't think so. I think this was very informative, and I hope that our board members and managers both maybe got some ideas on how they can use the AI differently. Um, and then also, you know, made them think about some ways that you know they need to take some precautions in using AI as well.

SPEAKER_00 1:00:29

Valerie, all closed. I asked ChatGBT today, towards the end, draft me one sentence where I could tell Valerie that she should let me be right every now and then. ChatGBT tells me that I should tell you, Valerie, just this once, let me be right till I can feel like I'm winning at something today.

SPEAKER_04 1:00:45

You're right, Ray. You're right.

SPEAKER_00 1:00:47

All right, I want to thank everybody for being here. Thanks, audience, thanks, panel. Cameron, thank you so much for inviting us. Thank you so much. Bye, everybody. See ya.

SPEAKER_03 1:00:55

Bye. Bye. Thank you.