Transcription
I have spent a number of years telling employees to assume every email, every text message, every social media post, recording, and more could become an exhibit at trial. Be careful what you put in writing. I never imagined I'd have to add ChatGPT to the list, but here we are. So, I'll just tell you the bottom line is that you should assume everything you write could someday become evidence. Technology changes, but the rule of law does not.
A federal judge recently issued a decision. I'm going to share what happened. A man was under federal criminal investigation and used an AI chatbot, in this this case it was Claude, to help organize information and prepare documents that he intended to share with his lawyer. Now, he prepared the documents, but his lawyer didn't know he was doing that. His lawyer certainly didn't tell him to use chat the chatbot to prepare the documents. And the person included information that he got from his lawyer and asked Claude how he should respond. He discussed legal strategy with Claude. Claude shared its comments, told him what, you know, he suggested or it suggested.
Now, when the FBI executed a search warrant at Hepner's home, they seized his electronic devices, which included all that information that he shared with Claude. And of course, they wanted to be able to use it at trial. I mean, that's that's golden. Now, of course, the prosecutors were, you know, gung-ho. It's like you have no privilege. We are entitled to to documents. And Halpner's uh attorneys argued, "No. This the these documents are protected by attorney-client privilege." And at first glance, you would absolutely think that's a no-brainer. Of course. When people come see me just for legal representation, and I don't represent them yet, they're coming to me to seek advice, uh find out whether they have a case, they are automatically protected by the attorney-client privilege. So, why not in this case? What What was the problem with this case?
Well, the judge said, "No. No attorney-client privilege attaches." And why is that? Well, first, the judge said, "You know what? He's not communicating with his attorney." And you know what? That's a really great point. Sometimes I have people copy me on emails. I don't know anything about what's going on, but they'll copy me on it cuz they want my name on it as an attorney, and thinking, you know, that's sufficient to have uh to to get the attorney-client privilege. But anyway, uh the judge said, "No, you're not talking to your attorney in this case. You're talking with Claude. You're communicating with an AI platform." And according to the court, once you voluntarily share information with a third party, you may lose attorney-client privilege. And that is absolutely true. The judge said, "You know, Claude is a third party. Claude is not your attorney. It's somebody that has nothing to do with this case." And that's very true. When I am communicating with a client, and sometimes they want to bring a friend in, um I say, "No, because you lose attorney-client privilege uh in that case." Now, that's um you know, sometimes I will If there's a mediation, and they want to bring a family member in, and I'm not concerned about any information that we're sharing, uh because we're going to be sharing it with the other side anyway, then that's okay. But, if there's something between you and your lawyer that you want to keep private, then you need to leave any other party out of it. And apparently, that includes AI chatbots. Think about that. People treat ChatGPT like it's their therapist, their lawyer, their diary, their best friend. The law doesn't necessarily see it that way.
So, the judge ended up showing that, um none of the requirements for privilege or attorney work product to attach actually applied in this case. Uh namely, he wasn't talking to his attorney. Uh he was not working at the behest of his attorney. He was having a conversation with somebody, namely Claude, not party to the litigation. So, none of the thing none of the requirements applied. Um it's, you know, it's an interesting decision. This is a developing area of law. Nothing is ever as cut and dry as it seem. Uh but, the thing you can take away from this case is, you know, it's dangerous to be talking too much to AI chatbots, Claude, ChatGPT, Grok, whatever. Be careful with that. I mean, I want to explain something that's even bigger than this case, the Helphner case, uh because that was a loser for that uh person, Helphner. You know, it's it's bad enough that you have to respond to what's going on in the case. You're a defendant. You're you're criminal defendant. But then, you put all that out there, how you might respond to the questions they have. That's just bad, you know, your your whole strategy is just out there. So, you can forget that. And if you were thinking in the least bit about being shady, um even if you change your strategy, that doesn't stop the prosecutor from sa- from bringing that up. So, that's just, you know, that that's just too much.
So, later on in the video, I'm going to share with you 10 things you absolutely should never share with chat GPT, Claude, Grok, any of the AI group. But before I do, I want to remind you to please subscribe to my channel because it means a whole lot. It helps a whole lot uh in getting information out there to uh employees. We need that information. Please also remember to go to amazon.com, check out my books. I have three. I have uh my most recent book, Let Me Think About It. It just tells you how how to pause uh and to not be pushed into making decisions and quite honestly, it applies outside of work, too. Check that book out. Uh I have my Document Everything book that's been there uh for a while. You need to be documenting everything that happens to you at work, and you don't need to just write a diary. Follow the uh steps that I suggest in this book, and um you'll create a nice insurance policy for yourself in case something goes wrong. And finally, I have my Regina's 22 rules for leveling up, the latest edition. Make sure you check that out. If you need to get in touch with me cuz you want me to weigh in on your case, the information for reaching out to me and scheduling appointments is in the description of this video. All this information is in the description of this video. So, let's continue.
Now, a lot of employees think that evidence includes only emails, um memos, that text messages, that type of thing. But let me tell you something. Discovery is incredibly broad, and I could give you so many nightmares about this. Yeah, let me just tell you one thing. Sometimes I get off on these tangents, but I have to tell you what I'm thinking of things. Sometimes employees will sue an employer after they've left and found another job, and that is a great way to do it. Um but let me tell you what some really nasty employers will do. They will start writing to your new employer. Now, they know good and well your new employer has no information that would help out the case against your former employer. But, they do it to put pressure on you because discovery is incredibly broad. And just to make sure we're on the same page, discovery is that point in litigation. It usually starts right after you file the lawsuit, and it goes on for 4 to 6 months, depending on whether you're in state or federal court, and it's the time where we seek information to support or defend our claims. So, if you're the plaintiff, you're telling the employer to turn over everything. Turn over personnel files. You are taking uh testimony from people. You are getting witness statements. You're doing everything to build your case, and they are doing the same thing. And so, it's discovery. That's what we call that period and the process of gathering facts, gathering witnesses to support um our case.
Now, when I say it's incredibly broad, I mean it's ridiculously broad. I have seen employers abuse the process in the worst way. You know, checking your past rental history, checking how well you pay your debts. I mean, it's ridiculous. And you have to really really be strong to stop employers from taking advantage of this. But anyway, um discovery is you know, sometimes can be your best friend and sometimes it can be your worst enemy. If you sue your employer or they sue you, lawyers can request text messages, Facebook messages, Instagram DMs, private or not, it does not matter. And I have personally been one of those lawyers just digging into um text messages and that type of thing because that's where all the true true information is. I mean, people will misrepresent facts because, you know, they're under pressure. But when you look at things that they um wrote at a time when they had no reason to be dishonest, that's why we issue subpoenas for social media posts. Even your private DMs because you're telling the truth. You have no reason to make up things. And so, you know, be careful what you put in writing. And this includes LinkedIn messages, Slack messages, Teams, WhatsApp, it doesn't matter. All of that. Any place where we can look to find out what you truly truly mean. You know, some people claim that they were going through all this emotional distress because a job created it. And then you start looking at their private DMs and you realize, oh, wait a minute, she's going through a divorce. And she's complaining about the divorce being the reason she suffered all this mental anguish. Well, there you go. That's how we dig into things like that. So, you really have to be careful. And remember, deleted messages can still be recovered.
I had a case where oh my gosh, the case was a hot mess. >> [laughter] >> I I um there was a female supervisor who was pursuing the girlfriend of her uh direct report. Now, I represented the two girlfriends. Um one was um African-American, one was white, and they were probably in their early 20s. The supervisor was a white female and she was in her mid-30s. So, this was kind of a love triangle of sorts among those three women. The female supervisor, white female supervisor, was obsessed with the young black girl who was in the relationship with her direct report. They all worked at the same place and it was a very challenging case. Everyone was difficult to work with, especially >> [laughter] >> my own clients. >> [gasps] >> Yeah, they were very young and the judge kept having to get on them for, you know, showing up late for court and sleeping through hearings when they did show up. It was a mess. Anyway, the supervisor kept, you know, refused to turn over her text messages. She kept saying that there were none, that they didn't exist. Now, my client told me that they did exist and that they would prove how much she was pursuing this other this young black female at work. After issuing a number of subpoenas to the phone company and everybody else, I finally got the text messages and they actually confirmed what my clients had been saying. So, that's really good that's really good evidence to have these text messages where you have a supervisor pursuing this young girl. She was in a different department, but they worked for the same company. And as a result of those text messages, the case finally settled. But, let me tell you something, it settled for a lot less than it should have because wow, what can I say? See, lawyers have to evaluate their clients and determine whether they would make good witnesses. I mean, we're looking at our clients, the other the lawyer for the employer is looking at the the witnesses. I mean, it is just you have to to evaluate constantly how they would present to a jury. I mean, I've seen people who had doctorate degrees just start lying. >> [laughter] >> Yeah, you just start making stuff up whole cloth. >> [laughter] >> People get under oath and just Oh man, it's it it yeah. >> [laughter] >> So, it it's not an education thing. It's not even an age thing. Sometimes people don't react well under pressure. You know, like I said, my my clients had difficulty showing up on time. They were getting on the judge judge's nerves, on my nerves. I mean, I was having to send people to pick them up. >> [laughter] >> Send an Uber to go get them and bring them while we I mean, we're sitting there waiting for hours. It was just, you know, that's how it was. And then, the final straw for me was, you know, litigation lasts a long time. So, remember, a lot of time is passing while the case is going on. But, the final straw for me was when the black girl in the relationship showed up to a hearing very pregnant. >> [laughter] >> I didn't understand. >> [laughter] >> And I'm like, okay. Here I am arguing that this love triangle which included three women, you know, that this supervisor did so much to destroy the relationship between these two girls and one of them shows up pregnant. And I'm like, okay, there's another person involved. >> [laughter] >> And it's not the supervisor. She didn't do that. >> [laughter] >> So, anyway, I finally just had to get the most I could to settle the case. Surprised that I could get anything, but I got it because of those text messages. So, what can I say? I have a lifetime of stories about interesting cases.
Even information sitting on your phone is dangerous when it comes to lawsuits. I mean, I have been in depositions, you know, where they're sitting under oath, and the lawyer says, "Do you have your phone with you?" Yes. "Can I see it?" >> [laughter] >> Now, of course, if you have a good lawyer, the answer is no. >> [laughter] >> But, don't think they won't ask. So, I've seen people turn it over. Turn over their phones. That's just, you know, that's nonsense. But, anyway, um be careful what you put in writing. And now here we are, AI conversations could become part of the discussion, too. I mean, you're putting it in writing, and um just like if you do a Google search, um all of that stuff is something that we can discover, um request during the discovery period, and get it. And usually, it does not help your case.
Here's the mistake people make, especially when they're talking to ChatGPT. I mean, you know, ChatGPT is my boy. >> [laughter] >> I get it. You become friends. They are always there when you need them. ChatGPT, Claude, Grok. I get it. But, people make the mistake of thinking the conversations are private. They're not. Well, keeping that in mind, here are 10 things you should never share with an AI chatbot, whether it's ChatGPT, Grok, whoever.
Number one, your secrets. Don't treat AI like your therapist, your diary, or your best friend, even though it might feel like it at times. If you wouldn't post it on a billboard, don't type it in a chatbot.
Number two, your personal information. Never share your social security numbers, bank account information, credit card numbers, passwords, driver's license numbers, passport information, your home address. Identity theft is hard enough. Don't make it easier.
Number three, photos and biometric data. Be very careful about uploading photos of yourself, your children, or your family members. Unlike a password, you cannot change your face. Well, maybe you can, but you cannot change your fingerprints. You can't change your voice. Biometric information is among the most sensitive data you have.
Number four, medical questions that require a doctor. AI can explain medical concepts. It cannot examine you. It cannot order test. It cannot exercise clinical judgment. So, if you're having a medical problem, see a licensed health care provider. And remember, when you put that information there, it's there.
Number five, legal problems that require a lawyer. This one should surprise absolutely no one coming from me. AI can help you understand the legal concepts. It should never replace legal advice. It can misunderstand facts and misstate them all together. Sure have seen it. Lawyers writing briefs uh or actually having ChatGPT or one of those AI chatbots write a brief and it's just wrong. I mean, some of the lawyers have gotten in trouble for that. I mean, seriously.
Number six, confidential workplace information. Never upload personnel files, internal investigations, customer information, trade secrets, business strategies, confidential company information. Once confidential information leaves your control, you can't always get it back.
Number seven, major life decisions. Don't let AI decide whether to quit your job, leave your marriage, invest your retirement, move across the country. AI can help you think. It shouldn't do the thinking for you.
Number eight, anything illegal. Don't ask AI how to commit crimes. Don't ask how to destroy evidence, how to hack systems, and how to evade law enforcement. Yeah, people are asking these questions. Don't be one of them.
Number nine, client information. If you're a lawyer, doctor, therapist, or anyone with confidentiality obligations, think twice before uploading client information.
And number 10, anything you wouldn't want to become an exhibit at trial. This is the biggest lesson. Discovery, once again, is very broad. Lawyers can and do request everything. We are so nosy. >> [laughter] >> But it's because that's where that's where we find the truth, in your internal private thoughts, and that's exactly what we go for. So again, assume everything you write could someday become evidence. Technology changes, the rule of law doesn't. See you next shift.