Transcription
Did you know that every word you type into ChatGPT, or Claude, or Gemini, or any public AI-facing tool means that your case is potentially at risk for being discoverable?
Anytime you do a search about your legal case in a public-facing AI tool, you are putting yourself at risk. This is not a prediction. This is actually a federal ruling from February of 2026. It's called the United States versus Heppner. I'm going to tell you all about it. I'm going to tell you what you can do instead.
Hey there, I'm Rebecca Zung. I'm a 25-year trial attorney, and I teach you all about how to build leverage. I'm known as the leverage lawyer. And if you're new here, welcome. I invite you to subscribe and hit that notification bell.
All right, so listen. If you are using ChatGPT, or Gemini, or Claude, or anything like that to research your case, research your divorce, research your business partnership dissolution, or your probate litigation case, or your defamation case, or your employment case, whatever it is, and you're trying to get it to summarize information, or write up events, or vent about your case, or anything related to your matter, I need you to stay with me because what you do in the next moments, the next move that you make on your computer, could decide whether those conversations are protected, or are they going to become evidence that is discoverable and potentially even used against you.
I've been doing this a long time, and I've seen a lot of different changes. I saw e-discovery come in. I saw, um, many, many changes, and this one is a huge one. AI is not coming, it is here. Courts know that it is here. Attorneys know that it is here.
And on February 17th, 2026, Judge Jed Rakoff of the Southern District of New York issued a written opinion in a case called The United States versus Hepner. The defendant, Bradley Hepner, had used a public version of an AI chat to research legal questions about his case and to prepare materials he later shared with his attorneys. He argued those AI conversations were protected by attorney-client privilege and the work product doctrine. A court disagreed. Judge Rakoff ruled that the materials were not privileged, were not work product, and were therefore discoverable.
The reasoning, in simplest terms, the AI was not his attorney. He used it on his own initiative and not at his counsel's direction. And the consumer-facing version of the AI tool he had used had terms of service that expressly allowed the company to train on his data and disclose it to third parties, including government regulatory authorities. There was no reasonable expectation of confidentiality.
This was a criminal case, but the privilege rules at issue are the same privilege rules that apply to any civil case in this country. Divorce, custody, business, employment, whatever it is.
So, if you are in a divorce situation, understand this. The Weinberger Law Group, Borden Smith, at least a dozen major family law practices have already issued formal warnings to their clients. The warnings all say versions of the same thing: Stop typing case information into public AI tools. Stop drafting emails to your opposing party in ChatGPT. Stop using public AI to analyze your settlement options or vent about the other person.
Because here is what most people don't realize. Your AI chat history is not just sitting on a server somewhere private. The companies that run public AI tools have contractual rights to access, retain, train on, and in many cases, disclose your outputs. Their terms of service expressly say so. And at the moment a divorce case becomes adversarial, or or any case becomes adversarial, your opposing counsel can subpoena that data. Not the same way that they subpoena bank records, a newer, more aggressive way. Because the Hepner ruling just established the precedent that those AI chats are not privileged in the first place.
So, if you've been venting to ChatGPT or Claude about the other person, and if, or if you've been drafting strategy in Claude, if you've been asking Gemini how to handle the next hearing, every single one of those conversations is sitting in a server that the opposing counsel may now be entitled to access. And what you said in those moments was almost certainly not what you would say if you knew a judge, or mediator, or the opposing counsel, or opposing party was going to be reading it later.
So, here are the four things that make a conversation discoverable. Here are the four conditions drawn directly from that Hepner reasoning that make your AI conversation discoverable.
Number one, the AI is not your attorney. Privilege only attaches to communications with an attorney or their authorized agent, and a chatbot is neither.
Number two, you use the tool on your own. Privilege requires that the communication be made at counsel's direction or in furtherance of legal advice. Researching your own case on a public AI without your attorney directing it does not meet that bar.
Number three, the platform's terms of service disclaim confidentiality. Every major consumer AI tool, ChatGPT, Claude, Grok, Gemini, has terms that allow data retention, training, and disclosure. By using the tool, you agree to those terms. You waived confidentiality the moment you typed.
Number four, you shared the information with a third party. Legally, the AI company is a third party. Once your information has been disclosed to a third party in circumstances that undermine confidentiality, that privilege is gone, and it does not come back. Sharing your AI-generated materials with your attorney later does not restore that privilege that you've already lost.
All right, four conditions. Every public AI conversation about your case meets all four, not just one, not just two, all four. And that's why this ruling has terrified attorneys.
So, right now, I want you to just drop in the comment below if you have used a public AI tool, ChatGPT, Claude, Gemini, any of them, to research, draft, vent about, or strategize on your case in the last 6 months. Be honest. Yes or no? The first step to protecting yourself is knowing your exposure.
All right, so that's the risk. Now, what do you do if you've already used a public AI tool for your case? Do you have a problem? If you need an AI to help you think through your matter, what's the alternative?
Here's what I tell our own clients. First, do not panic. Do not panic. The Hepner ruling does not retroactively criminalize anything you've done. It establishes that your AI conversations are not privileged going forward. The exposure depends on what you've typed, in which tool, and whether the opposing party will or can subpoena it.
Second, talk to your attorney. This is one of those moments where your attorney does need to know exactly what platforms you've used and what kind of content lives on those, um, those platforms in those chat histories. Not only because will they judge, I mean, they're not going to judge you, but they need to know because it may show up in discovery. The worst outcome is your attorney being sandbagged. Your attorney being ambushed. You don't want that to happen. It's so much better for them to know everything up front, and then they can prepare for it.
And then third, stop using public AI tools for anything related to your case. Not tomorrow, today. Every day you continue is another day of evidence being created against you.
And fourth, consider what kind of AI tool, if any, is actually safe to use. And I'm going to say, if any, but I do know of one, which I'm going to tell you about in a moment, because the answer is not no AI. The answer is the right kind of AI.
So, here's the critical distinction. Here's the part the news coverage usually misses. The Hepner ruling didn't say AI is completely incompatible with privileged use. It said consumer, publicly accessible, non-enterprise AI tools with terms of service that train on user data and reserve the right to disclose it. That's incompatible with privilege.
And this is a really important distinction because it leaves the door open for a completely different class of AI tools. Enterprise platforms with no training commitments, tools with contractual confidentiality protections. That's what you want. You want where there's an AI architecture in place that is a walled garden, private architecture, where it's not training using your information to train the model. And it, you want that data to be isolated into a structure, not bolted on after the fact.
That distinction is the entire reason why I built Slay AI in the first place, not as a public, as a patented platform, walled garden architecture design specifically for high-conflict legal matters with data isolation, no general purpose training on user inputs, and an architecture that anticipates exactly the kind of issues the Hepner ruling just made undeniable.
If you're going to use AI in a legal context, the choice is no longer convenience, it is structure. And the structure of the tool you use is now, after Hepner, potentially decisive.
If you're an attorney watching this video, you already know what Hepner means for your practice. Your clients are using ChatGPT. They're using it right now. And some of them are pasting strategy memos you sent them into a chatbot to ask them follow-up questions. "Oh, what does this actually mean?" Some of them are drafting their version of the events. Some of them are stress-testing your advice against an AI that has terms of service permitting disclosure to third parties.
The standard intake conversation has changed. Uh, law firms across the country are now adding AI usage to their client onboarding. Some are issuing formal written warnings as part of the engagement letter. Some are requiring attestation that the client will not use public AI for their case-related work.
So, if you're looking for a tool that you can recommend to your client that handles the structural problem Hepner exposed, Slay AI is built for exactly this. Professional tier access lets you bring the platform into your client relationship under your direction, which addresses the second prong of the Hepner reasoning directly. And you can see all of this at slaypro.ai, and you can get started there for free. slaypro.ai
The rules have changed, so so must we. Right? Here's the, the reframe. Every high-conflict case is now an evidence case. And every AI conversation is potentially that evidence. The era of casually using ChatGPT to think through your, your case, or your divorce, your business dispute, whatever it is, is over. That does not mean AI is the enemy. It means that the kind of AI that you use is now part of your legal strategy. The tool you choose is no longer about convenience or speed. It's about exposure, about discoverability, about whether the things you type in the middle of the night become evidence in the middle of your trial.
The Hepner ruling did not break AI for, for legal use. It clarified what kind of AI is actually safe to use. And anyone who is dealing with, handling, in the middle of a high-conflict matter going forward, whether you're an attorney or a client, you've got to take this distinction seriously.
If this just changed how you're going to handle your case, give me how it's going to handle it in the comments. Drop a comment below and let me know. And if you want an AI specifically built for high-conflict legal matters with data isolation and an architecture designed around the issues Hepner just made law, Slay AI is exactly that. It's patented. It's based on my 25 years of legal trial work. And you can start for free right now at slaypro.ai. For attorneys, check the professional tier. Check that out. It's the same address and the link is also in the description.
Nothing in this video is legal advice for your specific case. Talk to your attorney about your situation, but I want you to walk away with this. The ground shifted in February of 2026. What you type into public AI tools is no longer private. It is potentially discoverable. And anyone in a high-conflict matter needs to operate as if a judge will read every single word. Stop typing your case into ChatGPT. Talk to your attorney about exposure.
And share this, by the way, um, with anybody who is dealing with this and you know is using a public AI. They need to see this, so make sure you like it, give it a thumbs up, share it with anyone who needs to see this. Subscribe for more strategic breakdowns every single week. And if you're going to use an AI to help you build leverage, use one designed for the job.
The next video I want you to watch is why "document everything" is bad advice and what to document instead. That video is the tactical companion. This one tells you what not to type, that one tells you what to capture in the right structure. I'm Rebecca Zung, the leverage lawyer. Stop surviving, start leveraging.
If you've ever felt stuck trying to lose weight, you're not alone. Enter Weight Loss by Hers. It's designed to support you in reaching your goals. And Hers now offers access to an affordable range of FDA-approved GLP-1 medications, including the Wegovy pill and the Wegovy pen. With Wegovy at Hers, lose up to 20% or more of your body weight when combined with diet and exercise. It helps you regulate your appetite, eat less, and keep weight off. Ready to reach your goals? Visit forhers.com/leverage to get personalized, affordable care that gets you. That's f o r h e r s.com/leverage. forhers.com/leverage. Weight Loss by Hers is not available in all 50 states. Wegovy is a registered trademark of Novo Nordisk A/S. To get started and learn more, including important safety information, Wegovy clinical study information, and restrictions, visit forhers.com.
Lately, I've been wanting to just simplify what I wear. I want to have it feel effortless, but look polished, and that's why I've been loving Quince. Quince makes it really easy to refresh your everyday wardrobe with pieces that feel as good as they look, and everything at Quince is priced 50 to 80% less than similar brands because they work directly with ethical factories and cut out the middle men. I have this gorgeous leather jacket from Quince that instantly pulls any outfit together. My husband has this incredibly soft navy cardigan that looks tailored, but also still relaxed, and the quality is incredible, especially for the price. Refresh your everyday with luxury you're you're actually going to use. Head to quince.com/leverage for free shipping on your order and 365-day returns. Now available in Canada, too. That's q u i n c e.com/leverage for free shipping and 365-day returns. quince.com/leverage.