Don't Lose Legal Protection Using Public AI Chatbots Like ChatGPT #law #legal #lawyer

Here's what's happening right now in federal courtrooms across the country. Judges are being asked for the first time whether your conversations with chat GPT, Claude, or any public AI tool are protected. And the answers coming out of 2026 are ones every attorney and client needs to hear. Three landmark federal cases just drew the lines. First up to the plate is Warner vs. Gilbarco. A pro-save plaintiff used chat GPT to draft litigation materials. The other side wanted access to the logs. The court said work product protection survives. Why? Because AI is a tool, not a person. Using it is no different than legal research software. No waiver of privilege applied in that case. Second up is United States vs. HEPNER. A CEO, already represented by council, independently used the public version of Claude to generate defense strategy. Judge Raykov said, "No privilege, no work product protection." Why? Because Claude is not an attorney. And the public tool's privacy policy, which allows data collection and model training, destroyed any reasonable expectation of confidentiality. Production of the logs was ordered. Third up is Morgan vs. V2X. Another pro-save plaintiff, the court found that the work product was protected, but the court also ordered disclosure of which AI platform was used with confidential discovery materials. The judge also amended the protective order to restrict public AI use entirely. Here's the bottom line when it comes to litigation in U.S. federal courts. Public AI tools are not confidential. If you are a client, inputting sensitive facts, strategy, or legal questions into a free consumer AI, you may as well have just handed that to the other side. The fix? Use enterprise grade AI with contractual no-training guarantees. Or, better yet, talk to your attorney before you type anything sensitive into any AI platform.