Free and paid AI tools are not treated the same in a lawsuit, and a paid personal plan sits closer to the free one than most owners expect.
In February 2026, Judge Jed Rakoff of the Southern District of New York ruled in United States v. Heppner that a business executive’s chats with a public AI assistant carried no attorney-client privilege and no work product protection. The executive had used Claude on his own to analyze his legal situation and outline defense strategy. Federal agents seized about 31 documents of those conversations.
The judge looked at the provider’s consumer privacy policy. It allows the company to collect inputs and outputs, use them for training, and disclose them to authorities. The court treated typing into that tool as sharing with a third party. The AI is also not an attorney.
The same month, a federal magistrate in Michigan went the other way. In Warner v. Gilbarco, the court held that using ChatGPT to prepare litigation materials did not waive work product protection, and it treated the AI as a tool, not a person. The facts differ. The plaintiff represented herself. Courts have not settled this, so I plan as if chats can be requested.
Here is how the plan types compare.
Free, Plus, Pro, and similar personal plans run on consumer terms. Paying $20 a month changes the features, not the terms the Heppner court read.
Retention adds a second problem. OpenAI says deleted chats on its consumer plans are scheduled for removal within 30 days. In 2025, a court order in the New York Times copyright case forced OpenAI to keep consumer chats, including deleted ones. The order excluded ChatGPT Enterprise, ChatGPT Edu, and API customers on zero data retention. OpenAI says the duty to keep new data ended September 26, 2025. Deleting a chat does not control what a court can require.
Business and enterprise plans carry negotiated terms: no training on your data, confidentiality commitments, admin controls, and set retention periods. Law firms who have written about Heppner say these tools likely offer stronger protection. No court has confirmed that, so I do not call them safe.
Watch an example of how a litigation request about this topic can look:
▶ Watch the example video on YouTube
Here is what I tell business owners:
- Keep case facts, attorney advice, and settlement thinking out of every AI tool unless your attorney directs the use.
- If AI will touch a legal matter, let your attorney direct it, in a tool with written confidentiality terms.
- Use company-managed AI accounts, not personal ones. Turn training off. Know the retention setting.
- When a dispute starts or looks likely, stop deleting chats and tell your attorney they exist. Deleting after that point can look like destroying evidence.
- Put an AI policy in writing so staff know which accounts to use and what stays out.
At CIO Landing, we set up company-managed AI accounts with training off and retention settings reviewed, and we write the policy your team follows. Our rule for AI is trust but verify. A person reviews the final output.
I am not an attorney. Talk to yours before anyone on your team uses AI on a legal matter.
Does your team know which AI account they use today?