Why Legal Matters Should Not Be Shared With Public AI Platforms

With the rapid adoption of generative artificial intelligence (AI), many clients interact with these tools as though they were engaging in a private conversation. Some clients may turn to public AI platforms for legal guidance before contacting their lawyer, while others upload documents they receive from their lawyer for summaries and drafting assistance. However, sharing basic facts or legal materials with a public AI platform—such as uploading a legal memo from counsel to summarize it or convert it into a draft letter—may jeopardize attorney-client privilege and work-product protections. Once information is shared with a public AI tool, it may be treated as a communication with a third party—potentially placing it outside the scope of these protections.

The United States v. Heppner Decision

These risks were highlighted in a recent federal court decision, United States v. Heppner, No. 25-cr-503 (JSR), 2026 WL 436479 (S.D. N.Y. Feb. 17, 2026). There, the U.S. District Court for the Southern District of New York held that the government could review a criminal defendant's communications with a public AI platform, concluding that neither the attorney-client privilege nor the work-product doctrine applied.

Although courts have cautioned that the privilege protections are narrow, Heppner appears to be among the first to address how those doctrines apply to communications with public AI tools. The decision serves as a reminder that not everything connected to a legal matter is protected—particularly internet searches, communications with non-lawyers, and interactions with public AI platforms.

In Heppner, the defendant faced charges including securities fraud, wire fraud, conspiracy, making false statements to auditors and falsifying corporate records. While executing a search warrant, the FBI seized 31 documents reflecting the defendant's communications with the AI platform Claude. Those materials included documents the defendant asked AI to generate in preparation for discussions with his counsel, which he later shared with his attorneys.

Attorney-Client Privilege

The court clarified that these communications did not qualify for the attorney-client privilege. Applying the well-established test, the court explained that the attorney-client privilege attaches only to communications (1) between a client and an attorney; (2) that are intended to be, and in fact were, kept confidential; and (3) made for the purpose of obtaining or providing legal advice.

An AI platform is not an attorney, and the privilege depends on a confidential human relationship with a licensed professional who owes fiduciary duties and is subject to professional discipline.

The court further concluded that the communications were not confidential. The AI platform's privacy policy warned users that their inputs and outputs could be collected and shared with third parties, including the government. As a result, the defendant could not reasonably expect those communications to remain private.

The court also emphasized that the defendant did not use the AI platform at the direction of counsel or to obtain legal advice from counsel. In fact, the AI platform itself expressly disclaimed providing legal advice. The fact that the defendant later shared the AI-generated materials with his attorney did not cure the lack of protection, as the privilege does not attach merely because a document is later transferred to counsel.

The court further explained that even if the information the defendant entered into the AI platform had been privileged, the defendant waived that privilege by inputting the information the attorney provided into the platform, as doing so was no different from sharing the information with a third party.

Work-Product Doctrine

The work-product likewise did not apply. The work-product doctrine protects an attorney's mental impressions and materials prepared by or at the direction of counsel in anticipation of litigation or trial. The court concluded that the AI-generated materials were created on the defendant's own initiative, not by or at the direction of his attorney, and they did not reflect counsel's mental impressions or legal strategy at the time they were created.

Key Takeaways

Heppner flags the risks of treating public AI platforms as private conversation. Information shared with public AI platforms can become third-party communications, potentially placing those communications outside the scope of the attorney-client privilege and work-product protections.

So, before you ask a public AI tool how to handle a harassment claim, whether your patent is valid, whether you are complying with a regulation, or what the memo from your lawyer means, consider that those conversations could become an exhibit in a future lawsuit or government investigation. If you're inclined to use public AI platforms, you should take time to understand how the tools operate and the legal implications their use may carry.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.