United States v. Heppner: A Landmark Ruling on AI and Attorney-Client Privilege – What Every DUI Law Firm (DUI and Criminal Defendant) Needs to Know – the new jail recorded phone call

At George C. Creal Jr., P.C., Trial Lawyers, we have long embraced technology to deliver smarter, faster, and more effective representation for our clients in Cobb County and throughout Georgia. From advanced case-management systems to research tools that help us fight DUI charges, we stay ahead of the curve. But a recent federal court decision out of New York serves as a stark reminder: not all technology use in the legal field is created equal — especially when it comes to generative AI like large language models (LLMs) such as Claude, ChatGPT, or similar tools.

On February 17, 2026, U.S. District Judge Jed S. Rakoff issued a memorandum opinion in United States v. Heppner, 25 Cr. 503, slip op. at 2–3 (S.D.N.Y. Feb. 17, 2026). In what the court described as “a question of first impression nationwide,” Judge Rakoff ruled that written exchanges between a criminal defendant and a publicly available generative AI platform (Anthropic’s Claude) were not protected by either the attorney-client privilege or the work-product doctrine.

What Happened in the Heppner Case?

Bradley Heppner, facing serious federal securities and wire-fraud charges, used the free, consumer version of Claude to generate dozens of documents. These included legal strategy outlines, factual narratives, potential defenses, and analyses of his exposure. He later forwarded the AI-generated materials to his high-profile defense counsel at Quinn Emanuel, describing them on a privilege log as “artificial intelligence-generated analysis conveying facts to counsel for the purpose of obtaining legal advice.”

Federal agents seized the documents during a search warrant execution. The government moved to compel their production, arguing they were not privileged. The defense claimed protection under both attorney-client privilege and work product.

Judge Rakoff disagreed — emphatically.

The Court’s Key Holdings (Straight from the Opinion)

The court held that the AI documents failed at least two, if not all three, elements required for attorney-client privilege:

  1. Not a communication “between a client and his or her attorney.” Claude is not a lawyer. Period. “Because Claude is not an attorney…that alone disposes of Heppner’s claim of privilege.” The opinion emphasized that no attorney-client relationship can exist with an AI platform, and no fiduciary duties or ethical obligations attach.
  2. Not kept confidential. Claude’s own privacy policy (which users agree to) states that inputs and outputs may be used to train the model and may be disclosed to third parties — including governmental authorities — even without a subpoena. Once shared with a public AI tool, there is no reasonable expectation of confidentiality.
  3. Not made for the purpose of obtaining legal advice from an attorney. Heppner acted on his own initiative, not at counsel’s direction. Claude itself disclaims providing legal advice. Later sharing the outputs with lawyers did not “retroactively” cloak them in privilege.

The work-product doctrine also failed. The materials were not prepared “by or for” counsel or at counsel’s direction. Independent client use of AI, even if done in anticipation of litigation, does not automatically qualify as protected work product.

 Individuals using AI to ask legal questions or questions on how to commit crimes or avoid apprehension will be State’s Exhibit #2 right after their recorded jail phone call which as we have said before is State’s Exhibit #1. 

The AI documents were ordered produced to the government.

Why This Ruling Matters for Law Firms — Especially Criminal Defense and DUI Practices

This is not just a white-collar case. The implications ripple across all areas of litigation, including DUI defense here in Georgia. Defendants facing DUI charges, license suspensions, or criminal accusations increasingly turn to AI tools on their own to “brainstorm” defenses, draft statements, analyze police reports, or prepare for hearings. Heppner makes clear that doing so without proper safeguards or at least, at the direction of counsel can waive privilege and hand the government powerful evidence.

For law firms, the ruling is a wake-up call about how we and our clients use AI:

  • Public/consumer AI tools create real risk. Feeding case facts, client confidences, or strategy ideas into free versions of Claude, ChatGPT, Gemini, or similar platforms is the functional equivalent of posting them on the internet.
  • Attorney supervision is critical. Materials generated independently by a client (or even by a lawyer outside a supervised workflow) do not automatically receive privilege or work-product protection.
  • Data privacy policies matter. Most public AI platforms explicitly reserve the right to use your prompts for training and to disclose them to regulators or in litigation.
  • Enterprise or private AI solutions are the safer path. Tools with strong confidentiality agreements, no-training-on-client-data clauses, and SOC-2/enterprise security are far less risky — but even then, they should be used under direct attorney direction and documented carefully.

Practical Takeaways and Best Practices We Follow at George C. Creal Jr., P.C.

  1. Educate clients early. We now explicitly advise clients: Do not use public AI tools to analyze your DUI case, draft affidavits, or prepare for trial without first consulting us. What feels like a private “chat with AI” can become evidence for the prosecutor.
  2. Use AI the right way — under attorney control. Our firm uses vetted, secure AI tools only as an extension of our legal team. Every use is supervised, documented, and limited to non-confidential tasks where appropriate.
  3. Document everything. If AI assists in research, drafting, or strategy, we note the tool used, the attorney’s review, and the final human work product. This preserves privilege where possible.
  4. Stay ahead of the ethics curve. The Georgia State Bar and ABA opinions on AI competence and confidentiality are evolving rapidly. We treat Heppner as the new baseline for privilege analysis.
  5. Leverage AI responsibly to fight harder for you. At our Cobb County firm, we use AI to analyze body-cam footage faster, spot inconsistencies in police reports, and prepare more persuasive sentencing arguments — always with human oversight and full transparency.

United States v. Heppner, 25 Cr. 503, slip op. at 2–3 (S.D.N.Y. Feb. 17, 2026) does not ban AI in law firms. It simply reminds us that technology must serve the attorney-client relationship — not replace or undermine it.

At George C. Creal Jr., P.C., Trial Lawyers, we remain committed to using every legitimate tool available to defend DUI, criminal, and serious traffic cases — but we do so with the caution and strategic foresight this ruling demands. Our clients deserve aggressive representation that is also technologically responsible.

If you or a loved one is facing DUI charges or any criminal matter in Cobb County or metro Atlanta, contact us today. We’ll review your case, explain exactly how technology — and the law — will be used to fight for the best possible outcome, and make sure no preventable mistakes (like unprivileged AI use) ever jeopardize your defense.

George C. Creal Jr., P.C. — Fighting for You with Experience, Integrity, and Smart Technology.

Marietta, Georgia • DUI & Criminal Defense Trial Lawyers
(404) 333-0706 | www.georgecreal.com

Tags:
Posted to: ,