By Christine Leonard, Esq. — Florida Bar #1001612 · March 2026
The Heppner Ruling: AI Conversations Are Not Privileged
In February 2026, a federal court issued a landmark ruling that every person involved in legal proceedings needs to understand: conversations with AI chatbots are not protected by attorney-client privilege. The decision in United States v. Heppner (S.D.N.Y., Feb. 17, 2026) — handed down by Judge Jed S. Rakoff — has sweeping implications for anyone who has ever typed case details, legal questions, or personal information into ChatGPT, Claude, or any other consumer AI platform.
If you are going through a divorce, custody dispute, or any family law matter in Florida, this ruling directly affects you. Here is what happened, why the court ruled the way it did, and what you need to do to protect yourself.
What Happened in United States v. Heppner
Bradley Heppner was facing federal securities and wire fraud charges. After his arrest on November 4, 2025, and upon receiving a grand jury subpoena, he hired legal counsel to represent him. But Heppner also turned to an AI assistant — specifically Claude, Anthropic’s AI chatbot — to help prepare his defense.
Using Claude, Heppner prepared defense strategy reports, outlined potential legal arguments, and analyzed his criminal exposure. He later shared these AI-generated documents with his attorneys. When the FBI seized approximately 31 documents from his devices that were exchanges with Claude, Heppner argued these documents should be protected by attorney-client privilege and the work product doctrine.
Judge Rakoff rejected both arguments entirely.
Why the Court Rejected Privilege: Three Key Reasons
1. No Attorney-Client Relationship
The court’s reasoning was direct and unequivocal. Judge Rakoff wrote: “Because Claude is not an attorney, that alone disposes of Heppner’s claim of privilege.” Attorney-client privilege requires, at a minimum, a communication between a client and a licensed attorney. An AI chatbot — no matter how sophisticated — is not a lawyer. It is not licensed by any bar. It cannot form an attorney-client relationship. Without that relationship, privilege simply does not exist.
2. No Confidentiality
Even if an AI could theoretically form some kind of advisory relationship, the court found that the terms of service and privacy policy of consumer AI platforms destroy any reasonable expectation of confidentiality. These platforms expressly permit data collection, retention, and use for model training. When you type something into a consumer AI tool, you are voluntarily disclosing that information to a third party under terms that allow the platform to use it. There is no confidentiality — and without confidentiality, there is no privilege.
3. No Work Product Protection
Heppner also claimed the documents qualified as protected work product — materials prepared in anticipation of litigation. The court rejected this as well, because the documents were not prepared by counsel or at counsel’s direction. Heppner created them on his own using an AI tool, then later handed them to his lawyers. The work product doctrine protects materials prepared by attorneys (or under their direction) in anticipation of litigation. It does not protect a client’s independent use of AI, even if the output was later shared with counsel.
The critical principle: Privilege must exist at the time of the communication. You cannot create privilege after the fact by later sharing unprotected AI outputs with your lawyer. If the communication was not privileged when it was made, handing it to your attorney does not retroactively create protection.
The Privilege Waiver Problem: It Gets Worse
The Heppner ruling raised an even more alarming issue: sharing privileged attorney-client communications with an AI may waive privilege over the original communications themselves.
Here is why this matters so much. Suppose your attorney gives you confidential legal advice about your divorce strategy. You then type that advice into ChatGPT or Claude — perhaps to get a “second opinion” or to have the AI help you analyze it further. The government argued in Heppner, and Judge Rakoff agreed, that feeding attorney advice into a third-party AI platform constitutes voluntary disclosure outside the attorney-client relationship.
This means that using AI to analyze or strategize about what your lawyer told you could destroy the privilege protection over your lawyer’s actual advice. The original communication between you and your attorney — which would otherwise be fully protected — may lose its privileged status because you voluntarily disclosed it to a third party that has no duty of confidentiality.
This is not a hypothetical risk. It is the direct consequence of the Heppner court’s reasoning.
What This Means for Family Law Cases in Florida
The implications for family law — divorce, child custody, alimony, and property division — are particularly severe. Family law cases involve some of the most sensitive information a person possesses:
- Divorce cases involve highly sensitive financial information — bank account details, investment portfolios, income figures, business valuations, hidden assets, and debt obligations
- Child custody cases involve sensitive information about your children, parenting abilities, mental health, substance use, and family dynamics
- Alimony and support cases involve detailed income analysis, lifestyle documentation, and financial projections
If you shared any of this information with a consumer AI tool, it is now potentially discoverable by the opposing party.
And opposing counsel in a Florida divorce or custody case will ask. Discovery requests are already beginning to include questions like: “Did you use any AI tools to prepare documents, analyze your case, or research legal issues related to this matter?” If you typed your financial details, custody strategy, or your attorney’s advice into ChatGPT, Claude, or any other AI platform, all of that information is now fair game.
Florida family law discovery is broad. Under Florida law, anything relevant and not privileged is discoverable. Since the Heppner ruling confirms that AI conversations are not privileged, there is no legal basis to withhold them.
Why You Need a Qualified Attorney — Not AI
The Heppner ruling is a stark reminder of something that should be obvious but bears repeating: AI is not a substitute for a real attorney. Here is why:
- AI cannot form an attorney-client relationship. There is zero privilege protection for anything you share with it. Every detail you type is potentially discoverable.
- AI cannot appear in court, file motions, cross-examine witnesses, or negotiate settlements. When it comes time to stand in front of a judge and advocate for your interests, AI cannot do it.
- AI makes mistakes — serious ones. In Mata v. Avianca, Inc. (S.D.N.Y. 2023), attorneys submitted court filings containing completely fabricated case citations generated by ChatGPT. The cases did not exist. The attorneys were sanctioned. AI “hallucinations” are not a glitch — they are a fundamental feature of how these models work.
- AI does not know local judges, local procedures, or local opposing counsel. Family law is practiced in courtrooms, not in chat windows. Knowing how a particular judge handles custody disputes, or how opposing counsel negotiates, is the kind of practical knowledge that comes from real courtroom experience.
- AI cannot exercise professional judgment or adapt strategy in real time. A hearing can shift in seconds based on a witness’s testimony or a judge’s question. AI cannot read the room, pivot a legal argument, or make the judgment calls that win cases.
When you hire Christine Leonard, you get the full protection of attorney-client privilege, work product doctrine, and professional ethical obligations. Your conversations with her are privileged. Your legal strategy is protected. Christine is a trial-tested family law attorney who fights for families across Northeast Florida. She personally handles every case — your case is not passed off to a paralegal or an algorithm.
Practical Steps to Protect Yourself
If you are involved in a family law matter — or anticipate one — take these steps immediately:
- Do NOT type case details, financial information, or legal strategy into any consumer AI tool. This includes ChatGPT, Claude, Gemini, Copilot, or any other chatbot. None of it is privileged. All of it is discoverable.
- Do NOT feed your attorney’s advice or communications into AI platforms. Doing so may waive the privilege over your attorney’s original advice — the very protection you are paying for.
- If you have already used AI for case-related research, tell your attorney immediately. Your lawyer needs to know so they can assess the potential damage, determine what may be discoverable, and adjust your legal strategy accordingly.
- Use your attorney for legal research and strategy. That is exactly what attorney-client privilege protects. When your lawyer researches your legal issues, analyzes your case, and develops strategy, all of that work is protected by privilege and the work product doctrine.
- If you want to use AI tools, only do so under your attorney’s direction with enterprise-grade tools that have contractual confidentiality protections. Consumer platforms with open terms of service do not qualify.
Frequently Asked Questions
Are my conversations with AI chatbots protected by attorney-client privilege?
No. The Heppner ruling established that consumer AI platforms cannot form attorney-client relationships, and their terms of service negate confidentiality. Because the AI is not an attorney, there is no privilege — and because the platform’s terms permit data collection and use, there is no reasonable expectation of confidentiality.
Can the other side in my divorce see what I typed into ChatGPT or Claude?
Yes. AI-generated documents are discoverable. Opposing counsel can request them in discovery, and because there is no privilege protection, you cannot refuse to produce them. Any case details, financial information, or legal strategy you typed into an AI platform is fair game.
What if I only used AI for general legal research, not my specific case?
General research may be less risky, but any case-specific information you input is potentially discoverable. The safest approach is to let your attorney handle all legal research. If you used AI to analyze your specific financial situation, custody arrangement, or legal strategy, that information has no privilege protection.
Does this ruling apply to Florida family law cases?
The Heppner ruling is a federal court decision (S.D.N.Y.), but the underlying principles of attorney-client privilege and work product doctrine apply equally in Florida state courts. Florida courts are likely to follow the same reasoning because the privilege requirements — a licensed attorney, a confidential communication, and a purpose of obtaining legal advice — are the same under Florida law.
Can my attorney use AI tools on my behalf?
Yes, but with important caveats. When an attorney directs AI use as part of legal work, it may qualify for work product protection — but only with enterprise-grade tools that have proper confidentiality agreements, not consumer platforms. Your attorney has ethical obligations to protect your information and will use appropriate safeguards.
Your Case Deserves Real Privilege Protection
AI cannot protect your conversations. A real attorney can. Christine Leonard is a trial-tested family law attorney who fights for families across Northeast Florida. When you hire her, your communications are privileged, your strategy is protected, and your case is in the hands of a qualified professional — not an algorithm.
Book Your Consultation →Or call directly: (904) 392-4573