A recent federal court decision should give every divorce litigant pause before turning to an AI chatbot for legal guidance. In United States v. Heppner, a New York federal judge addressed a question of first impression: whether a party’s communications with a publicly available AI platform during a pending legal matter are protected by privilege or work product doctrine. The defendant, charged with securities and wire fraud, turned to Claude for help during a pending government investigation. He fed information he had learned from his defense counsel into the AI tool, generated 31 documents of prompts and responses, and then transmitted those documents back to his lawyers. When federal agents seized his computer and those documents, he argued they were shielded from disclosure.
The judge disagreed because AI is not an attorney — it has no law license, owes no duty of loyalty, cannot form an attorney-client relationship, and is not bound by confidentiality obligations or professional responsibility rules. Critically, the court also found that the applicable AI platform’s privacy policy — which provided that prompts could be used to train the model and that inputs may be disclosed to third parties — meant the defendant had effectively disclosed his communications to a third party, destroying any reasonable expectation of confidentiality.
While Heppner arose in a criminal prosecution, its reasoning could conceivably apply in civil proceedings, including divorce cases. Illinois discovery rules allow for broad discovery, and in contested divorce proceedings, parties routinely seek documents, communications, and electronically stored information pertaining to the issues in dispute — including finances, parenting fitness, and conduct during the marriage. If a spouse queries an AI chatbot about how to hide assets, undervalue a business, or strategize about custody, those chat logs could very well be discoverable. Any time a party poses a question to an AI tool to analyze legal issues, prepare for litigation, or seek legal advice, they may be creating an electronic paper trail ultimately discoverable by his or her adversary.
Even more concerning is the prospect of a client inputting an attorney’s otherwise confidential communication into an AI tool for analysis. Under Heppner, not only would the output be discoverable, but the previously privileged input as well.
The bottom line for anyone going through — or anticipating — a divorce in Illinois is straightforward: treat your AI conversations the same way you would treat any other written communication that could end up in front of a judge…carefully. AI platforms are an incredible resource, but any conversations that should not see the light of day should be addressed with an attorney alone.
Matthew D. Elster, Partner
For more on Mr. Elster, please visit: www.beermannlaw.com/team/matthew-d-elster.
