Your client’s use of AI for understanding the facts of their family law case is increasing astronomically. But are the written exchanges between clients and their AI platforms protected by either the attorney-client privilege or the work product doctrine? A New York federal court weighs in.

AI and Attorney-Client Privilege
Generative artificial intelligence tools have become increasingly prevalent across various domains of human activity. It has reliably been estimated, for instance, that more than 50% of all United States households have adopted AI in some form.
Generative AI focuses on creating the text, images, and music we use in our practice and personal lives. Generative AI systems, like ChatGPT and Claude, are the best-known subset of AI, but they are not the most common. The most widely used AI systems are Search Recommendation Engines, like Netflix, Amazon, and Internet Search Rankings, such as Google search, Bing, YouTube, and TikTok.
In one New York case, a man was arrested for securities fraud among other charges, and federal agents seized numerous documents and electronic devices, including over 30-pages memorializing communications the client had with Anthropic’s Claude generative AI platform.
The client used AI to prepare reports outlining his defense strategy, and outlined his arguments regarding the facts and the law that the government would use in charging him with securities fraud. At the hearing to exclude the documents, the client’s attorneys argued the reports the client prepared using Claude were used “in anticipation of a potential indictment” and for the purposes of speaking with his lawyers to prepare a defense.
The government moved for a ruling that the AI documents are not protected by the attorney client privilege or the work produce doctrine.
Florida Attorney-Client Privilege
I recently wrote an article about the increasing use of AI by lawyers and clients. The 2023 Future Ready Lawyer Report showed that 76 percent of legal professionals in corporate legal departments and 68 percent of law firms use generative AI at least once a week.
In Florida, a client has a privilege to refuse to disclose, and to prevent any other person from disclosing, the contents of confidential communications when such other person learned of the communications because they were made in the rendition of legal services to the client.
Additionally, the personal views of attorneys as to how and when to present evidence, their evaluation of its relative importance, their knowledge of which witness will give certain testimony, personal notes and records as to witnesses, legal citations, proposed arguments, diagrams and charts they may refer to at trial for their convenience, but not to be used as evidence, come within the general category of work product.
No Florida Court has published an opinion to date whether the independent use of public-facing generative AI in legal matters forfeits the protections afforded under the attorney-client privilege and the work-product doctrine, leaving such materials vulnerable to disclosure.
Is AI Use Privileged
The New York judge ruled that the client’s pages of AI-generated materials were subject to disclosure to the government because any privilege was waived when the client shared information with a third-party – the Claude AI platform – thereby failing to maintain the information’s confidentiality.
The trial court also rejected the application of the work-product doctrine, ruling the AI platform is not a lawyer, even though commentators have argued that whether Claude AI is an attorney is irrelevant because a user’s AI inputs, rather than being communications, are more akin to the use of other Internet based software, such as cloud-based word processing applications.
To the court, all “[r]ecognized privileges” require, among other things, “a trusting human relationship,” such as, in the attorney-client context, a relationship “with a licensed professional who owes fiduciary duties and is subject to discipline.
AI is not a lawyer and was not used at a lawyer’s direction. In other words, because the AI-generated materials were created independently by the client himself, the materials could not qualify as protected work product.
The recent New York ruling highlights that a client’s independent use of generative AI in a divorce or family law case may forfeit the protections afforded under the attorney-client privilege and the work-product doctrine. Clients should consider confidentiality before using AI and involve their lawyers early when using AI to analyze your legal case.
The opinion is here.