- within Technology topic(s)
- with Senior Company Executives, HR and Inhouse Counsel
- with readers working within the Business & Consumer Services, Insurance and Law Firm industries
AI platforms such as ChatGPT, Gemini, and Grok are quickly becoming an everyday tool – much in the way that "googling" something has become an integral part of daily life. Without commenting on the positives or negatives of its use, AI has become a trusted advisor for many individuals, often serving as the first source that they consult for advice.
However, there are important limitations to be aware of - every so often an event occurs to remind you that AI is not always your friend.
One such instance was for the defendant in a recent criminal case in the United States District Court for the Southern District of New York. With the prevalence of AI use for legal matters increasing exponentially, it was inevitable that courts would need to address the intersection of AI usage and the doctrines of attorney-client privilege and work product. In this case, the intersection was created by the client/defendant's use of AI to develop strategiesto subsequently discuss with his attorneys. The Court examined these issues and found, in a bench ruling, that the client/defendant's use was not protected and therefore the Court ordered disclosure.
While AI can be a valuable and powerful resource for the legal field, this ruling should serve as a clear warning for both clients and attorneys to exercise caution when utilizing it to develop legal strategies or otherwise assist in legal matters. This decision also provides a tentative guideline for AI usage – or its avoidance – in similar legal contexts.
But, of course, specifics matter. In the case at hand, United States v. Heppner, the client/defendant used AI to prepare reports and outline both a factual and legal defense strategy after receiving a grand jury subpoena, with the intention of implementing such a strategy if he were later indicted. He subsequently shared these reports with his defense counsel. In ruling that neither the attorney-client nor work product privileges applied to these reports, the court appeared to make its determination based on several specific factors:
- The reports were not prepared by an attorney and did not reflect the defense counsel's strategy, as they were prepared before the attorney and the client spoke.
- The AI tool used by the client/defendant was an open-source platform that included a non-confidentiality disclaimer, and thus the client had no expectation of privacy.
- Defense counsel did not direct the client/defendant to utilize AI in connection with its representation.
Open-sourced AI platforms - which allow anyone to use them – explicitly state in their terms and conditions that user inputs are collected and processed. Such information is, by default, not private or encrypted. Even if these inputs are purportedly used solely to improve future responses, there is no confidentiality guarantee comparable to attorney-client conversations. Moreover, even if you delete the chat, the company typically retains conversation data in its systems for future use. We expect that the government will now routinely subpoena these open-source AI platforms to gain critical insight into the thoughts of a criminal defendant or any litigant who defends a government lawsuit.
We are not so naïve as to advise clients not to use AI – its power and benefits are undeniable. However, in light of this ruling, clients should keep the following in mind: Submitting documents or information into public AI tools could jeopardize, or even forfeit, the protections of attorney-client privilege and the resulting materials do not constitute attorney work product. We strongly advise that you first discuss any legal questions or confidential information with your attorney before using AI for legal matters. Attorney-client privilege and confidentiality exists specifically to protect you. A client who uses AI, even during representation by an attorney, might find his entire defense exposed before he even get his day in Court. AI is not an attorney, and therefore no one should expect that their queries to AI and the AI-generated responses are protected.
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.
