The Perils of AI Confidantes: Lawyers Warn Against Sharing Sensitive Information with Chatbots

In an era where artificial intelligence chatbots are increasingly becoming digital confidantes, legal experts are issuing a stark warning: sharing sensitive information with these AI tools carries significant, often unseen, legal risks. Unlike the protected dialogues with medical professionals, lawyers, or therapists, conversations held with AI lack any form of legal privilege. This critical distinction means that what users consider private exchanges could, in the event of a lawsuit or government investigation, be compelled into evidence, exposing personal and confidential details to the public record. Therefore, individuals are strongly advised to exercise extreme caution, diligently review the terms of service of any AI platform they utilize, and fully comprehend the data retention policies to safeguard against potential legal entanglements stemming from their digital interactions.
The burgeoning trend of individuals confiding in AI chatbots, such as OpenAI's ChatGPT, for personal advice and even emotional support has raised alarm bells within the legal community. Lawyers specializing in AI-related legal frameworks emphasize that the absence of established legal confidentiality for these digital interactions leaves users vulnerable. Juan Perla, a partner at Curtis, Mallet-Prevost, Colt & Mosle LLP and a prominent figure in the firm's AI practice, underscored this point, highlighting that current legal frameworks offer no protection for users' chat data should it be sought in litigation. He stressed that people should think twice before revealing deeply personal information to AI, especially if they would typically reserve such discussions for a doctor, lawyer, or therapist.
This lack of legal protection extends to a broad spectrum of sensitive scenarios. For instance, any conversations with chatbots that touch upon workplace disputes, divorce proceedings, child custody battles, or even potential criminal activities could become discoverable evidence in relevant legal actions. Perla advises a simple but crucial rule: if a piece of information is something you would only share with a medical doctor, therapist, or lawyer, it should absolutely not be entered into an AI chatbot. Even attempts to anonymize or present information hypothetically may not entirely eliminate the risk. The most prudent course of action, according to legal experts, is to refrain from having sensitive conversations with AI chatbots altogether. While not every chat record is destined for public scrutiny, particularly those unrelated to potential legal disputes or crimes, the possibility remains a tangible risk that users must acknowledge.
Understanding how different AI platforms manage and retain user data is paramount for mitigating these risks. James Gatto, a partner at Sheppard Mullin and co-leader of the firm's AI industry team, pointed out that while some premium, paid versions of AI tools might offer enhanced privacy features, such as automatic data deletion, free public versions typically do not. He strongly recommends that users seeking to discuss sensitive matters choose tools that explicitly commit to deleting information, with such provisions clearly stipulated in their terms of service. Ultimately, the onus is on the user to conduct thorough due diligence and weigh the potential benefits against the inherent risks. Every digital record created—be it text messages or AI chats—carries a certain level of exposure. Therefore, individuals must ask themselves if they are truly comfortable with the possibility of their private information, once shared with an AI, falling into unintended hands.