A recent national survey conducted in October 2025 by Kolmogorov Law, a California-based legal firm, has brought attention to the growing use of AI chatbots such as ChatGPT for legal guidance and the associated legal vulnerabilities. The poll of 1,000 adults who have used AI chatbots found that 56 percent have sought legal advice from these platforms, yet 50 percent remain unaware that their conversations can be subpoenaed and used as evidence in court.
The findings point to a significant gap between user expectations and current legal standards. Attorney-client privilege, which protects confidential communications between licensed lawyers and their clients, does not extend to interactions with AI systems. No federal or state laws currently grant similar protections to chats with chatbots, meaning companies like OpenAI may be required to produce conversation records under a valid subpoena or court order. OpenAI CEO Sam Altman has publicly acknowledged this limitation, stating in 2025 that conversations lack legal confidentiality and can be subpoenaed if stored.
The survey showed that 67 percent of respondents believe AI conversations should receive the same legal protections as those with lawyers or doctors. This view reflects a common assumption that digital interactions carry inherent privacy, similar to professional consultations. In practice, however, AI providers operate under terms of service that allow data retention for various purposes, and courts have increasingly treated digital records as discoverable evidence in civil and criminal matters.
Additional results indicate shifting behavior once risks become clear. Fifty-one percent of AI users reported they would be much more likely to consult a human lawyer if informed about subpoena possibilities. Sixty-five percent expressed concern about their chats being used in court. Thirty-four percent admitted to sharing confidential business or personal information with chatbots, further illustrating the potential for unintended exposure.
Public response to these issues favors stronger safeguards. Seventy-six percent of respondents supported government regulation requiring AI companies to provide legal privilege for user conversations. Forty-seven percent advocated for prominent warnings before each interaction to alert users to potential legal risks, a concept some describe as “digital Miranda rights.”
The survey arrives amid broader discussions on AI in legal contexts. Cases involving digital evidence have highlighted how chat logs and data retention policies can intersect with court proceedings. For instance, ongoing litigation has addressed preservation orders for user data, underscoring that privacy policies do not always shield information from legal demands. Experts note that while AI offers speed and accessibility for tasks like explaining laws or drafting basic documents, it cannot replicate the accountability, accuracy, or confidentiality of licensed professionals.
Southern Maryland residents, like others nationwide, may encounter these issues in everyday scenarios such as family law questions, contract reviews, or small business disputes. Local courts in counties including St. Mary’s, Calvert, and Charles follow standard evidentiary rules under Maryland law, where relevant digital communications can be admitted if authenticated and pertinent. The absence of privilege for AI chats means users here face the same exposure as elsewhere.
This development reflects rapid technology adoption outpacing legal frameworks. The American Bar Association and various state bars have issued guidance on AI use by attorneys, emphasizing competence, confidentiality, and supervision to avoid waiving privilege when lawyers employ AI tools. For the general public, the takeaway centers on awareness: interactions with public AI platforms create records that lack built-in protections.
The Kolmogorov Law survey, distributed through various news outlets in mid-December 2025, underscores the need for informed decision-making. As AI continues integrating into daily life, users must weigh convenience against potential consequences in legal settings.