Lawyers have an ethical duty to understand how AI usage not only implicates their own practice of law, but also how AI impacts the attorney-client relationship. Clients use AI to summarize emails, draft outlines, brainstorm arguments, and attempt to address situations before ever calling a lawyer. That behavior has created a new intake risk before a client matter even begins. Increasingly, the first intake questions are no longer just “When did this happen?” or “Was that agreement in writing?” Intake questions must also include “Did you ask AI about this?”
Confidentiality and attorney‑client privilege have always depended on control over information. Historically, that meant being mindful of emails, documents, and conversations. Generative AI changes the medium, but not the privilege itself.
Many publicly available AI tools retain user prompts, permit internal submission review, use customer inputs to improve AI models, and operate under terms or conditions that defeat any reasonable expectation of confidentiality.
From a privilege standpoint, the problem arises before output is generated. Once sensitive facts are disclosed to a third party (which includes an AI chatbot) without safeguards, the legal status of that information may change.
Most AI ethics conversations focus on hallucinations, bad citations, or drafting errors. Those issues can sometimes be cured by attorney verification and diligence. However, once privilege is lost, it sometimes is impossible to restore.
When clients input internal investigation details, names of employees or whistleblowers, draft agreements or deal structures, litigation strategy or timelines, or regulatory exposure or compliance issues into an unprotected AI tool, lawyers may later inherit tainted facts at intake.
Importantly, recent federal decisions make clear that sending AI‑generated materials to counsel after the fact does not cure a loss of privilege where the information was originally disclosed to a third‑party AI tool without confidentiality protections.
Questions lawyers should consider asking clients, whether new or existing, may include the following:
When framed to a client that these questions are aimed at assessing risk and exposure, a lawyer both educates the client about AI usage and the dangers and sets expectations for the representation. The lesson is simple but uncomfortable: AI use by clients can create privilege problems before the lawyer is ever retained.
Kentucky lawyers must be aware of the ethical implications of AI’s increased prevalence and recognize that it is not just a purely client-driven risk. In March 2024, the Kentucky Bar Association issued Ethics Opinion KBA E-457, making clear that existing Rules of Professional Conduct already apply to the use of AI in legal practice.
Rules of professional conduct implicated:
Practice tips
When clients use AI before contacting counsel, ethical duties can become apparent at the intake stage. As KBA E-457 makes clear, ethical compliance in the age of AI begins earlier than many lawyers expect, often before the lawyer is ever formally retained.