Artificial intelligence (AI) tools are quickly becoming part of everyday business. AI note takers, transcription tools, and meeting assistants can be especially useful. They allow participants to focus on the conversation instead of taking detailed notes, they can generate transcripts, summaries, and action items after the meeting.
But when a meeting involves legal advice, these tools can create risks that are easy to overlook. The convenience of an AI note-taker may come at the cost of confidentiality, privacy, and attorney-client privilege.
Attorney-client privilege protects confidential communications between a client and their attorney made for the purpose of obtaining or providing legal advice. In practical terms, it can allow a client to refuse disclosure of protected communications in litigation, during a deposition, in response to a subpoena, or in connection with a discovery request.
The privilege, however, is not automatic in every situation. One of its key requirements is confidentiality. If a privileged conversation is shared with people outside the attorney-client relationship, recorded without proper safeguards, saved in a broadly accessible folder, or transmitted to a third-party vendor, or uploaded to an AI tool hosted by a third-party provider, the privilege may be weakened or challenged. This risk is particularly significant if the AI vendor’s terms of service permit data sharing, storage, or use for model training, especially when using low-cost or publicly available AI tools with permissive privacy policies.
AI note takers and recording assistants may seem harmless. Often, they quietly join a Zoom, Teams, or Google Meet call and operate in the background. But in a legal meeting, the tool may function like an outside participant. Depending on how the tool works, the meeting audio, transcript, summary, or notes may be uploaded, stored, analyzed, shared, reviewed, or retained by a third-party provider. That can raise difficult questions about who had access to the information and whether the legal discussion remained confidential.
Recent litigation highlights these concerns. In Brewer v. Otter.ai Inc., No. 5:25-cv-06911, filed in the Northern District of California on August 15, 2025, the plaintiff alleged that an AI notetaking application recorded and transcribed conversations without proper consent from all participants. The allegations in that case focus on privacy and consent, but the dispute illustrates a broader point: companies should understand exactly how these tools collect, store, use, and share meeting information before allowing them into sensitive conversations.
Courts are also beginning to address how privilege applies when individuals use public AI tools in legal contexts. In United States v. Heppner, the court held that materials a criminal defendant created using the free, consumer version of Claude were not protected by attorney-client privilege or the work-product doctrine. United States v. Heppner, 820 F. Supp. 3d 292 (2026). The court focused on traditional privilege principles, including confidentiality, attorney involvement, and whether the materials were created within a protected attorney-client relationship.
The takeaway is not that all AI tools are the same. They are not. Public, consumer-facing AI chatbots raise different issues than enterprise AI tools or closed systems that incorporate strong confidentiality protections. A fully local AI tool that does not transmit information outside the company may raise fewer concerns about disclosure to third parties. AI agents that interact with multiple platforms may require a more detailed review. AI note takers and recording tools fall somewhere in between: the privilege analysis often depends on the vendor’s terms, data practices, confidentiality obligations, access controls, and the purpose of the meeting.
Clients should be especially careful in common situations such as:
- an AI assistant automatically joining a meeting with counsel;
- recording or transcribing a legal strategy call without first discussing it with the attorney;
- automatically distributing a transcript or meeting summary to participants;
- saving AI-generated notes or recordings in a shared company workspace or other location accessible to individuals outside the attorney-client relationship;
- forwarding AI-generated summaries to colleagues, consultants, or other third parties who are not necessary to the legal representation; or
- responding to discovery requests, subpoenas, or government investigations that may require production of AI-generated recordings, transcripts, or summaries.
AI tools can be useful, but convenience should not come at the expense of confidentiality. Before using an AI note taker in a legal meeting, stop and ask: Could this affect privilege or expose sensitive information? When in doubt, consult your attorney first.
Disclaimer: This article is provided for general informational purposes only. It does not constitute legal advice and does not create an attorney-client relationship. For advice regarding your specific situation, you should consult a qualified attorney who can provide guidance tailored to your individual circumstances.
