Imagine joining a routine client call on Zoom. A few minutes into the conversation, you notice an unfamiliar participant: an AI notetaker.
Who invited it? Is it recording? Where will the transcript go? And who will have access to it?
AI notetakers built into platforms like Zoom and Microsoft Teams or offered through services like Otter.ai, promise real productivity gains. They let attorneys focus on clients instead of typing notes, generating meeting summaries, and capturing follow-up actions.
But for law firms, a recorded conversation is more than a convenience. It raises questions about consent, client confidentiality, attorney-client privilege, and potentially discoverable records.
Can law firms use AI notetakers safely?
Yes—with appropriate safeguards. The essentials are obtaining consent, protecting confidential information, reviewing AI-generated summaries, and managing records responsibly.
Here are four practical steps to get started.
First, Understand What an AI Notetaker Really Is
Think of an AI notetaker as three things at once:
- A recorder: It captures conversations, potentially triggering recording and consent laws.
- A technology vendor: Depending on the product and configuration, sensitive conversations may be processed or stored by a third party.
- A records system: It creates recordings, transcripts, and summaries that may become part of the firm’s electronically stored information.
This distinction matters. What looks like a simple productivity tool can introduce risks across all three areas.
Step 1: Get Consent Before Recording
California generally requires all parties to consent before recording a confidential conversation under Penal Code § 632.
The New York City Bar’s Formal Opinion 2026-2 also recommends obtaining consent from all participants before recording, including non-clients such as opposing counsel and witnesses.
What should your firm do?
- Disable automatic meeting recording and AI notetaker auto-join.
- Announce the notetaker and obtain permission before recording begins.
- If anyone objects, turn it off.
- Include appropriate recording provisions in engagement letters.
A simple consent script:
“I’d like to use an AI notetaker so I can focus on our conversation. We will review the summary before relying on it. Is everyone comfortable with that?”
Remember: A bot’s name appearing in the participant list is not informed consent.
Step 2: Protect Client Confidentiality
Not every AI notetaker handles information the same way. Some use external cloud processing, others offer enterprise-level protections, and retention and data-use practices vary.
Before approving a tool, your firm should understand where its data goes and who can access it.
The ABA’s Formal Opinion 512 and the California State Bar’s updated AI guidance emphasize lawyers’ continuing obligations to safeguard client information and exercise competent oversight.
Five questions to ask your AI notetaker vendor:
- Will our recordings or transcripts be used to train AI models?
- Where is our data stored, and who can access it?
- What security and confidentiality protections are provided?
- Can we control retention and request deletion?
- How are subpoenas, security incidents, and data breaches handled?
Get the answers in writing.
California attorneys should also prepare for SB 574, effective January 1, 2027.
The new law introduces statutory requirements for attorneys using generative AI, including protections for confidential information and verification of AI outputs. Firms should evaluate whether their notetaking tools fall within its scope.
Step 3: Do not Confuse a Polished Summary With an Accurate One
AI summaries can be remarkably convincing—even when they are wrong.
A notetaker might misidentify a speaker, incorrectly transcribe a settlement figure, or summarize legal advice that was never actually given.
Imagine a summary stating that your client agreed to a settlement amount when the client merely agreed to consider it.
That small distinction could become significant later.
The NYC Bar’s Formal Opinion 2025-6 emphasizes independent review of AI-generated transcripts and summaries before they are preserved and relied upon.
A practical rule: Never treat an AI-generated summary as the firm’s authoritative record until a person responsible has reviewed it.
Verify names, dates, figures, commitments, and especially statements attributed to counsel.
Step 4: Decide What to Keep—and What to Delete
Every retained recording or transcript creates another information-management obligation.
Depending on the circumstances, these materials may be subject to discovery, preservation duties, or privilege disputes.
In United States v. Heppner (2026), a federal court rejected privilege and work-product claims over an individual’s independent conversations with a consumer AI system. Although the case did not concern enterprise AI notetakers, it highlights the danger of assuming that information shared with an AI provider automatically remains protected.
Law firms should:
- Establish documented retention periods for recordings and transcripts.
- Limit access to authorized personnel.
- Delete unnecessary records in accordance with firm policy and applicable preservation obligations.
- Include AI-generated meeting records in legal-hold procedures.
The best time to decide your retention policy is before the first recording—not after receiving a discovery request.
A Real-World Warning: The Otter.ai Privacy Litigation
The risks are no longer hypothetical.
In In re Otter.AI Privacy Litigation, a federal class action filed in California, plaintiffs alleged that Otter.ai recorded or processed meeting conversations without obtaining appropriate consent from every participant.
On August 13, 2026, the court partially denied Otter.ai’s motion to dismiss, allowing portions of the litigation to proceed. The allegations have not been established as liability.
The lesson for law firms is straightforward: The people affected by an AI notetaker are not necessarily the people who chose to use it.
Your client, opposing counsel, or a witness may have had no say in inviting the technology into the conversation.
The GetAIReady Decision Checklist : Should You Record This Meeting?
Before turning on an AI notetaker, ask four questions:
- Consent: Have all participants agreed to the recording?
- Confidentiality: Is the tool approved and appropriately secured?
- Necessity: Is there a legitimate reason to record this particular conversation?
- Control: Who will review, access, retain, and eventually delete the recording?
If the answer to any question is uncertain, keep the notetaker off until the issue is resolved.
As a conservative default, avoid recording sensitive legal strategy discussions, witness interviews, settlement negotiations, and prospective-client consultations unless there is a considered reason and appropriate safeguards.
Your Law Firm’s Action Checklist
Start with these five actions this week:
- Identify all AI notetakers currently being used across the firm.
- Disable unauthorized autojoin and automatic recording.
- Establish a standard consent procedure.
- Identify meeting categories where recording should normally remain off.
- Assign responsibility for approving tools and reviewing AI-generated summaries.
Then develop a one-page firm policy covering vendor safeguards, retention, access, and staff responsibilities. Pilot the policy in one practice group before expanding.
The Bottom Line
AI notetakers illustrate a larger challenge facing law firms: capturing AI’s productivity benefits without introducing risks that nobody anticipated.
The technology is easy to deploy. Responsible adoption requires more thought.
At getAIready, we help legal professionals move beyond AI awareness to practical, responsible adoption through hands-on training and real-world workflows.
This article is for educational purposes and does not constitute legal advice. Attorneys should consult applicable jurisdictional rules and current legal authorities before implementing AI recording policies.