Your AI Meeting Notetaker Might Be a Legal Liability — What the Granola Lawsuit Means for Your Business

This article summarizes current litigation and legal analysis for informational purposes. It is not legal advice — any business using or deploying AI recording tools should consult qualified counsel about consent and privacy law obligations in its specific jurisdictions.

Key Takeaways

• A proposed class action filed July 30, 2026 in federal court accuses AI meeting-notetaker Granola of recording conversations without most participants' knowledge and using those recordings to train its own AI models by default.
• Granola is a venture-backed company valued at roughly $1.5 billion after a $125 million Series C in March 2026, with a customer list that reportedly includes Vanta, Gusto, Thumbtack, Asana, Cursor, and Mistral AI.
• This isn't an isolated case: Otter.ai has been defending a closely related consent lawsuit in the same federal district since August 2025, and Fireflies.ai faces a separate biometric-privacy claim filed in Illinois in December 2025.
• The legal theory at the center of the Granola case is specifically about "bot-free" tools — software that captures audio locally without ever visibly joining a call — which the complaint argues removes the one clear notice signal (a visible bot in the participant list) that earlier cases like Otter's have had to grapple with.
• For any business using these tools, responsibility for disclosure and consent currently falls on the person running the software in the meeting — not automatically on the vendor that built it.

What the lawsuit actually alleges

The case, Chamberlain v. Granola, Inc., No. 3:26-cv-07926 (N.D. Cal., filed July 30, 2026), was brought by Florida resident Tarra Chamberlain on behalf of a proposed nationwide class and a California subclass. It names two corporate defendants: Granola, Inc., a Delaware corporation, and Granola Labs Ltd., a UK-registered entity. The complaint alleges Granola's AI notetaker captures microphone and system audio during virtual and in-person meetings without disclosing this to most participants, then uses those recordings — by default, on the company's Free and Business plans — to train Granola's own AI models.

The specific design detail at the center of the case matters: unlike some AI meeting tools that join a call as a visible participant (a "bot" that appears in the attendee list), Granola's desktop application captures audio locally from the user's device without joining the call at all. Legal analysis published by Barnes & Thornburg attorneys frames this as a meaningfully harder consent problem than the visible-bot model: "The Granola complaint tests something narrower and, for the category, more awkward: whether a tool that never appears at all can rely on the account holder to disclose it," as one industry analysis of the case put it. Because the people whose voices are captured are frequently not Granola account holders themselves, they have no account setting to check and no way to opt out — the only person in the meeting with visibility into whether recording is happening, or control over it, is the Granola user.

Why the legal theory faces a real hurdle — and why that doesn't make the risk go away

The lead federal claim — a wiretap claim under the Electronic Communications Privacy Act (ECPA) — faces a structural obstacle described in legal commentary on the case: ECPA generally only requires one party to a conversation to consent for a recording to be lawful under federal law, and that one party can be the Granola user themselves. If the federal claim doesn't hold up, the case would rely on California state wiretap statutes, which generally require all-party consent — a higher bar, but one that raises separate questions about nationwide class certification given that the named plaintiff is a Florida resident, not a California one.

That legal uncertainty is precisely why this matters for businesses regardless of how the Granola case itself is ultimately resolved. This case is explicitly part of a broader litigation pattern, not an isolated dispute: Otter.ai has been defending a related consent claim in the same federal district since August 2025 with no ruling issued as of early August 2026, and Fireflies.ai faces a separate biometric-privacy claim in Illinois. Legal commentary tracking this trend has also connected it to a wider wave of AI-and-privacy litigation, including a 2025 jury verdict against Meta over health-data collection and a 2026 injunction in a related privacy case — signaling that courts are actively engaging with, not dismissing, these theories across multiple AI product categories.

What Granola says in its own defense

Granola's published help-center documentation, cited separately from the litigation itself, states: "It is best practice to make sure participants know when you use AI notetaking tools," and describes on-screen signals for virtual calls — a watermark on the user's camera feed and an automated chat message, both visible to other participants when the feature is enabled. The company's documentation also states that any user on any plan can opt out of having their data used for model training in settings, and that Enterprise customers are opted out by default with admin-level enforcement across an organization. As of early August 2026, neither Granola, Inc. nor Granola Labs Ltd. had publicly responded to the allegations on the merits — these remain allegations, not findings by any court.

What this means for your business right now

Whether or not your company uses Granola specifically, this case — and the broader litigation pattern around it — creates real, immediate exposure for any business using AI note-taking or transcription tools in meetings.

1. Don't assume the vendor's default settings satisfy your legal obligations. The core allegation across these cases is that consent defaults to "on" for recording and model training, placing the actual disclosure burden on whoever is running the tool in a given meeting — meaning your employees, not the vendor, may bear practical responsibility for informing other participants.
2. Set a clear internal policy on disclosure before, not after, a meeting. A simple, consistent verbal or written notice at the start of any meeting where an AI notetaker is active — "this call is being recorded and transcribed by [tool]" — is a meaningfully stronger position than relying on an on-screen watermark that outside participants may not notice or understand.
3. Check whether your organization's plan opts you into AI model training by default. Several tools in this category, including Granola on its Free and Business tiers per the complaint, reportedly use captured data to train their own models unless a user actively opts out — a setting worth auditing across your organization's actual usage, not just its written contract terms.
4. Treat multi-party and cross-state calls as higher-risk than internal, single-state meetings. Because consent law varies significantly by state — some require only one party's consent, others require all parties — a call involving participants in a state with stricter all-party consent requirements carries meaningfully different legal exposure than an internal call among employees in a single one-party-consent state.
5. Loop procurement and legal counsel into AI notetaker vendor selection, not just IT. As one analysis of the case put it for IT departments specifically: "no stored recording is not the same thing as no recording risk" — the interception itself, not just what happens to the data afterward, is where the current legal exposure sits.

Frequently Asked Questions

Is my business at legal risk just for using a tool like Granola, Otter, or Fireflies? The current litigation targets the vendors directly, not their business customers, and no court has yet ruled on the merits of any of these claims. That said, businesses that deploy these tools without clear internal consent practices may face their own separate exposure under state wiretap or privacy laws, independent of how the vendor-focused litigation resolves — which is why consulting counsel about your specific state and industry obligations is worth doing now rather than waiting for these cases to conclude.

Does a visible "bot" joining the call solve the consent problem? Not necessarily, and it's an open legal question. Otter.ai's tools generally join calls as a visible bot, yet the company has still been defending a related consent lawsuit in the same federal court since August 2025, with no ruling issued as of early August 2026 confirming that visible participation alone satisfies consent requirements.

What is the difference between one-party and all-party consent laws? One-party consent laws (the federal ECPA default, and the law in most U.S. states) require only one participant in a conversation to agree to recording — which can be the person using the recording tool. All-party consent laws, in effect in California and several other states, require every participant to consent, which creates a meaningfully higher compliance bar for tools that don't clearly notify everyone on a call.

Has any court ruled that AI meeting-notetaker recording without full disclosure is illegal? Not as of early August 2026. These are allegations in ongoing litigation, and companies named in these suits have not been found liable by any court. The purpose of tracking this litigation isn't to treat the outcome as settled — it's to recognize that the underlying legal theory is being taken seriously enough by courts to proceed past early dismissal stages in related cases, which is itself a meaningful signal for risk planning.

Sources & References

• Chamberlain v. Granola, Inc., No. 3:26-cv-07926 (N.D. Cal., filed July 30, 2026)
• National Law Review / Mondaq, "AI Note-Taking Tools, Silent Recording, and the Consent Problem: What the Granola Class Action Means for Companies Building and Deploying Conversation Capture Tools" (Barnes & Thornburg LLP)
• ppc.land, "Granola sued over bot-free notetaker and AI training use"
• Basil AI, "Chamberlain v. Granola: The Invisible AI Notetaker Wiretap Lawsuit"

Related Reading

For a broader look at how AI tools are creating new categories of business risk beyond this specific case, see PrimeWorldMedia's coverage of AI agents for business, which covers the related governance question of tracking which AI tools are operating inside your company at all.