AI Notetaker Lawsuits Widen as Granola and Fireflies Face Claims

The legal fight over AI notetakers is no longer about one company

For most of the past year, the legal risk around AI meeting assistants was attached to a single defendant: Otter.ai. That is changing. Granola and Fireflies.ai now face privacy litigation of their own, and a federal ruling in the Otter case gave plaintiffs a roadmap that could apply to the wider category of cloud-based notetakers.

The key moment came on August 13, 2026, when Judge Eumi K. Lee in the Northern District of California partially granted Otter's motion to dismiss in In re Otter.AI Privacy Litigation, a set of suits filed in August and September 2025 and consolidated that October.1 The claims that survived were the most consequential ones: the federal Wiretap Act claim, the California Invasion of Privacy Act (CIPA) claim, and two claims under Illinois's biometric privacy law. They now move into discovery.14

Why the Otter ruling matters

Otter's defense depended on how its bot should be characterized. The company argued that its Notetaker joins calls only when invited and acts as the meeting host's recording tool, essentially an extension of the customer who deployed it.4 On the CIPA claim specifically, Otter said the bot was an authorized participant and therefore a party to the conversation, not an outsider listening in.3

The court did not accept that argument at this stage. Plaintiffs alleged that Otter keeps the recorded conversations and voiceprints and uses them to train its own models. Judge Lee found that this was enough to plausibly cast Otter as a third-party eavesdropper with its own commercial interest.34 The ruling distinguished an earlier decision, Graham v., which suggests that how a vendor uses the data, not only how the tool is invited into a meeting, may decide liability.3

The limits of the decision should be stated clearly. It is a pleading-stage ruling and does not establish that Otter broke any law.3 Still, getting past a motion to dismiss changes the stakes. Patrick E. Zeller, general counsel of JetStream Security, warned that discovery itself creates problems for organizations that use these assistants. He argued that the companies most exposed will be those that never assessed the risk at all.4

Granola: invisibility becomes the allegation

On July 30, 2026, a class action was filed against Granola in the Northern District of California: Chamberlain v. Granola, Case No. 3:26-cv-07926-EMC.12 The case targets a different product design. Granola markets itself as "bot-free," meaning no visible participant joins the call. The 38-page complaint alleges that the tool captures every participant's speech without consent and uses transcripts to train AI models by default.2

That changes the argument. With Otter, the visible bot raised questions about who counts as a party to a conversation. With Granola, plaintiffs present the lack of any visible signal as part of the alleged harm.2 The case is at an early stage.1

Fireflies and the biometric angle

Fireflies.ai faces Cruz v. Fireflies.AI Corp. along with a second Illinois case, both dating to December 2025. They focus on voiceprints under Illinois's biometric privacy law, and neither has been settled.12 Competitors Fathom and Read AI had no identifiable lawsuits as of late August 2026.1

Where the analysis diverges

Observers broadly agree on what the Otter ruling says. They differ on what businesses should take from it. One analysis argues that "the bot was never the legal problem." In that view, the issue is what happens to the data after it is captured.1 Another goes further and says the only real safe harbor is architectural: transcripts that never leave the user's device cannot be intercepted or subpoenaed.2 That conclusion comes from a party with an obvious interest in on-device processing, so it should be read as advocacy, not settled law. Compliance-focused commentary is more measured and centers on consent practices and discovery readiness.4

The framing of a broader "litigation wave" also needs care. The documented cases involve Otter, Granola, and Fireflies. Claims about other major platforms being drawn into similar suits are not supported by the court record outlined here.

The takeaway

The most likely reading is that plaintiffs are testing every major notetaker design: visible bots, invisible capture, and voiceprint processing.2 The common thread is model training. Once a vendor uses meeting content for its own purposes, the argument that it is a neutral tool becomes much harder to make. Buyers should treat default training settings, participant consent, and data retention as legal questions, not just product preferences. Vendors should expect discovery in the Otter case to shape the rules for this category long before any verdict.