Otter.ai is facing four consolidated federal lawsuits over how its AI meeting bot records participants without explicit consent. The cases, filed between August and September 2025, aren't really about Otter. They're testing whether an entire product category, AI meeting recorders that auto-join and record calls, violates decades-old wiretap laws when participants never agreed to be recorded.
Fireflies.ai faces separate biometric privacy suits in Illinois. The plaintiffs in both cases share something significant: they never had accounts with either service. They were simply in meetings where someone else had enabled the recording bot.
The legal question affects anyone running recorded calls. If your current setup auto-joins meetings and starts recording without getting affirmative consent from every participant, these cases are testing the exact design pattern you're using.
What Is the Otter.ai Class Action About?
The consolidated case In re Otter.AI Privacy Litigation bundles four separate suits filed in the Northern District of California. The first, Brewer v. Otter.ai Inc., was filed August 15, 2025, by Justin Brewer, who alleges he was recorded during a February sales call despite never having an Otter account.
According to the complaints, OtterPilot (now rebranded as Otter Meeting Agent) syncs with users' calendars and auto-joins scheduled calls as a visible participant. It records audio, transcribes in real time, takes screenshots, and captures speaker voiceprints. Non-account holders receive no disclosure before recording starts.
The legal claims span the federal Electronic Communications Privacy Act, California's Invasion of Privacy Act, Illinois's Biometric Information Privacy Act, and the Computer Fraud and Abuse Act. ECPA allows damages of the greater of $10,000 per violation or $100 per day. California's statute runs to $5,000 per violation. With Otter's reported 35 million users processing over a billion meetings, the exposure is substantial.
A motion-to-dismiss hearing is scheduled for May 20, 2026. Judge Eumi K. Lee's ruling will be the first federal test of whether wiretap statutes reach AI bots that auto-join video calls without participant consent.
The Fireflies Biometric Privacy Lawsuits
Fireflies.ai faces two separate suits under Illinois's Biometric Information Privacy Act (BIPA). Cruz v. Fireflies.AI Corp., filed December 18, 2025, alleges the platform's speaker recognition feature generates voiceprints from meeting participants without consent or disclosure.
Unlike traditional wiretap claims, BIPA cases focus on biometric data collection. When Fireflies identifies different speakers in meetings, it necessarily creates voiceprints, which Illinois law classifies as biometric identifiers. The complaint alleges three specific violations: no public policy for biometric data retention, no written notice to participants about voiceprint collection, and no written consent from participants.
BIPA allows $1,000 per negligent violation and $5,000 per intentional violation, plus attorneys' fees. The statute doesn't require plaintiffs to prove actual harm, making it an effective vehicle for class action litigation. A second case, Fricker v. Fireflies.AI Corp., filed in March 2026, makes substantially similar allegations.
The biometric angle creates complexity beyond simple recording consent. Even if participants know a meeting is being recorded, they may not realize their voice is being converted into a permanent biometric identifier.
Why These Cases Matter for All AI Meeting Recorders
These lawsuits target a design pattern shared across the industry: auto-joining meetings and recording without explicit consent from every participant. The tool name is less relevant than the workflow. Any platform that starts recording based on calendar integration, without a consent mechanism that lets participants genuinely decline, operates in the same legal territory.
For sales teams, the risk is acute. External prospects haven't agreed to anything beyond attending a meeting. In a remote environment where participants can dial in from all-party consent states like California, Illinois, or Washington, the generic "this meeting may be recorded" disclaimer in a calendar invite doesn't satisfy legal requirements.
The consent gap isn't theoretical. I've been in dozens of sales calls where participants were clearly surprised by a recording notification. Some left immediately. Others stayed but seemed uncomfortable throughout. Neither outcome serves the meeting's purpose.
Recording Meetings Compliantly: The Technical Requirements
Compliant recording isn't about finding the right disclaimer language. It's about giving participants a real choice before recording starts, with a meaningful consequence if they decline.
The basic requirements work regardless of which platform you're using:
-
Collect consent before the recording begins, not during it. Verbal confirmation at call start creates social pressure. A pre-meeting consent screen is more defensible.
-
Make declining a real option. If participants can't say no without losing their seat in the meeting, it's not genuine consent.
-
Document everything. Know which calls were recorded, when consent was collected, and what your data retention policy covers.
-
Don't override refusals. If someone declines, the recording stays off for that entire session.
For Scriptivox users handling recorded content, the consent question typically comes before transcription. But if you're generating transcripts from meetings where consent wasn't properly collected, you inherit the same legal exposure. The transcript is evidence the recording happened.
Meeting Recording Consent in Different Jurisdictions
US federal law requires one-party consent, meaning one person on a call can legally record it. But all-party consent states including California, Illinois, Maryland, Connecticut, Pennsylvania, Washington, Oregon, Montana, and New Hampshire require everyone to agree.
In a remote environment, you don't always know which standard applies. Participants can dial in from anywhere. The simplest approach is universal consent collection, which satisfies the strictest requirements regardless of participant location.
European teams operate under GDPR, where recording constitutes processing personal data and requires a lawful basis before processing starts. Consent is the cleanest basis available, but it must be freely given, specific, and unambiguous. A calendar invite with fine print doesn't meet that standard.
For teams handling sensitive conversations, AI transcription legal compliance requirements extend beyond just consent to data handling, retention, and cross-border transfers.
The Current State of AI Meeting Recorder Compliance
As of April 2026, no court has ruled that using AI meeting recorders is illegal. But both the Otter and Fireflies cases highlight design assumptions that may not survive federal scrutiny.
The industry response has been mixed. Some platforms have added consent collection features, often buried in settings menus and disabled by default. Others maintain that terms of service language placing responsibility on account holders is sufficient protection.
What's clear is that the "record first, ask questions later" approach is under legal pressure. The May 20 hearing will provide the first indication of how federal courts view auto-recording without participant consent.
For anyone running recorded meetings, the practical question isn't whether your current setup is technically legal. It's whether you can defend your consent process if challenged. A genuine pre-meeting consent mechanism, where declining actually prevents recording, is the most defensible approach regardless of which jurisdiction applies.
Frequently Asked Questions
About the author
Abhishek co-founded Scriptivox and built its early optimization and scalability layer — the part that turns a working transcription tool into one that holds up under real load. Today he leads growth and marketing at Scriptivox. He writes about transcription accuracy, multi-language coverage, and what it takes to build an AI transcription product that stays fast and reliable as it scales.
