If you've turned on an AI notetaker for a sales call, a vendor meeting, or a job interview, you may have just made a legal decision you didn't know you were making. Two active lawsuits this year argue that popular tools like Otter.ai and Fireflies.ai record and analyze people who never agreed to it — and the legal theory behind both cases applies just as easily to a plumbing company's customer calls as it does to a tech firm's board meetings.
None of this means you should stop using AI to capture meeting notes. It means five minutes of setup now can save you a much worse conversation with a lawyer later.
Two lawsuits worth knowing about
In Brewer v. Otter.ai, filed in federal court in California, the plaintiff says Otter's meeting assistant recorded and analyzed a private conversation without the consent of everyone on the call, and argues the company uses those recordings to improve its AI without clearly telling non-account-holders what's happening. The case is still pending, but the theory is straightforward: if a tool records a call and only one participant — the person who installed it — agreed to that, every other person on the call may have a claim under state or federal wiretap law.
A second case, Cruz v. Fireflies.AI Corp., filed in federal court in Illinois, raises a different but related problem: the complaint alleges that Fireflies' speaker-identification feature creates a voiceprint of everyone who speaks on a call, including people who never created an account or agreed to any terms of service, which the plaintiff says violates Illinois' biometric privacy law. The plaintiff in that case says she was simply a guest on a video call hosted by someone else who had the tool turned on.
Neither case has gone to trial, and neither should be read as a verdict on whether these specific companies did anything wrong. What they establish is that the exposure is real, the plaintiffs are often people who weren't even the ones who installed the software, and the business that invited them onto the call is a natural target too.
Why this is a small-business problem, not just a tech-company problem
It's tempting to assume this is Silicon Valley's problem to sort out. It isn't. If your business uses an AI notetaker on calls with customers, subcontractors, job applicants, or anyone outside your own team, you are the one who invited that outside person into a recorded conversation — and you are the one who decided whether to tell them.
State recording laws split on this. Most states only require one person on the call to consent, which can be the person who turned the tool on. But a meaningful group of states — California, Illinois, and Florida among them — require everyone on the call to consent before it can be recorded. If any participant is sitting in one of those states, the safer legal assumption is that all-party consent applies to that call, regardless of where your business is based. A visible bot icon or a line buried in your AI vendor's terms of service is not the same thing as telling the other person you're recording them and getting them to agree.
How to use an AI notetaker without the legal risk
You don't need to give up AI meeting notes. You need a habit, not a lawyer on retainer:
- Say it out loud, every time. At the start of the call: “I'm using an AI notetaker to capture this conversation — is that okay with everyone?” A verbal yes, on the record, is far stronger protection than a silent bot icon.
- Put it in writing for recurring relationships. For a customer, vendor, or contractor you talk to regularly, a line in your intake form or engagement letter noting that calls may be transcribed by AI covers you without an awkward ask every time.
- Check what your notetaker does with speaker identification. If the tool creates per-person voiceprints to label who said what, that's the specific feature at the center of the Fireflies case. Look for a setting to turn off speaker recognition, or use transcription without it, if you're not confident every participant has agreed to it.
- Review the vendor's training-data default. Most AI notetakers default to using your calls to improve their models unless you opt out. Turn that off in account settings — it has nothing to do with getting your notes and everything to do with what happens to other people's conversations after the call ends.
- Set a retention and deletion policy. Decide how long transcripts live in the tool and delete what you don't need. A transcript you don't still have can't be the subject of a records request or a lawsuit.
The underlying appeal of AI notetakers isn't going anywhere — accurate notes without someone scribbling during the meeting are a genuine win. The fix here isn't to distrust the technology. It's to treat “turn on the AI notetaker” with the same five seconds of thought you'd give to turning on a tape recorder in a room full of other people — because legally, that's exactly what you're doing.