AI Notetakers and the Growing Legal Risk

Convenience Is Colliding With Compliance

AI notetakers have become a familiar presence in virtual meetings because they promise an obvious productivity gain: they can automatically join Zoom, Google Meet, and Microsoft Teams calls, produce live transcripts, generate summaries, and extract follow-up items. Tools from companies such as Otter.ai and Fireflies.ai are built around exactly that pitch, turning ordinary workplace conversation into searchable, structured records. The core tension is that the same feature set that makes these products useful also makes them legally sensitive, because the software is not just “taking notes” in the human sense — it is recording, processing, storing, and sometimes reusing conversations through third-party systems.

The Main Legal Flashpoint

The strongest example of that risk is the litigation now surrounding Otter.ai. In the California case Brewer v. Otter.ai / In re Otter.ai Privacy Litigation, plaintiffs allege that Otter’s meeting bot joined calls, recorded and transcribed conversations involving people who were not Otter customers, failed to obtain consent from all participants, and used captured communications to train its speech-recognition and machine-learning systems. The complaint also argues that Otter’s setup places too much responsibility on meeting hosts and users rather than on the vendor itself. Legal analysis tied to the case has framed the broader question plainly: whether AI notetaking can, in some situations, be treated as a modern form of wiretapping under existing state and federal law.

More Than a Privacy Problem

The exposure goes beyond consent alone. Legal experts have warned that AI-generated transcripts and summaries may become new categories of discoverable business records, raising questions about authenticity, accuracy, custody, and who ultimately controls the data when vendors retain access. In legal settings, the danger becomes sharper: routing attorney conversations, strategy sessions, or sensitive internal discussions through outside platforms can threaten attorney-client privilege and other confidentiality protections. Separate lawsuits have widened the concern further by arguing that transcription systems may create or store voiceprints, pulling AI notetakers into biometric privacy law as well. A complaint against Fireflies.ai alleges collection and retention of speaker-derived biometric data without the written policy and consent required by Illinois’ BIPA, while a separate complaint against Microsoft Teams similarly alleges that live transcription created voiceprints without proper notice, written consent, or retention disclosures.

“Disclosure and caution are the best interim fixes.”

That line captures the practical takeaway emerging from the current legal debate. Until courts and lawmakers draw clearer boundaries, the safest approach is not to assume these tools are routine office software, but to treat them as recording systems with legal consequences.

What Vendors Say

The picture is not one-sided. Otter.ai says it uses a proprietary de-identification process before training models on user data, and its terms say aggregated or de-identified data may be used for machine learning, analytics, and benchmarking. The company also describes administrative controls for workspace settings, auto-join behavior, retention schedules, and HIPAA-related safeguards for customers operating in regulated environments. That means the dispute is not simply about whether these products can be used responsibly, but whether existing consent notices, product defaults, data practices, and customer controls are enough to satisfy privacy, security, and professional obligations.

Why California Matters — Including the Central Valley

There is no direct Central Valley focus here, but the implications are significant for employers and professionals across California, including law firms, school systems, hospitals, public agencies, and agricultural businesses in places like Fresno, Bakersfield, Modesto, Merced, and Stockton. California is among the jurisdictions identified as requiring consent from all parties before recording, which makes the state especially important in the AI notetaker debate. For organizations in the Valley, the issue is therefore practical rather than abstract: if an employee, contractor, or outside participant adds an AI meeting bot to a call, the organization may be creating legal exposure even when the tool seems harmless or routine.

Why This Matters for AI and Technology

The larger technology story is that a fast-growing category of generative AI productivity software is running ahead of the legal frameworks that govern recording, consent, evidence, confidentiality, and biometric data. AI notetakers show how quickly a convenience feature can become an infrastructure issue for compliance, litigation, and workplace governance. They also reveal a broader challenge for AI adoption: businesses are not just buying software outputs, they are deciding who gets access to human conversation, how long that information lives, whether it can train future systems, and what happens when those records are later demanded in court. In that sense, the controversy around meeting bots is a preview of the next phase of AI regulation — one centered less on flashy model capabilities and more on the everyday systems quietly embedded in office life.

Central Valley AI is produced by the CVAI Newsdesk team and developed by Kaweah Tech, a regional firm that builds, deploys, and integrates AI solutions for businesses across California's Central Valley.

Source

https://www.nytimes.com/2026/05/09/business/dealbook/ai-notetakers-legal-risk.html