AI Meeting Assistants Face Privacy Lawsuits and Consent Backlash

A year of legal and institutional pushback

AI note-takers have spread quickly through video calls. They join meetings, transcribe them, and produce summaries. In 2025 that convenience ran into legal and institutional resistance, and the tools now look different because of it.

Three incidents stand out. Otter.ai became the target of a class action lawsuit alleging it recorded conversations without proper consent 1. Fireflies.ai was sued over the collection of biometric data 1. Chapman University barred Read AI from use, citing security concerns 1.

The privacy case against cloud-based meeting tools rests on three points [1]:

• Conversations end up stored on third-party servers.
• Those conversations may be used to train AI models.
• Recordings can happen without everyone in the meeting agreeing to them.

The short answer to whether these assistants are safe is "not all of them." That is a fair summary of where the category stands.

The market keeps growing anyway

The pushback has not slowed the product category. A 2026 ranking lists 18 AI meeting assistants and note-takers 2.

• Zoom AI Companion sits at the top, praised for in-meeting summaries 2.
• Zoom is also credited with a broader ambition: turning the assistant into something that works across its whole Workplace suite, so meetings shape the work that follows for days or weeks 2.

The two outlooks don't really conflict. They describe the same tension from opposite sides. One side is the pull to put AI into every conversation and connect it to everything downstream. The other is the growing legal and institutional risk of doing exactly that.

The more a meeting assistant integrates across a workplace, the more sensitive data it touches. That raises the stakes if consent or storage practices fall short. Zoom's vision of meetings feeding into weeks of follow-on work is appealing. It also means transcripts become longer-lived and more widely shared assets. That is precisely the exposure privacy critics warn about 12.

Consent is becoming a product feature

The clearest evidence that the lawsuits matter is in how vendors now describe their own products. Two examples from the 2026 rankings show it.

Notion AI's meeting notes. The tool requires users to confirm participant consent before it records or transcribes 2. The ranking lists this as a limitation, an extra step that some teams may find annoying 2. Seen against the Otter.ai lawsuit 1, that friction looks less like a flaw and more like the kind of safeguard the industry is being pushed toward. One reviewer's inconvenience is another organization's legal protection.

Krisp's bot-free design. Krisp is recommended specifically because it avoids the familiar model of a visible AI bot joining every call 2. The ranking says plainly that "bot backlash and consent concerns are real." It presents Krisp's approach as meaningful for teams that don't want an extra participant in every meeting 2. Krisp also offers audio cleanup, which helps remote and hybrid calls in noisy settings 2.

Bot-free capture changes the meeting experience. Whether it resolves consent questions is another matter. A bot announcing itself at least signals to participants that recording may be underway. Removing that visible participant reduces social awkwardness. It does not, by itself, answer whether everyone in the call has agreed to be transcribed. Teams adopting bot-free tools may still need their own disclosure practices.

What this means for buyers

The 2025 incidents point to several questions buyers should ask before adopting a meeting assistant:

• Data storage: Where does the data live?
• Model training: Can the vendor use your conversations to train its models?
• Consent: How does the tool obtain agreement from everyone on the call 1?
• Biometrics: The Fireflies.ai case adds a further issue. Voice data can carry biometric implications, which some jurisdictions regulate strictly 1.

The university ban signals that institutions handling sensitive information are willing to take a hard line rather than wait for lawsuits to settle 1.

Feature checklists still matter. The rankings note practical gaps, such as Notion's meeting notes not identifying speakers and its AI not reading files or images already on Notion pages 2. Superhuman, which approaches the space from email, is described as still building out its calendar and scheduling capabilities 2. But these limitations are now weighed alongside governance questions, not instead of them.

The bottom line

AI meeting assistants are not going away. If anything, the integration ambitions of companies like Zoom suggest they will become more embedded in daily work 2.

The 2025 lawsuits and the Chapman ban mark a turning point. Privacy is no longer a background concern but a competitive dimension 1. Explicit consent prompts and bot-free capture show vendors reacting 2.

The tools that win long-term are likely to be those that treat consent and data handling as core design choices rather than friction to minimize. For now, "safe" depends heavily on which assistant you choose and how carefully you deploy it.