Corporate lawyers have begun ejecting AI notetakers from meetings before they start, The New York Times reported. The reason is straightforward: automated transcripts turn routine business conversations into discoverable evidence in lawsuits and investigations.
Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.
yesSubscribe to our daily newsletter, PYMNTS Today.
By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions.
The tools have spread fast. Artificial intelligence (AI) assistants from Otter, Fireflies, Fathom and Read.ai join calls across Zoom, Microsoft Teams and Google Meet, while the platforms themselves ship transcription that some users switch on by default. A 2025 Fellow.ai survey of IT and business leadership found 3 out of 4 professionals using an AI notetaker in their work meetings.
The legal problem starts with what the transcripts capture. Human meeting minutes are curated. AI transcripts preserve offhand remarks, jokes and quickly corrected statements, and they appear in meetings that would never have been recorded otherwise. In litigation or a government investigation, that record can be requested wholesale, since discovery demands typically cover all documents and communications tied to a disputed topic, law firm Pillsbury noted in a May 11 piece based on the aforementioned New York Times article. Another law firm, Fisher Phillips, wrote that AI-generated electronically stored information (ESI) from notetakers, meeting summaries and chat assistants is becoming a core discovery battlefield in employment cases, with plaintiffs’ counsel now routinely requesting meeting recordings, transcripts and summaries.
The exposure compounds when lawyers are in the room. Sharing a privileged conversation with a third-party bot can void attorney-client protection, making discussions that would otherwise stay confidential fair game in court. The New York City Bar Association in December issued a formal opinion urging lawyers to weigh whether recording and summarizing is tactically wise and to warn clients of the downsides.
Law firm Mayer Brown noted in a June 3 article that transcripts containing sensitive data can surface through subpoenas, civil litigation, regulatory investigations or data breaches, and that inconsistent recording practices, where favorable meetings go untranscribed and others do not, can create evidentiary problems of their own.
The companies selling the tools are now defendants themselves. NPR reported last August that a lawsuit seeking class action status accuses Otter of recording private conversations without permission from all participants and using them to train its transcription models. The suit claims violations of federal and California wiretap and privacy laws, since Otter’s assistant joins meetings and captures non-users who never agreed to anything. The case has since been consolidated with similar suits as In re Otter.ai Privacy Litigation, pending in the U.S. District Court for the Northern District of California.
Advertisement: Scroll to Continue
Related claims against meeting-recording vendors raise biometric-data questions. In Cruz v. Fireflies.AI, filedin December 2025 in the Central District of Illinois, the plaintiff alleged Fireflies generated a voiceprint of her when she joined a meeting hosted by an Illinois nonprofit that had enabled the tool. She had never created a Fireflies account. The state’s Biometric Information Privacy Act defines voiceprints as biometric identifiers requiring express consent.
The pattern echoes earlier fights over Alexa recordings and Google keyword warrants, except the data pool is larger. Every meeting that runs through an AI assistant produces a timestamped, searchable document held by a vendor whose retention and training practices the participants rarely control.
The meeting assistant litigation sits within a wider pattern of AI-generated records reaching courtrooms. In March, the Delaware Court of Chancery issued an opinion in Fortis Advisors v. Krafton finding that Changhan Kim, CEO of South Korean gaming company Krafton, had used ChatGPT to plot a strategy for avoiding a $250 million earnout. Kim had deleted the conversations. They surfaced anyway, and the court quoted them directly in its opinion.
On the question of privilege, courts are now split. In February, Judge Jed Rakoff of the Southern District of New York ruled in United States v. Heppner that a defendant’s exchanges with a consumer AI platform carried no attorney-client privilege. One week earlier, the Eastern District of Michigan reached the opposite result in Warner v. Gilbarco.
The Meeting Bot Nobody Invited Is Now Exhibit A
Amazon Secures $17.5 Billion Credit Facility Amid AI Infrastructure Race
Anthropic Helps JFrog Govern the Software Supply Chain
Visa and OpenAI Unlock Agentic Commerce
Get PYMNTS Today, AI, B2B and more.
We’re always on the lookout for opportunities to partner with innovators and disruptors.
