At conferences across the country, companies are using artificial intelligence note-taking, technology that records, transcribes, and organizes business meeting conversations. AI note-taking increases efficiency, improves stakeholder focus, and improves accessibility for participants with different abilities and speakers of different languages.
However, the laws governing these “notes” are lagging behind. Businesses face a patchwork of state and federal laws, evolving court decisions, and real uncertainty. Against this background, the question arises as to how useful technology can be considered eavesdropping. Uncertainty about how AI note-taking will impact discovery further heightens the risk.
Is taking notes the same as eavesdropping?
To see this uncertain legal situation in action, consider the following: Otter.ai Privacy Lawsuit, The case is pending in the United States District Court for the Northern District of California.
Plaintiffs in the class action lawsuit allege that Otter.ai, an AI-powered transcription service provider, violated California wiretapping laws. The plaintiffs, who attended the meeting but were not Otter.ai account holders, allege that the company’s software “surreptitiously” infiltrated and recorded the conversations without the participants’ knowledge or consent.
Otter.ai’s response is expected to be filed in court early this year, but the wiretapping issue could take years to resolve if the district court’s ruling is appealed. Until then, the safest way to continue using notetaker is to make its deployment public and secure affirmative consent from all participants.
At Otter.ai It fits into a broader framework for understanding how AI note-taking can be eavesdropped.
Consent rules shape the wiretapping situation. The legality of wiretapping depends on a jurisdiction’s consent regime. All-party jurisdictions require the consent of all participants before recording, while single-party states require only the consent of one participant to permit recording.
The all-party field includes California, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania and Washington. Most other participants accept one participant’s consent. These jurisdictions include Texas, New York, and Washington, DC. Federal law is similar, but it does not supersede stricter state standards.
Vendor classification is complex. California’s invasion of privacy law is one of the most stringent wiretapping laws in the country. Vendor classification is an important unresolved issue in this plan.
Some courts treat software-as-a-service or AI vendors as tools used by parties to a conversation, such as a tape recorder. Other courts have equated vendors who retain the right or ability (even if not exercised) to access and use meeting data to be third-party eavesdroppers. of Otter.ai Plaintiffs share the eavesdroppers’ view, arguing that Otter.ai not only spies on accounts, but also makes money by eavesdropping on people other than account owners without their knowledge.
The complexity doesn’t end there. Vendors with data-mining rights may subject their customers to vicarious liability for the vendor’s CIPA violations.
It is unclear which laws apply to interstate issues. Whether working cross-office or remotely, many employees participate in online meetings from across the country. This reality has made the unresolved and inconsistent directives in which jurisdictions’ laws apply even more concerning.
Results vary because states use different analytical techniques to address the problem. Some courts use a variety of dispute techniques, while others avoid disputes altogether by treating the issue as evidence. Fluctuations in facts add to the uncertainty.
But in many states, the location of the interception is a deciding factor. However, states such as California apply the consent requirement of all parties when out-of-state recordings would be detrimental to Californians. The resulting hodgepodge makes agreement by all parties attractive.
Taking notes and discovering
AI note-taking creates an entirely new category of documents in the discovery process, and its presence raises discovery and evidentiary questions.
From a discovery perspective, the court will need to determine who owns, stores, and controls the notes (especially if a vendor has access to or retains the data) in order to determine what needs to be produced.
Work product protection may apply to meeting notes and summaries prepared in anticipation of litigation, but that protection is limited and fact-specific.
Given the proliferation of AI note-taking tools, it is only a matter of time before evidentiary issues such as authenticity and accuracy come up in court. in 2024 Dixon v. Royal Live Oaks Academy of the Artsa federal district court in South Carolina refused to consider an automatically generated transcript for summary judgment because it contained “errors.”
Lawyer ethics are unclear. Unreleased recordings are not unethical per se, as long as they are legal and not deceptive. However, private recordings remain illegal and unethical in bipartisan states because lawyers cannot ethically do anything illegal.
But even if allowed, the use of AI note-taking could draw the ire and skepticism of courts. Even when ethics opinions do not prohibit recording, some courts have emphasized candor, forcing disclosure or imposing sanctions.
Practical safety measures
Until courts and Congress answer important questions, two stopgap stands out.
The first is to require and explicitly disclose the consent of all parties whenever the AI takes notes. This may include a verbal script, pre-meeting calendar language, on-screen banners, roll call agreements, and procedures for late participants, appeals, and cancellations. of Otter.ai Plaintiffs praise similar steps by other AI note takers.
The second is planning usage rights and access rights in vendor contracts. The choice to retain these rights with an AI vendor increases the likelihood of potential liability and future discovery obligations.
These steps will help you leverage the benefits of an AI note-taker while avoiding the sharp edges of eavesdropping and ethical risks.
As AI innovation increases, even the most helpful technologies are likely to come into conflict with the law. Lawyers, judges, and lawmakers are working hard to resolve these issues, but disclosure and warnings are the best interim solutions.
This case concerns the Otter.AI Privacy Litigation, N.D. Cal., No. 25-cv-06911, Granted December 22, 2025.
A small amount of this content was drafted by generative artificial intelligence.
This article does not necessarily reflect the opinion of Bloomberg Law, Bloomberg Tax, Bloomberg Government, publisher Bloomberg Industry Group, Inc., or its owners.
Author information
Shayna Goldblatt Prowler is an attorney at Haynes & Boone LLP.
Tyler P. Young is a partner at Yetter Coleman in Houston, where his practice focuses on commercial litigation and antitrust law.
Please write to us: Author guidelines
