Recent Cases Spotlight Retention Lessons for AI Content from Video Recordings
Clear policies for managing collaboration platform videos and their AI-generated derivatives are becoming essential for enterprises navigating modern litigation risks. In this timely article, Phil Favro highlights two recent court rulings—Johnson v. Tuskegee University and Malhotra v. CVS Health Corporation—that provide valuable guidance on video retention practices. While these cases stop short of addressing AI-generated transcripts and summaries, their implications resonate deeply for cybersecurity, information governance, and eDiscovery professionals. As organizations increasingly rely on automated content from platforms like Zoom and Microsoft Teams, the need for defensible, strategic retention policies grows more urgent. This piece helps chart a practical path forward.
Recent Cases Spotlight Retention Lessons for AI Content from Video Recordings
By Phil Favro, Contributing Author for HaystackID
Questions abound on how to address retention of video feeds and corresponding AI-generated transcripts and summaries from video calls on workplace collaboration platforms like Zoom and Microsoft Teams. For certain enterprises, collaboration platform video recordings offer little in the way of business value, and they have established short retention settings accordingly. In contrast, some organizations appreciate recorded video meetings. In addition, they value how AI prepares automated meeting transcripts and summaries and provides analysis of meeting content for any number of business purposes. Those organizations have policies or practices with longer retention of recordings and related content. Other entities are still trying to determine the best method for handling the retention of recorded video content, transcripts, and summaries.
While they may not have the same approach for retaining collaboration platform video recordings and AI-derived content, organizations do want to understand how courts expect enterprises to handle this type of information in litigation. And yet, there have been few noteworthy decisions addressing the retention of collaboration platform videos—until recently.
In two cases from August 2025, courts entered orders to address spoliation motions regarding organizations’ retention practices over collaboration platform videos. The courts in those cases—Johnson v. Tuskegee University and Malhotra v. CVS Health Corporation—rejected litigants’ efforts to punish organizations that did not offer long-term retention of Zoom and Teams video recordings.[1]
While Johnson and Malhotra did not address retention issues relating to AI-generated transcripts and summaries, these cases are nevertheless instructive on the importance of establishing retention policies regarding AI-related content from collaboration platform video recordings. Considering the costs and risks of inaction on this issue, organizations should explore developing defensible retention policies that address such AI-related content.
Johnson: Zoom Video Recordings
In Johnson, the plaintiff served discovery seeking from the defendant university a video recording of a Zoom meeting from 2022 in which he participated with several people, including the university’s general counsel. In response, the university represented that it did not have the requested Zoom recording.
The university indicated that it had established what was effectively a 60-day retention policy for its Zoom video recordings. Under the policy, recordings would be kept for 30 days before being automatically deleted. Although “deleted,” the recordings would still be available for another 30 days until—pursuant to the policy—they were permanently eliminated. Under the operation of its policy, the university clarified that it had not retained any Zoom videos from the 2022 timeframe, including the video at issue. Moreover, the university had undertaken a search of individual employee emails and computers to identify a copy of the recording, though these searches proved unsuccessful.
Dissatisfied that the university failed to produce the requested recording, the plaintiff filed both a motion to compel the video and a motion for spoliation sanctions under Federal Rule of Civil Procedure 37(e) against the university. The court rejected both motions. While the plaintiff argued that the defendant should have preserved the Zoom recording for his lawsuit, the court disagreed. Instead, the court found that the recording was eliminated in accordance with the university’s standard retention policy months before the university learned about the plaintiff’s lawsuit and triggered its duty to preserve. Accordingly, the court refused to impose sanctions on the defendant or otherwise grant the plaintiff’s requested discovery relief against the university.
Malhotra: Teams Video Recordings and Messages
The court in Malhotra reached a similar conclusion to Johnson regarding a plaintiff’s contention that the defendants should have preserved relevant video recordings and messages from their Microsoft Teams collaboration platform. In Malhotra, the plaintiff sought Teams videos and messages to substantiate her discrimination and related employment claims against the defendants. While defendants produced emails reflecting copied Teams message content, they did not produce Teams videos or stand-alone Teams messages because such information no longer existed.
The defendants indicated they had implemented policies relating to the retention of the requested Teams data. For Teams videos, defendants had a zero retention policy: “unless specifically recorded by a user,” defendants would not record or retain Teams videos. Teams messages were retained for 90 days, after which they were deleted. Pursuant to these policies, defendants represented that they did not have the Teams data the plaintiff requested.
When the plaintiff sought relief from the court, she—like the plaintiff from Johnson—found an unreceptive audience. First, the magistrate judge issued a report and recommendation to the district court that recommended denying the plaintiff’s motion for Rule 37(e) sanctions regarding the lost Teams content. The magistrate judge found that the defendants’ duty to preserve the Teams data triggered only after the requested Teams videos and messages were deleted under the defendants’ retention policy. Accordingly, sanctions should not issue as defendants were under no obligation to keep the videos or messages when they were eliminated under the policy. The district court agreed, finding that the videos and messages—even under the best-case scenario for the plaintiff—would have been deleted at least two months before the defendants’ duty to preserve was triggered. Therefore, the court denied the plaintiff’s motion for sanctions.
Retention Lessons from Johnson and Malhotra on Video Recordings and Related Content
Johnson and Malhotra offer relatively straightforward guidance on a somewhat recent form of ESI: collaboration platform video recordings. The case holdings emphasize the importance of establishing an information retention policy that addresses recordings. By setting a reasonable policy that both meets business objectives and satisfies applicable industry norms, organizations can establish a defensible process upstream that can benefit them downstream in legal actions. These lessons of proactive information governance have been well known among enterprises for many years now.
The real question is whether organizations will apply these time-honored principles to AI-derived content such as video meeting transcripts and summaries. Even for enterprises that find value in this content, there is a strong business case for periodically deleting such information. Meeting transcripts and summaries offer immediate value in terms of automating the creation of meeting action items and providing the enterprise with an opportunity to analyze meeting content. While such information could have long-term importance, the immediate benefit of transcripts and summaries generally diminishes over time as action items are addressed, postponed, or otherwise discussed during subsequent calls. Stockpiling stale transcripts and summaries does nothing to advance business objectives. Worse, such information—multiplied daily by employee meetings across the enterprise—could significantly increase review costs in legal matters. In addition, retaining call transcripts and summaries—which may reflect personally identifiable information—can increase risks associated with data breaches.
Indefinite retention of AI-extrapolated transcripts and summaries presents other risks that organizations should consider when deciding whether to set retention policies. For example, transcripts and summaries may not be accurate. Whether such content memorializes unintended transcription errors or reflects hallucinated content, allowing potentially erroneous materials to remain long-term among the enterprise’s records may create unnecessary risks for organizations in litigation, regulatory investigations, or data breaches. And for every organization that has established a process to ensure the accuracy of such content, many others have not, and thus, they leave themselves vulnerable to substantive content inaccuracies.
Accordingly, while Johnson and Malhotra did not address AI-generated materials, the implicit rationale from these cases regarding the value of retention policies applies equally to AI content such as meeting transcripts and summaries. Indeed, organizations should consider implementing retention policies for meeting transcripts and summaries. Doing so should not be a one-size-fits-all approach. Enterprises should carefully explore the benefits and risks of such content and then set retention schedules accordingly. By so doing, they can more proactively establish defensible policies and practices regarding both collaboration platform recordings and related AI content.
About Phil Favro
Phil Favro is the founder of Favro Law PLLC, where he counsels clients on ESI, AI, and discovery issues and serves as a special master, mediator, and expert witness. Phil is nationally recognized for his expertise on ESI, discovery, and information governance, with courts acknowledging his credentials. See, e.g., Oakley v. MSG Networks, Inc., No. 17-CV-6903 (RJS), 2025 WL 2061665 (S.D.N.Y. July 23, 2025). This background makes Phil particularly well-suited to counsel clients and advise courts on information-related issues. As a special master, Phil is acclaimed for his collaborative approach, working with parties to find stipulated solutions to complex issues. For disputes that require adjudication, he is renowned for the clarity and vigor of his written dispositions, which are available on legal search engines.
About HaystackID®
HaystackID® solves complex data challenges related to legal, compliance, regulatory, and cyber requirements. Core offerings include Global Advisory, Cybersecurity, Core Intelligence AI™, and ReviewRight® Global Managed Review, supported by its unified CoreFlex™ service interface. Recognized globally by industry leaders, including Chambers, Gartner, IDC, and Legaltech News, HaystackID helps corporations and legal practices manage data gravity, where information demands action, and workflow gravity, where critical requirements demand coordinated expertise, delivering innovative solutions with a continual focus on security, privacy, and integrity. Learn more at HaystackID.com.
Assisted by GAI and LLM technologies.
SOURCE: HaystackID
[1] Johnson v. Tuskegee Univ., No. 3:24-CV-360-KFP, 2025 WL 2496312 (M.D. Ala. Aug. 28, 2025); Malhotra v. CVS Health Corp., No. 24 C 4769, 2025 WL 2453807 (N.D. Ill. Aug. 4, 2025).