AI-generated meeting transcripts are creating a new category of evidence for litigators and enforcement authorities. A recent Delaware Court of Chancery decision illustrates the risk: the court had both official board minutes and AI-generated transcripts of the same meetings, and cited the transcripts when assessing the board’s motivations. Although the transcripts did not decide the case, they provided an alternate—and potentially more revealing—record of what occurred in the boardroom.

The implications extend well beyond board litigation. AI transcripts from management calls, internal investigations, privileged discussions, and other sensitive meetings may be discoverable in civil litigation or regulatory and criminal investigations—and may preserve informal comments, hypotheticals, or shorthand that look very different when read later by a regulator, prosecutor, judge, or jury. Organizations also face questions about privilege, preservation obligations, recording-consent laws, vendor access, and the reliability of AI-generated transcripts and translations.

In this post, our colleagues examine the Delaware decision and practical steps organizations can take to manage these risks, including adopting risk-based policies governing when AI transcription is permitted, establishing retention and review protocols, training directors and employees, and extending AI governance to vendors and cross-border use. As AI notetakers become increasingly commonplace, organizations should consider these issues before an AI-generated transcript becomes evidence in the next dispute or investigation.

Click here to read more.