Skip to main content

The Role of Audio Evidence in E-Discovery

As discussion continues regarding the acquittal of George Zimmerman for the shooting death of Trayvon Martin, one of the jurors has publicly stated that she believed the screams for help on a recorded 911 call were those of the neighborhood watch volunteer, not the teenager.

The audio recording of the 911 call became a central issue in the second-degree murder trial in which Zimmerman claimed that he acted in self-defense. The recording of the emergency call captured the fatal struggle, though it was unclear whether Zimmerman or Martin screamed for help. The jurors heard the call multiple times during the trial as the prosecution and defense teams wrangled over the voice identification issue.

The Zimmerman trial highlights the rapidly increasing importance of audio evidence. Listeners of audio evidence frequently gain insight into a speaker’s true intent, relying on intonation to reveal context, emotions, and veracity. Recordings capture events as they unfold, and, unlike human recollections, are unchanged by future events: they remain unaffected by the passage of time, recollections that fade, or bias. Audio recordings can be powerful evidence.

Therefore, parties must not overlook the preservation of audio evidence during discovery and must focus on it as part of their legal hold efforts. If it is overlooked, parties risk sanctions, such as those imposed on the defendant in Banks v. Enova Financial, No. 10 C 4060, 2012 U.S. Dist. LEXIS 170000 (N.D. Ill. Nov. 30, 2012). In that case, the key issue was whether a company’s customer service employee had hung up on a complaining caller. The company recorded the call but subsequently destroyed it under its records retention policy (although unemployment and EEOC proceedings were underway). The court found the destruction amounted to gross negligence and sanctioned the company.

Even when preserved, audio recordings present another challenge during discovery. Traditionally, audio evidence has required someone to listen to the recordings, transcribe them, and search the transcription for relevant information. The transcribing process can be time-consuming, expensive and prone to human error. However, today, forensic search tools permit the rapid indexing of large volumes of audio evidence using phenome patterns. This technology is more accurate and consistent than dictionary-dependent, speech-to-text based audio searching, and it can reduce the cost of processing audio evidence by as much as 80 percent. It can also significantly reduce the volume of data that must actually be reviewed by human reviewers. As a result, parties can uncover critical evidence in days, not months.

Banks and other financial organizations, which must record all outbound sales and marketing conversations under the Dodd-Frank Act, are proactively addressing their audio data as part of their litigation readiness programs, and as part of their litigation and regulatory landscapes.  Staying abreast of new technologies in this area and understanding how they can reduce the time and expense involved in the review of this important data type is now more critical than ever before.

Gabriela Baron is senior vice president at Conduent. She can be reached at

About the Author