When facing impending litigation or regulatory investigations, many companies just focus their e-discovery efforts on written sources such as e-mails and text messages. The days of “I won’t ask for your e-mails if you don’t ask for mine” are long gone. And moreover, e-discovery is continually expanding into new areas. Audio, an up-and-coming dimension to e-discovery, has already become an important piece of the puzzle.
“Audio is a reality right now,” says Douglas Winter, of counsel at Bryan Cave and the head of its electronic discovery unit.
“You have that proliferation of data now for audio in the same way we’ve had it with text for e-mails. That has meant that in the types of lawsuits where business communication is important, people are looking more and more at audio information.”
Under the Federal Rules of Civil Procedure, audio is included as discoverable electronically stored information (ESI), and with the proliferation of Voice-over-Internet Protocol (VoIP) phone service, voicemails have found an increasingly important place in the discovery process.
As a result, companies using VoIP software need to be aware of the e-discovery liability they face. When voicemail becomes as easily discoverable as e-mail, counsel encounter a new set of risks and e-discovery requirements.
VoIP converts a person’s voice into a digital signal that travels over a broadband Internet connection. This system gives users the option of attaching voicemails to e-mails, just like any other e-mail attachment.
Many companies have begun using VoIP because it’s cheaper than traditional phone service, and it allows organizations to consolidate their phone systems with the computer software they already use.
A 2008 study estimated a 44.5 percent growth in business subscriptions to VoIP last year and predicted 25.4 million VoIP subscribers in the U.S. by 2012.
The obvious risk with VoIP, however, is that adding voicemails to e-mail causes massive data proliferation, increasing the pre-existing overload of discoverable ESI.
“I might receive a voicemail attachment, and then I’ll forward it to someone else,” says Deborah Baron, vice president of e-discovery at Autonomy, which makes audio search software. “So there are copies in the voicemail system, in my e-mail box and on my desktop. And now someone else has it, so there are four or five copies.”
Keeping track of multiple electronic–and discoverable–versions of voicemail can be especially frustrating, because in most cases having easily discoverable voicemail at all is a new challenge.
“You’re leaving a file somewhere,” Winter says. When you leave a conventionally recorded phone message, “you often leave a file that could be discoverable–but you leave it on a tape or a digital card, which is discoverable but not as easily dispersed.”
In addition to voicemail, audio files often come into play in financial services industry cases, where brokers must record conversations with their clients.
Faced with audio ESI, the first step is to have a well thought out archiving and retention schedule. Baron suggests making sure to dispose of the files in a timely manner, based on a well-planned and defensible retention schedule.
Once the time comes to actually review the audio files for an impending case or regulatory investigation, counsel must figure out approximately how many hours of sound must be searched. In a more typical situation, the traditional method–employing live humans to listen and transcribe the tapes–is easier than using a software solution. Because although sometimes there are hundreds of hours of recordings, attorneys generally need to look through fewer audio files compared with the millions of e-mails often reviewed in a single case.
“[Audio search programs] are not discrete software tools,” says Winter, explaining another part of the reason many companies don’t use cutting-edge technology in audio discovery matters. “They’re typically an add-on or an element of a particular search engine. You’ve got to be devoted to it.” Most businesses and law firms decide on a single e-discovery search software and stick with it–and often the program can’t handle audio.
But sometimes, attorneys have to review so many hours of audio that special search software becomes the most efficient option. Generally, this type of software uses phonetics to make the audio searchable like text, avoiding transcription entirely. Jeff Schlueter, vice president and general manager for legal markets at Nexidia, which makes audio search software, says the average matter his company handles has between 500 and 1,000 hours of audio.
“Lawyers don’t want to listen to thousands of hours–or even hundreds of hours,” Baron says. “They want to listen to the few minutes or 30-second clips where key custodians are speaking about substantive issues.”
With a diverse variety of keyword search strategies, employing audio search software becomes similar to keyword searching through written documents. Complex searches lead lawyers directly to clips in question and allow them to listen live.
Emerging technologies also allow audio redaction by inserting white noise into the file. And speaker identification has increasingly been added into audio search systems.
Reviewing audio the old-fashioned way can be tricky because poor recording quality can make tapes hard to understand. Regional dialects and varying pronunciation–among Americans as well as foreigners–can also create false positives.
‘Down’ can sound like ‘pound’ or ‘brown,’” Baron says. Additionally, it can take up to one hour for a person to transcribe 15 minutes of audio.
As a worst case example, Baron mentions a Wall Street law firm that farmed out thousands of hours of audio to a company to do live transcription, which came back less than 30 percent accurate. “They spent five-plus digits on it and were in a holding pattern for two weeks,” she says. “There’s got to be a better way.”
According to some experts, audio is just the beginning of the new discovery challenges. Video also is on the horizon.
In different capacities, video search software already exists. Nexidia has the ability to search the audio tracks on video files. Autonomy offers a similar service, as well as visual recognition–the ability to distinguish faces as well as objects such as license plates.
Video e-discovery in the corporate world is by no means prevalent–yet. But video does exist in corporate litigation. Baron says it most often arises in surveillance tapes relating to regulatory, employee and copyright issues.
“As people begin having a visual element to their communication, and that information is recorded, that’s going to become the next step in discovery,” Winters says.