Searching for Audio....Is it possible?
Rob Jones, Kroll Ontrack
Innovations in document review surface nearly every day. Now, new audio review technology makes it possible to search digital voice recordings for the occurrence of relevant keywords. This brings flexibility to an area which has previously been difficult to manage. Like ‘traditional’ documents, one of the key challenges is volume (forgive the pun!). But unlike paper or spreadsheets or word-processed documents, the relevance of an audio document is not always immediately determinable, taking several more seconds or minutes for a reviewer to appreciate the context of the recording than it does to skim read pages of text.
The length of time it takes to review an audio document has been estimated to be twice its duration (just think how many times you have had to replay important voicemails to catch important details) and carries no guarantee that all of the significant information has been recognised. Much depends on how well briefed and how alert the audio reviewers are.
Audio searching technology now makes it possible to look for keywords or phrases in a series of digital recordings and listen to the ‘hits’ within each recording.
This means that reviewers can easily revisit recordings if an important detail (such as a nickname or a euphemism) comes into the case after initially listening to the recording.
Advanced Technology
Significant efficiencies are achieved by the use of cutting edge technology which analyses the collections of sounds (called phonemes) that characterise speech. In digital recordings the phonetic structure of each spoken word can be recognised by a search engine, allowing the user to find all files which contain a potential match. To avoid listening to every recording in full, all of the ‘hits’ can be score-rated for confidence in the quality of the result. This means that lawyers can quickly identify relevant recordings, which may contain case-turning electronic evidence.
Furthermore, the control of such a powerful tool rests in the hands of the user. As the search is performed against sound rather than text, the user can modify the spelling of their terms to compensate for differences in the ways that people speak. Although most speech analysis systems are sophisticated enough to account for regional dialects, users can elect to use correct spellings (e.g. ricochet) or phonetic spellings (e.g. “rick-o-shay” or “rick-a-shay”) to define their terms. Additionally, the user may have the power to increase or decrease the volume of likely results by applying a simple confidence rating as part of the search criteria, which is based on a statistically calculated value given to each hit. This is helpful to lawyers because, like when searching text, the relevance of each ‘hit’ will depend on two things: (1) the quality of the recording and (2) the complexity of the search. Therefore the versatility to support creative and defensible searches is a key benefit.
Competitive advantage
The balance which must be struck is, of course, the need to be able to review audio evidence accurately versus being able to do so quickly and efficiently. As a tried and tested method of dealing with audio evidence, transcription and text-based review should guarantee full confidence in the output of the review. However transcription may be deemed inefficient and costly, as revealed in these two hypotheses:
In regulatory investigations, speed of response is often an important strategic consideration. In a large investigation, particularly with a financial angle, there could be hundreds of hours of material to review. The deadlines imposed for disclosure in these cases often seem unrealistic, needing fast reactions and military-style co-ordination, as well as the right tools to meet them. To wait for potentially numerous recordings to be transcribed before they are reviewed may be unacceptable.
Since the cost of transcription is directly linked to the amount of audio evidence, in modest sized cases in UK litigation, a party may claim it would be disproportionate to disclose audio evidence. However the reality is that transcription costs are not absolutely necessary and legal teams can categorise, annotate and control the playback of many different types of audio recording very easily.
In cases like these, audio review technology is a powerful weapon because lawyers can begin reviewing evidence earlier. Moreover, legal teams and can argue that important audio evidence should be disclosed because it is now easier and cheaper to review, making proportionality a non-issue.
Important Considerations
The opportunities to use audio review technology are wide and varied. Over the years, telephone calls and voicemails in particular have been used as ways of conveying messages that it might be undesirable or otherwise difficult to capture in writing. Therefore, in criminal, civil and regulatory cases, smoking gun evidence might be more likely contained in audio documents.
However, each case will raise important questions that practitioners should consider. Examples include the following:
- A recorded telephone conversation will raise questions of admissibility, particularly if the recording is obtained by interception of the communication during transmission (phone tapping).
- The increased use of unified messaging systems (which deliver voicemail to a person’s inbox), means that personal and case-related voicemails will be mixed together in the same collection of data. This in itself will necessitate a consideration of local laws relating to privacy and data protection.
- Finally, what happens to voicemails in the natural course of business may merit special consideration in any document retention plan.
Final Note
Audio review technology allows legal professionals to investigate audio material alongside regular documents such as e-mails and scanned images. This provides an efficient and cost effective way of reviewing potentially critical evidence while balancing comprehensive case preparation with proportionality. There are significant strategic and commercial advantages to be gained. However the collection and examination of audio evidence must be well planned.
Rob Jones is a legal consultant at Kroll Ontrack in London.
Disclaimer
This document is neither designed nor intended to provide legal or other professional advice but is intended merely to be a starting point for research and information on the subject of legal technology. While every attempt has been made to ensure accuracy of this information, no responsibility can be accepted for errors or omissions. Recipients of information or services provided by Kroll Ontrack shall maintain full, professional, and direct responsibility to their clients for any information or services rendered by Kroll Ontrack.
Copyright 2007 Kroll Ontrack, Inc & Kroll Ontrack Legal Technologies Ltd.
