Demystifying Voice Biometrics: The Future of Security is Available Today

A common error within the financial services industry is to compare an emerging technology, voice verification, to an existing technology, voice recognition. Simply put, voice recognition identifies what is being said, while voice verification identifies who is speaking. In other words, voice recognition interprets, while voice verification confirms.

Many believe voice verification is in the stage voice recognition was in five or more years ago. Financial institutions specifically are taking a “wait and see” approach as a result of the challenges and early trials associated with voice recognition. The common pitfalls and barriers to adoption associated with voice recognition solutions are now quickly becoming associated with voice verification solutions. This false perception, coupled with common misconceptions of biometrics, may indicate why despite support from the analyst community and numerous research studies confirming the viability of voice verification, the financial services industry has shown extreme caution when it comes to implementing voice biometrics as a security solution.

Given today’s current security climate, it is important to understand the fundamental differences between voice verification and voice recognition and dispel the common myths associated with voice biometric solutions used for authentication purposes. Financial institutions can’t afford to hesitate or be reactive when it comes to customers’ security. By then, it is too late.

Defining the Differences

Both voice recognition and voice verification technologies provide innovative and cost-effective solutions that can be implemented immediately within financial institutions of all sizes. Although they are closely linked, these solutions are distinctly different in terms of approach, outcome and overall strategic goals.

Voice recognition, commonly referred to as automatic speech recognition (ASR), is the process of interpreting human speech by a computer. Sounds, words or spoken phrases are converted into electrical signals, which are identified by a computer as words. This in turn initiates some type of actionable command. Essentially, voice recognition relies on speech patterns and algorithm-based programming to emulate phrases and commands for a broad population. Dating back to the 1930’s, voice recognition technology has a long, somewhat tumultuous history.

As with most emerging technologies, voice recognition was not without its initial technical problems, including speaker and dialect differences, resolution of ambiguity and background noise interference. The accuracy and convenience of voice recognition has been questioned since its debut. Voice recognition continues to face challenges in terms of its restricted capabilities, limited accuracy and lack of focus on security. Voice recognition solutions, however, have made significant improvements during the past five years and have become much more widely adopted in a variety of industries and verticals.

The same skepticism organizations once held for voice recognition is now falsely being transferred to emerging voice verification solutions.

In contrast to voice recognition, voice verification is the process of comparing a biometric sample voiceprint against a single reference of a specific user to confirm a person’s identity. Basically, it is a one-to-one comparison that confirms the identity of the speaker. Voice verification captures the unique characteristics, such as speed and tone, associated with an individual’s voice and creates a non-replicable voiceprint. Unlike voice recognition, which must identify a variety of phrases for a broad population, voice verification’s primary objective is to match a single voiceprint to a single user for identity authentication. A relatively new technology solution, voice verification is often compared to voice recognition. Many financial institutions are hesitant to implement a voice biometrics solution due to the challenges and pitfalls associated with earlier voice recognition solutions. These challenges, however, do not apply to voice verification.

According to Frost and Sullivan analyst Sapna Capoor, the global biometrics voice market is expected to grow rapidly over the next four years. “A primary driver for growth is the increased adoption of voice verification by financial institutions,” she says. “Key drivers will be the need for tighter security, enabling greater use of automated services, end-user convenience and cost reduction.”

Voice biometric technologies used for verification purposes are poised for immediate adoption and can be applied across an organization to secure multiple channels. Most of the reluctance or concern associated with voice verification is unfounded and is rooted in legacy problems associated with voice recognition or baseless myths and misconceptions.

Urban Legends of Voice Biometrics

Myth 1: Voiceprints can be recorded or duplicated.

“Hi, my name is Werner Brandes. My voice is my passport. Verify me.” This line from the 1992 movie, Sneakers, is a prime example of one of the most common misconceptions associated with voice verification. You may recall that during the movie a user’s voice was successfully recorded and played back into an access panel – giving someone false access. Duplicating someone’s voiceprint is perhaps one of the more widespread concerns with voice verification. However, a voiceprint is not a recording, but rather a mathematical representation of the physiological characteristics of an individual’s voice. A voiceprint is derived through mathematical modeling of multiple voice features and is nearly impossible to replicate. Unlike the depiction in the movie, a voiceprint is a secure method for authenticating an individual’s identity that unlike passwords or tokens, cannot be stolen, duplicated or forgotten.

Myth 2: Voice verification is cost prohibitive.

A biometric is a measure of physical traits that are unique to one person, giving biometrics an immediate advantage over other authentication methods. A common roadblock in adoption, however, is the misconception of the extravagant costs associated with biometrics. Biometrics is perceived as a complex, often expensive security measure. This may be true for other forms of biometrics, such as retina or fingerprint scans, which require in-person verification and expensive readers or scanners. Voice verification, however, is cost effective (virtually no hardware required), non-intrusive and can be used to verify easily and conveniently remote transactions. In fact, most experts agree the low-cost implementation will be the number one driver of growth in voice biometrics.

Myth 3: Voice verification is only for call centers.

Voice verification can be used seamlessly across multiple direct channels, including the Internet. For example, a financial institution could easily implement voice verification to identify customers in its call center and use the same solution to authenticate the same customer when accessing account information online. In fact, with tightening scrutiny and regulatory concern, financial institutions are being pushed to find better methods for authenticating customers who choose to process transactions remotely. Voice verification solutions offer a true strong multi-factor solution, while maintaining the privacy and added convenience customers are seeking.

Myth 4: Biometrics is too new and therefore a risky investment.

Biometrics has a longer history than is commonly assumed. In fact, in 1968 a Wall Street brokerage firm spent $20,000 for a vault that could only be opened with fingerprints. Voice biometrics specifically was first developed in 1970, and although it has become a sophisticated security tool only in the past few years, it has been seen as a technology with great potential for much longer. At a Gartner Group IT/Expo event held in 1997, Microsoft founder Bill Gates said, "Biometric technologies, those that use voice, will be one of the most important IT innovations of the next several years."

Conclusion

Turn on the television or the radio, open a newspaper or log onto the Internet – most likely you will read something about identity theft and a newly discovered security threat. Today’s consumers are demanding the convenience of remote transactions while maintaining a high level of security. Voice verification relies on something truly unique that cannot be duplicated or stolen – an individual’s voice. Despite the ever-present myths or misconceptions, voice verification is inherently more secure than most other verification methods, making an investment in this type of security measure worth every penny.