X

Tech Talk: Does Voice Biometrics Have A Place In Authentication?

Biometrics refers to metrics that are related to certain human characteristics – such as fingerprints, face recognition, iris or retina scanning, and even our voice patterns. Biometrics is used mostly for security purposes in identifying that an individual is who they claim to be and many times they use a combination of the above methods. Many times biometrics are used in conjunction with a password, lock pattern or some other more traditional security measure. Jason Stirling, SVP at Nuance Communications believes that voice biometrics will be used more in the future as a secondary source to help render a positive identification of an individual.

Stirling claims that it takes a mere nine seconds to identify a user by their voice pattern. He also points out that their system can distinguish between a live voice and a recorded voice using what they call a liveness detection score. The score is produced by combining, factoring in, and performing an arithmetical calculation on the outcomes of several tests. The biometric device will not always ask the same question, which makes it harder for thieves to use a recording of your voice. What is nice about the voice biometrics when compared to just saying or punching in a passcode, it could have prevented an earlier mishap with Alexa where a six-year-old girl ordered an expensive dollhouse and cookies that were delivered to their house. Being able to distinguish between voices in the same household is something Amazon is now working on for its Alexa-powered devices.

Nuance is not claiming that voice recognition is full proof or that it should be used by itself without a secondary layer of security – they believe that voice recognition biometrics is a solid way to add security to other forms of identification. A user may opt out of the second form of identification to get into their smartphone, but add another layer to open an app for bank or credit card, where extra security is warranted. Nuance admits that other forms of biometrics being worked on are also needed to heighten security. Japan’s NEC is working on using video face recognition, and so far, test results are very encouraging. Face recognition is a relatively easy biometric to solve and has been around for quite a while. It uses a stationary face from a photo or still shot and is used by law enforcement to located information on a person or simply to sign into our computers. Video face recognition can identify you from seeing your face in a moving video. If the authorities are taking live video of a crowd of people, they can focus in on you and the video face recognition can still identify you while you are moving in real time.

Face and fingerprint recognition is nice, but Nuance says there are unique instances where voice recognition works much better. They are looking at India and UIDAI’s Aadhaar (Unique Identification Authority of India). The Aadhaar is a 12-digit unique identification number issued by the Indian government to every individual resident of India – much like the US issues a Social Security Number (SSN) to each person born in the US. Many times fisherman go out to sea for weeks at a time, and when they return, their faces have become sunburned and bearded, and their fingers are cut up and swollen from pulling in the nets. This situation makes it more difficult for facial recognition software and fingerprint sensors to match up the individuals with their Aadhaar numbers at the bank, for instance. Voice recognition could play a significant role in helping to accurately identify them as an alternative security measure. With hacking on the increase, we can use as many forms of biometric security that the companies can come up with and it certainly looks as though voice biometrics can and will play a part in the future of securing our accounts. Having one form of security is just not enough anymore, and by using biometrics, it can add another layer of security making hackers jobs much more challenging. Certain types of biometrics are best used for difference circumstances, and voice recognition can play a role in our future fight to a secure login.