By Megan Bozman
Technology analysts predict we will increasingly use our voices to interact with devices. Currently, increases in voice interaction can be seen through both voice-to-text (my favorite way to text), as well as voice-only devices, such as the Google Home and Amazon Echo.
Natural language interactions clearly have numerous benefits for the speed and ease of operations. Perhaps the learning curve of touch-typing will become irrelevant, contributing to a decline in carpal tunnel syndrome. As I sometimes get achy wrists, I consider this welcome news. (And yes, I’m typing right now.)
Voice Recognition: Reliable and Secure
But our voices can also be a new asset in the ongoing battle for security within the IoT. Human voices are as unique as fingerprints and much less error-prone than passwords and PIN numbers. Passwords are an undeniable hassle. Difficulty remembering a constantly increasing number of passwords plagues even the sharpest minded tech users.
The biometric technology identifies individuals by analyzing over a hundred unique characteristics of their voice including the shape of larynx, vocal tract and nasal passage, alongside pronunciation, emphasis and speed of their speech. The systems can still identify customers when their voice has changed due to a cold or sore throat, or with age. Dr. Rita Singh of Carnegie Mellon University explains that even when individuals desire to conceal their voice, say by imitating an accent or changing modulation, it will likely fail to fool her algorithms.
UK banks, including HSBC and Barclays, are using voice recognition to authenticate telephone banking customers. Additionally, TalkTalk was the first UK business outside of the banking sector to roll out voice biometric technology in March of this year.
Voice recognition increases security, reducing a bank’s exposure to fraud. Since it also makes security process faster and easier for customers, the technology benefits all parties. When Barclays launched voice recognition technology to select customers in 2013, the time required to verify their identity fell from 1.5 minutes to less than 10 seconds.
Overcoming Customer Resistance
“Customers will have to register for voice security by using the telephone banking service. Barclays will collect a sufficient voiceprint over the course of approximately three phone calls, before allowing users to opt into to the voice security technology rather than using a password.” In contrast, setting up the feature “TalkSafe” with the TalkTalk Group requires only repeating a simple phrase three times to create a voiceprint.
Repeating a simple phrase three times on one occasion must be sufficient to create a unique voiceprint for secure identification purposes. The fact that Barclays requires three distinct phone calls to collect a voiceprint therefore appears to be overkill. I suspect they believe that the excess will appease customers and increase their comfort level with the technology. If my suspicions are correct, this seems wise due to the fact that customer resistance and mistrust are likely.
IoT Security with Voice Recognition + Convenience
With the expansion of voiceprint security, theoretically, I could speak to my Amazon Echo to order something and not have to provide any sort of additional PIN number or password. My voice itself both gives the command to place an order, and serves as my authentication. This sounds fantastic.