NXT-ID seeks patent for voice-authentication technology for making payments, other potential uses
Tunnell said the NXT-ID voice-directed method uses an algorithm that can identify a speaker’s voice, even if the user’s voice is altered by a health condition, such as a cold or flu symptoms.
“Believe it or not, our algorithm works really well, even if the person has some higher-frequency deficiencies,” he said. “If their voice is augmented, their voice is still their voice, and we can detect that. The only time we have a real problem is with laryngitis. With laryngitis, we simply give them an alternative means to access their wallet through other biometrics.”
Meanwhile, users should not be concerned that someone else—even professional speech imitators—will be able to steal access to private information, Tunnell said.
“They come in and say, ‘I can make my voice sound like yours.’ In reality, they can’t,” Tunnell said. “There are certain specific attributes to a human’s voice that can be detected that are unique to your voice—and to the word. As long as it’s generated by your vocal chords, there are specific attributes or parts of your voice that we can discriminate on to determine that you are who you say you are. That’s what algorithms do.
“And, even if they could duplicate your voice, they would have to know what the password is that you’re using.”
NXT-ID seeks a patent for voice-directed technology for making payments, but the approach also can be used for a number of scenarios requiring secure authentication for access to sensitive information, whether it be personal financial/medical information or the ability to query a criminal database, Tunnell said. In fact, voice-activated authentication could be used as a method for first responders to access the FirstNet broadband network, if the system design allowed for it, he said.
“This can be generalized to any private information, including passwords,” Tunnell said.