An interesting aspect of biometric security is how many technological advances were predicted ahead of time and are now commonplace in secure workplaces.
For example, whilst digital fingerprint scanning evolved from criminal analysis to biometric security, and iris scanners have seen increased usage in high-security areas, at one point both technologies were seen as highly advanced and bordering on science-fiction.
Another technology that is extremely similar is voice authentication, but despite also making regular appearances in films, at present the technology has not reached the same levels of use that other biometric systems have.
Why is this, and could voice authentication be the future of security?
What Is Voice Authentication?
A voice authentication system analyses a range of vocal characteristics, typically compared to a sample phrase or sentence. It then uses this information to create a voice print that can be checked against when someone tries to enter a restricted area.
The idea is that it is a simple system to use on the part of the user, but one that is remarkably difficult to forge since even impressions of a voice will have a slightly different cadence, pitch, tone or timbre.
It should not be confused with voice recognition, which is where a system is trained to understand specific words or phrases and use them as an input system, but instead is about discerning different voices from each other.
However, an alternate voice authentication system involving pass phrases sent to a voice control smartphone app have been tested as an alternative to keycodes and fingerprint scanners.
Has It Ever Been Used?
Voice authentication, also known as speaker recognition, has seen relatively limited use for account control, out-of-hours access and in locations where other forms of biometrics are unfeasible.
It was at one point used as part of telephone banking services as an alternative access control system that was easier to use than passwords, passphrases or security numbers.
Why Is It Not Used More In Security?
Unlike fingerprints and iris scanners, which involve biometrics that largely remain static throughout a person’s life outside of illness or injury, a person’s voice can differ on a daily or even hourly basis with such variance as to make it difficult for a system to find appropriate security markers.
Your voice will change throughout the day and could become unrecognisable if you are ill, have a hoarse throat, or your mood is different compared to when you took the voice print.
As well as this, background noise, microphone quality and the protocols for which you interact with the voice security system can be somewhat varied, which can lead to a lot of false negatives.
Reliability is a frequent issue with voice scanners, something that has been aided and deeply hindered by the rise of machine learning systems.
In one sense, AI can be used to improve the accuracy of voice authentication, but on the other hand, it is much easier to forge a voice print than it is a fingerprint or iris scanner.
An investigation in 2023 by Vice News found that an AI voice generation system could be used to break a voiceprint with just five minutes of voice samples.
Leave A Comment