Michael Gips, Chief Strategy Officer and Deepak Chandran, Technical Advisor, both for Emergence Technology Group, discuss the untapped potential of voice biometrics for security and safety.
Tens of millions of people have facial recognition on their smartphones.
Fingerprints, which can also unlock smartphones, have been used for access control for decades.
Iris recognition is deployed for national ID programs, high-security access control and border management.
Among other qualities that can be used for biometric technologies identified by The Biometrics Institute are ear shape, scleral vein, gait, hand geometry, gait, odor and keystrokes.
Voice is a well-recognized biometric, but it is overwhelmingly used for non-security purposes – think of your everyday commands to Siri or Alexa.
To be sure, banks and some customer support centers use limited voice recognition for authentication purposes and to detect fraud, but security uses are not widespread.
Is voice biometrics for security due for a breakthrough? Many experts think so.
Nacho de Marco of the Forbes Technology Council has dubbed voice biometrics “the next big thing”.
Grand View Research projects an almost 15% compound annual growth rate for speech and voice biometrics and recognition through the end of this decade.
The reasons are multifold: it’s touchless, non-invasive, simple, unique, easy to enroll users, cost-effective to integrate and familiar.
At the same time, the factors that have slowed its adoption are trailing away.
It is now comparable in accuracy to many other biometrics, incorporates “liveness detection”, can filter out background noise and can adjust to vocal changes (for example, illness, puberty and intoxication).
Speech recognition and voice authentication are distinct concepts that are often conflated.
While separate and distinct, the combination of both in a single technology can be extraordinarily powerful.
Speech recognition refers to the meaning of spoken words and converting them into written text.
It is commonly used in applications such as dictation software, voice assistants and call transcription services.
Voice recognition does not identify the speaker.
Voice authentication verifies the identity of the speaker.
It analyzes characteristics such as pitch, rhythm and tone to create a “voiceprint” that can be used for secure access control in various applications to include phone banking and smart home devices.
Unlike speech recognition, voice authentication does not necessarily comprehend the content of what’s being spoken.
Voice recognition can be compared to understanding the language of a book you’re reading.
You can grasp the meaning of the words and follow the story without knowing the identity of the author.
Voice authentication is akin to recognizing a friend’s handwriting in a letter before reading its content – you can identify their penmanship without knowing what they wrote.
Intelligent assistants have become an integral part of our lives, providing a hands-free way to control our devices, search for information and manage our tasks.
However, while these voice-activated/speech-recognizing assistants like Alexa and Siri are convenient, they also come with significant concerns – they store data in the cloud and cannot recognize individual users accurately.
New voice biometrics technology is emerging that fills the gap, offering a secure and reliable solution to these issues.
This technology employs an advanced on-edge voice capture algorithm to create a unique voice print for each user.
This voice print is a distinctive audio pattern the system uses to identify and authenticate the user.
The technology can store multiple prints of the same user to accommodate changes in voice per time of day, per environment, over time and so on.
New developments allow the technology to accurately identify a specific user’s voice even in noisy environments, regardless of their emotional state or physical health.
Coupled with speech recognition of particular words and phrases, the technology can verify that a specific person uttered a particular word or phrase.
This dual technology unlocks multiple potential uses.
Home automation: smart home devices can recognize who is giving the command, allowing personalized experiences.
For instance, when a user says, “Play my playlist,” the system will play the specific user’s collection of songs, not a generic one.
Security systems: technology can be incorporated into security systems for voice-based authentication, thus adding an extra layer of security, as the system will only grant access to specific phrases and registered voice prints.
Customer service: companies can authenticate customers during phone-based customer service, thus streamlining the process, as customers won’t have to answer multiple security questions.
Health monitoring: technology can be used in health apps to monitor patients’ health through their voice.
Any significant changes in the voiceprint could alert medical professionals to potential health issues.
Personalized learning: in education, voice biometrics be used to create personalized learning experiences.
For instance, an educational app could adapt its content based on who it recognizes is using the app.
Life safety: voice biometrics have the potential to enhance safety and security.
In terms of life safety, it could be used to activate emergency response systems when individuals are unable to trigger an alarm.
For instance, if a senior citizen falls, they could call for help using their voice as a means of identification.
Or, in the case of a home invasion, a trigger word could notify emergency response personnel, family members, friends, neighbors and others, geolocate where the victim is and send help on the way.
Hospitality: hotels could implement voice biometrics for check-ins and room access instead of dispensing key cards and codes.
This would allow guests to unlock doors, order room service or request hotel information by using their unique voice.
For example, simple oral commands could allow a guest to tune to a specific TV channel, turn on a particular light, close the shades or set a wakeup call.
Banks: banks benefit from voice biometrics by authenticating customer identities and authorizing transactions.
The technology provides a layer of security beyond frequently defeated passwords or PINs.
Additive manufacturing: voice commands can play a role in controlling industrial 3D printers and other equipment.
By recognizing authorized users through their voices this technology can control and monitor usage.
Call centers: call centers can streamline their verification processes by implementing voice biometrics.
Use of knowledge-based questions can help call centers identify customers through their responses and their voices.
Automotive: motor vehicles can leverage voice recognition technology to personalize settings and experiences for drivers.
Driver preferences, entertainment options and other features would be based on vocal characteristics.
This application provides an important safety function because it keeps the driver’s eyes on the road instead of on dashboard screens, knobs and buttons.
When considering potential applications for voice biometrics, users should ask themselves the following questions.
Speech and voice are exquisitely powerful tools with the potential to unlock doors, trigger alarms or call for help.
Exciting new developments are transforming the world of speech and voice-related biometrics.
While other biometric technologies will likely always have a place in security, the world is just starting to learn the power of its own voice.
Michael Gips is the Chief Strategy Officer at Emergence Technology Group, which develops voice recognition and authentication technology.
He is the former Chief Knowledge and Learning Officer and Chief Security Officer for ASIS International.
Deepak Chandran is Technical Advisor for Emergence Technology Group.
He has developed cutting-edge technologies for the smart grid, aquaculture, electric vehicles, lighting and many other applications.
This article was originally published in the March edition of Security Journal Americas. To read your FREE digital edition, click here.