The 2023 State of Voice Biometrics
As incidents of fraud continue to rise across the globe, the market for voice biometrics solutions has yielded huge success for speech technology vendors, and this is only expected to continue. In fact, research firm Insight Partners predicts that the worldwide market for voice biometrics will rise from $1.3 billion in 2021 to $4.8 billion by 2028, a compound annual growth rate of 20.6 percent.
“The fundamental driver hasn’t changed. We’re still seeing year-over-year increases in fraud,” says Brett Beranek, vice president and general manager of security and biometrics solutions at Nuance Communications.
Year in Review
Adoption of voice biometrics technology has been robust of late, particularly in financial services, insurance, government, and IT and telecom. Banks, according to Insight Partners, constantly remain on top of the best authentication methods to combat online fraud, and voice biometrics is the leading technology they deploy for that purpose. Voice biometrics, the firm says, simplifies market user identification and authentication while improving know your customer (KYC) management due to automatic calibration, active and passive authentication, liveliness detection, and panic detection.
Voice biometric authentication is based on vocal patterns, with each person having distinct phonetic and morphological characteristics that artificial intelligence can detect, Insight Partners says in its report. “The approaches are more convenient and speed up operations as the voice biometrics is contactless. Because there is a larger requirement for security than ever before, biometric authentication is rising to the top of the financial technology landscape.”
This coincides with the growing prevalence of mobile and online banking and e-banking, opening channels for cybercriminals to access user accounts and steal money, the report adds.
It’s more than financial services companies that are looking to protect their customers, though. According to Daon, 92 percent of consumers surveyed expect cybersecurity threats to outpace cybersecurity protection technology, so 81 percent are willing to take extra security measures, such as using voice biometrics to protect their information and accounts.
“Businesses that deploy advanced technology to prove and continually authenticate identities at every trust point across the customer life cycle will ensure a trust relationship with their customers,” Tom Grissen, Daon’s CEO, said in a statement.
Healthcare is another industry that has amped up its use of voice biometrics just within the past few years. Despite regulatory pressures and data security guidelines, breaches to hack medical databases are prevalent and can be very costly for both patients and caregivers. Voice biometrics can limit access to patient data to authorized personnel and patients only.
It’s not surprising, therefore, that MarketsandMarkets expects the healthcare and life sciences verticals to grow at the highest rate for the next few years. Other sectors that are investing heavily in voice biometrics include retail, government, IT, telecommunications, transportation, hospitality, and education.
Regardless of industry, one vector that is constantly under attack by fraudsters is the contact center. “Organizations that offer contact centers are getting hammered increasingly with fraud,” Beranek says. “That picked up during the pandemic and hasn’t died down.”
Daon, whose IdentityX voice biometrics solution recently became available on Genesys’ AppFoundry, has found that as much as three-fifths of all fraud losses can be traced to the contact center. That’s why the Genesys collaboration is so important for Daon.
With this partnership, Genesys’ AudioHook Monitor streams the call audio from the Genesys Cloud CX platform to Daon IdentityX to perform voice biometric authentication in real time.
“With contact centers being the heart of the activity, Daon is excited to bring our patented and proven voice biometrics authentication into Genesys’ industry-leading Cloud CX platform,” Ralph Rodriguez, Daon’s president and chief product officer, said in a statement.
Beyond the contact center, another growing threat is the evolution of synthetic voice attacks by fraudsters. This prompted the FBI last July to issue a formal warning about fraudsters using deepfakes and stolen personally identifiable information (PII) to apply for a variety of remote work and work-from-home positions. Deepfakes include videos, images, or recordings that are convincingly altered and manipulated to misrepresent someone as doing or saying something.
The remote work or work-from-home positions identified in these reports include IT and computer programming, database management, and software development, which could give fraudsters access to customer information, financial data, corporate IT databases, and confidential corporate information.
“Complaints report the use of voice spoofing or potentially voice deepfakes during online interviews of the potential applicants,” the FBI said in its warning. “In these interviews, the actions and lip movement of the person seen interviewed on-camera do not completely coordinate with the audio of the person speaking. At times, actions such as coughing, sneezing, or other auditory actions are not aligned with what is presented visually.”
One such synthetic voice attack was carried out by prisoners in Mexico, according to Beranek. “That was an aha moment for many in the industry. I’ve been predicting for some time that this was going to happen eventually. But it’s always hard to predict when in time this will arise,” he says.
Another similar attack, this time targeting contact centers, first emerged in October, Beranek says. “Fortunately for us and for our customers, we’ve been prepared for this, and we’ve been developing technology to detect synthetic fraud.”
The human ear has a hard time distinguishing state-of-the art synthetic voices that have made big advancements from the robotic-sounding synthetic voices of just a few years ago, according to Beranek.
A Look Ahead
Fraudsters are also starting to use artificial intelligence to increase the chances of success, Beranek says.
Such attacks highlight the need for top-of-the-line voice biometrics protection, he adds. “We’ve been developing technology to detect synthetic voices for over a decade. We’re here to help organizations out, to make sure that they are not vulnerable to those kinds of attack. We’re ahead of the fraud community, even those who are using the latest AI tech to support their fraudulent activity.”
Organizations that still rely on security questions, PINs, and passwords are falling behind in the battle against fraud, Beranek says. “They may be in a lot of trouble.”
Though Nuance, Daon, and others have offered voice biometrics solutions for several years, rising demand is bringing new companies into the market as well.
Among other leading vendors identified in an industry report by MarketsandMarkets are NICE, Verint, Pindrop, Phonexia, Aculab, Auraya, OneVault, Aware, SpeechPro, LumenVox, Uniphore, Sestek, Element, Trust Stamp, and AnyVision.
In August, Kardome expanded its voice technology offerings to include voice biometrics and wake word detection that work with the company’s 3D Audio Front End software.
Kardome’s 3D Audio Front End uses location-based source separation and noise reduction algorithms to improve automated speech recognition system performance substantially. Adding Kardome Wake and Voice ID further personalizes a secure voice recognition customer experience, according to the company.
The Voice biometrics solution verifies customers’ identities based on their unique voice characteristics, such as pitch and pronunciation.
In November, iPulse launched VoiceIQ.cloud, enabling customers to procure voice biometrics-as-a-service; it is hosted on the Microsoft Azure platform like the rest of the IQSuite.cloud product range.
Unlike most voice products in the market, the VoiceIQ.cloud charges on a per-transaction basis, without setup, licensing, or implementation fees. Through a standard application programming interface, companies can begin integrating the voice biometric technology immediately. The solution supports both active and passive options.
Voice biometrics is also gaining popularity in use cases outside the standard contact center and customer identification and authentication. Some of these include forensic voice analysis and criminal investigations, emergency management, workforce management, and building and facility access.
Another such use case is transcription. In a study presented in the journal Forensic Science International, voice biometrics-based comparison software did a better job transcribing phone calls than humans when significant background noise was present. Voice biometrics-enabled transcription also did a better job in distinguishing between multiple speakers in a conversation, which is important if a recording is going to be used as evidence in court proceedings.
Also in the criminal investigation and forensics environment, University of South Florida computer science and engineering professor Sriram Chellappan spent the past several years developing a patent-pending technology to accurately confirm individuals’ identities and locations using their voices. The technology, dubbed “Here I am,” creates an unforgeable, encrypted digital certificate on users’ cellular devices. The certificate is permanently verifiable and can be used for authentication in a variety of scenarios, such as employee accountability, protection for victims of domestic abuse, and criminal justice.
In 2023 and beyond, the voice biometrics industry will need to tackle continuing privacy issues, especially as recorded voice samples are still being stored locally and used for purposes beyond the reason for which they were collected. Accuracy is also an ongoing struggle, and although AI and neural networks have aided in this regard, voice biometrics are still not always successful.
Beranek says voice biometric technology providers will also need to pay close attention to the growth in AI-based spoofing attacks. To combat this, voice biometrics providers will need to use AI themselves.
“It’s going to be AI vs. AI, kind of like Terminator 2,” Beranek says. “It’s going to be commonplace. Companies that want to protect their stakeholders are going to look into this technology if they don’t already have it. It’s going to be exciting. The good guys are going to win.”
Phillip Britt is a freelance writer based in the Chicago area. He can be reached at spenterprises1@comcast.net.