Skip to Main Content

News Detail

May 20, 2015

Will Face and Voice Recognition Replace Passwords for Insurance Transactions?

by Ara C. Trembly

Thanks in large part to the science fiction genre and forward-looking films such as Minority Report (in which individual identification through retinal scans is the norm in a rather frightening future), the idea of biometric identification is not nearly as far-fetched today as it might otherwise have been.  

The insurance industry, of course, is not known to be on the bleeding edge of technology adoption, yet some in our universe are at least kicking the tires of biometrics in the form of voice recognition and/or face recognition to replace passwords in insurance transactions.  

Indeed, according to an article on Slate.com, Gartner estimates that 30 percent of companies will be using biometric identification on their employees by 2016. In January, Intel announced at the Consumer Electronics Show that it has come up with tools to use biometrics - in the form of facial recognition - to replace passwords.
A USA Today report notes that the company has introduced a new password manager app called True Key. A preview version of the app works with Windows computers and Android phones for facial recognition, and facial recognition will be added for Apple users when the app launches to the general public later this year.  

But are these new technologies a reliable method of identification for insurance transactions?  According to Jamie Bisker, senior analyst with Aite Group, “The voice verification market (leveraging the recognition of a qualified/enrolled user by means of a pass phrase or a brief conversation) has a number of offerings that show the maturation of this technology over the past several years. The ID mechanisms are strong and often make use of multiple characteristics to assure reliable and safe authentication via voice. 

“This is reliable enough to be used in the insurance industry,” he continues. “Smart devices such as phones and tablets contain sufficient processing power and much greater capabilities in terms of device features than ever before, including their speakers, microphones, and less visible features such as noise reduction and echo cancellation.” 

Like voice verification, Bisker notes, visual face recognition systems have improved over the years and are very reliable. “The concept is making its way into the insurance industry, although the uptake may be slower than the voice route,” he explains. “Facial recognition technology has benefitted from greatly improved smart devices. Our phones and tablets commonly feature cameras that include an increased number of pixels (8 megapixels are common now), high range of contrast, increased quality of optics, and greater color precision in their imaging sensors.”  

“I do think we’ll be using image/voice recognition for consumer identification sooner rather than later, but not for secure transactions,” says Jeffrey Goldberg, Vice President, Research and Consulting, at Novarica. “It will start as a way of speeding up a process or directing calls but then require further identification, or it will be used for crafting personalized experiences (i.e. advertising) but not for the actual purchase/binding/etc.”  

At the moment, Goldberg says there is “not nearly enough comfort with voice or facial recognition for use as a transaction identification. While these technologies do exist in controlled environments with sophisticated hardware, when it comes to use by general population on a wide scale via a phone call or a camera phone it’s not ready to be used in a secure way. It’s one thing to use voice recognition for navigating an IVR, and another to make official decisions about someone’s identity.”

Is it happening?

Are insurers today buying into these technologies and putting them into practice?  “From our conversations, this is not an emerging technology that insurers are considering a priority,” says Goldberg. “Even if they are exploring it, it’s not a significant part of their strategy.”  

Bisker, however, answers this question with “a qualified yes.” He notes that, “Carriers are investing in this technology, especially once seeing it in action in similar situations (e.g., SIG in Saskatchewan once they saw use of facial recognition in provincial driver’s license ID mechanisms; Insurance Corporation of British Columbia in Canada as well; and USAA using facial recognition in their mobile banking app).” 

While he specifies that these applications are relatively new, Bisker states that “Once this is accepted, the pace of adoption will pick up.” Voice verification has been instituted at various government organizations, he adds, citing a report that the Rhode Island Department of Labor and Training has implemented a new voice verification system that offers its Unemployment Insurance (UI) customers added protection against fraud and identity theft. Voice verification confirms a customer’s identity by comparing a caller’s voiceprint to an earlier digitized recording. Other government organizations, he says, will likely follow suit sooner, rather than later. 

“Voice is also offered for claims and CSR professionals and as a way to verify IVR users,” says Bisker. “Allianz is listed as a voice verification customer as well.”

What about security?

As with any new technology, security questions may arise.  According to Goldberg, biometric technology may be mature enough to work in some circumstances, for example, to secure an insurer’s data center so that only pre-approved employees can gain entrance. 
“However, in those cases we’re talking about dedicated biometric hardware that has been programmed with a small set of user voice or image information,” he notes. “When we broaden the use case to all insureds, gathering all of their voice/image information becomes very difficult and they will be interacting with normal consumer devices rather than specialty hardware. In that circumstance, the technology is just not ready to make accurate identification decisions.”  

Bisker says the primary security concerns include “corruption of the signal (the recording of the face or voice data) and inadequate secondary technology (the quality of the device, or its signal that is to provide the current voice or face at the time of the attempted authentication – dirty lens, shaking, atmospherics, or literally insufficient bandwidth at that point in time to deliver enough data for verification).” In terms of actual security, such mechanisms are relatively safe, he notes, but are sometimes paired with other security measures to provide even higher levels of assurance.  

Counting the benefits

When voice and facial recognition technologies are adopted, what can be expected in the way of benefits?  

“Biometric recognition systems will be faster to use, and there will be less hassle with user IDs and passwords being lost or stolen,” says Bisker. “We will also see recognition of recorded versus live data presentation, faster generation of new passwords, and less hassle than [having to come] up with a good password that one can remember.” Because the identification is biometric, the password doesn’t need to be written down, he points out.  

“At the moment, I don’t see the [new] functionality as being a major change in process, just a way to make an existing process more efficient,” states Goldberg.  

Looking to the future, he adds that, “There already is some movement to include fingerprint scanners on cell phones, and I think that’s pointing the way for this kind of tech as well. Once consumer devices have built in secure voice/image recognition technology, they will likely also provide APIs for access to third-party applications. This will allow insurers and other industries to take advantage of it.”