Market Spotlight: Security
AI Goes Beyond Voice Biometrics
Sometimes, when the stakes are high, even voice biometrics might not be enough to adequately secure information. And when fraud does occur, AI can help.
“Voice biometrics is an obvious first step in helping alleviate some risk, but AI can be applied in other ways to augment a biometric solution,” says Mars. “Conversational AI can be applied to make fraud reporting quicker and easier to navigate, as the current system (often through call centers) is time consuming and confusing. Conversationally enabling the fraud reporting process ensures that fraud is reported and resolved faster, while voice biometrics ensures that the process is safe and accurate.”
Fraud and security aren’t just concerns for businesses—with personal information often available on the web thanks to data breaches, everyone has to be worried about the state of their personal data. But with news of new breaches happening regularly, the average person can’t keep up. Mars says AI might be the solution for that as well: “Voice AI could also be applied to fight fraud in the form of a ‘fraud assistant’ that measures your risk, quickly alerts you of issues, and empowers customers to quickly react and respond. The benefit of a virtual assistant for security is that it is always ‘on’ and analyzing data, providing a 24/7, personal security system.”
Voice Biometrics in the Real World
Up until now, we’ve largely talked about the security applications of voice biometrics in the business world, on digital platforms, and in the call center. But there’s a great big world out there, and voice biometrics are being employed to keep us safe in all kinds of ways you might not expect.
For instance, Radisys is tapping into the Internet of Things (IoT) using real-time speech analysis to connect with emergency services. “Currently, Radisys is in trials with several large system integrators to enable real-time speech analytics and video analytics for smart cities, public safety, and corporate security command and control,” says Adnan Saleem, chief technology officer of software and cloud solutions at Radisys. “Radisys’s MediaEngine, along with network and RAN slicing, enables prioritization of resources over noncritical network traffic. This enables us to use real-time speech analysis to set up keyword alarms such as ‘fire’ or ‘gun’ or ‘call 911,’ as well as actual distress sounds such as the sound of gunfire or glass breaking. This can then initiate a number of security message alerts depending on the customer’s needs.”
So what does that look like in layman’s terms? Saleem adds, “Say someone is attempting to commit a robbery and says, ‘I have a gun.’ The IoT-enabled camera, activating on the ‘gun’ keyword, would allow the network to initiate both a call to local authorities, sending them the audio and video stream, and also to initialize real-time facial detection, providing additional information on the perpetrator and allowing for a more informed response.”
The folks at Radisys aren’t the only ones who have recognized how voice technologies can be used to fight crime. Interpol, the international police agency, launched the Speaker Identification Integrated Project (SIIP), in 2014. According to an announcement made at the time, “The partners in the SIIP project are working to develop cutting-edge speaker identification technology which will help law enforcement identify voices of unknown individuals.”
In 2018 the work came to fruition when the European Union’s website announced in 2018 that the EU-funded SIIP aimed to put an end to any doubts about voice recognition in the courtroom “with an innovative probabilistic, language-independent identification system. This system uses a novel speaker identification (SID) engine and a global information sharing mechanism (GISM) to identify unknown speakers who are captured in lawfully intercepted calls, recorded crime or terror arenas, social media, and any other type of speech source.”
It’s clear that as the technology continues to advance, voice biometrics and identification are becoming an integral part of the security landscape. Whether you’re trying to protect a mobile banking app or identify an international criminal, there may be a voice solution for you. But it is the uses and applications that have yet to be invented that are truly intriguing.
Theresa Cramer is the editor of Speech Technology magazine. She can be reached at firstname.lastname@example.org.
Apple has temporarily suspended its Siri response grading program over privacy concerns and says users will be able to opt-out in future iterations of its popular voice assistant.
Smart vendors will think beyond AI and algorithms and find ways to solve business problems
Voice biometrics is becoming popular with big corporations but is still out of reach for many
Customer experience is at the heart of every brand, but one of the most frustrating experiences for customers often comes when they're contacting you. Voice biometrics may be able to help streamline and improve their experience.
Voice biometrics just might be the next great hope in securing data and other content. But as people use their mobile devices for more and more, embedding speech biometrics on these devices becomes more important
Companies and Suppliers Mentioned