Every year, Mary Meeker presents the most influential report on internet trends from the previous year. Just out today, the information-packed 355-slide presentation can now be found online. One of the bigger trends it pinpoints? Thanks to mobile, voice interfaces are on the rise: Twenty percent of all mobile queries in 2016 were made through voice rather than typing.
Voice interfaces have looked like the next big interface for a while now, and this confirms their growing presence. It makes sense: Using your voice is a natural and organic form of communication that everyone uses every day, and using it for your phone means that your eyes and hands will stay free. Tech.Co located an industry expert — Vijay Balasubramaniyan, co-founder and CEO of voice security company Pindrop — to discuss the future of the interface, as well as one of the biggest threats to it: Voice fraud.
How big do you see voice interfaces getting in the near future?
“I see us going from voice interfaces being present on few devices like Alexa and Google Home to people designing new software/hardware systems with voice as the primary interface.
To bring it to life, in the not distant future, I’ll be on a trip and walk into a hotel that I’ve never visited before, and my voice will unlock the hotel room door. As I enter my voice-enabled hotel room, I will say “Alexa, or Ok Google” and it will say “Hi Vijay.” My voice is my identity and it seamlessly authorizes me to do things like play music from my own playlists, book a dinner reservation or log into my Netflix account on the hotel TV. As I head out for a dinner meeting, I order a self-driving car and I unlock the doors with my voice. The challenge is doing this accurately and securely. This is the future we see ahead of us — where our voice is the gateway to our entire digital world.”
So they could beat touchscreens as the dominant interface?
“Absolutely. Voice has three big advantages all leading towards making the interface completely invisible and shifting the focus to user action and intent.
First, your microphone as an input is typically less than 0.5mm (smaller than a pinhead) and can fit on devices with increasingly smaller form factors. Second, the speed at which you speak is four times faster than typing or gesturing. Finally, you have learned how to use and master your voice from your first words (command interfaces) to expressing complex thought (conversational interfaces).
You don’t need to learn new gestures to interact with a new device. Your human voice unlocks any device and brings it to life making the interface and the underlying technology completely invisible.”
How does voice fraud work?
“Globally, 100 billion phone calls are made every month into enterprises. Trillions of dollars including stock trades, product purchases, travel bookings and more are made over the phone. Criminals realize this and have been attacking telephone systems to obtain customers identities or access their financial accounts. Every year fraudsters steal $10 billion dollars by attacking call centers and is the fastest growing form of electronic crime with an increase of more than 110 percent from last year.
And now that voice is being used for IoT devices including your personal assistants, smart homes and connected cars criminals are looking at ways to break the voice controlled world. Voice services today largely know what you’re saying, but they don’t know who you are. For example. their current system of recognizing different voices is at its infancy and is susceptible to voice replay attacks (someone playing a recording of you saying ‘Ok Google'). Google even caveats this by saying: ‘A similar voice or recording may be able to access this info too' in its terms of service. We, therefore need to build security and identity into these devices as a first order primitive as compared to trying to fix it once attackers have compromised it.”
Could you share any tips that would help people keep their tech more secure from voice fraud?
“Set up personalization or identity services available on your device. For example in Google Home this requires you to say ‘Ok Google' and ‘Hey Google' twice. Clear any history of sensitive conversations. Don’t yet link important accounts to your devices.”
Read more about cybersecurity at Tech.Co
Photos provided by Amazon & Google Home