Hey! That sounded much more lifelike. Wait…did it just mimicked human’s chit-chattiness with that ‘Mm…hmm’ and ‘I Gotcha’?
Google Duplex: A breakthrough landmark or did we just opened doors for a disturbing future?
Google’s presence in our lives cannot be drafted better than this, “When LIFE gives you questions, Google gives you answers.” And Google just gave us a glimpse of our bright future at I/O 2018 Keynote at Shoreline Amphitheater in Mountain View, California. An annual developer conference of Google to introduce new hardware, software, and its apps update, Google’s I/O 2018 Concave is a paradise for major platform developers, engineers and executives to flock together to exchange views on all aspects of the android structure. This year Google’s I/O appeared to be all colored in Artificial Intelligence’s (AI) theme. From Uncovering details about Android P, digital wellness and fighting phone addiction wrapped with Machine learning technology and utilizing Artificial Intelligence to improved battery life, a redesigned Android launcher, the event was a technology treat. Built as a gesture-based navigation system, the key highlight of the event, Google’s new launcher is designed to take notes from a user’s apps usage routine to provide a preferred place for the apps. This new flavor, Android 9.0 P will be launched in August 2018 with the announcement of Public Beta on 8 May. Discovering more, Google Assistant AI service has been bestowed six new voices, both male and female so that users can choose their preference. Pop Artist John Legend’s Voice will also be included in Google Assistant’s voice pack. With a focus on making conversations more natural with Google Assistant, this ‘Continued Conversations’ feature will trim out ‘Hey Google’ or ‘OK Google’ phrases with a custom phrase. The keynote 2018 rolled out various thought-provoking features including Pretty Please, Gmail’s Smart Composer, and Google Photos updates, but Google Duplex’s demo by company’s CEO, Sundar Pichai captivated the light of the event as the main star of Google’s 2018 I/O conference.
Image credit: Hindustan Times
Google duplex came to light as a total surprise when Google CEO Sundar Pichai introduced it onstage as one of the most exciting demos with a pre-recorded phone call. Within fractions of time, the technology captivated the internet. Some applauded Google’s move as a new chapter cultivated by Automated Intelligence and some even crowned the concept ‘Creepy’!
So the big question arised- why this Kolaveri di?
Why the situation turned milky?
Blurring the line between human and Artificial Intelligence with Google Duplex
It’s obvious; our mind clearly detects those robocalls which are unnaturally efficient. To revolutionize this pattern, Google took a futuristic step. The recipe of this futuristic technology is still on the stove. Presently Google Assistant supports over 5,000 devices. With its testing beginning this summer, the final technology is yet to be rolled out as a furnished one with more features. It is estimated that by the end of the year, the same will be offered across 80 countries and 30 languages. Wowing the audience with its launch, AI-powered Google Assistant holds prowess to make calls on user’s behalf and could actually communicate to real people in more refined and natural way. CEO Sundar Pichai demonstrated Google Duplex with a pre-recorded call in which Google Assistant AI had a real conversation with a Salon employee to book an appointment. The concept behind Google Duplex is crafted to provide a hand in making calls for the appointments, making reservations. The most surprising of all was when the AI-driven voice call sounded just like a human and speaking human race’s favorite word ‘Mm…Hmm’!
Image Credit: Youtube
With Duplex, Google has established a landmark in Natural Language Processing (NLP), realistic as well as natural sounding communication process. The advancement will surely be an aid for people with hearing disability, the super busy ones and also resolving language barriers. Just imagine that you are running late for work and your mother asks to fix an appointment with her doctor. Before you say, “Mom, I will do it after reaching office” and make her nose go red with anger, you can ask AI powered Google Assistant to give you an extra hand and it will fix an appointment on your behalf. Voila!
“The technology is remarkable,” asserts John Havens, executive director of the IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems. “But I showed [the demo video] to my wife and she said, ‘Which one’s real?’ And there lies the rub.”
Initially, Google Duplex focus will be confined to three kinds of task: making restaurant reservations, scheduling hair appointments and finding out businesses’ holiday opening hours.
The futuristic technology
Conceptualized on recurrent neural network (RNN) and TensorFlow Extended, it’s the combination of text-to-speech engine and a synthesis TTS engine that gives human-like sound and varies the tone of the machine. A special feature of speech disfluencies like ‘um’, ‘hmm’ etc has also been added to the AI powered Google Assistant. The machine also shifts its gears in response time accordingly using low-confidence models or faster approximations. The whole training method utilized here has employed real-time supervision.
Here is a summary from Google AI Blog,
“Google Duplex provides you to get information that isn’t on the internet, acting as a bridge between Google Assistant and any business because it enables the Assistant to get information that isn’t available digitally. For example, you might want to know a business’s holiday opening hours but they haven’t listed it on their website, or you might want to know if a shop has a particular item in stock and it doesn’t have online stock availability. What RNNs like the one powering Duplex can do is process sequential, contextual information, and that makes them well suited to machine learning, language modeling and speech recognition. In the words of Google’s engineers Yaniv Leviathan and Yossi Matias, “It’s important to us that users and businesses have a good experience with this service, and transparency is a key part of that. We want to be clear about the intent of the call so businesses understand the context.”
Image Credit: Google AI Blog
Since the communication process will be generated through Google’s backend system, in spite of user’s phone, it has conceived a bouquet of doubts. What if someone makes a spam call harnessing the expertise of Google duplex? The realistic voice issue may lead to know vital information being disclosed by second party during the call, which the user might not be able to track. Also, can the receiver detect whether he is speaking to a virtual assistant or an actual human?
Google shared a note with CNET that the virtual assistant will “likely tell the person on the other end of the line that he or she is talking to a digital personal assistant.” But remained cloudy over the details about how that will happen.
Google Duplex nestles an entire gamut of millions of possibilities and what if we start exploiting the same?
Google has said that it is limiting the Duplex technology to very specific domains. But John Havens goes on the opposite pole. According to him, “Pretty soon it’s not going to be hard for someone to type in the words to have a virtual assistant break up with their boyfriend.”
A strong dilemma still stays with us; will the technology pull out our personal data to impersonate us?
According to a Google spokesperson, “We understand and value the discussion around Google Duplex — as we’ve said from the beginning, transparency in the technology is important. We are designing this feature with disclosure built-in, and we’ll make sure the system is appropriately identified. What we showed at I/O was an early technology demo, and we look forward to incorporating feedback as we develop this into a product.” “We are still developing this technology,” Pichai told the I/O crowd, and he admitted many calls in Duplex’s testing phase “didn’t quite go as expected,”