The annual I/O developers conference conducted by Google is an event awaited with interest and this year it was no different. Among the various announcements made and demos given, the one revealing the new developments to its artificial intelligence-based digital assistant took the cake.
Google’s research team has come up with voices for Google Assistant which sound extremely close to human ones. These voices will be made available for commercial use soon.
There are other improvements to Google Assistant as well.
Six New Voices Have Been Developed
Google Assistant users, particularly on their Google Home speakers, can activate any one of the six voices, including one of popular musician John Legend. At the I/O developer’s conference, a live demo of two phone calls being made using Google Assistant were played out on the audio track.
In one, a male voice is heard making a booking for a table reservation in a restaurant. The male voice sounded exactly like that of a human, and the person taking the booking at the other end would not have been able to realize that it is a robotic voice from the other end.
The Google Assistant-run voice could take questions and give answers and even pause between words to give out “umms” and other fillers as humans tend to do while making such calls. Google is set to be using a technology called WaveNet to create these near human-like sounding voices.
Can Manage Complicated Calls as Well
The other call went a step further. In that, there is a female robotic voice making an appointment with a hairdresser. Here, Google demonstrated the extraordinary capabilities of Google Assistant. The voice at the hairdresser’s line is heard telling that the appointment sought is not available and some other slot needs to be chosen. Google Assistant understands this situation, reacts to it and reschedules the appointment as per the timing available.
These should be considered amazing advancements in the use of artificial intelligence and machine learning. Google has named this feature Duplex, and what was demonstrated at the conference was just a first cut clip. The company is working on fine-tuning the technology and may include a built-in disclosure feature where the robotic voice will identify itself as a machine talking before going ahead and completing the task. This is being done to deal with the criticism on ethical considerations.
One Command and Multiple Actions
There is a feature called “Routines” within the Google Assistant app. It listens to a simple voice command from the user and will go ahead and execute a series of tasks.
For example, if you say, “Hey Google, good night” the machine understands that you are ready to go to sleep; it will switch off the lights and start up some soothing music or natural sounds to induce sleep and you can customize what you want to listen to.
You have to option to include other tasks as part of this routine and they will be carried out. You can do a similar routine for your dinner time and so on.
In another feature revealed, Google Assistant has been re-trained to accept multiple commands in one sentence itself. For instance, the technology will understand you when saying you wish to turn the TV off and reduce the heating in the room as well. You can tell Google Assistant to do these in one voice command; they will be complied with.
Some of Other Changes Brought About
Among the many other features introduced in Google Assistant, some are particularly interesting. One involves the attempt to teach some ethics or good manners for people, especially children. So, when a voice command is transmitted and the word “please” is added at the end, Google Assistant will respond by saying thank you for asking so politely. Google assigns names to each of these features; this one is called “Pretty Please.”
Visual responses apart from the audio replies have been added. “Continued Conversations” is another little addition. Here, instead of your having to say, “Hey Google” each time, you only need to utter it once you start. After that, you can simply talk to Google Assistant and it will listen to you and respond.
The important feature to note in this is that the machine is capable of making out if you are speaking to it or to someone else nearby.
Again, in all these, there are further improvements being worked on and one may witness constant updating of the technology in the coming months.
Technology vs. Ethics Debate Rages On
Recent developments in AI technology and its use to carry out many tasks have set off a debate among the puritans on the ethics of using technology to replace humans. In addition, as pointed out, there are those who question the practice of a machine being asked to speak to humans, as the chat bots do or as the Duplex feature does—their argument is the human should be told before the conversation starts that they are speaking to a machine or a robot.
They quickly allude to the Terminator movie-like situation, where the machines take over and the humans lose control.
Major technology companies like Google have the responsibility to assuage such concerns and the company did not fail to take this into account. Google announced that ethics in technology will be an issue included in the deliberations at the conference.
Google’s CEO, Sundar Pichai, made a mention even during his address to the developers that the company aims to bring transparency in technology, and the users should know what they’re getting into.
With all these new features and functions within Google Assistant unveiled by Google, one can expect that the other top players like Apple, with its Siri virtual assistant, and Amazon, with its Alexa, will be raring to come out with their own version of changes and improvements to make the competition intense in the field. The customer out there should indeed be the beneficiary of these developments.
How to Get Google Assistant’s New Voices?
If you have already heard about Google’s Duplex technology, you must be aware that it offers near-human like voices within the Google Assistant feature.
If you are keen to get one of these voices to use on your device, make sure to follow the steps below. (Remember that your device needs to be running on Android 6.0 or above.)
3. Find the three dots that take you to the detailed settings mode.
4. Choose the Preferences option.
6. In addition to the two voices already available, you can now see six new additional voices. You can get a preview of each and pick the one you like.