A.I. and machine learning era is constantly making our life easier and more comfortable.
Undoubtedly, areas which have been developing rapidly in last few years are Natural Language Processing (NLP) and
Natural Language Understanding (NLU) which have enabled development of advanced voice assistant devices.
However, speech recognition concept is not a recent development - first devices which were capable of recognizing
human voice had been presented in 50’s of 20th century by Bell Labs and later by IBM.
One of the breakthrough events in A.I. and voice recognition areas took place in February 2011,
IBM Watson - a supercomputer capable of understanding and answering natural language questions - took part in
popular quiz show “Jeopardy!” [1]. During two-game match it was competing with two best human contestants.
IBM Watson defeated them and finally won the main prize - 1 million dollars, which was later given to charities by IBM
That was the turning point in competition between the biggest tech companies in designing the best voice assistant system
Here is a brief summary of most popular ones
Apple Siri
In October 2011 Apple introduced Siri - a virtual assistant as a part of Apple’s xOS operating systems. Users can interact
with Siri using voice commands and perform different tasks like: finding information on the web by asking questions, calling & texting,
navigating, scheduling events in a calendar, playing music, changing device settings etc. In 2016 Apple decided to publish Siri API for
third-party developers allowing for Siri integration with a variety of mobile apps.
Recently, Apple has released HomePod which is a smart speaker powered by Siri.
Apple HomePod [5]
Amazon Alexa
A virtual assistant system introduced by Amazon in 2014 and initially deployed on Amazon Echo
and Amazon Echo Dot smart speakers. By default they are in the listening mode, waiting for the
keyword “Alexa” to be spoken, after which a user can ask a device to do something for him.
Similarly to other smart speakers it supports dozens of services and apps for playing music, checking
weather and calendar, navigating etc. Alexa can even order items from Amazon Prime online shop.
Alexa’s major advantage, compared to its competitors, is the “skills” mechanism. This feature enables
users and developers to create small, customized apps for interacting with Alexa-enabled devices,
which external service providers, like e.g: Uber, Lyft or Domino’s Pizza, can use to integrate their
products and services with Alexa. We can also use Alexa for Smart Home solutions to control and
monitor devices at home using voice (e.g. control your lights via Alexa and Philips Hue) There
is a good tutorial on official web pages about creating your own skill for Alexa:
One thing which Amazon engineers really need to improve is handling big numbers :)
Amazon Echo [6]
Google Assistant
Google introduced Assistant in May 2016 as the extension of Google Now search service. It is available on many
different platforms including Android devices, iOS devices and Google Home -a smart speaker by Google.
Google Assistant has similar capabilities as its main competitors: scheduling meetings in a calendar,
playing music or controlling smart home devices. On last Google I/O conference Google introduced new extension
to Google Assistant called Duplex which is capable of making phone calls
and holding natural conversation with humans.
They presented recorded phone calls with a hair salon and restaurant to show how Duplex can actually make a
haircut appointment or reserve a table in a restaurant on your behalf. Google claims that those phone calls had
not been staged beforehand.
different platforms including Android devices, iOS devices and Google Home -a smart speaker by Google.
Google Assistant has similar capabilities as its main competitors: scheduling meetings in a calendar,
playing music or controlling smart home devices. On last Google I/O conference Google introduced new extension
to Google Assistant called Duplex which is capable of making phone calls
and holding natural conversation with humans.
They presented recorded phone calls with a hair salon and restaurant to show how Duplex can actually make a
haircut appointment or reserve a table in a restaurant on your behalf. Google claims that those phone calls had
not been staged beforehand.
Google Home [7]
It is obvious that voice is becoming another interface in human-computer interaction and voice assistants supported by hyped
machine learning will be rapidly gaining popularity in next few years. According to Juniper Research [2] at least one smart
speaker will be installed in 55% of households in USA by the year 2022. Currently, we are hearing a lot of announcements
concerning possible applications of voice assistants. Here are two of them:
machine learning will be rapidly gaining popularity in next few years. According to Juniper Research [2] at least one smart
speaker will be installed in 55% of households in USA by the year 2022. Currently, we are hearing a lot of announcements
concerning possible applications of voice assistants. Here are two of them:
Healthcare
There are a number of use cases in healthcare industry where voice assistants can be successfully utilised.
There are already complete programmes for patients with a specific disease (e.g: diabetes) based on daily
interviews with patients, performed by s voice assistant.
There are already complete programmes for patients with a specific disease (e.g: diabetes) based on daily
interviews with patients, performed by s voice assistant.
Those interviews could cover questions about a patient’s condition, medications adherence, whether a
patient feels pain or not, etc. Voice assistants can also remind patients to take medications, take blood
pressure measurements or to check blood sugar level. Answers to those questions are regularly sent
to a physician in order to constantly monitor patient’s health.
patient feels pain or not, etc. Voice assistants can also remind patients to take medications, take blood
pressure measurements or to check blood sugar level. Answers to those questions are regularly sent
to a physician in order to constantly monitor patient’s health.
Another related area where voice assistants can help is elderly care, one of the examples is
ElliQ [4] - a voice-activated assistant for elderly people. ElliQ not only reminds to take medications or monitors health
but also helps elderly people to stay in touch with their family by setting up video chats and makes them feel less isolated.
ElliQ [4] - a voice-activated assistant for elderly people. ElliQ not only reminds to take medications or monitors health
but also helps elderly people to stay in touch with their family by setting up video chats and makes them feel less isolated.
Office
Voice assistants can also increase productivity in offices by automating many activities which are encountered on a daily basis.
One of the examples is “Alexa for business”, which can start meetings and control video conferences. Voice assistants can make
office workers’ lives easier by helping them in a more trivial way e.g.: by quickly finding information or an address, making to-do lists,
making calculations or scheduling meetings in a calendar.
One of the examples is “Alexa for business”, which can start meetings and control video conferences. Voice assistants can make
office workers’ lives easier by helping them in a more trivial way e.g.: by quickly finding information or an address, making to-do lists,
making calculations or scheduling meetings in a calendar.
Questions:
- Do you use any voice assistant on a daily basis ? If so, which one ? If no, do you consider using one ?
- What are the other areas where voice assistants would bring value to their potential users ?
- What are the possible disadvantages of dissemination of voice assistants ?
Links:
[2] https://www.theverge.com/2018/5/8/17332070/google-assistant-makes-phone-call-demo-duplex-io-2018
[7] https://www.newegg.ca/Product/Product.aspx?Item=N82E16881716002
Comments
The other area where voice assist is really worth it is car industry. Some time ago, Mercedes released their new A class with voice assistant at the level that I'd like to use but to be honest it's too expensive to buy this car to have high quality voice assistant.
I don't see may disadvantages except the fact that people who don't like to talk to other will start to talk with machines. It's dangerous.
As with any American film, it would be nice to have a voice system at home that would help you turn on the light or expose the louvre.
I think the biggest disadvantage of this system is that sometimes the machine does not understand what we are saying about it.
There are few problems. I guess today many people afraid if this technology because it can be used in different fraudulent scenarios, for example, what if some hacked/rogue voice assistant will call your bank by impersonating you and try to withdraw all money from your bank account. Such bots can cause big problems in the future and currently, there are no official regulations that can prevent abusive usage of AI technology.
Bixby and Google Assistant for quite some time. Generally everyone of them works good
enough but as was mentioned above their accuracy is still not enough to replace standard keyboard.
To my mind car industry should be the target of the voice assistants. It would be really helpful to have Siri or Google Assistant in your car. You don't have to interrupt to do something instead those assistants will do everything for you. I think this could significantly decrease the number of accidents.
To my opinion the biggest disadvantage of this system may be misunderstanding. It's not an easy task to recognize what person says and moreover what he/she meant.
What is more, e.g.: Amazon Alexa offers Voice Training feature by which you can teach Alexa your voice using 25 training phrases which in turn allow her to better understand your commands.
On a daily basis is a bad description of my voice assistance usage, as I don't use it that much. I've tried multiple of them (siri, google assistant, alexa) and I will definitely stick to google assistant.
2. What are the other areas where voice assistants would bring value to their potential users?
We could use it more often in work to speed some business processes up or to make notes during the conference meetings. We could use it for the same reason in court or in hospitals to create filled with crucial data reports. We could also use it in schools to learn real stuff and all boilerplate operations or information could be calculated or found by AI.
3. What are the possible disadvantages of dissemination of voice assistants?
I think we would get addicted to the voice assistants, which makes our data and private information vulnerable to the companies running these AI systems. Moreover, If we got addicted, we are much easier to control.
I think that voice assistant can help us with in driving, people should be focused on the road. We can replace buttons with voice command and it can decrease dangerous situations while driving. I think that disadvantages are only when technology is undeveloped. The biggest problem is bad word recognition because if system do it badly you have to make it on your own and voice assistant is unnecessary.
Actually I had my short adventure with Siri and I remember it to be fast and user friendly. I asked myself - why don't I use it then on a daily basis? I guess the main problem is our habits - we are used to type search queries in our browsers, instead of activating the search vocally with Siri. I guess one day, voice assistants will be the only option (touch interface will be obsolete) and we will see the true rise of voice assistants.
What are the other areas where voice assistants would bring value to their potential users ?
I think it is great for blind people. It helps them to cope with daily errands and let them use digital devices, not only telephone, but through IoT also smart houses and even order things from shops.
What are the possible disadvantages of dissemination of voice assistants ?
Everything that's happening due to dissemination of Internet and technology in general - lack of privacy, collection of personal data, the feeling of being overwhelmed and monitored and losing the ability to cope on your own without help of the technology, I guess.
2. I think in everyday life in house, if that voice assistant would be good enough.
3. I honestly have no idea. Sometimes i notice that they can recognize my speech clearly, but i guess it's because of my bad pronunciation)
Currently I don’t use any voice assistant, because one simple reason which is, any of the accessible on market no offer polish language. The idea of voice communication is in my opinion great but not yet for polish people. I consider to using one whenever will be available in polish. I have heard that Google currently working on polish NLP, and I can’t wait for that.
What are the other areas where voice assistants would bring value to their potential users ?
For me that will be cars, when the steering of car equipment will be not so distracting for drivers.
My second thought is that voice - computer communication will be good idea for call office which could reduce costs of employee.
What are the possible disadvantages of dissemination of voice assistants ?
It is already happened that one of the voice assistant send some conversation of its users to random contact, which was in my opinion a bug, but still they are listening us continuously and that could be not safe for people. It is huge privacy volition.
I think there are a lot of areas where voice assistants will bring value. It is hard for me to say something more about this because I don't know about this technology and I think that it is already used in many places.
In my opinion there are no disadvantages. Certainly they will find some, but none of them come to mind.
They have tremendous potential. As far as I am concerned, we can make them capable of solving tasks that personal virtual assistant can do.
It's quite scary that they might pretend to be real humans on the phone. That's what bothers me the most.