Voice Tech Like Alexa and Siri Hasn't Found Its True Calling Yet (recode.net)

An anonymous reader shares a report: As the holiday shopping season approaches, voice-powered smart speakers are again expected to be big sellers, adding to the approximately one-quarter to one-third of the U.S. population that already owns a smart speaker and uses a voice assistant at least once a month. Voice interfaces have been adopted faster than nearly any other technology in history.

While some of this will likely come to pass, the hype might be disguising where we really are with voice technology: Earlier than we think. About a third of smart speaker owners end up using them less after the first month, according to an NPR and Edison Research report earlier this year. Just a little more than half said they wouldn't want to go back to life without a smart speaker. While people are certainly enthusiastic about the new technology, it's not exactly life-changing yet. Today, voice assistants and smart speakers have proven to be popular ways to turn on the radio or dim the lights or get weather information. But to be revolutionary, they will need to find a greater calling -- a new, breakout application.

Smart speakers, like training wheels, are getting people more used to talking to their devices. However, the future of voice probably won't be on speakers at all. The major speaker makers have all added screens to their assistants. Samsung, smartly, is putting its voice assistant Bixby on its TVs, which have the potential to become the smart assistant hub of choice. The key element is the voice assistant, regardless of what device it resides in. Smart assistants will creep into every aspect of our lives and will be available at home and away.

Comments

True calling?

[110010001000]

Not sure what you are talking about. They are generating billions of dollars of revenue and profits. What do you think their true calling is?

Re: True calling?

[peragrin]

It isn't the profits but the usage.

They don't post profit by user numbers.

I listened to someone just yesterday go hey Siri and asked a question.

The question and resulting answers could be found faster by typing wiki subject name.instead the lady goofes off for 15 minutes attempting to get Siri to understand the question and then pull up the various sections of Wikipedia.

Voice assistants take a simple search or inquiry and lengthen it by minutes to get a response. That you end up having to read anyways

Re: True calling?

[Minupla]

Not sure I agree. We have one in our living room where we don't have any desktops. We often use it during dinner conversations to get facts to support a position, "Hey Google, when was France invaded during WW II?" or cooking "Hey Google, how long do you boil a potato?" or set a timer "Hey google, set a timer for 7 minutes".

None of those would be accomplished faster by going upstairs and bringing a system back from sleep and typing the question in.

And "Hey Google, let there be light!" is just fun :)

Min

Re: True calling?

[arth1]

Voice systems aren't optimal even when humans are listening.
Much of the automation and overall design we use in our daily lives became popular because it reduced the amount of talking we had to do.
It might have started with pulling a string instead of calling a servant, but it could have been even earlier, perhaps a library index allowing a monk to find the right manuscript without having to talk to their fellow menk.

Speech is inefficient and error prone. I predict that a few thousand years down the line, we will have evolved away from it.

Re: True calling?

[cykros]

For shopping, I've found it a LOT more useful for reordering things than buying them the first time. Trash bags, paper towels, toilet paper, etc, have all become as simple to restock as "Alexa, reorder ".

Buying things the first time with Alexa seems...suboptimal.

Re: True calling?

[Berkyjay]

None of those would be accomplished faster by going upstairs and bringing a system back from sleep and typing the question in.

So no one has a smart phone on them?

Re:True calling?

[Kjella]

No, but it has a lot of impact on how precise it is. Speech is inherently error-prone and ambiguous. People mishear each other all the time. If you really want to avoid misunderstandings, you write things down.

No, the main problem is that the spoken word is fleeting and poorly searchable. If it's not recorded you have no paper trail of what was actually said and even if you do it's is much easier to find the relevant bits with a transcript or better yet a summary that omits all the irrelevant bits, that is why we have long discussions and try to make short conclusions. But it's exceptionally rare that I actually mishear something in a conversation, like I literally didn't understand the words. It may be that I'm struggling to understand what they're trying to explain, but that would probably be the same over Slack or Skype. I will agree that if you're in effect editing a document then trying to come up with a non-ambiguous specification is much easier in writing. Though even the visual aids are more easily designed and debated on a whiteboard than with fancy electronic tools.

Re: True calling?

[lgw]

allowing a monk to find the right manuscript without having to talk to their fellow menk.

Can we all adopt "menk" as the plural of "monk"? Like polygoose as the plural of mongoose, it's obviously right in hindsight.

Re: True calling?

[Minupla]

Funny thing - before, yes. Now, not as often actually. I find myself tending to leave my phone in a charger and grabbing it when I go out.

And of course our 10 year old ("Hey google, how do you spell X") doesn't have a phone, and she often leaves her tablet on a different floor (or dead :)).

As for the flow of conversation (not your comment, but figure I'll save some electrons :)) - I find it helps for us. We'd get hung up on some question that's parenthetical to the main topic. Now we can google that and move on with the main topic.

Oh and "Hey google, tell ourgroceries to add butter" has saved so many runs to the corner store when something gets missed off the shopping list because one of us used the last of the butter and forgot to tell me before I run to the store. :)

Min

Re: True calling?

[Minupla]

I don't generally respond to ACs but just in case this point is useful to someone who hasn't though this through. If you're worried about spying, faraday bag your cell phone first. At least with my google home I can do network traffic analysis on it (hint, when its idle it sends very little). Try that trick on your cell phone. Well for starters, there's a whole level of your phone you don't have access to. (check out https://media.ccc.de/v/27c3-40... ) - spoiler alert: Silicon/firmware security hasn't gotten any better since then.

Source: I've been involved in cell network security.

True calling?

[Anonymous Coward]

Their "true calling" is collecting personal information from the users. And they are very good at it.

Siri is terrible

[Lucas123]

It feels like I'm alone in my opinion that Siri is a terrible voice-activated, virtual assistant technology. More often than not, Siri can't get simple commands right, often due to the iPhone's poor natural-language user interface. I don't think I've ever had a dictated message turn out correctly using Siri and any of my iPhones (generation 5-8). I'm actually a bit jealous when I see how easy my friend's Android phone understands voice commands and natural language dictation. Google's natural language processing, works nearly flawlessly.

I'm just throwing this out there because I'm wondering if anyone else thinks, for lack of a better criticism, Siri simply sucks.

Their true calling is already achieved

[DNS-and-BIND]

The entire point of these devices is not to be useful by speaking to them. The entire point is to get people to put a freaking microphone in their houses. They can listen to everything. How do you think the device knows how to respond to its name? But that's OK, they pinky swear they delete it all and it will never be used against you.

Down the road a few years people will get wigged out that they're in a home with no microphone. It will feel weird and unsafe. You'll get people refusing to allow their children to visit the houses of the microphone-less (a pejorative will be coined to describe these anti-progressive Luddites). When I read George Orwell's Nineteen Eighty-Four I thought that the telescreen that watched and listened to everything you did was an incredibly stupid idea that nobody would ever agree to voluntarily. We're already halfway there. How did it come to this?

Re:Their true calling is already achieved

[rogoshen1]

These tech gadgets are basically 'toys'. (Which granted, is a very unpopular opinion here.) Post war boom, consumerism and the resulting infantilism of society, fixing a 'need' through advertising yadda yadda.

But yeah, somewhere after 1950 the notion of 'growing up' and focusing on work/family was perverted into individualistic consumerism -- did grown men really play with toys to this extent before then?

Most of these gadgets that you think you need -- or make your life more convenient; really just make you lazy, neurotic, and stupid.

They certainly did

[ugen]

The true calling of the voice "assistants" is to collect and provide personal information that can be processed and used to better market goods and services by various corporations.

What they have not found yet is a plausible use case that would be universally acceptable and persuasive enough to get these devices into as many hands/homes as possible.

Re:"Caring" for lonely people

[jellomizer]

Well lets hope in the future, when you are old and begin to have trouble with once simple tasks, that you will have friends and family available to help you out, and live your later years with dignity.

I don't see caring technology, however I see technology doing the busy work giving people the time and resources to be caring to others. If the volunteer for meals on wheels isn't so interested in getting food to all his spots, they can actually take their time and talk with the people. Even it means 1 meaningful caring conversation once a week, compared to a couple of minute visit every day (To check to see if you are Alive, Healthy, and fed, then to the next house).

The industrial revolution help propel people to live beyond normal survival.
The technical revolution helped propel people access to information and learning.
Now today's revolution with AI and learning systems, is now opening a door to new opportunities and new risks (Just as the previous two was also abused to spread misinformation, and provide junk to people)

I would love to see the use of AI, and Robotics not a device to get rid of jobs, but making such jobs more focused on the customers and caring for people.

Who is using this stuff?

[sjbe]

Voice interfaces have been adopted faster than nearly any other technology in history.

Really? I find that curious because I almost never see anyone actually using them. Seriously, I just never see anyone using Siri or any of the others and I'm around people using smartphones and tablets constantly. Once it while I see someone do a search on their iPhone or dictate a text message. But if I see it happen more than once a week that's a lot.

I don't have any problem with the idea of them but in my experience they don't generally work very well outside of a few niche applications. It's almost always faster for me to type what I'm searching for because they screw up the transcription most of the time. (I have the most generic US midwestern accent you can imagine and no speech problems either) I also cannot imagine any practical use for something like Alexa in my house. Your mileage may vary of course but I don't really see the appeal. I have an iPhone and I find Siri nearly useless to the point of it actually being a hindrance at times. I've never used Cortana on any Windows 10 machine and see no point to it. I haven't played with the Google versions much but similarly I don't see much value in it. I also don't like the idea of announcing what I'm searching for in public even when it isn't anything sensitive.

Counterpoint - Siri works well for what people do

[SuperKendall]

I use Siri all the time and I find it works pretty well. Having friends with Google phones that also use voice I haven't seen any way they actually use the device so far with voice that is heads and shoulders above what Siri can do - it probably has better voice dictation but that's about it right now.

If I want to make a reminder or set an alarm, Siri works great.

If I want to ask for directions, Siri works great.

If I want to open an App, Siri works great...

I don't know what people are doing exactly where Siri does not work well for them, but for a lot of common tasks people do Siri seems to work pretty well.

Targeted Attacks

[DickBreath]

See Audio Adversarial Examples: Targeted Attacks on Speech-to-Text. And see the data.

Just imagine. A television commercial says: Alexa, what is the weather?

Now every human in the room heard that, and it sounds harmless.

What Alexa actually heard: Alexa, browse to evil.com

Pretty neato.

Or see this: DolphinAttack: Inaudible Voice Commands, and see this.

Hope that helps!

This is 2018

[Bobrick]

The last interaction I would want with a connected device is talk to it. I already patch my cameras (phone, laptop) with tape to avoid being snooped on, and at the risk of repeating what other commenters have said already, I have zero use for a surveillance and marketing device listening to me.
Shopping list? If it's really so long that I might forget something, use a pen and piece of paper.
Want to play music? Well, load up the playlist "manually", it will take a whole lot of 3 seconds.
Whatever the fuck else people use these for, I've never heard one example that didn't make me go "Why?".

Voice Recognition Shouldn't Needs The Internet

[BrendaEM]

I like the idea of voice recognition, but it's insane to trust any company to put a computing box in my room that records everything and sends out what information that company chooses over the internet. Own your own voice recognition.