this post was submitted on 29 Jul 2023
185 points (95.1% liked)
Technology
59582 readers
2645 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
A few things are very clear: 1. a phone with a voice assistant enabled has to listen all the time and 2. in order to train the voice assistant the data sometimes needs to be sent to the cloud and listened to by humans.
What is less clear is does this data ever get used for advertising. As you stated there are a number of reasons that make this unlikely.
Simple solution: disable your voice assistant. I do this today and I do not feel like I am losing anything. That said, with the pace AI is improving I can forsee a day when I would feel like I have to enable my voice assistant or I am losing some key functionality of my expensive smart phone service.
If manufacturers are to be believed, the only thing that our devices are always listening for is the trigger word. iPhones have a dedicated piece of hardware or circuit or smth that listens only for ‘hey siri’ and it doesn’t start keeping record until it’s heard that. After which it sends what you say to the cloud to understand what you said.
Yes, exactly. And in order to improve the ability to understand the wake word, they need to occasionally send data to the cloud when there is some indication there may have been a misunderstanding. Also, sometimes humans need to listen when the computer has low confidence.
And of course everything after the wake word goes to the cloud. And sometimes it thinks it hears the wake word when it did not. This goes to the cloud and a human may need to interpret it.
So, some things your phone hears will go to the cloud without the wake word. And humans sometimes listen to them. This is pretty clear. Is this malicious or nefarious? Probably not. But it is complex and hard for unsophisticated end users to understand. And the reality is your phone absolutely does 110% spy on you. Just not by listening to you. It is easy to understand why so many people refuse to believe their voice assistants are not spying on them.
This is generally wrong. Disconnect your device from the internet, and on most (for sure Siri/Alexa) will still activate if they hear the wake word. They won't activate it they don't. Both companies have basically said that the wake word functionality is hardware blocked, and that's not been disproven.
Second, not all assistants/companies are created equal. For example, Apple has made the process of involving human review opt-in. Apple also has no incentive to use this data for anything other than improving Siri. They're not an advertising company and if anything are fairly hostile to others using Apple customer's data for that type of thing without explicit consent. Contrary to Alexa/Google, which has an incentive to use your Voice recordings to advertise to you, EG: you ask your VA what the symptoms of food borne illness are, they show an ad or suggest a search for pepto.
This part is mostly correct. Again, in the case of Apple the phone isn't spying on you, but all of the shit you put on it is. All of those apps are collecting data and collating it in ways that people don't understand. So even though I have a burner Facebook account, since it's tied to my number or email (can't remember which) and I'm sure most of my social graph shares contacts with everything that asks, as soon as I created that account FB suggests to me a whole lot of people I actually know even though I gave it no other real data. People also don't realize that all of this data is often brokered through lots of services, so when you slow down buying tampons or something, another shopping app starts suggesting prenatal vitamins. This is a large part of the reason lots of major retailers have club cards or whatever.
I think you started derailing when speaking about data collection (by the likes of Meta). Most of their data comes from cookies that track nearly every website because nearly every website has Google analytics or Facebook comments embedded. Their bots and reach extend everywhere. On every site.
These kinds of businesses have entire infrastructures built just to get your info. Any info. They have working composites and models for nearly every human. I mean think about what you can learn about a person if you just followed them around every day. It actually gets pretty scary because that’s what’s happening. We are constantly being spied on, watched or monitored to a degree.
IPs also leak a metric ton of data, like geo location, ISP/carrier name, even your postal code.
These companies have myriad of ways to siphon data from the websites you visit, the social connections you make and the data they get from the services used.
It’s less about your phone and more about how many parties have access to the feature or service.
My opinion here is. Sure keep the valid stuff provided the user agreed to it. Have an opt-out though where data is analysed for whatever purpose then deleted. I don't know why they cannot keep data for a day, run analysis and delete on a rolling basis. The benefit of having old data to run improved analysis on is negligible when you're getting as much new data daily as they do.
But, regardless the excerpts it sends when it thinks you might have said the wake word which turn out to be false should be deleted. Do any short analysis for the why right away and delete. Because it really wasn't for the phone/personal assistant.
Yes, but this doesn't happen online. Your voice is not constantly being sent to Amazon or Google servers for activating the voice assistant. It's an offline feature that only reacts to the specific word. That's also the reason why you can't change the activation word.
No, the reason why you can't change the activation word is because each one is tied to a recognizable and highly marketable identity. Amazon does not want you to change the name of Alexa because they want you to repeat their marketing brand over and over and over again.
Funnily enough, you can change the Alexa wake word (edit: on certain devices, I can change it on my Show but not my Tap). It can be “Echo”, “Amazon”, “Computer” and for some reason “Ziggy” (someone is a big fan of Bowie or Sunday morning cartoons?)
I know you can’t change Siri, I’m not sure about Google but I assume that can’t be changed either.
Echo is another Amazon brand word, and Amazon, of course. Computer is so common of a word that nobody would ever want to change it to that.
Also, Ziggy was the computer on Quantum Leap.
Thanks for solving the “Ziggy” one for me! And I agree I feel like “Computer” was for elderly grandmas.
"Computer" is most certainly meant for Star Trek fans. But yeah, it's a pretty impractical choice
Google used to respond to "OK Jarvis" and maybe a couple other things back when it was new. I haven't tried it in a long time to see if it works anymore though.
I’ve never enabled voice shit and never will. Even if it’s all on-device on my phone. If I’m forced to, it’s back to a SideKick.
You don't have to even disable it, just turn off voice commands. You can still activate it by pushing a button. Personally Ive been using Bing chat over Google assistant for a few months now.
I've tried everything to try get rid of google's voice assistant on my phone, stripped it right back and disabled it and then went into the settings in google and my phone but still every time I turn on my wireless headphones it always pops up. It's doing my head in, why can't I stop it!